Waiting for Rain: Agriculture and Ecological Imbalance in Cape Verde 9781685856175

Citation preview

Waiting for Rain

Waiting for Rain Agriculture and Ecological Imbalance in Cape Verde

Mark Langworthy Timothy J. Finan

b o u l d e r l o n d o n

Published in the United States of America by Lynne Rienner Publishers, Inc. 1800 30th Street, Boulder, Colorado 80301 www.rienner.com and in the United Kingdom by Lynne Rienner Publishers, Inc. 3 Henrietta Street, Covent Garden, London WC2E 8LU © 1997 by Lynne Rienner Publishers, Inc. All rights reserved Library of Congress Cataloging-in-Publication Data Langworthy, Mark. Waiting for rain : agriculture and ecological imbalance in Cape Verde / Mark Langworthy and Timothy J. Finan. Includes bibliographical references and index. ISBN 978-1-55587-709-5 (hc : alk. paper) 1. Agriculture—Cape Verde. 2. Agriculture—Environmental aspects—Cape Verde. 3. Natural resources—Cape Verde. 4. Agriculture and state—Cape Verde. 5. Food supply—Cape Verde. 6. Cape Verde—Rural conditions. I. Finan, Timothy J. II. Title. S473.6.C25L36 1997 338.1'4'096658—dc21 96-44538 CIP British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library. Printed and bound in the United States of America The paper used in this publication meets the requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1992.

Para o povo caboverdiano tão sofrido, tão amável

Contents

List of Tables and Figures Acknowledgments Map of Cape Verde

ix xi xii

1

Sustainability in a Fragile Environment: An Introduction

2

Public and Private Resource Management: A Methodological Approach

13

3

The Natural Resource Endowment

35

4

Characteristics of Cape Verdean Rural Society

53

5

Rainfed Agriculture

79

6

Irrigated Agriculture

101

7

Household Management Strategies

131

8

Agriculture and Policy Choice

159

9

Conclusions: The Future of Agriculture in Cape Verde

187

References Index About the Book

197 204 212

vii

1

Tables and Figures

Tables

3.1 3.2 3.3

4.1 4.2 4.3 4.4 4.5 4.6

5.1 5.2 5.3 5.4 5.5

Annual Precipitation on the Four Major Agricultural Islands, 1970–1988 Groundwater Availabilities and Use on the Four Major Agricultural Islands Land Quality Characteristics for the Four Major Agricultural Islands Historical Record of Droughts, 1719–1996 Demographic Characteristics of Rural Population on Selected Islands Distributions of Rainfed Cropland on Santiago and Santo Antão Distributions of Irrigated Cropland on Santiago and Santo Antão Distributions of Rainfed Cropland by Landownership on Santiago and Santo Antão Distributions of Irrigated Cropland by Landownership on Santiago and Santo Antão Comparative Estimates of Cultivated Rainfed Land on Selected Islands, 1978 and 1988 Yields of Corn and Beans, 1975–1990 Characteristics of Land and Labor Use in Rainfed Agriculture on Santiago and Santo Antão Representative Budgets for Rainfed Crops Outputs and Net Values of Livestock ix

39 43 46 61 62 70 70 72 72

80 84 86 90 92

x 5.6 5.7 6.1 6.2 6.3 6.4 6.5 6.6 6.7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8

8.1 8.2

Tables and Figures

Economic Returns of Representative Rainfed Systems on Santiago and Santo Antão Annual Net Returns of Livestock Distribution of Irrigated Land on Santiago and Santo Antão Irrigated Land by Water Source Distribution of Average Irrigation Intervals Various Estimated Shares of Irrigated Area by Crop Production of Major Irrigated Crops, 1971–1990 Crop Budgets for Major Irrigated Crops Characteristics of Representative Farm Systems Distribution of Rainfed and Irrigated Land by Number of Adults in Household Rainfed and Irrigated Landholdings Average Household Labor Requirements and Availability, Annual and Peak Labor Season Regression Models of Household Labor Allocation Characteristics of Irrigated Farms on Santiago by Source of Water Distribution of Irrigation Intervals Among Farmers Within a Surveyed Irrigation Network, Santiago Household Income on Santiago and Santo Antão by Quintile Characteristics of Surveyed Households on Santiago and Santo Antão by Per Capita Income Quintile Cropping Patterns and Net Revenues for Irrigation Network Under Alternative Intervals Simulations of Potential Increases in Agricultural Incomes of Surveyed Households

93 97

102 106 108 109 114 120 127

134 135 137 138 143 144 149 152

174 177

Figures

1.1 2.1 7.1

Population and Agricultural Production Capacity over Time The Political Ecology of Cape Verdean Agriculture Household Labor Demands and Supply

7 15 155

Acknowledgments

This project has been a long labor of love, as endearing as it was enduring. Over the dozen years of research and travel that generated this work, we have benefited from the assistance and wisdom of many people who guided us over unknown waters, maintained saintly patience in the face of our disingenuousness, and provided the emotional support necessary to bring this endeavor to its fruition. In particular, we acknowledge Jimmy S. Hillman, project director of the Foods Crops Research Project and retired head of the Department of Agricultural Economics, and Carlos Véléz-Ibañéz, former director of the Bureau of Applied Research in Anthropology (BARA) at the University of Arizona. Both provided us with the motive and opportunity to finish this long-term project. Of the many Cape Verdeans who so kindly shaped our experience, we offer special recognition to Horácio Soares, Elísio Rodrígues, Milu Lobo, António Sabino, and other professionals who have dedicated their lives to the challenge of rural development. In our research effort we benefited from the friendship and knowledge of our manu Raul Varela Semedo, Tito Andrade, and Vladmiro Andrade; we shall forever cherish those exciting times spent in endless fieldwork. José and Zuleika Levy and their respective families, especially Dona Milu, provided professional support and friendship. In every rural community we visited over the years, local Cape Verdeans took us in, shared their lives, and made us aware of their awesome strength in the face of adversity. Their friendship will always be the richest reward afforded us. At the University of Arizona we received unerring assistance from our respective staffs—Connie McKay, Loretta Cosgrove, Linda Phipps, and Maria Rodriguez. From our hearths we received the quiet forbearance, encouragement, and warmth of Pat and Amélia. This book is as much their accomplishment as ours. —M. L. —T. J. F. xi

1 Sustainability in a Fragile Environment: An Introduction

The Republic of Cape Verde, an Atlantic archipelago off the Guinea coast of Africa, is small in both magnitude and population. Located 453 nautical miles due west of Senegal, the combined landmass of the 10 islands and 2 islets amounts to around 4,000 square kilometers. The six islands to the North (Santo Antão, São Vicente, Santa Luzia, São Nicolau, Sal, and Boa Vista) constitute the windward, or barlavento, islands, while the four southern islands (Brava, Fogo, Santiago, and Maio) make up the leeward, or sotavento, islands. The population of the country is currently estimated at about 380,000, and, as such, its population density (95/km 2) is among the highest of West African nations (World Bank 1993b). Although it is perhaps a minor participant in the overall flow of world economic and political affairs, Cape Verde nonetheless shares many defining ecological characteristics with its Sahelian neighbors and faces very similar challenges of resource management. The majority of its population resides in rural areas and derives its livelihood, at least in part, from agricultural activities. And like most of the Sahel, Cape Verde is arid, its farming subject to the vagaries of highly unpredictable and often miserly weather systems. Because of the common ecology and shared history that tie the archipelago to the continent, an understanding of Cape Verdean economy and society provides insights of a broader regional significance. Most evidence indicates that the islands were uninhabited until Portuguese explorers, venturing south from Madeira, arrived in the late 1400s and claimed them for the Crown. The motives for subsequent settlement were founded more on the central geographic location of the islands than on the richness of the local resource base. The proximity of Cape Verde to the Guinea coast and its far vaster and more diversified set of resources provided the Crown a dominant position in intercontinental trade and other commercial advantages related to centralized warehousing and transshipment facilities. 1

2

Sustainability in a Fragile Environment

As a consequence, Cape Verde soon developed a dominant position in the Guinea slave trade to Europe and the New World. Agriculture, conversely, never emerged as a viable economic alternative. The insular colony failed to meet the expansionist objective of the Crown, that of providing abundant agricultural land for Portugal’s impoverished peasantry—as Mozambique and Angola later did. During the early years of settlement, commerce flourished, leaving agriculture to occupy the role of a subsidiary subsistence activity. It is true that, on a minor scale, the early colonists cultivated cotton, from which the utilitarian trade good panos (dyed measures of cloth) was woven and exchanged for slaves. They also raised livestock to supply slave ships with meat and hides. But mostly the settlement farmed to feed itself. Despite the initial success of the Guinea flesh trade and the rapid demise of agriculture, history, with its ironic flair, would eventually determine other fortunes for Cape Verde. Through time, the trade advantages of the archipelago were eroded by changing political winds and innovations in sea travel, and as commerce diminished, agriculture emerged as the primary economic activity of the islands—somewhat by default. Today, Cape Verde is an agricultural society in an arid environment, and food production and provisioning constitute the most pressing and persistent social concerns. As a rather extreme example, Cape Verde might be considered a microcosm of the political ecology of the Sahel—a laboratory where complex coping strategies are conditioned by a skein of sociopolitical institutions and played out in a delicate and vulnerable environment. A rapidly growing population—with limited opportunities outside agriculture—struggles to eke out a subsistence livelihood by slowly exhausting a fragile, increasingly degraded natural resource base. From the perspective of the local rural household, survival is the central issue that drives shortrun decisionmaking. Government leaders also interpret their political success in terms of policies that direct agricultural activity toward broader goals of economic growth and national food security. Nonetheless, in this unrelenting effort to achieve critical short-run household and national goals, society runs the risk of undermining the productive capacity of the system as a whole. It is possible that household need and national intent have become incompatible with ecological reality, thus surpassing the sustainable limits of the productive base. The small scale of Cape Verde offers a dramatic fishbowl view of a human ecological dynamic that has unfolded in large parts of Sahelian Africa and elsewhere. In this regard, Cape Verde stands neither unique nor alone; rather, it shares the critical food-security dilemma that affects the entire region. There is widespread concurrence among both technicians and inhabitants that environmental degradation has occurred and is occurring throughout West Africa. The causes of this environmental decline,

Sustainability in a Fragile Environment

3

however, are hotly disputed and remain embroiled in a polemic armed with single-factor explanations. At its core, the complexity of this controversy contrasts the role of uncontrollable, cyclical phenomena such as climate against the role of human agency. Most scholars agree on the existence of patterned relationships among climatic events, rural population densities, and environmental degradation; however, the lines of causality are fervently debated. The frequent occurrences of severe drought over recent decades together with the declining trend in average annual precipitation throughout the Sahel have fed a global-warming hypothesis that suggests widespread climate change and minimizes the human hand in the causal process. Contrasting arguments assert that population growth and the expansion of agriculture into more marginal areas bear the responsibility for increasing desertification, thus reducing the role of climate as the independent causal factor. Other participants in the debate adhere to more complex, multileveled perspectives that examine the specific forms of human activity and their respective environmental consequences (see Gorse and Steeds 1987:7–8). This approach reasons back from the household and its decisionmaking patterns to explore the external forces that determine differing strategies of production, notably, the political pressures to expand cash cropping at the household level. In their influential book on African famine, Franke and Chasin (1980) interpret the ecological problem in the Sahel in terms of the widespread adoption of inappropriate technologies related to the regional relationship of dependence on the international economy. Within the context of this argument, Ball (1976:520) directly attributes ecological degradation to the colonial efforts to “integrate peasants and herders into a monetarized, commercialized economy” primarily through national policies to increase export and livestock production. The most comprehensive and cogent positions on environmental degradation in the Sahel draw eclectically upon the existing arguments. Ecological deterioration is a complex problem that cannot be understood solely in terms of natural events. As Boserup (1965) established, increases in rural populations result in the intensification of land use (with shorter fallowing periods), greater herd pressures on range vegetation, and the overexploitation of fuelwood. Historical realities as well as national and international forces further condition the strategies of agriculturalists and pastoralists by shaping the incentives (and disincentives) that influence their production decisions within the household. These decisions can aggravate and prolong the negative consequences of drought and create a downward spiral of environmental degradation that is difficult to reverse. In effect, the most credible models of environmental degradation must interrelate such macrolevel variables as climatic change, population growth, and international economic trends with such microlevel variables

4

Sustainability in a Fragile Environment

as household production strategies, patterns of access to agricultural resources, and technology choice. These models must account for potential contradictions in goals between household and nation-state, as well as for the inevitable fact that positive short-term returns may incur high long-run costs in terms of resource depletion. In Cape Verde, where food self-sufficiency is now remembered as part of some near-mythic past, the problem of environmental degradation displays all these levels of complexity. The resource base is fragile and under siege by an expanding population. Moran (1982), in a careful review of colonial history, presents a well-supported argument that the exploitative role of the Portuguese Crown in the formation of current agricultural patterns, especially the establishment of maize as the staple crop, created an agricultural economy inappropriate to local conditions. Today, soil conservation experts are unanimous in their assessment of the negative impacts of traditional rainfed agricultural practices (Lopes and Meyer 1993). To accept this position is not to deny, however, that climatic patterns do show a downward trend in rainfall while population growth has increased. For our purposes here, whether the root causes of the current ecological crisis lie in declining rainfall, historical patterns of resource allocation and use, the incentives presented by the international economy, or the short-run production strategies of the household itself, the critical enigma remains. How do farmers and policymakers manage resources so as to assure both the national food supply and the survival of the rural household when the expansion of agriculture could engender irreversible damage to that very resource base? At its core, the problem as we pose it is one of resource management in a complicated sociopolitical and physical environment.

Beyond the Ecological Balance

At one level, this book is about Cape Verdean food production and the welfare of farming households. As such, it offers a detailed and current description of Cape Verdean agriculture under both rainfed and irrigated regimes. Based on several years of fieldwork, these descriptions document current agricultural practices, variations in technical strategies of production, patterns of input use, crop choice and rotation, and forms of indigenous knowledge employed by local farmers. We also estimate the sources and magnitudes of rural incomes among representative rural households. No other such comprehensive documentation of Cape Verdean agriculture has been attempted since Ribeiro’s (1954) study of Fogo Island and Amaral’s (1964) landmark volume on Santiago Island, both works of Portuguese geographers based in Lisbon during the colonial period. At the

Sustainability in a Fragile Environment

5

same time, this book represents the first effort to document the impact on agriculture of Cape Verde’s independence and the dramatic social changes that occurred with this watershed event. This book is also about agricultural sustainability and natural resource management. Cape Verde has a resource endowment that defines a set of binding environmental constraints (i.e., climate, soil, topography) on agriculture. Just as critically, however, social institutions—wrought in history—also determine the characteristics of farming. Individual households gain access to existing resources and exploit them in accordance with two categories of man-made rules: those established by local communities (i.e., “traditional institutions”) and those imposed by nonlocal forces (i.e., “public institutions”). These sets of traditional and public rules shape resource access and influence the patterns of resource management, which in turn determine the short-run and long-run impacts on the environment. Not all farmers enjoy equal access to agricultural resources, and this inequality results in significant forms of socioeconomic differentiation in rural society, which inspires the hypothesis that different classes of households adopt different strategies of resource management. Thus, in this book, we intend to describe the natural resource base available to farmers, emphasizing the nature of physical constraints to production; then to document as systematically as possible the public and traditional rules that define resource access and management, accounting for the variation due to social differentiation. Always lurking throughout the following chapters is the shadow of the Malthusian dynamic, which is readily evidenced in Cape Verdean history. Over the first five centuries of colonial existence, the relationship of local population to the resource base was primarily determined in strict biological terms. Assailed by severe periodic droughts, population levels of the islands oscillated in direct proportion to changes in the productive capacity of the resource base. We learn from Carreira’s (1984) punctilious history of Cape Verde that the regular occurrence of pluriannual drought caused the production levels of staple crops to dip below minimum survival needs. The subsequent mortality due to widespread starvation lowered the person/land ratio, thus restoring the ecological balance. When the rains returned, agricultural production increased, and the population grew rapidly. Although it is true that traditional sharing institutions mitigated some of the regional consequences of drought by redistributing available food stocks, overall food supplies were cruelly short, and stark Malthusian principles ruled the day throughout most of Cape Verdean history. Under the colonial rule of the Portuguese governor, little was done to alleviate the horrors of drought. During these regular crises, the colonial administration vacillated between inadequate gestures of assistance and a laissez-faire approach consistent with reigning ideologies. Portugal itself

6

Sustainability in a Fragile Environment

suffered from chronic impoverishment and was without the financial means to mobilize substantial mitigation activities. Moreover, long-distance communications were notoriously slow, while starvation imposed its unforgiving timetable. In the final accounting, however, historical reports suggest that the Portuguese generally lacked the collective moral resolve to save Africans from famine. So the rhythm of environmental crisis wrought its devastating legacy of population control, even up to the mid-twentieth century. Beginning with the 1970s, however, the historical tandem of drought and decimation was severed. The wave of successful independence movements throughout Lusophone Africa, the fall of the Salazarist regime in Portugal, and the geopolitics of the Cold War all inspired a call for international intervention. As a result, drought-related starvation in Cape Verde became intolerable to both the international donor community and the many Cape Verdean emigrants residing in Europe and the United States. While the political climate improved over this decade of change, the physical climate worsened. The archipelago began to suffer its most widespread and extended drought—lasting, with only scarce interruptions, to the present day. Shored up by food-aid assistance, the population has nonetheless continued to grow since the 1970s, first at a 4 percent annual rate, then falling in more recent years to 2.4 percent (World Bank 1993a). Thus, in defiant reversal of the past Malthusian pattern, the Cape Verdean population expands while the productive capacity of the agricultural resource base declines. As the title of this book suggests, Cape Verde has moved into an ecological imbalance between sustainable production capacity of the resource base and the size of the population seeking to derive its livelihood from agriculture. This relationship, illustrated in Figure 1.1, casts its shadow over all aspects of Cape Verdean economic and political life. Over centuries, the population line (A) roughly accompanied the subsistence capacity line (B), determined by the amount and distribution of rainfall. The short-run changes in agricultural production capacity are due to variable rainfall patterns, while the long-run decline is a function of soil loss and reduction in soil fertility. After the last great catastrophe of midcentury (1946–1947) and the onslaught of the protracted drought period of the 1970s, the respective directions of the two lines have diverged, but not independently of each other. When a rural population exceeds carrying capacity for a prolonged period of time, there arises the danger of a permanent reduction in productive capacity that will not correct itself with the return of the rains. And in Cape Verde, there are few rural employment opportunities outside of agriculture to absorb the excess rural population. The threat posed by the continuing divergence of these two lines is that of irreversibility, which raises critical and fundamental questions about household food security, national self-sufficiency and sovereignty, and social and political stability.

Figure 1.1

Population and Agricultural Production Capacity over Time

8

Sustainability in a Fragile Environment

Agricultural sustainability implies resistance in the face of change and resilience under stress. Most scholars concur that sustainable activities are those that meet economic objectives without compromising biological ones. Thus, in theory, a sustainable food-production system utilizes a natural resource base without diminishing its availability to the next generation (see Repetto 1986; Beets 1990:16–18). As the complexity of the real world asserts itself, we see that the sustainable systems may not meet domestic food-security needs (Ruttan 1988), as is the case in Cape Verde. The complex nature of sustainability ultimately requires difficult decisions involving uncertain trade-offs, hence this book is also about food and resource policy. It describes past and current policies that Cape Verdean leaders have enacted to promote national and household food security in a fragile ecosystem. Here we focus primarily on two basic policy approaches through which Cape Verde has sought to redress the ecological imbalance. The first approach wagers a technological solution and seeks to identify and encourage alternative practices that reduce environmental degradation and increase the output efficiency of the resources currently being utilized—primarily soil and water. To achieve this end, the government has invested in research and extension activities and has played a major role in the management of natural resources. The second approach attacks the population side of the imbalance. The dilemma that Malthus so aptly described for Cape Verdeans presents little choice but to stabilize population growth by either reducing the birthrate or facilitating outmigration from rural areas. A combination of reduction in the number of rural people, improvements in the basic value of human capital through education and training, and, over the long run, the development of nonagricultural economic alternatives defines the ultimate strategy. National self-sufficiency in food production is not a realistic goal for a country in which 90 percent of its foodstuffs is either imported or donated by the donor community. On the other hand, the current ecological imbalance constitutes a critical threat to economic and political stability, and some level of sustainable agricultural production must eventually be established in order to avoid disaster. The overarching challenge facing policymakers in Cape Verde is how to narrow the gap between lines A and B (in Figure 1.1) and to achieve a sustainable balance over the long run. This imperative is depicted as lines P', P'', and P''', which represent the vectors of possible policy interventions that would conceptually work toward the establishment of a reasonable and sustainable system of production. The policy alternative depicted by line P' establishes a sustainable production level under current climatic conditions; line P'' represents a policy initiative that promotes rapid agricultural growth over the short run, but that ultimately results in the severe depletion of the natural resource base. The third alternative, line P''', seeks a policy set that would move gradually to lessen the food-deficit situation in a sustainable way.

Sustainability in a Fragile Environment

9

These alternative policy approaches will be discussed in the context of the public role in agricultural resource management. Over the short run, however, policymakers in Cape Verde face the urgent problem of feeding the rural population. The lack of nonagricultural alternatives leaves more people practicing agriculture than the resource base can support. At the same time, the government understandably fears an uncontrolled exodus of rural inhabitants to urban areas, which are already extended beyond their ability to provide either employment or basic infrastructural services. Moreover, the current short-run strategy that depends—almost entirely— on international largesse and emigrant remittances is clearly vulnerable. In essence, Cape Verdeans find themselves challenged to redefine the role of agriculture, to relocate its place in the general economy, and to confront the issue of sustainable resource utilization. But what are the specific policy issues that will shape the future of agriculture in Cape Verde? To address this question, we need to discriminate between dryland and irrigated farming.

Policy Issues in Dryland Agriculture

The majority of agricultural lands in Cape Verde are cultivated under a rainfed regime, and only a small minority of farmers has any access to irrigated lands. Numerous technical reports have directly associated farming practices with environmental degradation (Freeman et al. 1978; USAID 1984; IFAD 1990; Lopes and Meyer 1993). The widespread dryland cultivation of maize has been implicated as a major cause of soil erosion and high rainfall runoff rates. Soil and crop technicians argue that maize is an inappropriate cultigen for the environmental conditions of Cape Verde, which are better suited for more drought-tolerant and higher-producing crops, such as millet and sorghum. In addition, current population growth encourages the expansion of maize cultivation onto more marginalized lands, particularly the steep, volcanic hillsides of the valleys. The conventional argument denounces traditional production practices: farmers do not adopt systematic measures of erosion control, such as contour planting, and actually accelerate soil loss by removing all competing vegetation from the sloping fields. Nevertheless, maize, which was introduced by the Portuguese from Brazil, has been totally woven into the cultural fabric of rural society and has been the preferred staple of consumption. Under existing precarious conditions, yields are too low to assure household food security, and the responsibility for provisioning the rural population has fallen upon the public sector. Public work fronts are a common feature of rural Cape Verde and provide the employment that injects cash into the rural economy and reduces the attraction of rural-urban migration. The public work system is financed by food aid that has been

10

Sustainability in a Fragile Environment

monetized through the private market system and made available to rural customers. Thus, the major policy issues with regard to dryland agriculture focus on strategies to reduce erosion, increase productivity, and, at least for now, maintain a local safety net that keeps rural people on their farms. Over the longer run, the questions become the sustainable size of the rural population and the locations in which the excess population might be absorbed outside the rural areas.

Policy Issues in Irrigated Agriculture

On these arid islands, water is the critical resource not only for agriculture, but also for rapidly expanding urban populations. The availability of water for irrigation is highly variable and sparsely distributed, and the policy problem focuses on the need to increase the use efficiency of this scarce resource. Most experts agree that current groundwater supplies significantly exceed current usage levels, which suggests that capital investment in irrigation or reorganization of existing systems could increase highvalued production. A second issue addresses the equity of groundwater distribution. Policymakers perceive a critical trade-off in public versus private, or local, management of groundwater resources, in that larger, more efficient farms might serve some national production goals but might negatively compromise the situation of many rural households that rely heavily on their minuscule plots of irrigated land. Because of this same concern for equitable access, sharecropping tenure relations have been a target of policy interventions. A third policy dimension involves cropping patterns, because the majority of irrigated agriculture is devoted to the production of sugarcane for the purpose of distillation into the traditional alcoholic spirit, grogue, rather than to the food crops that would enhance national food security. In an economy with such low levels of domestic food production, the cash crop/food crop debate edges over toward one of national security and the ability of the country to finance necessary food imports should food-aid assistance be interrupted. In the aggregate, these problematic issues project a particular vision that guides policymaking in Cape Verde. Over the short-run, rural society must be made viable, first to avoid the disasters of the past and, second, to discourage rampant rural flight to ill-prepared urban centers. Over the long run, Cape Verde seeks to wean itself from international donor dependence and to discover its own sustainable basis for development both in and out of agriculture. To discover what policy alternatives will effectively achieve this vision constitutes the national challenge.

Sustainability in a Fragile Environment

11

The following chapters lay out the various pieces of the agricultural resource-management and policy puzzle. Initially, we present a discussion of the methodology that guides our analysis (Chapter 2), and we suggest that this approach is readily adaptable to any agricultural society where improved resource management requires a comprehensive understanding of existing production strategies. Chapter 3 describes the natural resource endowment of the archipelago, and Chapter 4 outlines the historical context of agriculture in Cape Verde. Detailed current descriptions of dryland (Chapter 5) and irrigated farming practices (Chapter 6) are followed by an analysis of private resource-management strategies adopted by farmers (Chapter 7). In Chapter 8, policy alternatives relevant to public resource management are explored. Although this case study of the Cape Verde archipelago highlights the problems and explores the management options that have been tried or appear feasible, we, alas, must warn that the Cape Verde story offers no facile solution to the essential dilemma that sets food security and environmental sustainability in a contrapuntal relationship.

2 Public and Private Resource Management: A Methodological Approach

The story line we establish throughout this book interprets the current reality of Cape Verdean agriculture within both the physical and the policy environment. In this chapter, we present the overall methodological approach that integrates the natural resource base, the agricultural activities of rural households, and the impacts of national policy in Cape Verde. The goal of the analysis is to identify those critical factors that influence the management of productive resources at the disposal of rural households. With an understanding of the incentives and constraints that direct agricultural and nonagricultural activities, we can then assess the current welfare of rural families in Cape Verde and identify the potential for public actions to improve conditions of rural residents. Toward this end, we construct a framework that depicts the relationships between the combination of productive assets—labor, land, water, and capital—available to households and the outputs they are able to generate with those resources. As an analytical basis, this framework allows us to estimate actual income levels of households and to predict expected changes resulting from specific policies designed either to increase household access to productive resources or to improve the technologies that transform these resources into valuable outputs. We then outline the types of information necessary for our analysis and the data-gathering strategies that generated the empirical foundation of this book.

The Political Ecology of Cape Verdean Agriculture: A Conceptual Framework

Borrowing from the recent social-science literature, we refer to the complex relationships that link household strategies to the physical and institutional environment as the political ecology of Cape Verdean agriculture 13

14

Resource Management

(depicted in Figure 2.1). In this conceptual diagram, we identify three essential components: a specific resource endowment (the physical environment), a basic management unit (the household), and a set of institutions that defines the means by which households can mobilize and utilize these resources (the institutional environment).1 The overall resource base is established by the geophysical, climatic, and biological conditions within the archipelago. In Cape Verde, the critical aspects of the physical resource base are the extreme scarcity and variability of rainfall, the sharp topographical relief that promotes erosion and the consequent low soil fertility, and the small amount of total land available relative to the existing rural population. In short, the agricultural productivity of land is very low and extremely variable over time and location, and land is scarce relative to the number of households that depend on agriculture for their livelihoods. Households are the empirical units that manage the natural resource base so as to assure livelihoods for their members. Following Netting, Wilk, and Arnould (1984), households are residential units of variable structure that perform a set of essential functions, including production, reproduction, and transference of resources.2 In contrast with polygynous societies of Sahelian Africa, the Cape Verdean household, as a social norm, follows a rather straightforward, monogamous, nuclear pattern, although the empirical reality reveals a high frequency of single-parent, womenheaded families. Although the structure of the residence unit may extend to include a lineal or collateral relative, the production and decisionmaking processes are concentrated around the household head, rather than dispersed among different household subunits, as with the Sahelian concession households. The private ownership of agricultural land is endorsed under Cape Verdean social and institutional structures, and both law and custom sanction the partible inheritance of family property, including land. Family members constitute the core pool of labor services to work household lands. Thus, each household has at its disposal a mix of available resources, which varies according to the composition of the family and the set of assets that has been accumulated through ownership or other institutional forms. A large body of literature analyzes the factors that may affect the production and consumption decisions of agricultural households as economic units.3 In the context of incomplete market integration, such decisions may depend more on specific household characteristics (the combination of available resources relative to the number of consumers in the household) than on external market prices (Lopez 1986). For example, in developing economies, households generally do not have ready access to markets for all resources and commodities, so they articulate with markets in a complex fashion. In the effort to meet domestic food needs, households decide between consuming their agricultural production and marketing the output

Resource Management

Figure 2.1

The Political Ecology of Cape Verdean Agriculture

15

16

Resource Management

to acquire cash income for the purchase of necessary consumption items, including food. Without markets to determine the prices of the whole range of inputs and outputs, the relative values of cash and food depend in part on household-specific characteristics that affect consumption patterns and production possibilities. Perhaps the most important example of incomplete market integration in rural Cape Verde is that of labor. Individuals have very limited opportunities to employ their labor in rural areas. The two traditional alternatives have been the public work fronts and emigration. Public work fronts do not operate throughout all of the country, and many households do not have access to this opportunity simply because of their geographical location. In addition, the public work program restricts the number of members from a single household that may participate, and it closes down during the rainy season. Emigration, on the other hand, is a very costly and risky employment strategy, especially for individuals without contacts in the country of destination. Thus, particular household circumstances—namely, having friends or family currently abroad—strongly affect the attractiveness of this employment alternative. These household-specific differences in access to public work fronts and emigration are not due to differences in market wage rates, but they can significantly affect labor-allocation decisions in agricultural activities. Furthermore, farmers in many parts of the world commonly devise complex complementary interrelationships among productive activities, particularly between cropping and livestock production (McIntire, Bourzat, and Pingali 1992). In Cape Verde, for example, many farmers use corn straw to feed animals, and the value of the straw in some years surpasses that of the grain. This interplay between production and consumption decisions arises when households are not freely able to buy or sell all inputs, outputs, and factors of production in markets. In such an absence of complete market integration, household resource-management strategies depend on household characteristics and access to resources as well as on relative prices. It is thus necessary to understand the operational characteristics of all the institutions, including markets, that regulate or define the transfer of resources among agents within an economy. The particular terms that define transfers—for instance, forms of payment, forms of access—determine transactions’ costs. Because of the various factors that either influence the prices faced by different producers or restrict producers from participating in market transactions, agricultural systems exhibit significant variation with respect to adopted technologies, proportions of different inputs used in farming activities, and production responses to changes in economic incentives.4

Resource Management

17

According to our framework, households operate within an institutional environment comprising all the socially sanctioned “rules of the game” that regulate human interactions (North 1990:3). More specifically, rural households acquire access to or dispose of available resources through enabling institutions that define the conditions under which these transactions may take place and the form of payment (if any) needed to acquire use rights. Furthermore, these institutions define the distributional patterns of use rights among specific groups or individuals within the community. For example, agricultural land in Cape Verde is transferred across households by partible inheritance, and local customs define the specific process by which such intergenerational transfers take place, namely, how resources will be shared among offspring. In addition to inheritance, land can also be transferred across households through market institutions—either purchase, rental, or share tenancy agreements. Individual households may mobilize one or several of these avenues to build their land endowment. For analytical convenience, we divide the institutional environment into local or private institutions, which are formed and monitored within the rubric of community traditions, and state or public institutions, which are embodied in the political, bureaucratic, and legal structures of the nation-state. The distinction between these two forms is based on the notion that the management of resources is more direct and personal in private institutions, whereas public institutions require formal structures (government bureaucracies) to undertake the role of resource stewardship. Our conception of private institutions corresponds closely to Bromley’s definition of autonomous and self-enforcing conventions, which are “autonomous because they arise from the internal dynamic of . . . several agents, and they are self-enforcing because no independent agent has any incentive to modify the convention” (1991:90). Public institutions, in contrast, represent wider interests outside the local rural communities. Although such public institutions as organized churches and international assistance organizations also operate within rural Cape Verdean communities, our analysis focuses on state institutions because they have the most direct impact on resource-management decisions of households. The activities of other public institutions must be sanctioned within the national legal system and, in fact, operate in close association with governmental agencies. The productive resources of rural households are governed by an overlapping network of private and public institutions. For example, much of the daily management of irrigation networks is undertaken directly by the farmers who share water from a specific water source, but some of their activities are regulated by the national agency responsible for groundwater management (and urban supplies). Similarly, household labor may be

18

Resource Management

exchanged with neighboring households through local institutional mechanisms or may be hired out through national or even international labor markets. Land is also affected by a range of local customs, as well as by national legislation defining legally sanctioned forms of land transfer. In some cases, this intersection of influence among several institutional levels can be complementary, as in the case of water management where public and local institutions are highly integrated. In other cases, however, state institutions clash strongly with the interests of rural households and local institutions. In the early 1980s, land-reform legislation outlawed sharecropping, but this land-transfer institution persists because it is a preferred mechanism for many landlords and tenants. Within this web of interacting institutions that enables resource access and control, households face complex management decisions, particularly when various institutions impose conflicting constraints and incentives on resource use. From where do such institutions come? In general, both public and private forms have evolved over time through political negotiations among interested parties seeking to address particular problems related to the distribution of resources among members of the community. Private institutions reflect the interests of constituents within local communities as well as adaptive responses to specific environmental conditions. Public institutions, on the other hand, are shaped by outside (in many cases, even international) interests and may have little regard for the particular concerns of members in local communities. Local institutions can adjust rapidly when all members in the community can agree on a mutually advantageous change; however, local institutions are often slow to respond in situations where individuals have conflicting interests. Those who benefit under the status quo will tend to resist any changes that are prejudicial to themselves but improve the overall management of the resource base within a community. By contrast, public institutions may totally ignore the interests of local communities, or they may successfully address situations of conflict among groups or individuals in an equitable manner. The specific institutional environment operative at a given point of time represents an amalgamation of many diverse elements forged over a period of time and reflects shifts in relative power positions of different constituencies, as well as the dynamics of resource management under changing circumstances. Some of the institutions that govern rural life in Cape Verde are holdovers from the colonial period and were fashioned when the agricultural resources were less scarce in relation to the rural population, and they have been adapted, with varying degrees of success, to current conditions. Our analysis considers the institutional environment to be exogenously determined for rural households at any given point in time. We do, however, recognize the adaptive and evolutionary nature of social institutions (Barth 1966; Bailey 1969). Although households and local communities

Resource Management

19

must acknowledge the rules established by public institutions, they are often able to devise actions that circumvent the effects of these external institutions. Consequently, for public institutions and policies to be effective, they cannot diverge too radically from local interests, or avoidance behaviors will become widespread and efforts to ensure compliance will become too costly and unwieldy. Our analysis helps to identify some of the long-run pressures for change that will come to bear on the institutional environment, but we do not attempt to explicitly model these political processes. In this analytical framework, households have an initial endowment of resources that is determined by the individual circumstances of the household and its particular articulation with the institutional environment. At any given point in time, individual households will allocate their existing resource base among alternative productive activities. In order to have an empirically tractable model, we assume that households make these decisions to maximize total household income, which comprises cash income (from agriculture and other activities) and the monetary value of agricultural output consumed directly by the household. Income is generated either by utilizing household resources in productive activities or by selling or hiring out resources. Thus, household land and labor can be utilized to produce agricultural and livestock outputs. Alternatively, land can be sold or rented, and household labor can be employed off the farm to receive wage income.5 Figure 2.1 illustrates the interrelationships among these three elements. The natural resource endowment of the islands forms the overall productive base on which rural households earn their livelihoods. At the same time, the institutional environment defines specific household resource endowments. In the figure, the household moves its resources in one of two allocative directions—toward either agricultural or nonagricultural activities. Under the agricultural category, we include crop and livestock activities, off-farm agricultural wage labor, and exchange labor. The nonagricultural category is comprised of nonagricultural wage labor (primarily public work front employment), artisan production, petty commerce and other entrepreneurial activities, domestic maintenance, education, and emigration. Household managers are assumed to make their decisions based on their perceived trade-offs of the benefits and costs of the alternatives they face. This decisionmaking process determines for each household its allocation of available labor to agricultural, domestic, and off-farm activities; specific crop choices and technologies employed in agricultural production; investment levels in physical and human capital; patterns of consumption and savings; and participation in community affairs. From these allocative decisions, the family obtains some level of income. Over time, households are assumed to undertake constant evaluation and modification of their resource-management strategies in order to enhance their well-being.

20

Resource Management

At the same time, household resource-management strategies incur direct environmental impacts that affect the sustainability of the resource base, as Figure 2.1 suggests. In conceptual terms, households can rationally adjust their strategies if prevailing agricultural practices function to degrade the production base, for example, by relying more heavily on nonagricultural income sources. In rural Cape Verde, however, hard reality asserts itself in two ways. First, it is not clear that farmers consciously weigh the trade-offs between short-term welfare and long-term sustainability. Second, even when the consequences are known, the immediate lack of opportunities outside agriculture severely restricts the household ability to opt for long-run environmental benefits if they mean a reduction in agricultural activity. Within limits, it is likely that households do alter their patterns of resource management to the extent that they perceive and directly incur the long-run costs of their actions, but survival occupies the position of priority for allocative decisions. Yet government leaders may not share the same sense of urgency that households do, and their policies reflect a priority concern with the environmental consequences and their social costs. In sum, a wide range of overlapping enabling institutions defines the rules and conditions under which households may acquire, use, and transfer agricultural resources. These institutions are neither deterministic nor immutable but, rather, face continual pressures for change. The particular allocative strategies that households pursue may themselves impose pressures on the enabling institutions and their preservation through time. As the dynamic between human intervention and the natural resource base unfolds, the trade-offs between household production strategies and resource sustainability may reduce the effectiveness of some institutions and increase the importance of others. Changes in household behavior will have interactive impacts on both the natural resource endowment and the institutional environment. As one of its principal advantages, this framework explicitly recognizes the variability across households with regard to access and use of resources toward income generation. Rural households vary along measurable and patterned dimensions. Neither populations nor productive resources distribute themselves evenly over the physical landscape, and all individuals do not enjoy equal endowments of talent or equal access to opportunities. Thus, in the social environment of rural Cape Verde, we expect to find differences in scale, in forms of access to resources, in technology, and in wealth. These differences establish the bases for patterned variation in resource management—both directly in terms of farm household strategies and indirectly in terms of the impact of public policies and institutions. The first methodological challenge is to depict the existing variation in a relevant, accurate, and interesting manner. We have chosen to employ

Resource Management

21

the concept of the representative household, or farming system, synthetically constructed on the criteria of access to land and water, cropping patterns, and the type of household management structure—either jointly managed or female managed. All policies, whether well-intentioned or insidious, exact their impacts differentially on households, and we have attempted to develop a classification of Cape Verdean farm households that face significantly different “opportunity frontiers” and thus respond differently to policy interventions. The second methodological challenge is to explain the observed variation in such a way that identifies the critical variables affecting resource allocation. If we first describe the range of strategies that households devise within the context of Cape Verde’s resource endowment, it is then necessary to discern why these particular strategies have been selected. At this point, the analysis becomes useful to policymakers concerned with both national food security and sustainability of the resource base.

The Economic Analysis of Rural Cape Verdean Households

As our conceptual framework argues, the household is the operational unit that makes management decisions about agricultural resources based on the relative costs and returns of alternative management strategies. For this analytical task, we construct and compare representative budgets of the crop and livestock activities of discrete household types. The representative models incorporate information about the prices of inputs and outputs, the production technologies utilized, and the resource endowment of each household type. This approach constitutes an alternative to standard econometric models that posit the uniformity of household response to economic signals, namely, relative prices. Most empirical economic analyses of allocative behavior assume that markets exist for all outputs, inputs, and productive resources, and that all producers have identical access to these markets and face the same prices. Under these assumptions, the decisions of producers are driven only by relative prices between inputs and outputs, and all producers will respond similarly to changes in relative prices. Furthermore, all producers will use the same technologies and apply the same relative proportions of factors to their production activities. Econometric estimates of supply functions, profit functions, and cost functions measure the relationships between quantities supplied to the market and prices of alternative outputs and inputs. Fixed factors, such as land and capital, may be included in the models, particularly in estimations of profit or cost functions, but usually only as a measure of available capacity. Thus these models do not assume different supply responses for producers

22

Resource Management

with different allocations of fixed factors; rather, econometric models typically compress the observed input/output decisions of all producers into a single average production function for the entire sample. As a result, differences in responses across different types of farms are concealed.6 It is precisely our interest in the variation among farm types that makes this econometric approach less appropriate. In the Cape Verdean context, the assumptions of free access to markets and identical prices across producers do not accurately describe the agricultural reality. Many resources are not allocated through markets, and the institutions that govern nonmarket transactions can impose different opportunity costs upon different individuals. For example, the cost of irrigation water varies significantly across farmers in Cape Verde, but not because of differences in price. Water is not allocated through markets where individuals are able to purchase desired amounts at any point in time. Rather, allocations are either made through negotiations among members within a network, or, in the case of publicly managed networks, a national irrigation agency determines patterns of distribution among the members. The physical capacities of different irrigation systems and the particular institutional mechanisms that schedule water deliveries to users create significant variations across individual producers in terms of access to irrigation water over the course of the growing season. Similarly, households face very different opportunities with regard to off-farm employment. Households residing near an operating public work project will have more ready access to this alternative than those households more distant from any public project. Likewise, households with family members in other countries have much easier access to emigration opportunities. These are very important differences across households that cannot be captured by price differences alone, either because markets (and thus prices) do not exist, or additional, nonprice constraints restrict market participation. Representative Household Budgeting

Our analysis of household resource strategies incorporates more information than just the relative market prices between inputs and outputs. In particular, this approach requires a quantitative understanding of variations in technologies and access to resources across farms within the agricultural sector, as well as a qualitative assessment of patterned variations in producer access to markets and other resource transaction institutions. The use of standard budgeting techniques allows us to depict this institutional variation. Budgeting assumes a fixed-coefficient technology and requires specific information about the magnitudes of the physical input/output coefficients. Market prices for all outputs, inputs, and factors of production are also required. This framework may be used to evaluate differences in

Resource Management

23

technologies (input/output coefficients) and prices by constructing separate budgets with appropriate changes. Budget results indicate the relative economic incentives of alternative resource strategies and changes in those incentives that result from changes in prices or technologies.7 A number of factors constrain the income-generating opportunities available to Cape Verdean families. Most obvious among these are the limited quantity and quality of resources at their disposal. In addition to the overall restricted land area accessible to most households, the geological and climatic conditions on the archipelago influence the productivity of both cropping and livestock activities on rainfed land. The scarce supply of developed subterranean water sources poses limits on the total amount of irrigated land. Moreover, institutional constraints determine resource use rights of households, the mechanisms of resource transference, and how returns from resources are distributed among households with overlapping ownership and usufruct claims. These institutions define and limit the way in which households may use their available resources. The representative household budgets explicitly incorporate these constraints by specifying both the representative combinations of resources for the different household types and the associated factor opportunity costs. The representative budgets provide a basis for the evaluation of alternative policy and resource-management initiatives. The budgets incorporate the impacts of the enabling institutional environment in Figure 2.1 in two ways. First, the budgets that model representative household types capture the variation in resource combinations resulting from specific policy decisions. Second, the prices included in the budgets may also reflect the impact of institutional interventions, such as the cost of water. To the extent that a given policy can be associated with a change in access to resources, the adoption of a new technology, and specific pricing structures, the budgets generate relevant economic consequences at the household level. If information is available on the relative frequency of a given representative household in the rural population, more aggregated impact estimates are possible. This use of budgets as a foundation for policy analysis has an established tradition in development economics (Monke and Pearson 1989). Representative household types are interrelated systems of individual activities. They are defined in terms of a specific mix of available resources and agricultural production patterns. In our Cape Verdean example, a given farm could be comprised of a certain amount of rainfed land and, perhaps, an irrigated plot with a specific crop mix and yields that vary with access to supplies of water. This farm also has a given livestock component. All cropping and livestock activities are represented by individual standardized budgets, which are combined in specific proportions to model the representative household system of resource allocation. These models

24

Resource Management

thus explicitly incorporate the representative household’s access to a particular combination of resources, as well as the multiple and interrelated activities undertaken by the household. The representative households are assumed to share the overall objective of increasing net returns to household resources, or household income, defined as the value of outputs minus the cost of all inputs and resources that are not owned by the household. In particular, household labor and owned land are not counted as costs in the calculation of net returns. Since the household does not have to purchase these owned resources, the net returns from all household activities can be interpreted as household income, or the payment to family-owned factors.8 This analytical strategy allows us to avoid the problematic valuation of household labor at current market wage rates (Barlett 1980) and also provides a derivative estimate of the value of the critical domestic labor (child care, food processing, etc.) that supports productive activities but is seldom measured. The budgets estimate family income from representative combinations of cropping and livestock activities. However, Cape Verdean households also allocate labor to nonagricultural sources of income. To reflect this reality, the analysis also considers the dependence of rural households on off-farm sources of income, primarily public work front labor, agricultural wage labor, nonagricultural wage labor, emigration remittances, and pensions. With this approach, any change in prices, in access to productive resources, or in wages can be evaluated in terms of the impacts on household income, that is, on family welfare. Some qualifications must be acknowledged with respect to resourceallocation decisions. First, the effects of risk associated with different resource-management strategies require careful attention, since households incorporate risk perceptions into their assessments of competing allocative strategies.9 Poorer households are likely to be particularly adverse to any strategy that has a significant probability of occasionally generating an unacceptable low return, even if, on average, expected returns are higher than the traditional and less risky alternative strategy. Several authors have discussed differences in production decisions depending on the level of risk associated with access to irrigation water over time (Easter 1977; Reidinger 1974; Wade 1988). The tail-end users—those most distant from the water source—are generally subjected to greater levels of risk in access than are those closer to the water source. Bromley, Taylor, and Parker (1980) cite case studies indicating that farmers with more insecure access to water will select more drought-resistant crops with lower expected returns and will reduce application rates of other inputs in comparison to farmers with more reliable water supplies. Time preferences may also vary systematically with income levels. 10 Poor households may be expected to discount possibly higher future income

Resource Management

25

more heavily than wealthier households, since the more vulnerable farmers have fewer incentives to make investments with substantial delays in payoffs. These poorer households are thus more likely to pursue nonsustainable strategies, since they cannot afford to wait for long-term benefits that could accrue from alternative resource uses. These differences in time preferences are independent of the higher opportunity cost of external (investment) funds for poorer households, but the effects are the same: a reduction in the attractiveness of long-term investment. The economics of the farm household, then, are modeled on synthetic budgets that are in turn used to construct a typology of farms built around specific combinations of resources. The analysis is static in the sense that the results are tied to the specific assumptions about price and technical relationships, as well as to the institutional environments that are changed only exogenously to examine potential impacts of policy interventions. At the same time, the budgets permit an effective comparison of the incentives underlying alternative uses of resources. In the analysis, we employ the budgets to establish a base, then alter the assumptions to examine the impacts, and finally, interpret the results against the accumulation of data acquired over a multiyear research period.

The Data-Collection Process

The focus on households as the unit of analysis exacts a heavy demand for detailed and systematic information about the socioeconomic conditions and behavioral strategies of rural families. The information set for this study was collected in a series of field activities over a five-year period, during which time a specific research framework was developed. Our fieldwork methodology, which we call sequential rural research, is an iterative, mixed-methods approach that draws inspiration from the methodological innovations associated with the farming systems research paradigm (Finan 1996). Under this strategy, a specific sequence of research methods is employed, with each stage addressing a different dimension of the research problem and building, in an interactive manner, upon the corpus of knowledge accumulated in the prior stages. The sequential rural research strategy is based on two basic principles. First, some types of information are more costly or difficult to obtain than others. For example, descriptive information, particularly that which can be directly observed, such as the number of people in a household, is much easier to obtain than information about the whole range of factors that influences households’ resource-allocation decisions. Second, the relative benefits of different types of information depend on the extent to which they address specific policy issues and offer a basis for continued empirical

26

Resource Management

research. For example, the distribution of individuals’ eye color is relatively easy information to obtain, but it is not relevant to any policy decision. Distributions of age and sex within a population may not be directly relevant to policy decisions, but these variables become extremely important to the extent that household welfare is associated with the dependency ratio or the gender of the household head. The sequential rural research program in Cape Verde unfolded as part of a five-year Food Crops Research Project with the Cape Verdean National Agricultural Research and Development Institute (Instituto Nacional de Investigação e Desenvolvimento Agrário [INIDA]), funded by the U.S. Agency for International Development (USAID). The overall objective of the socioeconomic analysis component of the project was to identify the most important constraints to increased agricultural production faced by rural households. This information was then directed toward the establishment of research priorities in the various departments of technical and socioeconomic research in INIDA. The long life of the project facilitated the development and implementation of this sequential rural research approach. The initial stage in this methodology was a series of rapid rural appraisal (RRA) visits designed to acquire background information on farm households within the project’s target regions.11 RRA techniques are often recommended for situations where the existing knowledge base is minimal—an accurate assessment of the Cape Verdean case. Here, the RRA activities attempted to identify broad patterns of variation in several areas of general information including: (1) household access to productive resources and the range of agricultural and nonagricultural activities in which those resources are employed; (2) technical information about agricultural practices, including cropping calendars, input/output relationships, and access to input and output markets; and (3) institutions that provide opportunities for households to acquire additional resources or to sell and hire out owned resources. The RRAs also provided invaluable access to the local idiom of agriculture—that is, indigenous knowledge that ranges from ethnobotanical terminology and local weights and measures to social values and expectations. The RRAs were initiated on the two most important agricultural islands, Santiago and Santo Antão. The information obtained in these activities then provided the basis for a further series of data-collection exercises designed to document the observed variation in a systematic and representative fashion. Major household surveys were organized and carried out first on Santiago then on Santo Antão and São Nicolau. These surveys were designed to simultaneously elicit information that is relatively easy to obtain in a formal interview and provide a useful empirical basis for further analyses. They focused on family demographic characteristics, access

Resource Management

27

to various types of agricultural resources, off-farm employment across households on the two islands, and other factors thought to constitute constraints to the improved well-being of households. The need for quantitative information about the distribution of resources across households required a statistically representative sample of rural households. On both Santiago and Santo Antão Islands, the research team designed random samples of households stratified according to major agroclimatic zones. In this two-stage sampling process, local communities (localidades on Santiago, zonas on Santo Antão) were selected from complete lists obtained from local representatives. Then local community leaders provided the names of all families with agricultural resources, specifying the sex of the head of the household and whether or not they had irrigated land. A sample was then drawn from each of these household lists. The household surveys, using questionnaires administered in the crioulo language, elicited three general types of information. First, an inventory of household members was compiled, including specific information on each individual’s age, sex, and level of education; time spent in agricultural activities; emigration experience; and finally, off-farm employment. The total number of individuals in the household provided an indication of the aggregate expenditure requirements necessary to meet the needs of all family members, and the number of working-age individuals provided an estimate of the supply of labor services available within the household, which could then be compared with actual time spent in both on- and off-farm activities. The second category of information was an inventory of agricultural resources to which the household had access in the year prior to the survey. This category included rainfed land, irrigated land, livestock, purchased inputs, and water. Land area was broken down into owned, rented, and sharecropped, providing an indication of household dependence on different institutions of land access. For those households with irrigated land, the average interval between irrigations and their source of water were also obtained. The third category concentrated on the utilization of family resources, including cropping patterns on rainfed and irrigated land, the use of family labor in agriculture, the hiring out of family labor, and sales of crops and livestock products. The surveys also inquired about the time needed to perform basic domestic chores such as gathering fuelwood and securing water for domestic use, both time-consuming activities. Labor allocation was discriminated by gender and age when relevant. The analysis of these survey data permitted the establishment of a basic typology of representative farms based on access to resources, cropping patterns, and gender of the household head. Since agricultural production

28

Resource Management

processes are complex and dynamic, involving biological, physical, and social variables, the creation of a representative typology is a daunting exercise that requires careful judgment as well as accurate data. The essential difficulty was to discover a level of representativeness that adequately reflected rural differentiation. At one level, production strategies vary among all households, and it would have been possible (in fact, easier) to establish a unique type for each household—clearly an unacceptable analytical solution. At the other extreme, collapsing all households into a single representative model would have precluded the evaluation of differences across households—equally unacceptable. Thus, the analysis was challenged to identify a defensible set of criteria with respective value ranges that discriminated an accurate typology, but also controlled for reasonable patterns of variation around any single type. As a pragmatic rule of thumb, the variations within any type were judged to be of secondary importance with respect to the overall analysis and to the evaluation of factors that policymakers can influence. The use of formal survey data to generate household typologies also permitted preliminary identification of relational patterns between different variables. For our sample of farmers in Cape Verde, analysis of the survey information suggested relationships between gender and cattle ownership, cropping choices and source of irrigation water, farm size and tenure mechanism, among others. Although the survey information can reveal patterns and correlations among variables, this information alone seldom provides an adequate causal explanation for these observed patterns. With the survey, the researcher gains a certain level of understanding of the patterned relationships among social facts, but an iterative process of verification (or contradiction) through subsequent discussions with farmers is necessary to more fully fathom the rural reality. With the results of the formal survey and the household typologies in hand, therefore, we identified a sample of farmers for periodic monitoring. The sample was purposive in the sense that representatives of each household type were included. The sampling procedure was also self-selective and participatory in the sense that we sought out households displaying a willingness to participate in the research (after lengthy explanations of project design and objectives). It is a basic research fact that some individuals have a better understanding of important processes and are better able to communicate with outsiders (including Cape Verdean researchers). For such deep and often introspective information, random selection of the interview sample is often unsuccessful; therefore, we sampled households that met a definitional set of characteristics and displayed an understanding of the project, a desire to participate, and the ability to articulate household strategies of production and well-being.

Resource Management

29

The technical assumption orienting this sampling procedure was that, within each type category, the variation of response would be insignificant for the kinds of information that were sought. For example, we assumed that among farmers with a given pattern of irrigated crops, their explanations of the relationship between cropping patterns and source of water would be acceptably consistent and uniform. Field interviews were undertaken with the farmers in this smaller sample to obtain the information necessary to construct activity budgets for rainfed, irrigated, and livestock activities. Because this information requires understanding of complex management processes, we employed open-ended interviewing techniques that permitted follow-up questions and the flexibility to follow particular lines of questioning tailored to the respondents’ particular areas of knowledge. This data-collection procedure maximized the amount of local indigenous knowledge integrated into the design of the crop budgets. There are two advantages associated with this approach (Norgaard 1984). First, crop growth is influenced by an extremely large number of variables in the farming environment. Weather and soil conditions and application of inputs by the farmer all influence yields. The extreme complexity of experimental designs to isolate the impacts of all these variables prohibits this avenue of research in countries such as Cape Verde, which have limited resources to devote to applied agronomic research. Nevertheless, farmers have vast experience in observing relationships between inputs and outputs. With this experience, they can make informed judgments about, for example, the effects of varying the interval between irrigations on crop yield, assuming all other conditions are normal. This indigenous knowledge base of farmers provides a low-cost alternative to extensive experimental programs. Admittedly, farmer responses cannot be accepted unquestioningly, but they provide the best guesses to use as working hypotheses until experimental data do become available. The second reason for using farmers’ estimates of technical relationships is that the goal of the analysis is to understand the decisionmaking process of farmers. This process draws on the knowledge base of the farmers themselves. (We assume that the knowledge base of farmers is accurate at least with respect to phenomena they directly observe.) Thus, understanding farmers’ choices of alternative strategies requires in-depth knowledge of their perceptions of trade-offs between alternatives. Experimental data about input/output relationships are not available to most farmers in Cape Verde (indeed, they are not available to researchers) and therefore are not directly relevant to understanding farmer resource-management strategies. Several assumptions are implicit in this technique for characterizing agricultural production technologies. First, there are assumed to be general

30

Resource Management

similarities in production techniques across households, so that information gleaned from one farmer can be generalized to other farmers in similar circumstances. Preliminary observations from the RRAs confirmed that agricultural production practices are quite homogeneous in Cape Verde. Crop budgets also incorporate assumptions of fixed-coefficient technologies and constant returns to scale. In order to model variations in production techniques for a given crop, such as differing numbers of weedings or applications of chemical fertilizers, separate budgets must be constructed that reflect these differences. For example, several separate budgets were constructed for each of the irrigated crops, each with a different irrigation interval and relevant variations in labor inputs and output yields. The budgets with more frequent intervals have higher yields. The assumption of constant returns to scale is appropriate for the technologies found in Cape Verde, since there are no significant fixed costs to be averaged over the quantity produced. The most important capital costs are the maintenance of erosion-control structures and the establishment of perennial crops (e.g., sugarcane, bananas). These costs vary directly with the scale of operations, so their unit costs are constant. Analysis of the crop budgets provided insights into farmers’ resourcemanagement decisions, but it also pointed to specific areas for further fieldwork and analysis. In particular, our attempts to identify representative production techniques and crop budgets highlighted the importance of access to irrigation water as a constraint for farmers with irrigated land. This discovery prompted us to examine more closely the factors that determine access to irrigation water. A detailed survey of an irrigation network was organized, with the objective of understanding in more detail the operation of the water-scheduling institutions, the impact of farmers’ access to water, and the relationship to their strategies for utilization of irrigated land. This survey provided detailed information about the schedule of water deliveries to all network members over the course of a year and the cropping patterns of each member. Hence comparisons could be made between water deliveries to individuals and their cropping strategies. Also, detailed discussions with the network manager and many of the members provided information on the operational minutiae of network management and the constraints faced by individuals to secure water at a particular time. The final analysis incorporates information from the survey, the budgets, and the ethnographic study of the irrigation networks into the construction of the household models. Estimates of total household income from all sources were made for the households in the survey sample. The sample provided information about the quantity of resources available to each household. Net incomes calculated from representative agricultural activities—rainfed land, irrigated land, and livestock—were then attributed

Resource Management

31

to households on the basis of their available resources. Nonagricultural income was estimated by multiplying the number of hours of off-farm work by the average wage. These results provide a basis of comparison for the welfare levels among different household types, as well as the potential impacts of changes in resource endowment for specific households. Moreover, the aggregate effects of such changes can be estimated by reference to the survey data and the relative frequency of a given representative type. While the multiyear research period provided us with the opportunity to complete the sequence of tasks described above, the fact that field information spans a number of years also presented a problem of comparability of results obtained from different years. Two important sources of annual variation were accounted for in our analyses. First, because of the great variation in growing conditions from year to year, the agricultural production and productivity information obtained from any single year cannot be used to extrapolate future conditions. For this reason, our estimates of yields used in the crop activity budgets are intended to represent normal or average growing conditions and are based on several years of historical data as well as the recommendations of local agronomists and farmers. By using these synthetic yield estimates, our calculations do not refer to the specific conditions of a particular year, but represent what can be expected in a normal year. The second major factor to account for in comparing results from different years is inflation—the general tendency of all prices in an economy to increase over time. Because of inflation, prices from different years are not directly comparable unless the effects of inflation are removed. We have used 1990 as the basis for all financial calculations. All prices are adjusted for inflation to be in real 1990 price levels. In 1990 the exchange rate of the Cape Verdean escudo (CVE) was 80 to the U.S. dollar (USD). In sum, the sequential rural research process consists of an ordered set of information-gathering activities in which each succeeding round of information collection is directed toward addressing issues or points of uncertainty identified in previous analyses. In the case of Cape Verde, fieldwork was initiated with RRAs in order to acquire a qualitative understanding of the problems facing rural households and to gather the background information necessary to design an appropriate questionnaire for a formal survey. The next step was to obtain information about the distribution of resources across rural households through a statistically representative survey. This information, in turn, was used to define a small number of representative farm types, and farms from each category were selected for more in-depth surveying. Detailed information necessary to define crop and livestock activity budgets was obtained through interviews with farmers in this smaller sample. In this interactive research sequence,

32

Resource Management

the socioeconomic analysis of the different representative farm types raised a number of more specific issues that came to direct further research. As presented in the following chapters, three particular issues formed central themes of this book: (1) the need to better understand the institutional mechanisms by which farmers receive access to irrigation water; (2) the importance of off-farm employment to household welfare and the need to examine the factors that affect or limit household access to off-farm work; and (3) the desirability of a more detailed analysis of the economic incentives for corn and bean production from the perspective of households given low reported returns to rainfed farming.

Summary

Our analysis of resource utilization by rural households is based on a methodology that posits that households attempt to maximize the income from their available resources. With this methodology, we are able to estimate the incomes that households are able generate with their existing resources given the available technologies to use these resources and the institutional environment that defines how households can acquire, dispose of, or hire out the services of their productive resources. In addition, the methodology permits analysis of the effects of alternative policies on household incomes; thus, incentives for changes in resource-management strategies can be inferred. Rural households form the unit of analysis in our examination of the agricultural resource-management strategies in Cape Verde. These households face a wide variety of constraints with regard to access to productive resources for their own use and access to markets or other outlets for their resources. Analysis must be directed toward understanding the variations in access to resources among households; the ways in which these variations lead to different agricultural strategies, and ultimately, the overall incomes of households. To obtain empirical information about rural households in Cape Verde, a sequential rural research approach was developed and applied on Santiago and Santo Antão Islands. This methodological approach is iterative and interactive, and it is particularly effective when little previous knowledge of rural society is available. The particular form that the analysis takes depends on specific characteristics of the agricultural sector being studied and the questions being addressed; however, three categories of information can be used as a checklist at the beginning of any empirical exercise of agricultural systems: distribution of resources among households (or relevant decisionmaking units); characterization of agricultural (and nonagricultural) technologies that households can employ with their available

Resource Management

33

resources; and understanding the opportunities, or lack thereof, provided by markets and other resource-transfer institutions to which households have access. As more knowledge is obtained about a particular system, the critical problems will fall into one or more of these areas. In the case of Cape Verde, the key problems have been identified as the absolute lack of access to productive resources for most households, inefficiencies in the way that irrigation water is provided to farmers, and restricted access to off-farm employment opportunities for the workers in rural households.

Notes 1. This framework is inspired by the seminal works of Binswanger and Rosenzweig (1986) and Binswanger and McIntyre (1987). 2. We recognize that households are made up of individuals and that household strategies for the utilization of available resources are, in fact, outcomes of some sort of an institutionally constrained bargaining process among household members. Income and the products generated from the household resources are similarly allocated to individual family members through complex institutional processes. The methodological approach followed in this book abstracts from the explicit analysis of allocations among individual family members to focus on household-level resource-management strategies. 3. Chayanov (1966) was one of the first writers to focus on the factors that affected resource-allocation decisions within peasant households. More recently, Becker (1965, 1981) applied neoclassical economic production and consumption theories to examine household decisions. Subsequently, many economists have continued to develop this theoretical approach (Pollack 1985; de Janvry, Fafchamps, and Sadoulet 1991). A number of authors have applied this theoretical framework in quantitative analyses of production and consumption decisions within agricultural households (Barnum and Squire 1979; Singh, Squire, and Strauss 1986). 4. Berry and Cline (1979) find a strong inverse relationship between land productivity and area farmed in a wide range of developing countries and argue that this reflects differences in perceived prices of factors across different sizes of farms. 5. It is not our intention to ignore the empirically observed scope of intrahousehold bargaining or conflict. The meaning of consensus here is that either the bargaining process has been successful and a compromise has been reached or that one or another household member has imposed his or her will over the other member(s). 6. An exception to this is the analysis of short-run and long-run price responses of large, medium, and small dairy producers by Adelaja (1991), who found that both short-run and long-run price elasticities were different across these three groups. Another study, by Barnum and Squire (1979), of producers in northwest Malaysia also incorporated farm characteristics into the model, in this case, farm size and tenancy status. 7. Budget analysis, however, does not provide quantitative information about the aggregate market-level response to such changes in prices. Econometric analysis permits more general assumptions about technologies; in particular, it allows for substitution and complementarity among inputs, as well as for economies of

34

Resource Management

scale. The specific input/output relationships do not have to be known independently. However, a large number of observations on prices and market quantities is needed in order to obtain statistic estimations. 8. To compare the economic efficiency effects of using resources in alternative activities within an economy, the opportunity costs of family-owned resources must be included in the profitability calculations (Monke and Pearson 1989). However, as will be argued in later chapters, the opportunity costs of most resources belonging to rural families in Cape Verde—predominantly rainfed land and household labor services—have very low opportunity costs. Thus, economic profitability calculations that incorporate the opportunity costs of family resources would not diverge significantly from the net revenue figures. 9. Antonovitz and Green (1990) compare alternative means of incorporating risk into econometric models. Based on empirical evidence from the U.S. fed beef industry, they conclude that no single formulation is preferred to the others, and thus expectations are heterogeneous rather than homogeneous. 10. Using panel data from households in the United States, Lawrance (1991) finds that households with lower permanent incomes and education levels have significantly higher rates of time discount than households with higher permanent incomes and more education. 11. The literature on RRA is rich and varied. In our research, we adopted the principles expressed in Hildebrand (1981) and van Willigen and Finan (1990).

3 The Natural Resource Endowment

Cape Verde forms the westernmost part of the semiarid Sahelian swath that traverses sub-Saharan West Africa. Its climate is characteristically Sahelian and subject to periodic droughts of greater or lesser severity. At the same time, Cape Verde is somewhat different from its West African neighbors in terms of its geomorphology. The archipelago is volcanic in origin, its present landform dating to the middle of the Tertiary period. The islands that emerged with more recent volcanic activity (i.e., Santiago, Fogo, Brava, Santo Antão, and São Nicolau) have a distinct topographical relief with peaks rising to over 1,000 meters in elevation. Steep hillsides and, in some places, nearly vertical walls shape valleys (known as ribeiras) that are narrow in their upper reaches but then widen toward the sea. The major ribeiras with their branches form large watershed systems that encompass many square kilometers and are recognized as distinct residence zones. The vertical topography of these islands increases their vulnerability to erosive processes caused by rapid surface runoff and wind. However, these are also the islands that accommodate agricultural activity. The remaining older islands (Maio, Sal, Boa Vista, and São Vicente) have been worn flat through the millennia and no longer have well-formed ribeiras. Consequently, their potential for agricultural development is limited to isolated spots of irrigated farming. Rainfed agriculture depends entirely on seasonal meteorological patterns that are highly erratic and unpredictable in terms of intensity and timing. Since Cape Verde has no stored surface water in the form of lakes, rivers, or reservoirs, irrigated farming is also constrained by the amount of annual precipitation that recharges groundwater aquifers. In contrast to other parts of arid West Africa, Cape Verde has no vast rangeland that permits extensive nomadic herding, virtually no wildlife to support hunting activities, and little in the way of native vegetation that offers quality supplements to local diets. Subsistence systems based on hunting and gathering have never existed in the archipelago. 35

36

The Natural Resource Endowment

The objective of this chapter is to describe the natural resource endowment and to establish its fragility and vulnerability to population pressure. Since our concern lies with Cape Verdean agriculture, we focus on those resources directly related to agricultural production, primarily water and soil.1 With regard to water, the relevant issues become its general availability for farming and the efficiency of its use in irrigated farming. In the case of soil, the sustainability of the resource is discussed in relation to the threat of wind and water erosion. Central to the theme of this book, we relate the variations in resource availability to the dynamics of resource use in agricultural activities. The chapter also presents the official policy concerns about the role of agriculture in the process of environmental degradation and describes the corrective measures that have been designed to conserve the resource base.

Rainfall and Climate

Climate is the most dominant yet unpredictable force in Cape Verdean agriculture. For the majority of the population, the annual rainy season is a dramatic, somewhat awesome uncertainty in the rhythm of everyday life. In any given year, the annual rainfall pattern determines a situation of plenty or of hardship for poor families and sets the emotional tenor for the year. As the rainy season unfolds during the summer months, families begin to perceive what nature has dealt for that year. In the event of inadequate rains, there is apprehensiveness and a sense of urgency in that some alternative subsistence strategy must be rapidly devised; when rains are bountiful, there is collective relief.2 The rains in the archipelago are extremely variable and highly unpredictable in terms of both total quantity of precipitation and spatial and temporal distribution. The advent of the rainy season in Cape Verde is determined by the dominant meteorological phenomenon that produces annual storms throughout all Sahelian West Africa. The intertropical convergence (ITC), an atmospheric turbulence consisting of a moisture-laden, unstable air mass, migrates northward from the south Atlantic bringing the monsoons that deposit rainfall throughout the region (Amaral 1964:50– 53). Specific meteorological variables, such as wind, temperature, and barometric pressure, determine the annual variations in the size, path, and velocity of this migration, which in turn determine the length and the intensity of the rainy season. Since the Cape Verdean archipelago lies somewhat north of the usual migration pattern, the islands are vulnerable to the dramatic results of slight shifts in the normal weather pattern. Interannual rainfall varies widely around the long-term average, and, in some years, the islands are totally bypassed. Freeman et al. (1978) report that annual

The Natural Resource Endowment

37

precipitation levels in the capital city of Praia fell below the long-term average (250 mm) in 62 of the last 102 years. Of greater concern, the country recorded 58 official drought years from 1719 to 1947, that is, an average of one every four years (Moran 1982:71). This year-to-year variation is not predictable, and farmers have few prior indicators—either traditional or scientific—to orient their adaptive responses. In recent times, the available pluviometric data reveal a further disturbing reality: annual mean precipitation levels have been declining since about 1952 (Freeman et al. 1978:43; INGRH 1993:8). And, since the late 1960s, the archipelago has been assailed by near-constant drought, spurring a reevaluation of what constitutes a “normal” year of precipitation. Another consequence of the ITC is the marked seasonality of rainfall and the extreme intra-annual variation. Between 60 and 80 percent of Cape Verde’s rainfall occurs between August and October, and precipitation between March and June is very rare. The actual number of days that register precipitation is low (from about 20 days in the more arid regions to about 50 days in the higher elevations). Nevertheless, individual storms can be extremely violent with rainfall exceeding 100 millimeters per day (Moran 1982:67). It is not uncommon for most of the annual rainfall to occur in one or two stormy episodes so intense that older farmers unable to read calendars often date life history events by reference to these particularly fierce periods. The concentration of rainfall within a very narrow time span creates several problems for farmers. The uneven distribution of moisture throughout the growing season can result in what is locally known as the seca verde (green drought). Early rains summon agricultural activity and promote strong initial vegetative growth, but the subsequent absence of moisture during critical periods in the plant cycle inhibits fruit or seed development. Thus, in some years, above-average annual precipitation levels still do not prevent poor agricultural outcomes. At the same time, torrential storms deposit more water than the ground can effectively absorb, which results in the accelerated, high-volume flooding that uproots young crops and unearths ungerminated seeds. Such downpours also exact a severe long-run toll in terms of soil loss. The vertical land-relief characteristic of the agricultural islands produces a pronounced orographic effect and wide spatial variations in rainfall during any given year. Whereas there are unpredictable interisland variations in rainfall due to varying storm tracks, the regional differences in annual precipitation on any island are significant and generally predictable. The prevailing winds in Cape Verde come from the north and northeast, and the rain-bearing clouds release their moisture upon contact with vertical landforms. Precipitation increases at a nearly constant rate with increases in elevation, while the coefficient of variation around mean annual rainfall decreases. Consequently, in most years, the orographic

38

The Natural Resource Endowment

effect dictates that the higher elevations facing the northeast will receive more moisture. Conversely, all the coastal areas, as well as the ribeiras that debouch toward the southwest, are characterized by greater levels of aridity. Table 3.1 provides a quantitative demonstration of this orographic effect on the four principal agricultural islands from 1970 to 1988. On Santiago and Fogo, the leeward islands, annual averages can vary from about 600 millimeters at higher altitudes to less than 200 millimeters at lower levels. The differences—while somewhat reduced—are also dramatic on the agricultural windward islands. Spatial variation has a corresponding impact on local vegetation and agricultural production potential, thus distinguishing three distinct agroclimatic zones: humid, subhumid, and semiarid. This standard classification is recognized in most agricultural planning and intervention strategies. Other climatological characteristics show less variation relative to rainfall. For example, temperatures average about 25 degrees Celsius during the year, with an average maximum of about 34 degrees and an average minimum of approximately 16 degrees (Freeman et al. 1978:53). Although prevailing winds blow from the north and northeast, a seasonal wind (the harmattan) off the African continent occurs between January and May and carries sight-obscuring dust particles concentrated enough to disrupt normal air travel schedules. Humidity and insolation are mostly uniform throughout the year. Considered in total context, the Sahelian climate of Cape Verde shapes and constrains the nature of agriculture. First of all, Cape Verdean farmers operate under conditions of extreme year-to-year risk. There seems to be no proven method—neither scientific knowledge nor local lore—for accurately anticipating a drought. Even in a wet year (on record), the rainy season may be dangerously confined to one or two violent storms that wreak substantial damage on the countryside but do not provide sufficient moisture throughout the agricultural growth cycle. Second, there is highly variable local distribution of rainfall, not only from island to island and region to region, but also from one side of a ribeira to another. Over the centuries, rural people have developed, and continue to revise, local management strategies for adapting to these conditions. Nonetheless, the downward trend in rainfall since the 1960s has placed further pressure on an already precarious livelihood by reducing the productivity of rainfed agriculture, limiting irrigated agriculture, and lowering the quality of livestock grazing areas.

Water Resources

The geohydrology of Cape Verde is as complex as its varied landforms and still awaits the systematic research necessary for a comprehensive evaluation

275 426 31 236 320 782 799 106 1,207 1,101 886 463 514 398 711 673 732 744 608 580

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1970–1988

Source: MDRP 1989.

Humid

158 285 0 216 194 233 434 11 435 389 328 333 129 300 308 204 421 466 332 273

Subhumid

Santiago

31 92 4 51 101 418 275 — 146 337 245 93 136 112 166 92 369 388 308 187

Semiarid 842 396 81 69 — — 247 139 699 480 681 239 — — — — 686 1,123 593 483

Humid 453 187 99 130 111 276 185 185 399 417 436 200 334 119 690 335 412 591 447 316

453 187 99 130 111 — — — 24 276 414 52 — 4 36 77 21 51 25 131

Subhumid Semiarid

Santo Antão

221 222 19 175 97 396 84 129 49 117 515 136 168 50 501 187 371 375 282 215

Humid 234 113 0 96 131 242 63 111 73 92 111 84 120 28 404 — 788 242 130 170

Semiarid

São Nicolau

748 552 209 371 73 748 95 — 325 1,105 1,105 425 541 454 1,260 672 1,031 1,288 954 664

Humid

268 307 61 291 — 103 42 — 162 289 604 321 321 224 379 228 847 781 328 327

Subhumid

Fogo

148 40 8 229 91 519 71 — 58 187 177 123 184 36 226 211 202 258 159 163

Semiarid

Annual Precipitation on the Four Major Agricultural Islands, 1970–1988 (by zone in millimeters of rainfall)

Year

Table 3.1

40

The Natural Resource Endowment

of available water resources. Various sources (Freeman et al. 1978:69; Amaral 1964; Republic of Cape Verde 1982:11–12) describe the general hydrologic morphology in terms of three volcanic formations that occur in varying combinations on the different islands: • A dense, essentially impervious formation called the Complex Ancien, consisting of consolidated and cemented pyroclasts, constitutes the bedrock supporting overlying formations • A Miocene-Pliocene Série Intermédiaire, consisting of massive basalts, is the major aquifer of the islands • The Série Supérieure, consisting of pyroclastic deposits as tufts, represented in gently sloping or plateau (achada) formations, is highly permeable with most water percolating into the Série Intermédiaire Groundwater collected in the Série Intermédiaire is stored in springs or seeps into the river alluvium and flows to the sea. The alluvia of the ribeira bottoms constitute the most accessible irrigation sources on the islands, especially where they are underlain by the impermeable bedrock. The morphological characteristics of each watershed, while derived from similar volcanic events, are unique, and, as a result, the recharge and retention capacity varies widely. Within each watershed—in fact, each ribeira—the quantities of available water differ, as do the principal mechanisms of mobilizing this groundwater for human use. For example, in one ribeira, the most accessible water is found in the alluvial bottom, while in another, the water emanates from springs (nascentes) or horizontal handfashioned tunnels (galerias). Since there exist neither projects nor structures that transfer groundwater across watersheds or ribeiras, irrigation activities are currently limited by the localized groundwater sources. The quantities of fossil groundwater—that is, the deposits of groundwater that were established before human agency became a factor—are not known or referred to in the available literature. Existing hydrological studies tend to focus on the complex relationship between available surface water (i.e., rainwater) and the quantity that can be used or stored in subterranean aquifers. The most detailed analyses to date have estimated surface water quantities (Fernandopulle 1977) and groundwater supplies (BURGEAP 1974), and we draw upon these studies and several summaries (Freeman et al. 1978; SCET-AGRI 1985a–d; INGRH 1993) to make our assumptions about total availability of this resource. Cape Verde has no permanent surface stock of water, and the lack of appropriate sites for a large-scale reservoir creates a near-total dependence on groundwater supplies for domestic and agricultural water use. Moreover, the morphological and climatic conditions of the islands do not foster efficient aquifer storage of rainwater. During a rainfall, the water can

The Natural Resource Endowment

41

be either absorbed by vegetation, returned through evaporation, trapped in different subterranean media, or lost through runoff. The authoritative Fernandopulle report (1977) estimates that approximately 33 percent of total rainfall in a normal year, or 330 million cubic meters, is lost to the sea as surface runoff. With more recent and slightly different information, the National Institute for Water Resources Management (Instituto Nacional de Gestão de Recursos Hídricos [INGRH]) has estimated that in a year that averages 227 millimeters of rainfall throughout the archipelago, about 20 percent, or 180 million cubic meters, is lost as runoff; more than half the moisture returns to the atmosphere through evapotranspiration; and the balance, only about 13 percent of total precipitation, or 118 million cubic meters, accrues to aquifer recharge (INGRH 1993:8–9). Variable factors such as intensity of rainfall (quantity per episode), slope, surface permeability, vegetative cover, and subsurface retention capacity significantly affect runoff loss and, consequently, the rate of aquifer recharge. The quantity and velocity of surface runoff not only determine the replenishment of critical aquifers, but can also affect the success of agricultural endeavors in both positive and negative ways. During a rain episode, runoff from the steep slopes of the ribeiras generates a gravitational flow toward the alluvial bottoms, where most irrigation is practiced. In the case of a storm event, flood levels are quickly reached as water accumulates from the upper reaches of the watersheds and moves toward the sea with increasing force and volume. The floodwaters may carry large debris loads, including animals and buildings, depending on the characteristics of the watershed and the intensity of the storm (Freeman et al. 1978:67). Within a period of several hours, the initial flood subsides to a streamlet that meanders across the alluvium bottom and may flow for several days. Farmers direct this stream by gravity toward fields cultivated along the edges of ribeira bottoms. This traditional farming activity is referred to as temporary irrigation (regadio temporário), and the water is primarily applied toward the production of short-cycle food crops such as beans and squash. In a year of well-distributed rains, temporary irrigation can produce significant quantities of food. When there is too little rainfall, the small amount of runoff precludes significant production. On the negative side, torrential rainfall can cause widespread destruction, damaging fields and irrigation wells in the ribeiras as well as nearby dwellings. The loss of human life also occurs with unsettling frequency. Over the long run, heavy runoff strips unprotected land of its scarce topsoil and permanently reduces productive capacity. The major challenges for comprehensive watershed management are to direct the maximum amount of runoff toward irrigation or aquifer recharge and to control the potential erosion and flood damage. Government efforts have been concentrated on the construction of physical structures and

42

The Natural Resource Endowment

vegetation to impede and collect runoff, maximize the infiltration of surface waters into the groundwater aquifers, and direct floodwaters away from fields and wells. These efforts include the widespread construction of rural structures, such as bunds, retaining walls, terraces, and catchment dams, along with an ambitious program of reforestation, both of which are discussed in more detail below. Permanent irrigation depends on access to groundwater over the course of a crop growing cycle. There have been several studies that attempt to estimate hydrological balances and the availability of water for irrigation (BURGEAP 1974; SCET-AGRI 1985a–b; Republic of Cape Verde 1982). Since different empirical databases are employed, the results of the analyses and the conclusions do not always concur. Nonetheless, there are important areas of agreement. First, the studies demonstrate that groundwater for irrigation varies widely on both an annual and a seasonal basis. Annual variation in groundwater availability is a function of the rainfall distribution and recharge characteristics, so that from one year to the next, irrigated agriculture is as dependent on the amount of rainfall in a given region or zone as is dryland farming. As a consequence, the trend of decline in annual precipitation since the 1970s has drawn down aquifer reserves. The INGRH (1993) estimates that overall availability of groundwater reserves has diminished by more than a third due to extended drought conditions. Equally variable is the seasonal access to groundwater. When rain occurs in concentrated storms, aquifers are recharged only during that period. Aquifer dynamics, however, are similar to surface runoff in that groundwater also flows downstream through the ribeira alluvia to the sea. Consequently, water for irrigation is most available during the period following the rainy season (October to May), when the cultivation of most annual irrigated crops takes place. During the summer months, the supply of groundwater diminishes, and its extraction becomes more costly. The traditional hand-dug wells may actually dry up. In response to this reduction in groundwater, farmers either leave irrigable land fallow or substantially increase the time interval between waterings. The area of irrigated farming is estimated to recede by up to half the maximum level attained in October (Republic of Cape Verde 1982:44). The second area of general consensus is twofold: more groundwater supplies could be utilized in agriculture, and the quantity of irrigated land could be expanded. Table 3.2 summarizes the current estimates of irrigated area on four agricultural islands, as well as the changes in area during the nine-year period between the last two censuses. Approximately 2,200 hectares of irrigated land are cultivated throughout the archipelago, with the vast majority concentrated on the two islands of Santiago and Santo Antão. The table also presents a set of estimates of existing groundwater

43

The Natural Resource Endowment

supplies and the potential for expanding irrigable land. Although these studies suffer from a lack of consistent and systematic data (and a possible tendency toward speculation), they do concur that the agricultural islands, under conditions of normal rainfall, have the potential to increase their irrigable lands considerably, particularly Santiago and Fogo.3 The discrepancies between the actual and the potential for groundwater utilization in agriculture reflect a combination of technical and management factors. Due to specific geohydrological formations and available technologies of extraction, large quantities of water trapped in aquifers, estimated to be as much as half the existing groundwater, are currently not exploited. With more systematic hydrological assessment of the islands, it may become technologically possible to develop the underutilized aquifers. Improvements in methods of groundwater extraction and more efficient local systems of distribution and use could also increase the availability of irrigation water, and greater investment in the necessary land structures (e.g., terraces, floodplain protection) could enhance aquifer recharge thus expanding irrigable area. The SCET-AGRI study (1985a–b) argues that in some ribeiras—for example, in the Ribeira Janela on Santo Antão—the amount of available water exceeds the quantity of irrigable land and suggests that interribeira water transfers might increase irrigated agriculture. Thus the range of interventions that might direct more groundwater to irrigation includes defined roles for both public and private participation. The challenge is to determine what combination of public and private initiatives would encourage the investment and technological change thought to be most appropriate (i.e., sustainable) and most effective.

Table 3.2

Groundwater Availabilities and Use on the Four Major Agricultural Islands Principal Agricultural Islands

Groundwater Availabilities and Use Irrigated area, 1979 (ha)a Irrigated area, 1988 (ha)a Average recoverable recharge (106 m3/year)b Present use (106 m3/year)b Development potential (106 m3/year)b Potential irrigated area (ha)c

Santiago

Santo Antão

701.0 677.3

840.1 1,357.1

0 60.4

20.0 24.8

29.0

27.0

4.5

21.0

13.9

11.5

0.8

1.5

15.1 1,568

14.6 1,004

3.3 169

18.3 320

aFrom MDRP 1989. bFrom Fernandopulle 1977, in Freeman et al. 1978. cFrom SCET-AGRI 1985a–d.

São Nicolau

Fogo

44

The Natural Resource Endowment

More recently, irrigated farming has encountered a new threat: the contamination of freshwater due to seawater intrusion at the mouths of the larger ribeiras. This problem is particularly acute on Santiago Island where the competition between urban and agricultural interests has become firmly established. The population of the capital city, Praia, has expanded rapidly, along with the subsequent demand of water for urban consumption. At the same time, the growth in urban building has created a need for the sand found at the mouth of the island ribeiras. The large-scale mining of sand facilitated seawater intrusion, and the highly productive farmers near the mouth have complained of rapid salinization of their lands. This emerging urban/agricultural conflict will likely constitute a major policy issue in resource management over the next decade.

Land Resources

The central theme of this book argues that a major disparity has arisen between the size of the rural population and the productive capacity of the land to support this population. The previous sections of this chapter have discussed the availability of moisture as a major factor that determines the productivity of the land; now we turn our attention to the quality of the land itself. Most public officials in Cape Verde point to degradation of the rural ecosystem as a major concern in long-term development planning (Trigo de Abreu and Soares 1983; Republic of Cape Verde 1992:111– 118). From a public perspective, the combination of high rural density, agricultural practices, climatic patterns, and topographic features creates a potentially disastrous formula for permanent damage to the critical land base. In this section, we address the issue of land degradation and the steps that are being taken to counteract the perceived danger. Of the 4,000 square kilometers in land area that comprise the Cape Verde Islands, approximately 10 percent is cultivated. The soils are derived from volcanic or igneous origins and, due to their recent formation, are coarse textured and shallow (Freeman et al. 1978:59; Moran 1982:70). Several interrelated factors influence the productive quality of land in Cape Verde, including soil fertility, density of stones, and steepness of slopes. Following Freeman et al., the most extensive agricultural soils are brown to reddish brown, sandy or clayey loams formed from rocks rich in calcium and found primarily along the rolling achadas of the Série Supérieure. They have a neutral pH and high organic matter content and comprise the principal base for dryland farming. Irrigable lands are mostly comprised of alluvial and coarse colluvial soils located in the ribeira streambeds and are partially replenished with finer sediment during normal rainfall years. On every island, agricultural activity is constrained by the

The Natural Resource Endowment

45

extreme density of stones—either basalts or pryoclasts. Even along the ribeira floors, surface runoff distributes material so coarse that it must either be removed or cultivated around. Retention of the relatively rich, but shallow, topsoil is critical to the sustainability of agriculture on the islands. The imbalance between land resources and population, however, has forced the expansion of agricultural activities onto the more marginal lands, particularly up the extremely steep slopes where wind and water erosion carry off finer sediment and organic matter. The permanent cultivation of slopes with a height rise of over 30 percent is considered a major contributor to the degradation of the land base. On islands with deeply cut, steep ribeiras, like Santo Antão and São Nicolau, farmers must terrace their lands to hold plants and soil in place, for both irrigated and rainfed agriculture. On Santiago, however, the largest agricultural island, terracing of rainfed lands is not a common practice. Table 3.3 presents several estimates of the distribution of agricultural soils among the four primary agricultural islands of Santiago, Santo Antão, São Nicolau, and Fogo. The quantity of cultivated land derived from the 1988 agricultural census is compared with estimated areas for different classes of land, according to the criteria of soil quality and steepness presented in an earlier report (SCET-AGRI 1985a–d). Although a comparison of these figures reveals some inconsistencies due to the lack of both definitional exactness and good primary data, certain unambiguous patterns can be discerned. Santiago Island, with its more extensive achadas, has a greater share of cultivated land relative to total area (about 20 percent), while Santo Antão and São Nicolau have the smallest shares (8 and 5 percent, respectively). Santiago also has the largest proportion of quality land with a slope of less than 15 percent (approximately one-quarter of all cultivated rainfed land), while Fogo, still an active, cone-shaped volcano, has the least (5 percent). Nearly one-half of the rainfed land on Fogo and Santo Antão is classified as poor quality, due either to fertility, stone density, or declivity. In the case of Santiago, about 20 percent of the cultivated rainfed lands are considered to be marginal. These estimates, even allowing for a certain margin of error, suggest two incontestable conclusions about Cape Verdean farmland. First, the quality of rainfed cropland, like the rains themselves, is highly variable. Numerous in-depth interviews with farmers have repeatedly affirmed that land quality changes dramatically from one ribeira to another, from one parcel to another, and even from one end of a given parcel to another. In fact, a recent survey carried out in the upper reaches of Ribeira Seca on Santiago (Finan 1993) reveals that most parcels are positioned vertically up the side of the ribeira, often incorporating a significantly steeper slope at the higher end and a much flatter slope at the bottom.4 The higher end of the parcel is generally less fertile than the bottom end due to the cumulative

46 Table 3.3

The Natural Resource Endowment Land Quality Characteristics for the Four Major Agricultural Islands Principal Agricultural Islands

Land Quality Characteristics Irrigated land, 1988 cultivated (ha)a Land with irrigation potential (ha)b Cultivated land, 1988 rainfed (ha)a Rainfed good-quality slope 15% and 25% (ha)b Rainfed poor-quality (ha)b variable slope Good rangeland (ha)b Total island area (ha)

Santiago

Santo Antão

São Nicolau

Fogo

677

1,357

69

12

1,365

881

60

6

20,155

6,401

1,806

5,730

5,100

518

432

312

7,755

1,789

404

2,078

4,135

877

103

1,127

3,901 20,665 99,100

3,437 5,910 77,900

488 3,170 38,800

2,765 7,230 47,600

aFrom MDRP 1988. bFrom SCET-AGRI 1985a–d.

effect of erosive processes, which is clear to an observer during the agricultural season. Table 3.3 also provides irrefutable evidence that Cape Verdean farmers cultivate a large proportion of rainfed cropland with poor productive potential and high susceptibility to erosion. These lands are frequently cited by public planners and technicians as being inappropriate for traditional agriculture under current technologies (SCET-AGRI 1985a–d; Trigo de Abreu and Soares 1983). For their part, farmers clearly recognize the variable potential of their parcels and adapt their crop mixes to this reality. The most flat and fertile cropland is found in the irrigated areas. On Santiago, irrigated lands are primarily located in the alluvial ribeira bottoms or in the proximities of artesian springs found at the higher elevations. In the upper reaches of any watershed, there is less irrigated land, but as the ribeiras widen toward the sea, irrigated activity becomes more concentrated. On Santo Antão, the majority of irrigated agriculture is practiced on contoured terraces ( pilares) laboriously constructed horizontally across the steep contours of the canyonlike ribeiras. The fields are irrigated by highly complex networks that draw water from galerias or from underground check dams (captações) set in the ribeira alluvia at higher elevations. In São Nicolau, virtually all irrigated land is concentrated within a major government project area where water is extracted from galerias in a specific ribeira.

The Natural Resource Endowment

47

It is extremely difficult to accurately measure the amount of irrigated land at any given time because of the extreme variation in water availability. According to the 1988 agricultural census, Santo Antão has 1,357 hectares of irrigated land, even though a previous study estimated that the potential area for irrigation was smaller by about one-third. Santiago has almost 700 hectares of irrigated land, an area significantly smaller than the estimates in the 1979 census. Most experts think that Santiago has the greatest potential to expand its irrigated agriculture. The importance of pasturelands has been overlooked in most references to Cape Verdean agriculture. Nonetheless, livestock provide a major contribution to the subsistence strategies of rural households. On each island, large areas considered inappropriate for agriculture are designated as pasture areas and generally are managed as commons. As Table 3.3 illustrates, rangelands comprise about 8–20 percent of the total area on the main agricultural islands. Many of the government’s reforestation programs utilize these lands and have introduced multiuse strategies of range and fuelwood management. In general, however, the major pasture areas have suffered rapid decline with the extended drought and the increase in the goat population. While the size of the domestic cattle herd has diminished by nearly one-third since 1970, the goat herd has more than doubled (MDRP 1992). With the growing rural population, recent tendencies point to an even more intensive use of the native vegetation on these lands.

Soil Erosion and Conservation Measures

Soil loss due to wind and water erosion is considered Cape Verde’s most critical long-term environmental and agricultural problem. Although comprehensive quantitative measurements are not yet available, isolated smallscale efforts to assess the impact of water erosion have indicated severe levels of soil loss (Lopes and Meyer 1993). The principal cause of soil erosion is the high-velocity flooding that occurs with torrential storms, moving not only precious topsoil, but also coarser material and even boulders, downstream. Even without the benefit of scientific measurement, the effects of erosion—gullying and flood debris—are readily apparent. And following a rainfall, at every point of debouchment into the sea, the normally blue-green seawater turns brownish yellow for as far as one’s vision can stretch. It is widely accepted among officials that agricultural practices associated with rainfed farming exacerbate erosive processes. Due to population pressure, rainfed lands are never fallowed, and during land preparation and weeding, farmers commonly remove all vegetation that might compete with crops for scarce moisture. This lack of vegetative cover, particularly on

48

The Natural Resource Endowment

slopes above 30 percent, increases the exposure of topsoil and the velocity of surface runoff after rains. Except where terracing is necessary (especially on Santo Antão and São Nicolau), farmers do not customarily construct water-retaining structures, nor do they use contour techniques in land preparation or planting. Rather, we have observed that farmers, following customary weeding techniques, create minicatchments around each cluster of plants in an attempt to harvest available runoff and reduce soil loss. While any cultivation of unprotected, steep-sloped land clearly exacerbates soil loss, the specific relationship between actual cultivation techniques (e.g., weeding) and erosion is not yet adequately researched. The government in Cape Verde has supported a long-term program of soil and resource conservation as a centerpiece of its agricultural policy. The major conservation projects have been directed toward integrated watershed management, with emphasis on antierosion and flood-control measures. In particular, the two predominant strategies of soil conservation have focused on the construction of rural structures that interrupt or dissipate surface runoff and the widespread reforestation of poor-quality soils, steep slopes, and semiarid rangelands. The funding for these projects has been obtained through international donor programs, and, for the most part, they have been implemented as public works projects that provide rural employment and reduce the tendency toward rural-urban migration. Conservation Structures

Much of the public investment in water and soil conservation has been channeled into the construction of rural structures designed to retard sediment flow and to increase infiltration into the soil. Erosion and floodcontrol projects have been instituted in all the major ribeiras of Santiago, as well as on the other islands. The type of rural structure varies with the specific characteristics of watershed (Lopes and Meyer 1993:53). Along the slopes of the ribeira, particularly at the upper reaches, the major structures are contour rock walls (arretes), contour furrows (banquetas), and microcatchments (caldeiras) built around plants as water-harvesting devices. In the gullies and ravines that feed into the ribeira floor, check dams (diques de correção torrencial) have been constructed to intercept and slow surface flow and to establish new agricultural land with the sediment that accumulates behind the dam. At points along the ribeira bottom, catchment check dams, captações, have been built to catch the subsurface flows through the alluvium. Some of these structures then channel water into irrigation systems, while others store the water for recharge into the aquifer. The catchment dams are a particularly important source of irrigation water for farmers on Santo Antão.

The Natural Resource Endowment

49

Private investment in resource conservation on rainfed lands has been largely insignificant. In fact, one of the major constraints to the success of the public projects is the reluctance of farmers to maintain the structures erected on their lands. Flooding from rainstorms often inflicts substantial damage on these rural structures, especially the arretes, which resemble low New England stone fences. Farmers often do not repair these walls, much to the chagrin of project managers and international donors. In the irrigated areas, the level of private investment is higher, particularly on Santo Antão, where terracing is a necessary condition for high-value irrigated farming. The construction of these structures is extremely laborious, since the terrace walls are often constructed on the slope, then soil from the ribeira alluvium is gathered to fill in the area behind the walls, thus forming the terrace structure. Terraces located closer to the ribeira bottom frequently suffer annual damage from the floods, which farmers must then repair as preparation for planting. Despite more than a decade of major investment in soil and water conservation, the impact on agricultural production and rural well-being is not yet adequately documented. On-farm testing has been used to demonstrate the advantages of rural structures in terms of increasing water retention, and qualitative interviewing among farmers has elicited their perception that protected slopes increase production in years of sufficient rainfall (Finan and Harshbarger 1996); however, the corpus of research on this critical topic remains surprisingly thin. Researchers and public officials acknowledge this information gap, and future impact studies should provide guidance for future investment policies. Reforestation

Low, erratic rainfall and periodic drought have always restricted arboreal cover in Cape Verde. As is common in Sahelian West Africa, the vegetation tends to be low and thorny, with the prevalence of several acacia species in the arid and semiarid zones of Santiago and a wider range of trees in the better soils (Ficus, Ziziphus, Mangifera, Calotropis, and Tamarindus). In the humid zones of Fogo and Santo Antão, the moister climate has favored the development of larger species, including pine (Pinus) and eucalyptus. Since trees anchor and protect the soils from violent downpours, especially at higher elevations of the watershed, the reduction of forested area also affects agriculture and the productive capacity of the soil. At the same time, an estimated 80 percent of the Cape Verdean population derives its cooking energy from fuelwood or charcoal, and some tree species also provide valuable forage for domestic animals. Thus, trees have traditionally played an important role in this delicate human ecology.

50

The Natural Resource Endowment

From independence in 1975 through 1991, nearly 60,000 hectares of private and public land (almost 15 percent of total land surface) were reforested (MDRP 1992), mostly with international support channeled through public work projects. In recent years (1986–1990), the annual rate of reforestation has averaged about 5,700 hectares of covered area, with Santiago receiving about 80 percent of this amount. Santo Antão (about 440 hectares annually) and Fogo (540 hectares) also have significant reforestation programs. An estimated 85 percent of the seedlings used in reforestation are comprised of small-stature, drought-resistant species of Acacia, Prosopis, and Parkinsonia, which are planted in caldeiras dug along the contours of slopes. In the designated public forests, native pasture varieties have been sown in an attempt to recapture the carrying capacity of these forested rangelands. Some soil-conservation projects have strongly promoted the substitution of pigeon pea (Cajanus cajan) for the traditional corn-bean intercrop on the more severe slopes. This cultigen is perennial, produces both an edible bean and forage material for small ruminants, and, in its later years, is utilized as a source of fuelwood. As a perennial, pigeon pea provides more permanent vegetative cover and is recommended from a sustainability perspective. Fully cognizant of the potential advantages of pigeon pea, rural populations have proven reluctant to substitute the corn-bean mix to any large extent. From household research in localized regions of Santiago, we learn that farmers have argued that pigeon pea is appropriate for specific kinds of microniches that are determined on the basis of such characteristics as soil quality and prevailing wind direction (Finan 1990). They also point to the fact that pigeon pea produces less forage than the traditional mix, which is critical for the livestock component of the domestic farming system.

The Declining Natural Resource Base

This inventory of Cape Verde’s natural resource endowment does not conjure up visions of an agricultural paradise. The scarcity of land and water relative to the existing population would argue that far too many people are currently involved in farming activities, and the unpredictability of rainfall makes all agriculture, but most immediately, rainfed agriculture, a precarious and high-risk enterprise. Even assuming the most optimistic estimate of potential surface runoff and groundwater use, irrigated land will not be sufficient for all those who currently farm or are employed in agriculture. The existing data on resource availability, while sometimes inconsistent and nonsystematic, also support concerns that the quality of the resource

The Natural Resource Endowment

51

base has declined since the mid-1970s. Rainfall demonstrates a downward trend, and erosion rates appear high relative to those of other countries. Pastures have deteriorated, and total irrigated area has decreased. Perhaps more telling are the subjective impressions of the local populations. Farmers generally concur that in recent decades, there is clear evidence of less production, sparser vegetation, fewer animals, and less water. Even with the human tendency to romanticize a dimly remembered past, the local assessment of a decline in the natural resource base is consistent and persuasive. It is also sadly ironic that the lack of rainfall bears the blame for the current threatened situation. Although the paucity of rains does create the immediate crisis, more bounteous rainy seasons might but hasten the arrival of the long-term degradation crisis that population pressure seems destined to precipitate. Despite the rather dire reality of a poor and declining resource endowment, there are few alternatives to agriculture at the present time. Furthermore, the disparity between rural population and the resource base has created the danger of an irrevocable degradation in the quality of the existing resources, precisely because of this lack of nonfarm opportunities. Thus, the dominant challenge over the short run is to optimize the use of land and water in a sustainable way until such time that a longer-run solution can be found and implemented. It is indeed the terms of that challenge that we describe in more detail in the following chapters.

Notes 1. By no means do we intend to minimalize the importance of biodiversity on the islands. The number of plant and mammalian species was of sufficient quantity to attract Darwin to Santiago during the voyage of the Beagle, and the preservation of unique species of bird and plants is a concern of INIDA and Cape Verdean leaders. In this chapter, however, we focus on species that contribute to food security of the population. 2. The word for rain in crioulo, tchuba, has near-mystical qualities and is a theme that is reiterated in many cultural forms. The mostly Catholic Cape Verdeans often refer knowingly to the bountiful rainfall that coincided with the visit of John Paul II in January 1987, the first ever by a pontiff. Rain has virtually never occurred during this month. 3. Fogo Island presents unique characteristics. The island is, in essence, an active volcano and has the highest elevation in the country (over 1,000 m in altitude). There are no distinct ribeiras as there are on the other agricultural islands, and the geohydrology of the island is such that most rainfall escapes to the sea through underground fissures. German donors have invested in the creation of a small irrigation perimeter (Monte Genebra) that uses water pumped from deep wells. The success of this project is extremely limited by the high costs of this irrigation water. 4. In local crioulo, farmers refer to the steepest lands as those that require a monda sekedu, or an upright or stoopless weeding posture.

4 Characteristics of Cape Verdean Rural Society

Although not immune to the urbanization trends that continue to alter the demography of West Africa, Cape Verdean society remains essentially rural in terms of population distribution and economic source of livelihood. Most recent estimates show that approximately 68 percent of Cape Verdean inhabitants live in rural zones and that nearly 60 percent of the economically active population is employed in agricultural activities (World Bank 1993a).1 To say, however, that Cape Verdean society is rural does not mean that it is homogeneous, isolated, or impervious to change. On the contrary, the interplay of natural and historical forces in Cape Verde has worked to forge a complex, dynamic society externally integrated into world economic currents and internally structured by systematic social and economic differentiation among islands and within each individual island. This differentiation displays itself in several forms: wealth is not evenly distributed across the rural population; landowners coexist with landless farmers; some households are managed jointly, many are headed by single women; some families enjoy a broad array of economic choices, while others find themselves hopelessly constrained in their options. These measurable differences in household structure and economic strategy reflect the influence of the local institutions that Cape Verdean society has devised and continues to test, to modify, and to bequeath to subsequent generations. As our description of the agrarian political ecology suggests, these institutions channel resources and define the opportunity horizons available to different types of households, thus leading to social differentiation. Wolf (1982) has cogently presented the theoretical argument that human populations are not impassive absorbers of external pressures but, rather, active agents that “shape” society from the raw materials provided by the physical and sociopolitical environment. To comprehend fully the current structure of society, it is necessary to understand the particular historical context of human agency, where over time the definitive institutions have 53

54

Cape Verdean Rural Society

been wrought. In the case of Cape Verde, the prime working clays of history were climatic diversity (i.e., drought), slavery, and Portuguese colonialism. The historical interaction of the local population with these fundamental realities engendered the set of institutions that have come to characterize and differentiate contemporary rural society. In this chapter, we examine the structure of rural society, primarily drawing from our empirical household research on the islands of Santiago and Santo Antão.2 We begin with a review of the historical context of drought, colonial administration, and the unfortunate legacy of slavery, then trace the subsequent development of the socioeconomic and cultural institutions that determine household access to the basic productive resources and their allocation toward alternative subsistence strategies.

The Settlement of the Archipelago

By all credible accounts, the Cape Verde Islands were uninhabited when first discovered by the Portuguese explorer António de Noli in 1460. The Crown, having set its design on the potential resources of the Guinea coast, sought to exploit the strategic locational advantages of the islands and sent settlers to establish a colony on Santiago, the largest island. These pioneers established the first European city in Africa—complete with social classes and cathedral—at Ribeira Grande (now Cidade Velha). Lisbon recognized early the insurmountable environmental and logistical constraints to pursuing the agricultural development model that had proved successful on the other Atlantic islands, principally Madeira (Amaral 1964:172–174). Without a feasible agricultural economy (based on exports) to attract potential settlers, the Crown laid the economic foundation of settlement with trade and shipping incentives. As Carreira (1982, 1983), Cape Verde’s foremost historian, has reported, the European inhabitants in Cape Verde (both Portuguese and Genovese) were granted trading rights over sections of the Guinea coast, and Ribeira Grande became a market entrepôt that linked the flow of goods between the Guinea coast of Africa and Europe (including its colonies). Settlers moved up and down the continent from present-day Senegal to Sierra Leone establishing complicated trading networks in gold, ivory, iron, cloth, spices, and shells. The commercial item of highest value, however, was the African slave, first acquired by capture, then through active local slave markets based in the river systems. Resident merchants in Cape Verde created at Ribeira Grande a holding point and market clearinghouse, where slaves were trained and Christianized as they awaited transshipment, primarily to the New World. In the early days of the colony, while monopolistic conditions prevailed, the transaction costs of trading were low, and the value of slaves high, the Cape Verdean economy and society flourished (Carreira 1983).

Cape Verdean Rural Society

55

The direct and indirect consequences of the slave trade at Ribeira Grande, more than any other factor, influenced the early formation of Cape Verdean society. There were few European women available to the men on the island, and miscegenation between whites and slave women became an accepted and common pattern. The offspring of these relationships (crioulos), along with free blacks (negros forros), contributed the middle categories to a color-determined social-class structure that maintained privileged whites at the top and slaves at the bottom.3 The crioulos were often acknowledged as the sons and daughters of the colonists and were raised with filial status. Along with outcast whites, the crioulo sons constituted the pool of Cape Verdean traders (called lançados or tangamaos) who frequented the coastal rivers representing island merchants (Carreira 1983:53). The Portuguese-based creole language, crioulo, emerged as a lingua franca used as the idiom of coastal trade. 4 Crioulo is the language that is now spoken in Cape Verde (as well as Guinea-Bissau) and has assumed the symbolic importance of distinguishing Cape Verdeans—regardless of their skin color—from outsiders. Just as the crioulo (the person) represented a genetic admixture of great phenotypic diversity and as crioulo (the language) represented an amalgamation of Portuguese and African linguistic elements, the culture of Cape Verde also borrowed freely and eclectically from coastal and European cultures. From marriage patterns to dance form and culinary preference, Cape Verdean society has demonstrated the remarkable human ability to creatively meld seemingly disparate meaning systems into a logical, coherent and functionally independent hybrid. Slavery soon generated political consequences as well. The early settlement of Cape Verde was a significant chapter in the European colonization of West Africa insofar as the thème du jour was the extraction of wealth rather than the promotion of sustainable society. The Portuguese Crown initially supported trade and distributed commercial concessions as incentives to induce settlement, but soon discovered that this strategy did not control nor easily expropriate the flow of wealth through the islands. As long as local merchants were free to trade, they could easily avoid administrative fees and taxes, for supervision was far too costly and generally ineffectual. Even before the close of the fifteenth century, the Crown began to restrict the privileges formerly granted to the islands in an attempt to capture more directly the rents from trafficking. In 1468, exclusive trading rights over Sierra Leone were leased to the influential Portuguese adventurer Fernão Gomes, and the island colonists (along with their lançados) were prohibited, under severe penalty, from practicing commerce along this part of the coast. The expansion of this leasing strategy during the first decades of the 1600s increasingly restricted the trading area available to the island merchants. The Crown also published a “protected” list of commodities (called mercadorias defesas) that were made the exclusive trading

56

Cape Verdean Rural Society

domain of government merchant houses, ostensibly for reasons of “national security.” These items, such as cloth, shells, and iron, were the main barter commodities in the slave trade along the coast, and their prohibition in effect stanched the flow of slaves through Cape Verdean hands. The cultural impacts of these restrictions on economic activity resulted in the gradual emergence of what might be called a national consciousness and cultural identity built around crioulo language and culture. Carreira points out that the geographic proximity gave the lançados a significant advantage over government agents in maintaining regular commercial ties along the coastal rivers, and, in essence, monarchial prohibitions diverted trade into clandestine contraband activities. This quiet defiance against Crown policy generated a sense of cultural unity that usurped class distinctions and that would reaffirm itself even more clearly during the difficult times in colonial history. High profits and the expansion of the New World slave market brought the English, Spanish, and French into direct competition, and often conflict, with the Portuguese. As the foreigners increased their presence on the Guinean trade routes, Cape Verde’s role as commercial entrepôt was diminished. At the same time, enemy corsairs increasingly attacked shipping along the coast during the sixteenth and seventeenth centuries, as Portugal saw its military and economic might reduced. In 1712, Ribeira Grande was plundered by the French, after which the city experienced irreversible decline, and island government offices were moved to today’s capital, Praia. During the eighteenth century, the Crown imposed an even more restrictive policy on the local economy with the institution of trading monopolies that effectively retained in Lisbon total administrative control over all economic and commercial activity on the islands. The unpopular monopoly, the Companhia do Grão Pará e Maranhão, set up by the Marquês de Pombal, managed the slave trade, set prices for imports and exports, and exacted taxes from local merchants (Carreira 1982:24). Entering the 1800s, Cape Verdean society found its access to trading revenue vastly reduced, its domestic economy in severe recession from lack of currency in circulation, and its local production in decline. In sum, the archipelago had assumed all the characteristics of an economically marginalized, peripheral society in the world economic system.

Agricultural Beginnings

The commercial demise that resulted from both international and Portuguese pressures forced the colony toward a greater reliance on agriculture. During the heyday of trade, agriculture supported commercial activities by

Cape Verdean Rural Society

57

providing both food for the urban population and tradable commodities that could be used in slave trafficking. Cotton and sugarcane were the primary cash crops traded both in Europe and along the African coast (Amaral 1964:174; Ribeiro 1954). A widespread local cottage industry transformed cotton into panos, the orchid-tinged cloths that became a standard barter commodity exchanged for slaves and other utilitarian goods. 5 The original settlers also introduced various breeds of livestock, including cattle, horses, and mules; but goats were the animals that adapted most easily and flourished not only on Santiago, but also on the drier islands (e.g., Boa Vista and Maio). Carreira (1983), citing various sources, refers frequently to the coastal trade in live animals, hides, tallow, skins, and salted meat. The latter was also sold to ships that docked in the Cape Verdean harbor of Mindelo to restock their supplies and refuel from the Britishcontrolled coal station. The organization of agricultural production was influenced by the different settlement processes on the leeward and windward islands. On Santiago, the Crown imposed a traditional feudal model under which the island was divided into two capitanias, or extensive land grants, issued by royal decree to the original discoverer of the islands, António de Noli, and to the administrator of Madeira, Diogo Gomes (Carreira 1983:27–29). Along with rights over the land, the grantees enjoyed wide privilege with regard to the administration of their holdings. Fogo Island was also divided into largeholdings and distributed among the loyal and favored of the Lisbon court. On both islands, the original land distribution evolved into a system of feudal land institutions called morgadios and capelas under which usufruct rights were granted in perpetuum to noble families (Carreira 1982:5). The lands themselves were indivisible and indissoluble, while the administrative rights and responsibilities resided with the firstborn male, the morgado, in each generation (i.e., an inheritance system based on primogeniture). Actual day-to-day management was often relegated to a representative of the morgado, the feitor, a person who has come to sport an unflattering reputation in rural cultural history. The labor to cultivate these lands was provided by slaves. This system of agricultural production proved through time to be economically nonsustainable, and it began to evolve toward more contemporary patterns. Traditional labor supplies were curtailed first when local access to the slave trade was reduced and then as pressures toward abolition mounted. Into the nineteenth century, the morgados had largely become an idle class of absentee landowners who left administrative duties to the often-oppressive feitores (Carreira 1982:27). And many slaves, seeking to escape the penury of their situation, fled into the remote regions of Santiago or to other islands, where they established small-scale units of subsistence production.6 Along with a disappearing and more costly labor pool,

58

Cape Verdean Rural Society

the cyclical onslaught of drought regularly hobbled the rural economy and created severe financial crises for the landowners. Under the combined burdens of drought and labor scarcity, the morgadios and capelas became nonfunctional, shell-like institutions, and the large properties began to disintegrate into small management units—with parcels sold off to more affluent families or farmed under rental and sharecropping arrangements by landless families, mostly former slaves or crioulos. These historical processes generated a highly concentrated ownership structure consisting of a small number of large landholdings, but a highly disperse management structure comprised of numerous small sharecroppers and renters, as well as a smaller contingent of independent subsistence farmers. From the beginning, the windward agricultural islands (Santo Antão, São Nicolau) were less affected by transatlantic trade, and the Crown exerted little influence over the occupation of agricultural lands. Santo Antão, for example, was settled about the middle of the sixteenth century by pioneers from Santiago who sought land to expand agricultural activities. The island was extremely isolated (and in many ways is so today) with little infrastructure and no natural harbor. Without easy access to slaves and commerce, the early landholdings were generally small, and the morgadio system did not establish itself in the ribeiras of this island. The lack of an established labor pool and the difficult physical access of the island restricted the operational farming units. Large landowners could not attract sufficient labor supplies (and the costs of supervision were very high) to manage all their arable lands and thus turned to sharecropping as a mechanism to exploit their agricultural holdings. This resulting pattern of relatively small, independent farmers and sharecroppers has prevailed to the present. The first systems of production were developed through the local experimentation with different crop and livestock alternatives imported from both Europe and the New World. The early settlers found the traditional European cereal crops (e.g., wheat, barley) unsuitable to local conditions. During the 1500s, maize and beans were introduced via Brazil and became the established staple foods of the population. These two food crops, with the cash crop cotton, comprised the primary rainfed system. In the moister ribeira bottoms, cane, yams, manioc, and sweet potato were produced along with bananas, figs, and other fruits. At the higher elevations on Fogo and Santo Antão, farmers experimented with wine grapes and coffee. From Europe, the settlers brought livestock, especially goats, which quickly adapted to the local rangelands and the semiarid climate. The historical record of agriculture’s contribution to the Cape Verdean economy has escaped the detailed scrutiny of scholars, whose inquiries concentrate more on the problems related to commerce and trade. It appears that agriculture exercised a more dynamic role in the national economy

Cape Verdean Rural Society

59

during the early years of settlement. We receive passing mention in Carreira’s work that agricultural products were commonly exported to foreign markets (e.g., meat to Lisbon, maize to Madeira) and used as barter (cotton panos along the Guinea coast). During the first century of settlement, agricultural trade appears to have been limited not so much by constraints on production, but by misguided trade policies and the associated absence of appropriate incentives from the Crown. By the beginning of the nineteenth century, however, exports had begun to decline significantly. Cotton production had disappeared almost completely; and the dependence on food imports had begun to increase. The causes of decay in the Cape Verdean rural economy await a definitive explication. Carreira (1982:39) suggests that the sometimes violent transformation from a slave-based production system to one based on the exploitation of smallholder tenants and sharecroppers precipitated the emergence of the undercapitalized subsistence system that predominates today. It is also apparent that agriculture in Cape Verde faced a relentless enemy in periodic drought that eventually disheartened agricultural efforts. Missing among these social and climatic factors is Boserup’s argument that population increases effectively reduced the fallow period of lands, resulting in a general decline in the productive capacity of the resource base.7 Although we are unable to find references to specific agricultural practices, Carreira (1982:28) does cite nineteenth-century accounts that imply the existence of “uncultivated” (either virgin or fallowed) lands. Since presently virtually all available cropland is cultivated annually without a fallow reprieve, it seems plausible that agricultural intensification followed a Boserupian path—modified by the ravages of famine—into the current century. As the population expanded, more marginal lands (of which there are many [see Chapter 3]) were brought under cultivation, and the periods of fallow were reduced. Without widespread access to more intensive technologies, such as manured fields, it is likely that the sustainable “carrying capacity” of the land was reached during the nineteenth century, after which time the declines in land productivity became pronounced.

The Impacts of Drought and Migration on Rural Society

Recurrent drought not only influenced the development of agriculture in Cape Verde, but it also contributed to the very formation of rural society. Whereas the nagging persistence of current drought patterns inspires serious concern about global climate change, drought itself is not a recent phenomenon. The first deadly occurrences were recorded in the early 1500s (Carreira 1984), and periodic “crises” have assailed human and animal

60

Cape Verdean Rural Society

populations into modern times. Table 4.1 summarizes this chronology of major droughts with their respective tolls in terms of human life. Even as late as 1946–1947, the consequences of drought proved catastrophic in human terms, as an estimated 20 percent of the population perished from drought-induced starvation. During the most recent crisis, which has extended with little relief from the late 1960s to the present, national and international relief efforts have avoided the widespread mortality of previous crises. The drought phenomenon in Cape Verde has generated long-run impacts that extend far beyond the immediate tragic losses. Most prominently, drought and its specter of famine have made Cape Verde a nation of emigrants. Since the nineteenth century, emigration has constituted the principal economic strategy of households that cannot meet their subsistence needs. Our rural surveys in 1984–1985 revealed that in 54 percent of Santiago rural households, in 44 percent of the Santo Antão sample, and in 38 percent of the São Nicolau sample, at least one member had either currently or previously emigrated (see Table 4.2, p. 62). Although emigration has been a continuous process, the rates of rural exodus have always increased dramatically during periods of drought. Emigration in Cape Verde began with the arrival of U.S. whaling ships in the first decade of the nineteenth century (Carreira 1982:43). Cape Verdeans recruited as sailors on ships that stopped for provisioning sought employment along the eastern coast of the United States (primarily Rhode Island and Massachusetts) as the ships returned from their voyages. Emigration was accelerated by Portuguese colonial policy of forced migration in the twentieth century, which channeled Cape Verdean labor to colonies suffering labor shortages (e.g., São Tomé and Príncipe) and then to Portugal itself as the Portuguese labor force sought higher-paying jobs in France and Germany. Emigration has followed regular routes by island, although the correspondence is not perfect (Richard 1983). The inhabitants of Brava and Fogo have tended to emigrate to the United States, particularly to the eastern seaboard, while the rural poor from Santiago have provided the majority of workers bound for menial jobs in Portugal. The inhabitants of Santo Antão have created a strong base in Holland, where they tend to work on ships or in shipbuilding industries. In São Nicolau (and, to a lesser extent, Santo Antão), an Italian religious order has set up a program of controlled emigration for single Cape Verdean women to Italy, where they work under contract as domestic servants in wealthy homes. With the exception of female emigration to Italy, most emigrants have been young and male (Richard 1983: 116). As a result, rural society has traditionally had a predominance of females in the prime productive-age categories, although recent studies suggest an upward age shift in that demographic pattern (Finan 1993). Thus, from a socioeconomic perspective,

61

Cape Verdean Rural Society

Table 4.1

Historical Record of Droughts, 1719–1996

Year Begun

Affected Islands

1719 1747 1754 1764 1773 1790 1804 1810 1825 1830 1845

Santiago All All Boa Vista, São Nicolau All Barlavento, Brava All All Santo Antão All Santiago, São Nicolau, Santo Antão Sotavento Sal, Boa Vista Barlavento, Fogo Maio, Brava, Santiago All Santiago, Santo Antão All All All Maio, Brava All Sotavento All All Santiago, Fogo Fogo São Nicolau, Santiago, Fogo All All

1850 1853 1854 1858 1863 1875 1883 1885 1889 1892 1896 1897 1900 1920 1931 1935 1940 1946 1965–1996

Duration (Years)

Mortality (% of Population)

Aid Provided

1 3 2 1 3 1 2 1 1 3

Unknown Unknown High Unknown 44 Unknown Unknown High Unknown 42

Very little Unknown Unknown Corn CVE 30,000 Unknown Unknown Very little Very little None

1 1 1 3 3 3 1 1 1 1 1 1 1 3 2 1 1

Unknown Unknown Unknown High Unknown 40 Unknown Unknown Unknown Unknown Unknown Unknown Unknown 15 16 Unknown Unknown

Unknown Unknown Unknown CVE 77,000 Unknown Unknown Unknown CVE 60,000 CVE 105,000 CVE 134,000 Unknown CVE 50,000 Unknown Very little CVE 4,250,000 Unknown Unknown

1 2 —

15 18 Very little

High CVE 50,000,000 Massive foreign assistance/emigrant remittances

Source: Adapted from Moran 1982:71.

rural society has for a long time had a large number of female-headed households as well as households dependent on outside remittances for their sustenance. The major forces of Cape Verdean history—slavery, colonialism, and drought—laid the foundation for the development of contemporary rural society. Slavery and miscegenation produced a variegated society wherein, with the important exceptions of the islands of Fogo and Brava (Meintel 1984), color distinctions do not generally correlate with class distinctions. The slave trade also produced a unique language (crioulo) and incorporated many distinctive elements of different African cultures into the social

62

Cape Verdean Rural Society

Table 4.2

Demographic Characteristics of Rural Population on Selected Islands

Demographic Characteristics

Santiago, 1984 Santiago, 1993 Santo Antão, 1985 São Nicolau, 1985 (n = 239) (n = 177) (n = 176) (n = 100)

Household size (no. of members) Sex ratio (percentage of females in the sample) Dependency ratio (ratio of dependent household members [65 years] to independent household members) % population % female-managed households Average age of household head Household emigration (% of households with a current or past emigrant member)

5.7

6.1

6.8

5.5

56

54

47

48

.97 36.6

1.49 44.3

1.00 35.7

1.22 31.6

38

32

9

24

53

54

54

56

54

35

44

38

Sources: Finan 1993; Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986; Langworthy, Varela, and Finan 1986.

fabric. The colonial land-grant policies created a small, privileged landowner class and a large class of landless operators, and the vestiges of these distributive patterns are still apparent today. Portuguese colonialism also brought Catholicism and the widespread power of the local padre, still very much in evidence. On the other hand, the cumulative effects of periodic drought and famine have worked to erode class differences to the extent that vast inequities in accumulated wealth are no longer as apparent in rural Cape Verde as in other countries with a similar colonial past, such as Brazil or even Mozambique. Drought also encouraged widespread emigration, which, being mostly male, has resulted in an imbalance in the rural sex ratio. Many households are managed by women who are either single mothers or whose husbands have emigrated (Finan and Henderson 1988). This unique interaction of historical processes has given Cape Verdean rural society a particular blend of European and African traditions that remains its trademark.

Demographic Characteristics of Rural Society

A central theme in this study of Cape Verdean agrarian ecology examines the impact of heavy population pressure on fragile environmental resources.

Cape Verdean Rural Society

63

Since independence in 1975, the birthrates have remained high, while mortality rates have dropped with expanded access to health care. Furthermore, emigration has diminished significantly since the mid-1980s because of decreasing access to employment opportunities in the international economy. A 1993 study of Santiago households demonstrated a significant reduction in the percentage of households with emigration experience (see Table 4.2). Moreover, to prevent massive population flight to the major cities of Praia and Mindelo, both of which are inadequately prepared to receive such influxes, the government has implemented a program of rural employment tied to investment in rural infrastructure. As a result, despite a rapid rate of urbanization (particularly in Praia, where growth is due as much to interisland relocations as to rural-urban flows), the rural population continues to expand. With projections for further population growth through the end of this century, it is likely that rural pressure on the land will remain strong over the near future. The structure of the Cape Verdean rural population is a broad-based pyramid that reflects the preponderance of the young (under 15 years old) and the differential impacts of gender-specific emigration, particularly within the most productive age group (20–50 years). Comparison of survey results in Table 4.2 reveals that more than one-third of the population was under 15 by 1985; and the more recent data suggest the movement toward an even younger population structure. A related descriptor of this population, the dependency ratio, has important economic implications. This measure expresses the proportion of household residents under 15 years of age or older than 65 years relative to those in the productive years of life. The higher numbers reflect a greater “responsibility” placed on these household producers. Again, the data from the surveys reveal that in 1985 every “able-bodied” individual provided for about one (other) person, with a somewhat higher ratio for Santo Antão and São Nicolau. The more recent data suggest an even greater burden on the productive-age categories, an expected result from a population becoming increasingly younger. The percentage of females in the population is directly related to the emigration phenomenon. On Santiago Island, the proportionately greater number of male emigrants (to Portugal) has produced a ratio of about 115 females for every 100 males. On the islands of Santo Antão and São Nicolau, where emigration programs have been directed at young women, the data reveal about 90 females for every 100 males. These impacts of out-migration are even clearer when sex ratios are differentiated by the age categories within which emigration occurs most frequently (Finan 1993; Richard 1983). A plausible interpretation of this demographic structure would suggest that under the likely assumption of fewer emigration opportunities, the rural population is poised for continued future growth with a tendency toward a more equal sex ratio. In other words, current trends portend greater, rather than less, population pressure on the natural

64

Cape Verdean Rural Society

resource base, unless the natural growth rate of the population is curbed through education and family-planning efforts.

Rural Settlement Patterns

The residential patterns in Cape Verde vary from island to island, reflecting both historical population pressure and the spatial distribution of productive resources. In general, rural settlement is concentrated within the ribeiras and close to the alluvial bottom because of the easier access to water for both agricultural and domestic use. There are, however, many population clusters not found along the ribeira, but located in the vicinity of small artesian springs (nascentes) that erupt to the surface at higher and more isolated elevations. Especially on Santiago, where settlement patterns significantly reflect the historical impact of slaves fleeing their masters, large segments of the population live in very remote and barely accessible regions of the island. Popular wisdom asserts that these distant, almost perched settlements are occupied by either the descendants of those runaway slaves who escaped to the mountains and began farming or, less dramatically, those families who value the constant refreshing breezes and the panoramic view (even though the securing of water is an all-day task). Along whatever trail of reason one wanders, it is clear that rural Cape Verdeans have not adopted the villagelike settlements characteristic of much of West Africa. Rather, kin-based clusters comprised of several households (joined through traditional inheritance practices) are distributed sporadically along the hillsides of the ribeiras or along the highest crests, safe from the severe floods associated with heavy rainfall. Access to urban centers still varies widely throughout the archipelago. On Santiago Island, the majority of the population has access to Praia or to two of the three smaller urban cities: Pedra Badejo, Tarrafal, and Assomada. These small towns provide higher-level markets and administrative services but little employment. On Santo Antão and São Nicolau, and, to a lesser extent, on Fogo, access to urban areas is difficult because of the mountainous terrain and lack of physical infrastructure. There are significant numbers of isolated settlements that can be reached only by footpath or by sea. The rural population is also located within an administrative structure organized hierarchically into different levels of economic and political function. Each island is divided into a small number of concelhos, the centers of which are found in larger urban concentrations. The concelhos are divided into freguesias, and an urbanlike nucleus (vila) is usually associated with each freguesia. The smallest unit in the administrative structure

Cape Verdean Rural Society

65

is the zona, within which is commonly found the localidade, a loose concentration of households around a few retail stores and perhaps a primary school. On some islands, inhabitants recognize smaller residential clusterings called lugares, which may be comprised of one or two extended kin groups. To the observer, then, rural households are dispersed widely over the landscape in settlements that range from 20 or 30 homes clustered around a common water source to a thin distribution of individual farms across the hillcrests. Although this settlement pattern lends itself to an initial impression of spatial isolation and a loose, individualistic social organization, such is not the case. In reality, physical distance in Cape Verde does not directly correlate with social distance. Each household identifies with a clearly bounded community (both lugar and localidade), which are focal points of social and economic integration. Even the households in the most isolated zonas have friends or relatives in other areas, including major urban centers, and these networks are maintained with periodic contact.8

The Structure of Rural Households

The Cape Verdean rural household is nuclear in core, but close intrafamily cooperation often results in three-generational co-residence patterns. In an ideally normative sense, households are initiated through formal marriage or cohabitation and expand with the arrival of children. Adult children leave the parental household only when they marry and form their own households. This basic pattern is often altered to meet the subsistence exigencies of individual households. On Santiago, for example, it is common for a grandmother, in need of more labor, to raise one of the children of her daughter, and it is equally common for a family to bring in an elderly parent or other collateral relatives. Overall, households tend to be large: family size varies from about six per household in Santiago to seven in Santo Antão, as shown in Table 4.2. Children acquire farming skills by working with their parents; and, as in any difficult environment, detailed knowledge of the land and its production potential is a fundamental survival tool. For this reason, sons (and single mothers) find it advantageous to establish households where basic indigenous knowledge is maximized and a pool of labor is available; consequently, marriage residence patterns tend to be patrilocal, that is, the bride joins her husband at a site near his father’s home. In large part, this practice accounts for the settlement pattern of small, kin-based clusters that are empirically documented throughout the islands. In this predominantly Catholic country, ceremonial marriage is a strong and prestigious social institution. Since, however, the ritual involves

66

Cape Verdean Rural Society

significant economic costs and implies wider family responsibilities, many poorer households are not formally sanctioned and consecrated by the Church. Household formation through either formal ritual or casual agreement normally occurs at an age ranging from late teens to early twenties, until which time children reside in their natal households under parental supervision. Thus, economic and social independence, especially for women, is generally achieved only with the constitution of a new household. One of the predominant characteristics of the rural household structure is the high incidence of female-headed families.9 Table 4.2 illustrates that more than one-third of Santiago households and about one-quarter of São Nicolau households have female heads. This category of household is primarily comprised of single mothers without husbands, but it also includes wives of emigrant husbands and widows. Invariably, the single-mother and widowed households tend to be smaller and significantly poorer. Their asset base—measured in land and livestock—is decidedly lower than the national average for rural households. It is common to find older femaleheaded households with two or more generations of resident single mothers and their children. Universal primary education in rural Cape Verde was ushered in with independence. Under colonial rule, access to education was restricted to those families with financial wherewithal, which limited the economic alternatives for rural youth. Today, most young people—even in rural areas—finish the four-year primary program (which recently has been expanded to six years) and are able to read and write, while most preindependence adults are illiterate. According to the 1988 agricultural census, over 60 percent of the total population aged 11 and older can read and write, but only about 40 percent of household heads are literate. Until the age of 20, equal proportions of males and females are literate (MDRP 1990:76–77); in the older age categories, males have had relatively greater access to education. At independence, there were only two high schools in the country; now, virtually all vilas have access to a high school, and Cape Verde is rapidly developing models of vocational and technical education and will soon introduce its first system of higher education. This investment in human capital among the younger segments of the rural population will certainly engender important positive changes as the impacts percolate throughout society.

Agricultural Land Institutions

As part of their livelihood strategies, households retain and allocate resources: land, labor, capital, livestock, and so on. In this section, we describe the public and private institutions that enable access to the critical

Cape Verdean Rural Society

67

rural resource, agricultural land (see Chapter 2). Not all the islands on Cape Verde have agricultural land. Santiago, Fogo, and Brava, on the leeward side of the archipelago, and Santo Antão and São Nicolau, on the windward side, are the islands with the necessary climatic conditions that allow farming. These five islands share the same critical topographical characteristic—namely, vertical relief of sufficient altitude to trigger the release of rain from moisture-bearing clouds. On the two major agricultural islands, Santiago and Santo Antão, significant amounts of rainwater are trapped in the silt alluvia of ribeira bottoms or in basaltic aquifers, thus making irrigated agriculture possible. On the other islands, limited areas of irrigated land have been developed through government investment projects. The following discussion focuses primarily on Santiago and Santo Antão, which together account for 72 percent of total rainfed cropland and 94 percent of the total irrigated area in the country. Rainfed and irrigated agriculture are very distinct pursuits, employing different technologies, crop mixes, and institutional arrangements. For example, rainfed agriculture is perceived by farmers as a subsistence activity meant to supply household food needs, while much of the irrigated output is aimed at urban markets. Nearly all rural households (and many urban families) cultivate some rainfed land to produce the important staple crops, corn and several varieties of beans. The agricultural cycle for rainfed agriculture begins in July and August, prior to the expected start of the rainy season. If rains are adequate, the harvesting of corn and beans is initiated in December and is generally completed by February, after which time rainfed activities cease until the following year. In the case of irrigated land, only about one-third of the farms in Cape Verde have access to water for irrigation. Where water supplies permit, irrigated farming can be a year-round activity, although there is often a slack period during the rainy season when farmers turn their attention to the rainfed fields. In the context of heavy rural population pressure upon the land, institutional mediation of access assumes a predominant and determinant role in rural society. Cape Verde has no additional land to clear and no agricultural frontiers to expand. Society, instead, must devise the formal and traditional mechanisms by which the existing land resources are distributed and redistributed among households and between generations. Rural people can acquire access to productive land through several different institutional channels. First, a farm family can obtain ownership rights to farmland (conta própria) through inheritance or, to a lesser degree, market exchange. By law, Cape Verdean society follows a principle of partible inheritance, in which the patrimony of one generation is divided in equal parts among both sons and daughters of the subsequent generation. By custom, however, male children are favored in the distribution of agricultural resources. Since independence, illegitimacy of birth has been

68

Cape Verdean Rural Society

legally disclaimed, and all heirs both within and outside formal marriage arrangements have rights to the family estate, including agricultural land. As in other semiarid regions, the land market in Cape Verde is very weak, especially for rainfed lands that present low returns and high risks. Most market transactions involve the more predictable and profitable irrigated lands. The second institutional form of access to land is sharecropping (parceria), which may include tenant residency for the poorest families. The agreement between landowner and sharecropper is informal and tends to be long-lasting, even involving intergenerational transfer of usufruct rights to the children of the sharecropper; however, no formal or legal institution guarantees security of tenure. Although not a common occurrence, the feitor (or landowner) can demand return of the fields at the end of a harvest without penalty. The terms of sharecropping vary under individual agreements, but the general custom is to divide final output in halves, with all production costs borne by the sharecropper.10 The third institutional avenue of acquiring farmland is the rental contract (arrendamento), by which landowner and renter enter into formal written agreements protected by law. Under this contract, the rental fee is a fixed annual payment based on the economic value of the parcel and not on a given year’s production. The agreements are often multiyear and thus guarantee stability of tenure. With the 1982 land-reform decrees of the Partido Africano da Independência de Cabo Verde (PAICV) government, the nature of land ownership and land use became a highly controversial issue bathed in political rhetoric and ideological symbolism. Sharecropping was prohibited and condemned as a colonial, capitalist tool of exploitation. Some lands were expropriated or purchased by the state and allocated to small farmers (through a formal mechanism called posse útil). More important, the law insisted that sharecropping relations be transformed into rental contracts managed by a government land-reform office. In many parts of rural Cape Verde, the land-reform measures incited widespread discontent and, in some instances, violence. Since the political demise of PAICV and the introduction of a multiparty political system, much of the land-reform legislation has been revoked or simply ignored, although some elements have endured. Rainfed agriculture is a quintessential subsistence activity, and the rural community recognizes the informal and unspoken, but nonetheless moral, right of individual households to the resources necessary to pursue a subsistence livelihood. The traditional land institutions feed this sense of morality and function to assure that all households gain at least minimal access to productive land. In response to specific stages of the domestic cycle, most households combine different access mechanisms to build a land base. The process of household reproduction described by Segurado

Cape Verdean Rural Society

69

(1983) begins when a father provides his son at the time of marriage a homestead (assento de casa) upon which to construct a house. With this homesite comes the right to a designated parcel for subsistence production, which the son first cultivates as part of his own household and eventually inherits. When this initial quantity of land becomes insufficient as the new family increases in size, the son can complement his existing stock of land through sharecropping or rental arrangements. For households in later stages of the domestic cycle or with a land-to-labor imbalance, sharecropping represents a flexible means of divestment. It is not uncommon for sharecropping agreements to exist among relatives or neighbors within a given localidade. Thus, the institutions available to households are highly adaptive, low-cost mechanisms of redistributing land to meet the fluctuating needs of the community of households. The same institutions operate very differently with respect to irrigated lands. In contrast to the situation with rainfed land, there is no implicit “moral imperative” within the community stating that all households should have access to an irrigated plot. Since the production potential of irrigated land is high, the demand is more intense. Larger landowners use sharecropping (and rental contracts) to overcome the labor constraints that limit the direct cultivation of irrigated lands under current technologies. While sharecropping still prevails in many irrigated regions, the landreform legislation forced a significant number of landowners to switch from sharecropping arrangements to formal rental contracts, as specified in the reform law, and this transformation has tended to resist political changes. Table 4.3 summarizes the distribution of rainfed cropland on farms on Santiago and Santo Antão. Farmers measure land in liters, and 1 liter of land represents the area that can be seeded with 1 liter of corn seed, approximately 0.1 hectares. The average size of a rainfed farm on Santiago is about 13 liters (1.3 ha) and about 17 liters (1.7 ha) on Santo Antão. As a subsistence activity, rainfed agriculture is affected by the number of consumers in the household and by existing technologies. As a normal pattern, as family size increases throughout the developmental course of the domestic cycle, the household head seeks access to additional amounts of cropland. And large landowners prefer not to manage their lands directly, but to rent and sharecrop out smaller plots to others. As a result, while the ownership of land may be highly concentrated, the distribution of land by operational unit is not unequal by most conventional standards. On Santiago, about one-half the farms are between 5 and 15 liters in size, with that group retaining more than 40 percent of the land. The largest farm, by contrast, is about 40 liters of cultivated area. The distributional pattern is somewhat more uneven on Santo Antão, where the categories of very small farms and (relatively) very large farms are more numerous. While it

70

Cape Verdean Rural Society

is true that on both islands the larger farms account for the majority of the land, the differences between the largest and the smallest are not great. The major source of wealth in Cape Verdean agriculture is land with water. Access to irrigated cropland provides a much securer base of food production and opens opportunities for expanded participation in the market. Irrigated land in Cape Verde, as we have seen, is extremely scarce. The proportion of Santiago farmers with access to water for irrigation is less than 30 percent, while on Santo Antão—with its greater abundance of water—about 58 percent of the farmers practice irrigated farming. The typical size of an irrigated farm is greater on Santo Antão (8 liters) than on Santiago, where the average size is only 1 liter. Table 4.4 summarizes the results of our agricultural surveys of these islands and effectively illustrates

Table 4.3

Distributions of Rainfed Cropland on Santiago and Santo Antão Land Size Categories in Liters (1 liter = 0.1 hectare)

Island Santiago % of farms % of area Average size (l) Santo Antão % of farms % of area Average size (l)

0–5

6–10

11–15

16–20

>20

Total

16.0 4.6 3.7

32.1 21.1 8.4

21.8 21.2 12.6

18.1 26.1 18.1

12.0 27.0 28.7

100 100 12.7

22.8 3.7 2.8

28.7 13.5 8.2

12.0 9.2 13.4

12.0 11.9 17.3

24.5 61.7 43.8

100 100 17.1

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

Table 4.4

Distributions of Irrigated Cropland on Santiago and Santo Antão Land Size Categories in Liters (1 liter = 0.1 hectare)

Island Santiago % of farms % of area Average size (l) Santo Antão % of farms % of area Average size (l)

0–0.5

0.51–1

1.1–4

4.1–7

7.1–10

58.7 19.2 0.3

21.1 18.7 0.9

15.8 32.4 2.0

2.9 13.6 4.7

1.5 16.1 10.0

3.3 0.1 0.3

7.6 0.8 0.9

43.5 14.4 2.7

18.5 12.5 5.4

8.6 9.0 8.4

>10

Total

0 0 —

100 100 1.0

18.5 63.2 27.5

100 100 8.1

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

Cape Verdean Rural Society

71

that patterns of access to irrigated land are more skewed than those found in rainfed land. On Santiago, 80 percent of the farms cultivate less land than the average land area, and they account for only 38 percent of total available irrigated land; on Santo Antão, 73 percent of the farms are less than the average size and account for 28 percent of the total. The overall scarcity of irrigated cropland is a principal issue in Cape Verdean agriculture. Tables 4.5 and 4.6 show the relative importance of ownership versus sharecropping and rental contracts as institutions of access to land for both rainfed and irrigated farmers on Santiago and Santo Antão. On Santiago, over 40 percent of the farmers own none of their rainfed cropland, and 61 percent of the irrigated farmers do not own the land they cultivate. That twice as many rainfed farmers on Santo Antão as on Santiago own all their cropland reflects the differences in colonial policy between the leeward and the windward islands. Preliminary data from the 1988 agricultural census further suggest a trend toward reduction in the class of landless farmers either through out-migration or through land-transfer mechanisms (MDRP 1990a).

Agricultural Labor Institutions

All agriculture in Cape Verde is labor-intensive. The principal tool is the long-handled hoe, and animal traction—where found—is almost exclusively used for transportation. Substantial variation marks the patterns of labor use in rainfed and irrigated agriculture. In rainfed farming, the demand for labor is highly seasonal and intense for certain tasks such as land preparation, seeding, and weeding. For irrigated agriculture, labor demand is more dispersed throughout the year, with occasional increases in intensity, for example, for land preparation and during the sugarcane harvest. Because the timely completion of agricultural tasks is critical in semiarid and arid agriculture, the recruitment of labor is a major concern around which several labor institutions have emerged. As was the case with land, these alternative institutions are distributed differently between rainfed and irrigated agriculture (Finan 1988). Cape Verdean farmers generally prefer to employ family labor to the extent that it is available. As Binswanger and Rosenzweig (1986:519– 521) point out, family labor remains the first order of preference because it incurs lower supervisionary costs and is more reliable. Even those larger landowners with plots either rented or sharecropped to others often attempt to create familylike relationships in order to reduce the burden of supervision. In most cases, however, the supply of family labor does not adequately

72 Table 4.5

Cape Verdean Rural Society Distributions of Rainfed Cropland by Landownership on Santiago and Santo Antão Landownership Shares (%)

Island

None Owned 0–25

Santiago % of farms % of area Average size (l) Santo Antão % of farms % of area Average size (l)

25–50

50–75

75–99

All Owned

42.8 39.9 11.8

10.7 12.9 15.2

7.5 8.7 14.7

8.1 8.2 12.8

5.6 8.0 18.2

25.3 22.3 11.2

26.3 19.0 12.6

6.0 7.9 23.3

6.6 9.0 23.9

6.0 8.1 23.5

1.8 4.5 43.3

53.3 51.5 16.8

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

Table 4.6

Distributions of Irrigated Cropland by Landownership on Santiago and Santo Antão Landownership Shares (%)

Island Santiago % of farms % of area Average size (l) Santo Antão % of farms % of area Average size (l)

None Owned 0–25

25–50

50–75

75–99

All Owned

60.7 73.8 1.2

1.1 2.6 2.5

2.5 2.3 0.9

4.9 3.5 0.7

0.0 0.0 0.0

31.0 17.8 0.6

32.6 24.9 6.2

7.6 7.3 7.8

5.5 3.0 4.4

3.3 1.5 3.7

2.2 10.2 38.0

48.9 53.0 8.8

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

meet the peak labor demands in the agricultural cycle. Rainfed farmers have developed two widespread sharing institutions to solve the labor problem: djunta-mon and djuda. Irrigated farmers, in contrast, more frequently rely on the wage labor market to adjust their supplies. Most African rural societies practice some form of exchange labor in their subsistence agriculture systems. In Cape Verde, two traditional, or private, exchange institutions mediate access to the labor pool. Djuntamon (literally, a “joining of hands”) is the more common and involves short-term, reciprocal exchange within a network comprised of friends, relatives, and neighbors. As its distinguishing characteristic, djunta-mon adheres to a rigid, albeit informal, accounting standard. A day’s labor

Cape Verdean Rural Society

73

requires the prompt repayment of a day’s labor of equal value: women reciprocate for women, and men for men. It is, in practice, common for a household head to send her or his resident children to perform a djuntamon task in return for equal quality labor from the exchange partner. Since djunta-mon usually involves groups of workers, there is a definite social air surrounding the execution of these tasks, as young men and women have the opportunity to meet and to impress one another with their work skills.11 Djunta-mon normally provides the necessary complement for family labor during peaks in demand, particularly for land preparation and weeding, which require timely attention. It is an especially effective solution to short-term labor bottlenecks in an agriculture characterized by strong microecological variations within a limited geographical area. For example, different parcels of land may germinate at slightly different times, and rate of weed growth is not consistent throughout the local region. Djunta-mon allows the fuller utilization of existing labor supplies by matching the immediate availability with the immediate need. Such an institution functions successfully only when residents possess both a detailed knowledge of the local environment and a strong ethic of community. Both these criteria are met in the case of rural Cape Verde. Djuda (literally, “help,” in crioulo) constitutes a transgenerational exchange mechanism also solidly grounded in a sense of community sharing. In any rural zone, a certain number of households have neither the family labor nor the access to other forms of labor to meet the requirements of subsistence agriculture. In reality, these households tend to be those of the elderly, the widowed, the single mother, the invalid, and the infirm. With djuda as the enabling institution, local networks of relatives and neighbors provide the necessary labor to tend the rainfed fields, without any rigorous accounting or any expectation of immediate reciprocity. This form of labor sharing occurs for both agricultural and nonagricultural tasks (such as house construction and repair). Although no repayment is expected, djuda—from a community perspective—represents an exchange across generations, since the elderly are the most frequent recipients. Thus, those who provide the “free labor” may at some later point in their lifetimes need the benefits of djuda, and they can reasonably expect that the younger generation will provide that assistance. Wage labor is found throughout rural Cape Verde and is concentrated in irrigated agriculture, although some wealthier farmers or those with large irrigated holdings may also hire laborers for their rainfed plots. The labor demand for sugarcane occurs at harvest time, when the cane is processed into grogue, the distilled spirit. For vegetables and bananas, the demand is more constant throughout the growing season. Day labor is the most common form of wage work in agriculture. On the supply side, agricultural

74

Cape Verdean Rural Society

labor is provided by poor, landless farmers who may sharecrop or rent a small parcel of rainfed land or, more frequently, by young adults who welcome the opportunity to earn scarce cash. Agricultural wages on Santiago averaged CVE 250 per day for men and CVE 150 for women during the late 1980s. In most cases, the employer is expected to provide a generous meal replete with grogue, the worth of which is estimated at about CVE 30 per person. On Santo Antão, the wage rates averaged CVE 170 for men and about CVE 120 for women. Except in those cases where a wealthier farmer employs a small number of permanent workers, agricultural labor does not provide an adequate subsistence wage throughout the year, primarily because of the seasonality of employment. Thus, income from agricultural wage work usually must be supplemented by some other source. The primary nonagricultural employment alternative is wage labor on public works projects. Labor-intensive work fronts—now called FAIMOs (frentes da alta intensidade de mão-de-obra)—predate independence as a colonial strategy to alleviate the horrors of drought; however, under the Republic, rural work projects (financed by revenues derived from international assistance) have become the centerpiece of governmental investment policy in rural Cape Verde. International food aid—primarily in the form of corn and rice, vegetable oil, and milk powder—is monetized through its sale to the state marketing agent EMPA (Empresa Nacional de Abastecimento), which resells it to private retailers throughout the archipelago. The generated monies are deposited in a national investment account and revert back to the rural areas in the form of salaries for the FAIMO fronts. The fronts provide the labor for public investment projects, especially in soil and water conservation, reforestation, and rural infrastructural development. Under current policy, each family has access to public employment on these projects. Wages are established according to classified skill levels (e.g., mason, laborer); and, at the time of this research, daily wages for FAIMO activities averaged CVE 150, the amount that we apply to the analysis in subsequent chapters. Public work employment does not include the traditional meal and grogue. A major attraction of the public works program to policymakers is that rural people remain in rural areas and do not flood urban areas where the ability to provide services has already been severely strained. At the same time, the government acknowledges that public employment is a temporary emergency strategy that can last only as long as the international community supports it. At the beginning of the rainfed agricultural cycle, public work may be suspended so that farmers can return to their fields. From the rural perspective, however, project work has assumed an air of permanence in the sense that rural households depend on this employment and plan their household decisions around it. A further complication arises

Cape Verdean Rural Society

75

from the fact that most rural workers prefer public employment to agricultural labor. Despite the inferior wages, public work is not seasonal and provides a steady income. Furthermore, the work is not rigidly supervised, the workday is only six hours, and farmers can find time to incorporate their personal agricultural activities into the public work schedule. For this reason, larger landowners have complained of the competition from public works projects for available rural labor supplies. Some minor off-farm employment opportunities are offered in the private sector. Women are the primary marketers (rabidantes) of agricultural products throughout Cape Verde. As described elsewhere in the literature on West Africa (e.g., Clark 1988), these market activities are eminently small-volume and labor-intensive; however, the income contribution to the household is often significant. For those who live close to the coastline, artisanal fishing and small-scale fish marketing provide a small income source. In other areas, women load beach sand (for construction) onto trucks. The rural surveys did not quantify off-farm employment on Santiago and Santo Antão, but observations suggest that these diverse sources of informal minientrepreneurial activity are significantly underreported. Many household tasks—both agricultural and nonagricultural—are differentiated by gender, although the traditional patterns vary from island to island. On Santiago, for example, women perform all agricultural tasks except those that require heavy effort, and they use the long-handled hoe in such activities as weeding and planting. On Santo Antão, women do not use the hoe, but transport most agricultural inputs and outputs between access roads and fields. In the extremely vertical topography of Santo Antão, this task is as formidable as it is important. On both islands, women also bear the responsibility for domestic household tasks such as cleaning, child care, food processing, and food preparation. Young children also have their standard chores that contribute to the maintenance of the household, despite the fact that the government actively supports primary education in rural areas and most adolescents attend the local schools until the age of 14. School-age boys usually have the responsibility for specific tasks, such as care for livestock and protection of newly planted fields from pests, while young girls typically care for younger siblings, carry food and supplies to the fields, and assist in cooking and cleaning. The younger children are also charged with the daily chore of obtaining water for the household, often an arduous and time-consuming task. After primary school, the young are integrated more fully into the agricultural labor cycle of the household and carry the load of adults. Thus, on all the islands, rural inhabitants tend to combine some agricultural and nonagricultural labor strategies to enhance their survival under the precarious and unpredictable circumstances that mark rural Cape

76

Cape Verdean Rural Society

Verde. Labor opportunities other than emigration are limited, and only in recent years has the rural worker begun to look toward urban areas for sources of possible employment. The public work front is the major institutional employer in rural areas, and it has become enrooted in the survival strategies of rural households. All agricultural and household tasks are well defined, as are the relative roles and contributions of the different members of the household.

Sharing Institutions

As Amaral (1964:197) has so cogently described it, survival in rural Cape Verde has always challenged the ingenuity and innovativeness of the local population. The land and labor institutions discussed here represent sharing strategies that function to distribute the means of production. The social value of sharing is fully enmeshed in other aspects of community life, such as the exchange of seeds and consumption items (see Geertz 1956 for an ethnographic parallel from Indonesia). The cultural significance of sharing is captured in the crioulo term ferta (literally, “an offering”). In an environment characterized by the relative proximity of diverse microclimates, production outcomes vary substantially across time and space. Farmers retain their own corn and bean seed from one harvest season to the next planting season, and households in a given zone or year may lack seed at the appropriate time. It is common for neighbors to “offer” seed to the more unfortunate farmers. Strict reciprocity does not regulate this behavior, although there is the understanding that the benefited recipients will help other disadvantaged colleagues as the situation commands. A similar sharing custom exists for sweet potato vines (for planting). On the consumption side, ferta can occur on a daily basis. The food items commonly shared include the traditional rural fare, cachupa, and the principal protein sources: pig meat, goat cheese, or cow milk and butter. In the case of households with close sharing ties, a plate of cachupa from one family pot may be offered to a neighboring family, while the latter shares a plate from its own pot. Every family raises pigs, goats, and chickens. When a pig is slaughtered, parts of the meat are often distributed to neighbors, friends, and relatives. When eggs and goat milk are available, similar sharing occurs. The networks are informal and flexible, but nonetheless concrete and enduring. All households seem to participate in one or more exchange groups. These sharing practices can be interpreted from several perspectives. On the one hand, many of the shared consumption items are perishable and appear in quantities beyond the immediate household needs. Thus, from a community level, ferta represents an efficient way of avoiding spoilage

Cape Verdean Rural Society

77

and waste of perishable products. On the other hand, the thickness of these exchange networks represents a community investment in the survival of individual households, and the sharing of scarcities helps maintain a minimum survival level for the largest number of people. This perspective is supported by the fact that scarce protein sources are frequent exchange units. Thus, families with only irregular and occasional access to protein produced on site can benefit from more regular protein availability on the community level. Gift giving, sharing, exchange, and communal activity all fit consistently into a pattern of survival in rural Cape Verdean society. Individuals identify closely with their communities because the thick skein of relationships within a community is a form of insurance. Under the harsh and unpredictable environment of Cape Verde, the fortunes of individual households may vary drastically from one year to the next. The communal sharing strategy functions to keep temporarily disadvantaged households at or above the minimal subsistence line (Wharton 1971), because this year’s recipient may well become next year’s donor in this highly integrated communitarian system. This argument should not imply the absence of self-seeking or even exploitative behavior in rural Cape Verde. Finan (1988) has suggested that the sharing strategies apply most generally in the context of rainfed agricultural activities, where scarcity is most acute and the moral imperative of survival obtains. In contrast, households practicing irrigated agriculture are as much attuned to market opportunities and profit maximization as to neighborliness. In fact, the sharing of commodities in an irrigated agriculture network is much less frequent than in a rainfed one, and the behavior appears to be much more individualistic. As the next two chapters focus on specific rainfed and irrigated systems, there will be frequent reference to these widespread rural institutions and their role in mediating access to resources in this particular agrarian ecology.

Notes 1. The estimate of the workforce involved in agriculture is likely underestimated, since even large segments of the urban population sow rainfed fields in the rural zones during the rainy season. 2. We must acknowledge that the social structure of Fogo and Brava Islands is different from that of the rest of the archipelago. The local institutions that define access to productive resources are essentially the same; however, class and race differences are much more pronounced determinants of the social structure of these islands (see Meintel 1984). 3. In the formative social structure of Cape Verde, there was also an “underclass” of whites comprised of degredados, or exiled outcasts from Portugal. This group included Jews and those punished for criminal or prohibited behavior. The

78

Cape Verdean Rural Society

members of this underclass are often associated with the crioulos in the sense that they seem to share the same rung on the social ladder. 4. Crioulo is not generally understood by a Portuguese speaker. Like most creole languages, crioulo borrows heavily from the base language in terms of its glossary, but the phonetics and the grammar have assumed their own logical form. A good structural description of crioulo is found in Veiga (1982). 5. Panos were dyed cotton cloths primarily worn by women around the waist or used to secure children on the back. These cloths were commonly traded along the Guinea coast for slaves. Today, this artisanal tradition has almost disappeared. 6. From this phenomenon, we have the crioulo term badio, which means literally “to wander” or “to roam.” The runaway slaves occupied the most inaccessible of points in any ribeira and began to practice subsistence agriculture—mostly unmolested. Today, badio is the general gloss for someone who was born on Santiago Island. 7. This model of intensification is, of course, associated with Esther Boserup and her seminal discussion of agricultural intensification (Boserup 1965). 8. The loosening of import restrictions associated with the current liberalization of the economy has resulted in the widespread acquisition of private passenger vans (generically called Hiachi in crioulo, after the maker) that shuttle people from the ribeiras to urban areas. This system has vastly accelerated rural-urban contact even from the most isolated areas. 9. Finan and Henderson (1988) have examined this social characteristic at length, concluding that the most cogent explanation lies in the access to opportunities for single women. On Santo Antão and São Nicolau, where the percentage of households run by single women is considerably lower than on Santiago and Fogo, females have been able to emigrate in significant numbers. 10. Some parcels have economically valuable trees species (such as mangueira and zimbrão), and the distribution of fruit is negotiated separately from the agricultural products of the land. 11. During the labor-intensive weeding period, men and women proudly show off the upturned palms of their hands and the “trophy” callouses that have been earned as a result of their labor.

5 Rainfed Agriculture

On the agricultural islands of Cape Verde, rainfed farming marks the rhythm of rural life with almost sacred rigor. Since virtually all rural households, as well as many urban residents, cultivate some rainfed cropland, there is a collective sense of urgency as families pool their labor to wait for this year’s rain. Farmers look expectantly for signs of tchuba (rain), and a general angst—formed throughout the tragic history of past droughts—seems to dominate the countryside. All society seems focused on the wait for rain; all rural households—their fields clean and ready— are poised for action. The onset of the rainy season defines the beginning of the agricultural year, and with the rains, rural society turns its full attention to the task of land preparation and seeding. Activities on the public work fronts yield to the demand for farm labor and cease temporarily. The wait for rain continues after seeding, since moisture is needed throughout the crop cycle to guarantee production and a harvest. In this chapter, we will describe the characteristics of rainfed agriculture as the predominant farming system found throughout the archipelago. Our discussion touches on prevailing technologies, patterns of access to essential productive resources, and the critical integration of livestock into this system. At an analytical level, we present representative budgets created to quantify the economic benefits of rainfed farming to the household. We remain aware, however, that the annual agricultural campaign is much more than an economic activity: it is social drama reenacted year after year and associated very deeply with survival itself. As we have seen, rainfed agriculture is restricted to the leeward islands of Santiago, Fogo, and Brava and to the windward islands of Santo Antão and São Nicolau. The total amount of rainfed cropland in the country is a matter of some debate because the estimates from the two most recent agricultural censuses (Table 5.1) were derived from different methodologies, and the results must be interpreted with caution. The apparent 79

80 Table 5.1

Rainfed Agriculture Comparative Estimates of Cultivated Rainfed Land on Selected Islands, 1978 and 1988 Rainfed Cropland (ha)

Island Santiago Santo Antão São Nicolau Fogo Brava Cape Verde

1978

1988

% Change

21,576 5,060 1,644 5,483 265 36,702a

20,155 6,401 1,806 5,730 392 36,601

–7 21 9 4 33 –1

Sources: MDRP 1979, 1988. aDoes not include Sal and São Vicente Islands.

reduction in Santiago cropland is possibly related to extended drought conditions and some urban migration; however, the large increases on the other islands more likely reflect discrepancies in the data-collection techniques (MDRP 1990b). The comparison of these data does allow two reasonable interpretations. First, one can conclude that during the last ten years of drought, there has been no significant reduction in total cropland cultivated under rainfed regimes. Second, since the rural population has increased over the decade and cropland has not expanded, it can be deduced that virtually all arable land is in fact farmed in most years.

Cropping Activities in Rainfed Agriculture

Throughout Cape Verde, the major rainfed activity is the annual cultivation of the primary subsistence staples: corn and beans. These two products prepared together form the basis of the traditional Cape Verdean stew, cachupa, the common repast of rural households.1 Corn, first introduced via Brazil, has deep cultural significance for local households (de Pratere 1986), and, from the household perspective, its annual availability constitutes a significant local indicator of family welfare levels. Under normal consumption patterns, the average household of six requires about 3 liters of corn and 0.5 liters of beans per day (along with 1 cl of cooking oil) to maintain the traditional staple diet and to meet local standards for nutritional adequacy.2 The agricultural cycle begins in June or July with land preparation, when farmers remove most dry vegetation from the fields, leaving them bare for sowing. In most areas, seeding occurs after the initial rainfall has

Rainfed Agriculture

81

moistened the ground sufficiently to permit germination. 3 Often a second seeding is required to fill empty spaces caused by washout, poor germination, or bird and animal damage. The traditional variety of corn is a regional, small-kernel white corn, preferred by rural consumers for its taste and ease in processing and preparation. Farmers select the most attractive seed cobs from each harvest and store them for the subsequent planting season. Hybrid varieties have had little success under the precarious rainfed conditions (Amaral 1964:276) and are not commonly used. Several types of beans are sown in the same planting hole with the corn seed, including feijão pedra (Lablab niger), bongolon (Vigna unguiculata), sapatinha (Phaseolus vulgaris), fava (Phaseolus lunatus), and others. These varieties display different levels of tolerance to drought (or, conversely, different responses to moisture availability) and to pest infestation. For certain kinds of predator insects, one bean variety can serve as a preferred host, thus helping to protect the others. As reported elsewhere in West Africa (e.g., Richards 1987), this diversification strategy—built on accumulated indigenous knowledge—enhances the probability that one or another variety will produce under the highly variable and unfavorable conditions of Cape Verdean agriculture. In some regions, farmers have introduced pigeon pea, or feijão congo (Cajanus cajan), either intersown among the corn plants or in separate stands. This cultigen is a perennial legume well adapted to semiarid conditions and provides a range of usable outputs including food, animal fodder, and fuelwood. It reaches maximum production levels in its third year and continues to produce for up to five years or more. The more mature plants require little attention other than the harvesting of beans and the pruning of woody branches for fuelwood. On Santiago, it tends to be cultivated on the steeper, less fertile soils, where the traditional corn-bean intercrop is less productive. Pigeon pea has also been integrated into local diets both as an ingredient in cachupa or as a separate dish. The freshly harvested green seeds are particulary relished. The first rains usher in an immediate, seemingly miraculous greening of the islands. While aesthetically comforting, this vegetation competes with the field crops for scarce moisture, and farmers hasten to eliminate weeds. The first weeding, or monda, takes place soon after germination in a flurry of activity that often mobilizes the djunta-mon exchange networks. A second weeding, ramonda, follows about a month later, and during this task, earth is mounded up around the plant mixture. After the corn has been pollinated, the tassels are removed (corta-flor) for animal fodder and, according to farmers, to channel vegetative “force” into seed production. In December, the harvest sequences begin and continue on through February. The dry stalks and bean straw, along with any remaining spontaneous vegetation are harvested in loosely bound bundles (feixes) and fed to

82

Rainfed Agriculture

domestic animals or sold to livestock owners. There is a very active market for straw, and families with few animals may also sell their unharvested residue on site for a negotiated price, thus avoiding the labor costs of bundling and transportation. As we shall see, the value of fodder may often exceed that of the grain produced in drought years. While the presence of intercropped corn and beans appears ubiquitous and unvarying throughout the archipelago, producers may take advantage of specific microclimatic differences to expand cropping alternatives. For example, in certain regions of Santiago that receive more moisture and have somewhat cooler temperatures, farmers cultivate a range of crops, including groundnuts, sweet potato, and Irish potatoes (including seed potatoes) in addition to the standard corn and beans intercrop. In the humid zone of Santiago and in the upper elevations of São Nicolau, some farmers cultivate pigeon pea in pure stands. On Santo Antão, the upper northeastern reaches of the island permit rainfed cultivation of pigeon pea, sweet potato, cassava, and even a rustic cherry tomato variety. As frequently reported in agricultural studies (Bellon and Taylor 1993), farmers know intimately the fertility characteristics of their fields and attribute to each parcel an aptitude for a specific range of crops. For example, it is common to have a corn-bean intercrop on one section of a single parcel—usually the flatter area—and a pigeon pea–corn intercrop on the steeper section of the same parcel. The technology practiced in rainfed agriculture is labor-intensive and based on family and exchange pools of labor. The principal tools used for cultivation are the pointed hoe, the broad-faced hoe, and the machete. On Santo Antão, the short-handled hoe is more prevalent because the verticality of the terrain requires less stooping, while on Santiago, the long-handled tool is preferred. Mechanical traction and animal traction are not utilized (other than for transportation), and manure, as well as other forms of organic matter for sustaining soil fertility, is not widely available. In our numerous surveys of the agricultural islands, we have not yet identified a farmer who applies chemical fertilizers to rainfed crops. The absence of modern inputs does not signify a backward technology. Farmers rely on stores of accumulated experience to cultivate their rainfed lands. The detailed characteristics of individual parcels have been incorporated into planting strategies, and farmers select and experiment with seeds as systematically as do plant breeders. Timing of agricultural activities, such as planting and weeding, are particularly critical, and farmers must monitor crop production cycles closely. Good farmers also practice soilconservation and water-harvesting techniques to enhance yields. In sum, the prevailing technologies are complex and based on the only capital available to these resource-scarce households: indigenous knowledge.

Rainfed Agriculture

83

The rainfed fields are sometimes subject to severe insect infestations, particularly locust attacks, which can devastate vegetation on a widespread scale. The government maintains a mobile brigade service to provide chemical control measures for the most damaging pests, but the teams can seldom meet the need during a period of attack. Farmers recognize a relationship between certain rainfall patterns and the appearance of locust swarms; however, they are generally helpless to prevent the ravages. Thus, even in the years that drought does not exact its dispiriting toll, farmers still face the risk of harvest loss due to insects. It is difficult to assess the meaning of an “average” yield in a context of wide variability in output. Surprisingly, little research exists on the factors that affect food crop production. In its 1985 studies, SCET-AGRI estimated that during an average rainfall year on both Santiago and Santo Antão, corn yields range from 300 kilograms per hectare in the arid zones to around 700 kilograms per hectare in the humid zones. In “good” years and “bad” years, however, these yields may vary from 30 percent above average to 80 percent below average. In very bad years (1994, for example), production drops to almost nil. As a corroborating source of information on yields, national statistics on total production of corn and beans can also provide a general idea of the annual levels of productivity. In Table 5.2, the annual estimated production figures of these two crops for the years 1975–1990 are divided by the total rainfed area reported in the 1988 agricultural census to give estimates of annual average yields. The results emphasize not only the extreme variability in yields from year to year, but also the lack of a clear upward trend in yields, except perhaps during the years 1986–1988, when rainfall patterns improved somewhat. The overall averages (189 kg/ha for corn and 150 kg/ha for beans) further suggest how severely domestic production has failed to meet estimated consumption needs of the average household in recent decades. Livestock ownership is closely integrated with rainfed agriculture since crop residues from rainfed fields are a major source of animal feed. Animals play several important roles in the economic strategies of households. First, they are able to convert biomass (such as plant stalks and scraps from food preparation) that is not directly consumed by humans into high-quality sources of protein and other nutrients. In addition, livestock serve as important assets that households can retain as accumulated wealth or can liquidate to meet unforeseen expenditure requirements. Manure is used to fertilize soils, primarily on irrigated lands, or is marketed to others. Finally, cattle are used as sources of power to run the sugarcane mills (trapiches), and donkeys and mules are used for transportation of crops, fuelwood, and water. In general, rural households consume goat milk and cheese, cow milk and liquid butter (manteiga), and eggs, depending

84 Table 5.2

Rainfed Agriculture Yields of Corn and Beans, 1975–1990 (in kg/ha)

Year 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Average

Corn

Beans

Combined

41 138 3 28 248 234 83 121 74 69 36 333 584 455 268 315 189

83 55 7 37 69 248 14 82 61 149 58 165 386 391 496 100 150

124 193 10 65 317 482 97 203 135 218 94 498 970 846 764 415 339

Source: MDRP 1992.

on their resource base and the local island customs. Meat, especially beef, is only an occasional treat for most families, and our surveys have identified many households where meat is served but once or twice a year at important rituals. Almost all rural households keep some animals. A study of 234 families in nine Santiago localities (Varela, Langworthy, and Finan 1990) indicated that 40 percent of the households owned cattle (1.3 animals); over 70 percent owned goats (1.8 animals); nearly 90 percent owned pigs (1.5 animals), and over 70 percent owned chickens (3.7 birds). With the exception of cattle, which are concentrated on Santiago Island, this livestock ownership pattern generally applies throughout the archipelago (MDRP 1990b). Wealthier families have the capital to invest in cattle and to maintain them, while poorer families are restricted to the ownership of smaller livestock and poultry, which have low acquisition and maintenance costs. For example, goats are particularly efficient converters of the arid Cape Verdean vegetation and require only minimal tending. Pigs and chickens survive mainly on household refuse and by-products of food processing (e.g., corn bran) and are tended in the proximities of the household dwelling. Cattle require greater attention, since feed must be obtained from the field or the animals must be pastured. Young children often take the responsibility of tending herds away from the homestead, and there is little communal care of the animals. An exacting task is the watering of animals, and, again, household members either bring water from the communal wells or take their animals to these sources to drink. During periods of

Rainfed Agriculture

85

drought, watering is a particular burden in those areas where water must be delivered and rationed. When all animal care is accounted for, survey responses suggest that approximately 90 person-days of labor are required to maintain the average herd.

Household Resource Allocation

The two major factors of production in rainfed agriculture are land and labor. In Chapter 4, we distinguished between patterns of landownership and operational size of farming units, arguing that the primary institutional mechanisms of access to land—that is, sharecropping and rental agreements—allowed households to adjust their farm size to changing subsistence needs. During the initial century of settlement, the combination of the morgadio inheritance system and labor-intensive technologies permitted absentee landlordism and widespread exploitation of the landless and powerless. The periods of prolonged drought, together with recurrent financial crisis and economic decline, reduced the size of landholdings as largeholders abandoned agriculture or partially divested their assets as a response to economic difficulties. With the abolition of slavery, the historical trend moved toward a more dispersed ownership pattern and the widespread adoption of sharecropping. In modern times, independence ushered in concerns of social justice, and the postindependence PAICV government decreed severe restrictions on absentee ownership, which expressly prohibited sharecropping and, in fact, expropriated a modest number of landholdings. These processes did not, however, create the desired smallholder class of peasant owners, nor did they destroy the capitalist landowner class that derives at least part of its wealth from the labor of the rural landless. The ownership of rainfed lands in much of Cape Verde is still concentrated, as evidenced by the fact that 66 percent of the cropland on Santiago (44 percent on Santo Antão) is cultivated by sharecroppers or renters (MDRP 1990b:90). Furthermore, about 30 percent of the national farmland (rainfed and irrigated) remains in the hands of absentee landowners (ibid.:108). At the same time, there is little market demand for rainfed lands and very few instances of market transactions. The low production potential and high risk of rainfed agriculture has created a land-tenure situation in which subsistence farmers may prefer sharecropping (or renting) to purchasing. In fact, many sharecropper relationships are highly personalized through family and other forms of social ties, and sharecropping functions less as an institution for extracting rent and more as a flexible mechanism that distributes cropland according to the shifting consumption needs of the household throughout its domestic cycle. This interpretation,

86

Rainfed Agriculture

however, does not extend to irrigated lands, where the higher-value agricultural potential attracts more investment. Patterns of access to rainfed land are portrayed in Table 5.3. The average amount of rainfed land per household on Santiago is approximately 13 liters (1.3 ha) and about 17 liters (1.7 ha) on Santo Antão. The variation in the amount of rainfed land per household is higher on Santo Antão than on Santiago, as demonstrated by the coefficients of variation of land per household of the two islands. On Santiago, 20 percent of the households with the smallest rainfed holdings farm an average of 4 liters of rainfed land, and 10 percent of the households farm more than 25 liters. The largest rainfed farm sampled had an area of 43 liters. On Santo Antão, the distribution of rainfed land across households is more skewed. The smallest 20 percent of the surveyed farms cultivated an average rainfed area of 1.5 liters, while 10 percent of the farms had more than 35 liters, and 5 percent of the farms had over 50 liters of rainfed land. The average per capita access to rainfed land is approximately 2.8 liters on both islands, but again the distribution is more skewed on Santo Antão, where one-third of the households have less than 1 liter of land per person, as compared to only 13 percent of the households on Santiago. The average of approximately 3 liters of land per household member is insufficient to meet the food-security needs of an individual in all but exceptionally favorable production years. It is commonly assumed in Cape Verde that the annual per capita consumption needs based on minimum calorie

Table 5.3

Characteristics of Land and Labor Use in Rainfed Agriculture on Santiago and Santo Antão

Characteristics Average farm size (l) Coefficient of variation Land per household member (l) Coefficient of variation Total labor utilized in rainfed cultivation (person-days) Per unit labor utilization (person-days/l) Direct family labor (% of total rainfed labor) Male (% of family labor) Female (% of family labor) Child (% of family labor) Djunta-mon (% of total rainfed labor) Salaried labor (% of total rainfed labor)

Santiago

Santo Antão

12.7 0.62 2.8 0.83

16.5 1.32 2.8 1.16

124 10

75 7.4

75 31 58 11

65 69 24 7

22

7

3

28

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

Rainfed Agriculture

87

intake requirements are about 185 liters of the staple combination of corn and beans (approximately 0.5 liters per person daily). At this basic level of biological maintenance, the combined yield of these two crops would have to be about 600 kilograms per hectare for the average family. Referring back to Table 5.2, we see that this level of productivity was attained only three times between 1975 and 1990. The average combined yield over this period was only 57 percent of the minimum needed to meet the subsistence needs of the households. In fact, in most years, the majority of farm households are deficient in grain production and must rely on other income sources to supplement their agricultural production, as well as to provide for nonfood necessities. Table 5.3 also provides a summary of labor use, the other critical factor of production. The total amount of labor that goes into rainfed cultivation is far less than the total availability of labor in the average household (as judged by the number of members aged 15 to 65 years). On Santiago, the average household spends about 124 days in rainfed activities, while on Santo Antão, the average drops to 75 days. These figures represent roughly one-quarter to one-half of full-time annual employment for one adult. The composition of labor sources also varies significantly between islands. On Santiago, three-quarters of all the labor in rainfed activities is provided directly by household members, while djunta-mon, the laborexchange institution described in Chapter 4, accounts for more than 20 percent of the labor. Since djunta-mon must be repaid with family labor, this category represents an indirect use of household labor services. Salaried labor for rainfed agriculture is insignificant on Santiago. On Santo Antão, the household provides a smaller proportion of total labor needs, and djunta-mon does not figure as prominently in rainfed labor activities. Salaried labor, on the other hand, represents over a quarter of the total. What factors explain these distinctive differences in the allocation of rainfed labor supplies? On Santo Antão, the relative abundance of highvalued irrigated land (when compared with Santiago) creates a high opportunity cost for labor applied to rainfed agriculture. Thus it is likely that the quantity of labor allocated to rainfed plots (and away from irrigated plots) is less than that found on Santiago. Also unique to Santo Antão is the distance of rainfed lands from the majority of the rural population located within the steep-sided ribeiras of the island. Difficult access to many of the fields significantly increases travel time, and farmers tend to visit their fields sporadically to carry out specific tasks. In essence, then, it is possible to conclude that the rainfed fields on Santo Antão are less intensively managed and operated. Moreover, among the larger irrigated landowners, there is a tradition to hire labor on their rainfed as well as their irrigated plots. The Santiago community ethic of sharing described in

88

Rainfed Agriculture

Chapter 4 does not exist in rainfed agriculture, since on this island, marketoriented sugarcane cultivation drives social custom. In contrast, on Santiago, large landholdings are generally not managed directly by the owners, but sharecropped or rented out into smaller operational units. This leads to a more uniform distribution of access to rainfed land across households and more suitable conditions for the development of exchange-labor mechanisms. The sexual division of labor is also markedly different between islands. On Santiago, women and men perform similar agricultural tasks. Since the working-age rural population is skewed toward women, female labor provides the majority of family labor (more than 75 percent) and contributes a larger share of the services offered through djunta-mon. On Santo Antão, where rainfed lands tend to be located at great distances from the household, women cannot easily move between the farm and field, resulting in time conflicts between domestic and agricultural responsibilities. Consequently, during the rainy season, women do not usually work the fields directly; rather, they are responsible for the even more arduous task of transporting meals, materials, and outputs up and down the steep-terraced ravines. On neither island do children contribute significantly to the physical tasks associated with rainfed agriculture. We hasten to note that average estimates of total labor demand conceal the seasonal variation in rainfed labor allocations. On both Santiago and Santo Antão, virtually all farms cultivate rainfed lands, and much of the labor demand is concentrated in the period between July and September (depending on the arrival of the rains). Timing is critical to the success of rainfed agriculture—especially for the activities of planting and weeding—and it is during this period that households seek labor sources outside the family. A major shift in labor-allocation strategies can be noted as families reduce irrigated activities and the public work fronts close down. Finally, the estimates of total rainfed labor reported in Table 5.3 do not include the other ancillary tasks associated with the management of a household. The processing of corn with the wooden mortar and pestle (coche o midju) and the preparation of cachupa occupies about three hours of household time per day. The provisioning of water and fuelwood is another task that is highly variable in terms of time requirements, and, in some isolated communities, the process of obtaining water can occupy one person for an entire day. Animal care also requires two or more hours of labor per day. These tasks are part of the daily routine of the household and are usually performed by the women and children of the family.

Costs and Returns to Rainfed Agriculture

Since rainfed agriculture is a subsistence strategy, its output tends to be consumed by the household or converted into animal protein. However, estimation

Rainfed Agriculture

89

of the economic costs and returns to rainfed farming provides insights into the incentives for households to undertake these activities, even though households generally do not accrue monetary returns from them. These calculations assume that households may use their available resources to produce food for their own consumption, which reduces the need to purchase food from the market, or they may use their land and labor resources in alternative activities that could generate income, but would then imply a greater need for food purchases. An estimation of the costs and revenues of subsistence-oriented activities provides a valuable framework for evaluating the relative trade-offs of these alternative resource-management strategies. Because so few households are food self-sufficient, policymakers tend to assume that rainfed agriculture is a poor use of domestic factors, particularly household labor. In an environment of resource scarcity, however, the more important consideration is the availability of alternative allocations of this labor. If it can be demonstrated that the imputed “wage” to family labor compares favorably with its opportunity cost (i.e., what this labor would “earn” in an alternative application), then the policy concerns can be expanded to incorporate relevant nonagricultural issues. Thus, the initial challenge is to value the output from rainfed farming and to examine the returns to the resource investment of the household. Detailed interviews with farmers provided the basic information to construct representative budgets for the different cropping activities in rainfed agriculture. These budgets provide estimates of the household labor requirements for each agricultural task and the necessary purchased inputs for production, including seeds and tools. On the other side of the coin, the budgets attribute a value to all products and by-products of each crop. Table 5.4 summarizes the technical and monetary input/output relationships assumed for these representative activities and the resulting returns to household labor. Based on yield estimates from national production figures reported in Table 5.2, extended discussions with agricultural researchers at INIDA, and our own fieldwork, we have incorporated a per hectare corn grain yield of 200 kilos for Santiago and 250 kilos for Santo Antão. These yield assumptions are conservative interpretations of a widely variable distribution. It is recognized that in several years in favored climatic zones many farmers may attain yields three or four times greater, while in other years, as in 1994, yields may drop to nil. The labor estimates and costs associated with other inputs were derived from survey data and more extensive interviews with farmers about their agricultural practices. These budgets reveal several points of interest. First, they effectively underscore the labor-intensive nature of the technologies currently employed in subsistence agriculture. The total amount of labor required for each activity, however, only ranges from 12 to 25 days per liter of land per year, which is substantially below the labor supply available to the average

Source: authors’ calculations.

25 18 30

liter liter bundle

6 9 8 11 100

20 15

liter bundle

liter liter kg kg kg

2 3 2.5 3.5

20 15 25

Yield (per liter of land)

liter liter kg kg

liter liter bundle

Unit

70 50 3 5 35

20 40 40

40 20

70 50 2 4

20 40 40

Unit Price (per liter of land)

Representative Budgets for Rainfed Crops

Santiago Traditional intercrop Corn Beans Straw Pigeon peas/corn Pigeon pea Grain Green Dried Straw Wood Corn Grain Straw Santo Antão Traditional intercrop Corn Beans Straw Pigeon peas/corn Pigeon pea Grain Green Dried Straw Wood Sweet potato

Island

Table 5.4

3,500

945

2,420

1,409

2,000

Gross Revenue (CVE)

—

28

43

48

43

Purchased Inputs (CVE)

—

100

76

101

76

Capital (CVE)

3,500

817

2,302

1,260

1,882

Net Revenue (CVE)

24.5

12.7

12.5

12.7

12.5

Labor (persondays)

143

64

184

99

151

Returns to Labor (CVE/person-day)

Rainfed Agriculture

91

household with the average amount of cropland. Although specific tasks, such as weeding, may tax the household labor pool over a short period of time, the modest total requirements of rainfed agriculture imply that at least part of the household is underemployed in agriculture. Second, the cash costs incurred in these activities are virtually nonexistent. Some farmers may purchase seeds for the planting season, but generally seeds (or sweet potato vines) are either selected and saved from a previous harvest or borrowed from a friend or neighbor. The purchased inputs category also includes an estimate of annualized capital costs for tools and for working capital, both of which are insignificant. In effect, rainfed agriculture is carried out without recourse to input markets. The third insight derived from these budgets focuses on the positive returns from the principal activities of rainfed agriculture. Formal surveys (e.g., Finan and Belknap 1985) have estimated that only about 15 percent of the rural households have access to sufficient rainfed land to completely meet their food requirements under normal growing conditions. Nonetheless, while rainfed activities do not fully attend the consumption needs of most households, they do provide household labor an employment opportunity with a positive return ranging from about CVE 64 to 184 per person-day of work. As stated above, this level of remuneration is primarily determined by the amount and distribution of annual rainfall, which in a good year may increase several times. But even under our conservative yield estimates, these returns were competitive with the only other realistic alternative opportunities for labor in rural households: namely, FAIMO work front employment or agricultural wage labor on irrigated cropland. Therefore, this analysis suggests that the problem with rainfed agriculture is not so much its low profitability but, rather, its low potential for generating rural employment. Simply put, rainfed agriculture cannot absorb existing rural labor supplies due to the scarcity of agricultural land relative to the size of the rural population. Since agriculture and livestock management are complementary components of the same rainfed system, Table 5.5 presents estimates of outputs and of the net values of the different types of livestock kept by households in Cape Verde. More detailed information on the estimates of net values can be found in the appendix to this chapter (see Table 5.7, p. 97). The net value of an animal represents the total value of the animal and its by-products (milk, eggs, manure, etc.) minus the costs associated with purchasing and feeding the animal. Cows and goats live for several years, but while they must be fed throughout their lives, they produce milk and offspring only as adults, and their meat is obtained only when they are culled from the herd. In our calculations, for example, cows are assumed to have a useful life of seven years. The cow is assumed to bear calves and provide milk

92

Rainfed Agriculture

in its third through sixth years and is culled (either sold or slaughtered) in its seventh year. Because the costs and returns of cows are spread out over time, a summary figure that represents the average annual value of owning a cow is needed. This estimate is arrived at by adding up the discounted net values (returns minus costs) over the seven years of the cow’s life to derive the net present value of the cow. This net value must then be divided equally over the seven years. The standard accounting convention is to estimate the annual payments that would pay off the principal and interest on a loan equal to the net present value of the cow. Thus, the stream of net benefits obtained from owning a cow over the course of seven years is equivalent to receiving CVE 56,340 today, which in turn is equivalent to receiving equal payments of CVE 9,737 per year over seven years. Similar calculations are made for goats. Because farmers typically do not keep pigs and chickens for more than one year, a simple calculation of the net value can be made by subtracting the feed cost (and acquisition cost in the case of pigs) from the value of the outputs. Pigs provide meat, while chickens provide both meat and eggs as valuable outputs. Rural surveys of Santiago and Santo Antão (Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986) provided the empirical bases for constructing representative rainfed farming systems on these two agricultural islands. Each system integrates mixed crop and livestock activities that constitute the representative agricultural component of household resource allocations. In the diverse reality of rural Cape Verde, actual

Table 5.5

Outputs and Net Values of Livestock

Livestock

Useful Life (yrs)

Outputs

Cows

7

4 calves (yrs 3–6) 1 cull cow (yr 7) 600 liters milk/year (yrs 3–7) cows: 1 ton manure/yr (yrs 1–7) calves: 0.5 tons manure/yr (yrs 3–6)

Goats

5

Pigs Chickens

Net Present Value (CVE)

Annualized Present Value (CVE)

56,340

9,737

5 kids (yrs 1–5) 1 adult (yr 5) 90 liters milk/yr (yrs 1–5)

7,547

1,743

1

1 pig

4,320

4,320

1

1 chicken 75 eggs

440

440

Source: authors’ calculations.

93

Rainfed Agriculture

farms vary with regard to land size and access to the particular microniches that permit a broader crop choice. In terms of technology, however, the basic subsistence cropping practices display a remarkable degree of uniformity. For this reason, a single rainfed system was judged adequate to reflect the technical representativeness for each island, although possible sources of variation are indicated where relevant. The technical characteristics and the economic returns of each rainfed system are presented in Table 5.6. Santiago Representative Rainfed System

The typical system constructed for Santiago has 13 liters of rainfed land (the average cropland area for a family of six), approximately 1 liter of which is allocated to the pigeon pea–corn intercrop. Policymakers consider pigeon pea to be a highly adapted crop and sometimes wonder why this particular mix is not more pervasive. In a survey carried out in 1990 (Finan 1990), farmers insisted that pigeon pea does well only on “poor” land, because on the better soils, the plant produces vegetative mass but does not yield adequate levels of grain. Since corn produces more forage and grain per area than pigeon pea, farmers reserve the better lands for the primary corn-beans intercrop and the poorer lands for the pigeon pea–corn alternative (see Table 5.4). Thus, although some regions may devote larger proportions of available cropland to pigeon pea, over the island as a whole, the crop mix in this system generally reflects farmer decisionmaking. With

Table 5.6

Economic Returns of Representative Rainfed Systems on Santiago and Santo Antão Santiago

Total cropland (l) Cropping patterns (l) Corn-beans Pigeon pea–corn Pigeon pea Sweet potato Livestock (no.) Goats Pigs Chickens Net revenue (CVE) Agriculture Livestock Labor requirements (days) Net revenue per person-day (CVE/day) Source: authors’ calculations.

Santo Antão

13

16

12 1 n.a. n.a.

12 n.a. 2 2

2 1 3

3 1 3

23,700 9,000

36,200 10,700

253

314

129

149

94

Rainfed Agriculture

respect to livestock, the representative Santiago system is attributed two goats, one pig, and three chickens as the basic inventory. The annualized value of each animal, reported in Table 5.5, is included in the revenue calculations. The combined Santiago system of mixed-crop and livestock activities provides estimated annual net revenues of about CVE 33,000. Livestock outputs provide approximately one-quarter of the total revenue. With regard to the returns to household labor employed in these activities, the average remuneration per person-day (not separating male, female, or child labor) is estimated at CVE 129. The total labor requirement for this system is 253 days, which represents only about 25 percent of the labor supply available to the average household in rural Santiago. This observation, however, must be tempered by two important factors. First, the amount of time allocated to domestic activities, such as the processing and preparation of food and the obtainment of water and fuelwood, has not been included in system calculations. In some cases, these activities can employ a full-time household member, especially when water sources are distant. Second, seasonal variations in the agricultural cycle will distribute that labor demand unequally across the calendar, which implies that during specific periods, household labor is fully occupied. For example, the mixed-crop component in this system requires almost 20 person-days per month from June through September and an estimated 60 person-days during January, when harvest activities occur. During the rest of the year the labor intensity is substantially reduced. Santo Antão Representative Rainfed System

The representative cropping patterns for Santo Antão rainfed agriculture differ slightly due to higher altitudes, greater amounts of atmospheric moisture, and more varied microniches in the primary rainfed areas. Thus, most farmers have some access to land judged appropriate for pigeon pea and sweet potato cultivation. This representative system has 16 liters of cropland (the island average), comprising 12 liters of the corn-beans intercrop, 2 liters of pure pigeon pea, and 2 liters of sweet potato. This particular crop mix generates almost CVE 47,000 of net revenue per year. Food and income derived from livestock activities are critical to Santo Antão’s rainfed farms. Very few households own livestock, but, as on Santiago, nearly all raise some goats, pigs, and chickens. The representative Santo Antão system includes three goats, one pig, and three chickens. As with the Santiago rainfed system, the Santo Antão representative farm spends about CVE 3,000 per year on purchased feed and utilizes about 90 days of labor. The estimated value of all livestock products is calculated at CVE 10,700, which constitutes about 23 percent of the total household agricultural income.

Rainfed Agriculture

95

The economic characteristics of the Santo Antão rainfed system again reflect a low-input, labor-intensive agriculture oriented toward providing the household with its consumption needs. The returns to the labor employed in agricultural activities are CVE 149 per workday, almost equivalent to 1990 FAIMO average wage rates and about 60 percent of the estimated wage rate in irrigated agriculture—a relatively narrow labor market. The total amount of labor used in this mixed crop-livestock system is slightly higher than on Santiago since the Santo Antão farm has more cropland. But, again, the total demand for agricultural labor is far beneath the potential supply available to the average household (of seven members), even when allowances are made for domestic chores and high transportation times.

Issues in Rainfed Agriculture

This analysis defines a sobering reality with regard to Cape Verdean economic development. The majority of the population lives in rural areas and practices agriculture, and most households do not have access to irrigated land. This incontrovertible fact implies that the primary issue in rainfed agriculture is its potential to provision the consumption needs of rural households. If corn, the Cape Verdean “noble” crop, can be used as an indicator or measure of basic subsistence requirements, the average Santiago household of six must produce 1,100 kilograms of corn per year in order to survive. Under current technologies and yields, this quantity of total production would require the cultivation of approximately 55 liters of cropland. Less than 1 percent of the rural population has access to this much land. Furthermore, a rainfed farm of these dimensions would also imply access to 700 person-days of labor, an amount well beyond the capacity of the household to provide (given animal labor needs, domestic chores, and seasonal labor bottlenecks). Even if a more generous yield assumption is accepted (i.e., 500 kg of grain/ha), the average household still has to cultivate 22 liters of land, which would require major structural changes in land access on both Santiago and Santo Antão. In sum, the current combination of land, labor, and technology constraints in Cape Verdean rainfed agriculture makes self-sufficiency in food production—of even the rural population—a very remote policy objective. Only in rare years of optimal rainfall could Cape Verdean farmers expect to feed themselves or even part of the growing urban population. A second, and related, issue involves labor use. On both islands, the returns to labor allocated to agriculture are equal to or better than its opportunity cost in off-farm pursuits. The problem is the low level of rural employment that rainfed agriculture affords. Even if some inevitable underestimation of labor tasks is accepted, the large discrepancy between the

96

Rainfed Agriculture

overall agricultural demand and the household supply strongly suggests a surplus labor force in the rural areas. This underemployed pool of labor poses a potential threat to social stability through uncontrolled rural-urban migration, and the government has used public work front employment to root the rural population in rural areas. In fact, the public work program, originally intended to be a drought-related emergency measure, has become a primary component of the rural household’s survival strategy. As a third major issue, this analysis reaffirms the ecological imbalance in Cape Verde. The current situation of too many people chasing too little land is, simply, unsustainable. As this research reveals, economic opportunities in the rural areas are not sufficient to meet the needs of all rural residents, but the urban centers are also not prepared to offer attractive alternatives. The decline in emigration has further diminished a historical source of important employment opportunities away from the farm. With nowhere to go, the large rural population continues to practice rainfed agriculture on marginal, poor-quality lands. The government fears that environmental degradation is the inevitable long-run consequence of rainfed agriculture under current technologies and the existing level of population pressure. Although the persistent drought has certainly limited production on both cropland and pastureland, a systematic evaluation of environmental damage has not yet been undertaken. Thus, rainfed agriculture in Cape Verde raises critical policy questions that inextricably involve technical, social, and ecological considerations. The governmental response to the rainfed problemática has focused on rural investment that would conserve soil and water and reduce the threat of ecological disaster. With support of the international community, Cape Verde has implemented widespread programs of soil-erosion control through the construction of networks of physical structures (dikes, bunds, retaining walls, etc.) and a major reforestation program. Again, the impacts of these programs on the environment or on yields have not been measured in any detailed fashion. In sum, rainfed agriculture presents a Herculean challenge to the policymaker. The technological solutions are limited, but nonagricultural solutions have not yet appeared. This, however, is not the story of irrigated agriculture, as we will see.

1 2 3 4 5

Year

Goatsc

1 2 3 4 5 6 7

Year

Cowsa

Table 5.7

Costs

500

Acquisition

15,000

Acquisition

Costs

2,000 2,000 2,000 2,000 2,000

Feed

14,600 14,600 14,600 14,600 14,600 14,600 14,600

Feed — — 15,000 15,000 15,000 15,000 40,500

Animals

500 500 500 500 2,000

Returns

— — 18,000 18,000 18,000 18,000 18,000

Milk

Returns

Animals

Annual Net Returns of Livestock (in CVE)

1,000 1,000 1,500 1,500 1,500 1,500 1,000

Manure

3,600 3,600 3,600 3,600 3,600

Milk

Appendix

–2,500 2,100 2,100 2,100 3,800

Net Value

–28,600 –13,600 19,900 19,900 19,900 19,900 44,900

Net Value

7,547

Net Present Value

56,340

Net Present Value

(continues)

1,743

Annualized Present Value

9,737

Annualized Present Value

Feed Cost 2,880 Value of Animal 250

Acquisition Cost

400

Feed Cost

560

continued

750

Value of Eggs

7,600

Value of Animal

440

Net Value

4,320

Net Value

calf at CVE 1,000/ton. Feed requirements: 7,500 kg straw/yr at CVE 2/kg. bGoats: Prices of kids: CVE 500; adults: CVE 1,500. Milk production: 90 l/yr at CVE 40/l. Feed requirements: 1,000 kg straw/yr at CVE 2 /kg. cPigs: Prices of piglets: CVE 400; adults CVE 7,600. Feed requirements: 1 l no. 2 corn/day for 180 days at CVE 16/l. dChickens: Prices of chickens: CVE 250. Egg production: 75/yr at CVE 10/egg. Feed requirements: 35 l no. 2 corn/yr at CVE 16/l.

aCows: Prices of calves: CVE 15,000; cull cows: CVE 40,500. Milk production: 600 l/yr at CVE 30/l. Manure production: 1 ton per adult, 0.5 ton per

Chickensd

Pigsc

Table 5.7

Rainfed Agriculture

99

Notes 1. In some parts of Cape Verde, cachupa is being replaced by a rice-bean staple, arroz pintado, due to the increased availability of rice and its shorter cooking time. 2. With regard to diet, rural consumers frequently add a piece of fish or, less frequently, meat to the cachupa pot, and they also have access to squash and manioc tubers on occasion. Thus while the basic diet is corn (or rice), beans, and cooking oil, the consumption of these minor foods make an important contribution to overall nutritional intake. 3. In some areas of Cape Verde, the seeding of the fields occurs before the arrival of the rains (sementeira em pó). Dry planting is common where labor supplies are scarce and where the steepness of the fields under wet conditions restricts mobility.

6 Irrigated Agriculture

Irrigated agriculture offers an extremely attractive alternative to the widely variable and generally meager returns from rainfed agricultural production. Crops grown on irrigated land in Cape Verde have much higher yields and are more profitable than the rainfed combination of corn and beans. In contrast to the subsistence staples, many of the irrigated products are cash crops, and substantial shares are marketed to provide a source of income. To the extent that irrigation water is available throughout the growing season, risk and variability in production are much lower than is the case for rainfed agriculture, which depends upon the caprice of climate at critical times in the crop growth cycle. The total amount of irrigated land on all the islands of Cape Verde is extremely limited, and irrigated plots are not equitably distributed among rural households. Most farmers have no irrigated land at all, and the majority of households with access to water have only a very small area. Limitations in the physical infrastructures and the means by which water is distributed to individual fields restrict the timeliness of access to water for many farmers. These conditions presently constrain the benefits that agricultural households can obtain by growing irrigated crops. In an otherwise bleak agricultural picture, the improvement of access to irrigation water has the potential to dramatically increase returns to agricultural households. This chapter examines irrigated agriculture as currently practiced on the islands of Santiago and Santo Antão and documents the distribution of irrigated land among rural households, patterns of access to water, common cropping patterns, and the range of production technologies currently employed in Cape Verde. As in the previous chapter, budgets for the major irrigated crops and several representative crop rotations are presented to indicate current economic returns to irrigated agricultural activities, the nature of existing constraints, and the potential benefits that could be 101

102

Irrigated Agriculture

obtained by increasing the access to water. A final section focuses on the policy issues related to irrigated farming.

Access to Irrigated Land

Irrigated land is geographically concentrated in a few areas throughout the archipelago, primarily within the windward ribeiras on Santiago and Santo Antão Islands. Because of the overall scarcity of water, most farmers in Cape Verde have no access to any irrigated land, as indicated in Table 6.1. Our Santiago survey found that only one-quarter of the sampled households had irrigated land. For geological reasons, water is relatively more abundant on Santo Antão, and about half of the households there have access to irrigated land. Even among those with access to irrigated land, the distribution across households is very skewed, with major differences between Santiago and Santo Antão. Table 6.1 provides the average irrigated area farmed by each quintile, or subgroup of 20 percent of the sample of irrigated farmers, for Santiago and Santo Antão. On Santiago, the smallest fifth of irrigated farms had an average irrigated area of 0.12 liters, which corresponds to a small garden-sized plot of 11 meters on a side. At the other extreme, the largest fifth of irrigated farms had an average area of 3 liters. On Santo Antão, the average irrigated area of the smallest quintile was about 1 liter, and the largest quintile had an average of over 25 liters. The quintile breakdowns of irrigated areas provide two important insights about the access to this resource on Santiago and Santo Antão. First, the absolute amount of irrigated land per household is much smaller on Santiago than on Santo Antão. Overall, the average amount of irrigated

Table 6.1

Distribution of Irrigated Land on Santiago and Santo Antão

Land Distribution Pattern Households with irrigated land (%) Average irrigated area per household (l) Quintiles of households with irrigated land (l) 0–20% of sample 21–40% of sample 41–60% of sample 61–80% of sample 81–100% of sample Households with irrigated area less than sample average (%) Median irrigated area (l)

Santiago 24.9 1.0 0.12 0.25 0.5 0.9 3.0 70.0 0.5

Santo Antão 52.3 8.1 1.1 2.5 4.0 6.5 25.6 77.2 4.0

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

Irrigated Agriculture

103

land per household is 1 liter on Santiago, compared with around 8 liters per household on Santo Antão. On Santiago, almost 70 percent of all irrigated farms have less than 1 liter of irrigated land, and the largest irrigated farm in the Santiago Island survey had an irrigated area of 10 liters. On Santo Antão, only 11 percent of the surveyed households had less than 1 liter of land, and conversely, 7 percent of irrigated farms had more than 30 liters of irrigated land. Second, on both islands, there is a large relative difference between the largest and the smallest farms, with the average irrigated area of the largest quintile about 24 times that of the smallest quintile in both cases. In addition, most irrigated farms have relatively small irrigated fields, with a much smaller number of households farming significantly larger irrigated areas. This can be seen in the large jump in the average irrigated areas between the fourth and fifth quintiles; on both islands, the average irrigated area of the largest quintile is well over three times that of the fourth quintile. One implication of this skewed distribution is that the average irrigated area per farm is not a good representation of the typical farm. The average irrigated area per household (of those with irrigated land) on Santiago is slightly less than 1 liter. However, 70 percent of the surveyed households reported irrigated areas less than the sample average. On Santo Antão, 77 percent of the households farmed less than the sample average of slightly more than 8 liters. In light of such widespread variability, the median perhaps offers a more accurate indicator of a typical irrigated farm 1: one-half of the irrigated farms on Santiago have an area of 500 square meters or less, while on Santo Antão, the median value is 4,000 square meters, or 4 liters. With the exception of 17 state farms and cooperatives, all agricultural land in Cape Verde is privately owned. In addition to direct ownership, indirect forms of access (i.e., rental and sharecropping relationships) provide farmers with use rights to irrigated land. Information about private landownership is not readily available in Cape Verde, since the agricultural census only canvases farm operators; however, informal discussions with farmers throughout Santiago and Santo Antão suggest that the ownership of irrigated land is concentrated in the hands of a relatively small number of proprietários. For example, in many zones, a significant number of neighboring farmers rent or sharecrop land from a single landowner. Thus, ownership is undoubtedly even more concentrated than overall access to irrigated land, and indirect forms of tenure afford landless families the opportunity to undertake irrigated agriculture, even if on a small scale. On Santiago, only one-third of those farmers with irrigated land owns all their land, and almost 60 percent are fully dependent on indirect forms of tenure for access to irrigated plots. On Santo Antão, a larger portion of farmers— almost one-half—owns all their irrigated land, while one-third depends entirely on indirect forms of land tenure.

104

Irrigated Agriculture

In terms of a broad pattern, most major irrigated ribeiras have several traditional resident families that own significant amounts of irrigated land (and on Santiago, often a water source, such as a private well), but they sharecrop or rent out much of their land to large numbers of smallholder farmers. This pattern is particularly evident in the large Santo Antão ribeiras where family names are commonly associated with different zones in the watershed. The actual operational sizes of the farms are much reduced because of the cost of supervising nonfamily labor (see Binswanger and Rosenzwieg 1986). Thus institutions such as sharecropping and renting arrangements have emerged because of the inefficiency of large-scale farming under Cape Verdean conditions and the availability of a large segment of the rural population seeking irrigated land. The cost of irrigated land varies substantially. The MPAAR (Ministry of Fisheries, Agriculture, and Rural Extension, formerly the MDRP) Office of Land Reform, which monitors agricultural land transactions, reports prices ranging from CVE 50,000 to 100,000 per liter of land, with an average price of CVE 70,000. The market for irrigated land compared to rainfed land is quite active. Rental rates for irrigated land are fixed by this office with values ranging between CVE 3,000 and 8,000 per liter, depending on the office’s evaluation of the plots. The average level of observed rental rates for irrigated land was CVE 3,300 per liter. Sharecropping arrangements consistently call for the equal division of the product, with all production costs borne by the sharecropper.

Access to Irrigation Water

Cape Verde has no surface water except for the small amounts of runoff that appear temporarily during the rainy season and flow at most for only two to three months during the year. Although appropriate agricultural land near the runoff channels is often limited, there are stretches of cropland in the ribeira bottoms where flood-recession agriculture (regadio temporário) can be practiced. The 1988 agricultural census reports, for the country as a whole, that 11 percent of irrigated land utilized temporary surface water. On Santiago Island 3 percent of all irrigated land was classified as temporário, while on Santo Antão 16 percent of irrigated lands utilized such temporary surface water. Most irrigation water is provided from subterranean sources, either aquifers within the basaltic rock formations or in alluvial flows. Several technologies are used to capture subterranean water supplies. The least demanding technology is to tap into groundwater where it approaches the surface. Throughout the higher elevations of Santiago and Santo Antão, low-flow springs called nascentes appear at various points that can be

Irrigated Agriculture

105

readily identified by their lushness of vegetation in comparison with the surrounding aridity. Here, local residents are able to capture the flow with rudimentary earthen structures and direct it toward fields in the vicinity. These nascentes, if the volume of water is minimal, may provide only drinking water and irrigate a few hectares; however, in the upper reaches of Santo Antão, the nascentes furnish water for canals that irrigate relatively large areas downstream. Water is also close to the surface in the alluvium bottoms of the ribeiras. Here, small hand-dug wells called poços are sunk into the alluvial soils to an average depth of 5 meters, although many older ones from colonial times are deeper and cement lined. Currently, these wells are installed and managed by individual farmers and provide sufficient quantities of water for 3 to 5 hectares of land. In the past, there were no limitations placed on the number of wells, but since the early 1980s, the government has restricted the introduction of new wells in areas that already have a significant pumping capacity. Tubewells (furos) were first introduced by the Portuguese colonial authorities, and the national government continues to drill new boreholes and install wells. These wells reach much deeper into the aquifer and may provide water for 15 to 20 hectares of land. The government also constructs captações, doublewalled subterranean dams that stretch across the ribeira floor to capture alluvial streams. The water is stored in the area between the walls, then diverted into canals (levadas) that follow the contours of the valleys and distribute the water to downstream fields. These networks may provide water for quite long distances, up to several kilometers along the course of a ribeira, and may irrigate dozens of hectares. Finally, one of the ancient technologies of water capture brought in by the Portuguese is the galeria, a horizontal tunnel dug into hillsides, sometimes up to several kilometers in length, to tap a basaltic aquifer. Common in northern Portugal, these small, narrow tunnels demand arduous stoop labor to construct and then maintain. These galerias, having reached the source of water, generate a small stream on the tunnel floor that is then directed to small reservoirs or to networks of levadas that carry the water to fields downstream. The relative importance of these different subterranean sources of irrigation water differs markedly between Santiago and Santo Antão, as shown in Table 6.2. Wells, both poços and furos, are much more widespread on Santiago, providing water to over half of the total irrigated land. On Santo Antão, furos and poços supply only 5 percent of the irrigated land, while captações, nascentes, and galerias provide water to over threequarters of all irrigated land. Almost all irrigation water sources provide water to several farmers arranged in networks. Some poços supply water to a single farmer, but in most cases, even these smaller wells supply several neighboring smallholders. Since most irrigation involves networks of users, some means of

106 Table 6.2

Irrigated Agriculture Irrigated Land by Water Source Cape Verde

Water Source Poço Furo Nascente Temporário

Area (ha) 344 110 1,251 226

Santiago

Santo Antão

% of All Irrigated Land

Area (ha)

% of All Irrigated Land

Area (ha)

% of All Irrigated Land

17 6 61 11

204 100 231 17

34 17 39 3

59 — 969 206

5 — 76 16

Source: MDRP 1990b.

scheduling access to water for all members over the course of the irrigation season is required. On the privately operated poços, arrangements are made informally among the users themselves. Often, the well and pump belong to a large landowner, while the other users of the system are renters or sharecroppers. In this situation, the landowner retains priority access to the well, and the remaining members have access to the water only when the landowner is not irrigating his or her fields. The tenant farmers may pay the owner directly for water, or water use may be included in the rental or share payments. Flow rates of the shallower wells vary widely during the course of the dry season, with water availability decreasing in the late spring and summer months. Extended periods of below-normal rainfall have also reduced average annual flow rates in recent years, thus diminishing the area of irrigable land. Many wells near the coast have been abandoned because of saltwater intrusion caused by the reduced downstream flow of alluvial freshwater. All subterranean water supplies fall under the jurisdiction of the INGRH. This institute is charged with development and management of underground water resources to meet all needs, including drinking water for human consumption and irrigation water for agricultural use. In its capacity, the INGRH attempts to evaluate available recoverable water supplies in the watersheds throughout the islands and installs and administers networks of wells in order to meet all rural and urban demands without causing long-term overdraft of the water table. During times of extreme scarcity (e.g., during 1993–1994), the INGRH manages the provisioning of water to rural populations for their domestic use. On Santiago, the INGRH directly manages the irrigation networks supplied by furos. Each network has a pump operator hired by the INGRH, usually one of the local farmers, and it is the operator’s responsibility to run the pump, conduct routine maintenance on the equipment, and record the number of hours of access each network member receives. Members are charged a nominal fee for the number of hours that water is drawn

Irrigated Agriculture

107

from the network, and they must also provide fuel for the pump while it is operating. Water is allocated to members on a fixed-time rotation. Under this system, each member is allocated a specific number of hours of water per rotation. In principle, the length of time to complete a single rotation is fixed, but because of pump breakdowns and other unforeseen occurrences, the intervals between rotations may vary substantially. Almost all irrigation water on Santo Antão is drawn from subterranean captações or nascentes and is distributed down the ribeiras in cement-lined or earthen canals. The INGRH is responsible for the construction and maintenance of the canals. Each network has a ditch master, called a merim, who diverts water from the levadas into the individual fields. These networks also distribute water to individuals on a rotation basis, but, unlike the Santiago pump-driven systems, they do not operate on fixed-time rotations. Under this traditional system, called tornador-em-tornador, each member may divert water onto his or her fields as long as the member wishes during his or her turn in the rotation. This difference in operating procedure can be explained by the fact that the flow rates in the captaçaõ systems fluctuate much more than do pump-driven systems over the course of a growing season. As a result, there is no fixed relationship between the time of access and the physical quantity of water received. A fixed-time allocation scheme would thus result in fluctuations in the amount of water received by a farmer over the growing season. Under the open-ended system used on Santo Antão, the interval between irrigations for any individual farmer depends on the irrigation decisions of all other network members. Uncertainties about the timing of future deliveries create incentives for all farmers to apply as much water as possible when they receive access. As individual users adopt this strategy, the time required to go through the rotation increases, and the interval between irrigations is prolonged for all members. Access to irrigation water varies substantially over time. There are seasonal changes due to a decline in underground supplies toward the end of the dry season and year-to-year changes caused by fluctuations in recharge from rainfall. Seasonal variations are reflected in cropping patterns, with farmers growing more water-demanding vegetables in the late fall and winter, and tubers in the spring and summer. In fact, many irrigated fields are fallowed in the summer months, when temperature increases coincide with diminished water availability. With the long-run decreases in water availability in aquifer flows, salt intrusion has become a serious problem in many coastal areas.2 The appropriate timing of irrigations over the course of a growing cycle is critical to obtain good crop yields. Because of seepage through the root zone and evapotranspiration, water must be applied periodically to replenish the moisture needed to maintain growth. Given soil, temperature,

108

Irrigated Agriculture

solar radiation, and wind characteristics, as well as plant-specific water needs, farmers require water for their fields within limited and generally inflexible intervals. With longer than optimal intervals between irrigations, the necessary amount of water available to plants is reduced, resulting in water stress and lower yields. In our surveys of Santiago and Santo Antão, farmers reported a wide range of average intervals between irrigations during the dry seasons of 1984 and 1985 (see Table 6.3). On both islands, a majority of farmers reported intervals greater than two weeks. The average interval between irrigations was 26 days on Santiago and 36 days on Santo Antão. The observed variations in access to water across individual users depend to a large extent on the source of the irrigation water. Pump-fed networks generally provide water on a shorter interval than do networks supplied by springs or captações. The longer average interval between irrigations on Santo Antão can be explained by the fact that irrigation water on this island is traditionally supplied through the open-ended tornadorem-tornador system, which encourages members to water heavily during their turns in the rotation, thereby increasing the irrigation interval for the entire system. In addition, the flow rates in these networks are more highly influenced by rainfall patterns. For example, our surveys on Santiago and Santo Antão were conducted after a succession of several years of very low rainfall, and farmers stated that flow rates in the aquifers had been adversely affected. These irrigation intervals, because of the constraints they impose on crop choice, constitute a critical element in our story of Cape Verdean agriculture and will be referred to frequently in the following sections.

Principal Irrigated Crops

The range and combination of crops grown on irrigated land in Cape Verde are primarily determined by conditions of access to water. Four general

Table 6.3

Distribution of Average Irrigation Intervals % of Irrigated Land

Irrigation Interval (days) 1–8 9–15 16–22 23–30 31–40 >40

Santiago

Santo Antão

4.7 19.7 7.2 53.4 12.4 2.6

5.7 6.8 13.6 40.9 10.2 22.7

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

109

Irrigated Agriculture

categories of irrigated cultigens make up the common cropping patterns found throughout the archipelago. The first is sugarcane (cana sacarina), which is the most widely grown irrigated crop in Cape Verde. Sugarcane is unique in that it absorbs large quantities of water during a single irrigation but is able to withstand long intervals between waterings, even up to several months, and still produce. The second major category is comprised of tuber crops: sweet potato, cassava, and yams (inhame). These important food crops also withstand relatively long periods between irrigations, although their yield response to more frequent waterings is high. Finally, bananas and vegetables are high-valued crops grown primarily for urban markets, and they require water more frequently than sugarcane, cassava, or sweet potato. In Table 6.4, we present various estimates of the area allocated to these different cropping alternatives. As this information reveals, more than 70 percent of all irrigated land in the country is planted with the more drought-resistant crops of sugarcane and tubers, a percentage that corresponds closely with the distribution of irrigation intervals. Sugarcane

From the farmer’s perspective, sugarcane is the only crop that can produce a significant level of output with irrigation intervals greater than four weeks. All sugarcane grown in Cape Verde is transformed into a local variety of rum called grogue for sale in the national market.3 Cane is cut by hand and transported to the local trapiches, or crushing mills, where the stills (alembiques) are also located. Very few cane growers own trapiches; normally, one of the larger farmers in any locality rents his or her trapiche out to neighbors on a negotiated schedule. At the trapiche, the juice is extracted, most frequently using oxen-powered crushing equipment, and placed into large drums to ferment. When sufficiently fermented, the juice is heated and run through a traditional copper still cooled by water trough, and the distilled grogue is then placed in 20-liter bottles called garrafões,

Table 6.4

Various Estimated Shares of Irrigated Area by Crop % of Irrigated Land

Irrigated Crop Sugarcane Tubers Bananas Vegetables

Freeman et al.

SCET-AGRI

61 13 9 11

56 14 10 14

Sources: Freeman et al. 1978; SCET-AGRI 1985a–d; MDRP 1990b.

MDRP 43 27 15 11

110

Irrigated Agriculture

which can be stored by the farmer or sold directly to marketers or consumers. An increasingly common response to the drop in production associated with extended drought has been to add refined sugar to the cane juice during the fermentation process, thus producing more grogue per liter of cane juice. This practice—attributed to grogue producers on Santiago—produces spirits of much lower quality (and price), and it has been roundly condemned by the government, although attempts to prohibit this activity have not been successful. Despite the potential legal consequences and the concomitant notoriety for bad grogue, the economic incentives to add sugar are great, and the practice is widespread, albeit clandestine. Crop residues from sugarcane and the crushed bagasse are also used as animal feed and as fuel for the mill furnaces. Sugarcane is a perennial crop, and plantations remain in the ground for periods of 10 to 15 years before being uprooted and replaced with new cuttings. The stalks are harvested once a year, between January and the end of June. Before cutting, the stalks are stripped of leaves, which are tied into bundles to be burnt or used as animal feed. The cane is then taken to the trapiche and transformed into grogue. After a field has been harvested, any remaining plant material is collected, and manure is incorporated into the ground around the root bunches. On Santo Antão, virtually all sugarcane is grown on terraced fields, and maintenance of the terraced structures is also performed after the annual harvest. The fields are irrigated throughout the dry months—November through August—and watering is curtailed one month before harvest in order to maximize sugar concentrations. According to farmers, the preferred irrigation interval for sugarcane is two to three weeks, although some production can be obtained even with intervals of over two months. Too-frequent waterings promote vegetative growth rather than sugar production. Santo Antão has the largest proportion of irrigated land in sugarcane production and is renowned for producing grogue of the highest quality. Fifty-five percent of irrigated land on Santo Antão is planted in sugarcane, compared with 43 percent for the country as a whole. The fluctuations in water availability from the captações and nascentes help explain the predominance of cane. Given the uncertainty of access to water and long intervals between irrigations in the traditional tornador system, the droughtresistant characteristics of sugarcane make it the most viable crop alternative. The price of grogue has been fairly stable over recent years, although there are significant price differentials based on quality. Low-quality grogue distilled from a mixture of cane juice and refined sugar sells for CVE 100–150 per liter, while grogue from pure cane commands a price of CVE 200 or more. Traditionally, commercial grogue was never identified by source with a label, nor was it sold in standardized bottles (for lack of bottles). Grogue wholesalers, or the producers themselves, guaranteed the

Irrigated Agriculture

111

genuineness of the source and the quality of the product. Recently, however, there have emerged several commercially oriented producers of grogue on Santo Antão, the center of “bom grogue,” who pay close attention to quality control throughout the fermentation and distillation process and who pronounce their reputation of quality with a labeled and standardized bottle. Since 1992, a small export market of labeled (“château,” as it were) grogue has been introduced to international markets where Cape Verdean immigrants and descendants are concentrated. This market appears to be growing. Tuber Crops

Tuber crops are an important source of starchy staples in the diets of both rural and urban consumers. Sweet potato and cassava are grown in all irrigated areas of Cape Verde. They are either prepared and consumed alone or added to cachupa. Inhame is a more localized crop, found mainly along the wet ribeira floors on Santo Antão. On Santiago, sweet potato and cassava are usually intercropped in mounded rows. They are planted at the same time, normally between the months of September and April. Cassava stalk cuttings are set approximately 1 meter apart, with sweet potato vine cuttings placed in between. Most farmers retain their own cuttings from the previous year or exchange them with neighbors. Sweet potato is harvested five months after planting, while cassava remains in the ground for approximately a year, although harvesting can be delayed for several months as a means of storing the tubers. Sweet potatoes are also grown on some rainfed lands, particularly at higher elevations that have more humid microecological conditions. Harvesting and planting can take place at any time of year, so there are no great seasonal fluctuations in market availability. In recent years, the production of sweet potatoes (and Irish potatoes) on Santo Antão has been decimated by a ravaging infestation of millipedes (mil-pes). Sweet potatoes are particularly susceptible to attack by these insects because of their soft skin. However, the mil-pes are not able to penetrate the tougher cassava tubers, so this crop is commonly grown in pure stands in many parts of Santo Antão. As an interesting variation in intercropping, cassava is also often grown within the retaining walls of terraces. Cuttings are placed horizontally among the rocks of the wall face, from which the plants emerge and then grow vertically. When the tubers are mature, the adjacent rocks are pulled away, the tubers removed, and the wall rebuilt. Irrigated sweet potato and cassava can produce very high yields if irrigated at intervals of one or two weeks, but they also provide some production even with irrigation intervals of three to four weeks. Due to the large urban demand, market prices for both sweet potato and cassava are

112

Irrigated Agriculture

relatively stable throughout the year. As a result, production of these crops under favorable irrigation conditions can be quite profitable. Inhame requires very moist growing conditions and is grown in the streambeds of ribeiras in the months when surface water is running at maximum levels. Production is seasonal, with harvest taking place between the months of March and May. Most production of inhame is consumed on the farm, although some appears in urban markets. Bananas

Bananas are grown primarily as a cash crop and, in effect, have been Cape Verde’s only significant agricultural export commodity. They were introduced on Santo Antão and Santiago during colonial times as an export crop for the Portuguese market. At independence, one of the largest plantations on Santiago, owned by a Portuguese family, was taken over by the Cape Verdean government and has been operated as the Justino Lopes state farm.4 There are also numerous private producers of bananas, both large and small. Following independence, shipments of bananas from Cape Verde decreased substantially, as Portugal turned to other suppliers, primarily Ecuador; however, during the 1980s, exports to Europe began to increase once more. Until 1988, the public agency Fomento Agro-Pecuário (FAP) handled all banana exports, but since that time, private traders have also been permitted to export bananas. Unfortunately, the banana export market has declined sharply in recent years, and most bananas are directed toward the domestic market. Bananas are propagated vegetatively. The root system continually puts out new fruit-bearing shoots for a period of 10 to 15 years before it has to be replaced. Each shoot takes approximately one year to mature and bear fruit, and after the bunch is cut, a new shoot from the same root system is ready to harvest in eight or nine months’ time. Thus, a single root system normally produces three bunches of fruit within a two-year period. Bananas are usually planted along the ridges of rectangular basins—approximately 2 meters wide and 3 meters long—to a depth of 50–60 centimeters. Due to the space and soil requirements, banana production is concentrated on ribeira bottoms, usually more toward the coastal areas of the islands. Bananas have quite substantial water requirements and need to be irrigated often in order to obtain good yields. Farmers indicated that one- to two-week intervals between irrigations are preferred, although some production can be obtained even with intervals of up to three weeks. At fourweek intervals, the perennial plants survive, but bunch production is minimal. Bananas also require relatively fertile soils, and almost all growers manure their plants annually. Many also apply chemical fertilizers.

Irrigated Agriculture

113

Due to a combination of historical particularities and climatic conditions, highly commercialized banana production tends to be limited to specific regions on the islands of Santo Antão and Santiago. On Santiago, Ribeira Seca and Ribeira dos Picos (on the northeastern side of the island) produce most of the bananas. On Santo Antão, Ribeira Paúl and Ribeira da Torre (also on the northeastern side of the island) have the highest concentration of banana trees. During colonial times, production was much more widespread, particularly on Santo Antão, but at the beginning of the current drought, farmers began to replace bananas with sugarcane as water supplies decreased. This reduction in cultivated area was accelerated when access to Portuguese markets became more restricted. Cape Verde clearly produces sufficient quantities to meet the domestic demand for bananas, and the potential for expansion is dependent on external market conditions. Vegetables

Commercial production of vegetables has grown rapidly in recent years, following an increase in demand from urban markets. The most important vegetables are onions, Irish potatoes, tomatoes, cabbage, and garlic—all destined for the interisland domestic market. Table 6.5 shows the dramatic increase in vegetable production since independence. Produce is sold in market centers either by the farmers themselves or by rabidantes, the ubiquitous and fiercely independent market women who are so critical to the distribution of goods throughout the archipelago. The scale of vegetable production is very small, due to land and water scarcity and perceived market risks. Some producers, however, are able to harvest two vegetable crops within a year, planting first in October through December and again from January to March. Most farmers cannot plant later in the spring because of widespread water shortages in the summer months. Production of vegetables is severely restricted by limited access to water, and since they need to be watered every week to two weeks, vegetable production is concentrated in areas where water is abundant and scheduling is flexible. On Santiago, only farmers with their own wells or members of irrigation networks with few scheduling constraints are able to grow vegetables. Production is similarly restricted on Santo Antão. The region of Alto Mira has abundant water provided from captações, and the network of users has developed a scheduling strategy that distributes water in intervals appropriate for a dynamic vegetable-producing activity. Parts of Ribeiras Paúl and Janelas also enjoy favorable conditions for some farmers to grow vegetables. The vegetables grown on Santiago are primarily for the Praia market, but some are also shipped to the islands of Fogo, Brava, and Maio. Large numbers of rabidantes purchase farm vegetables in very small volume, then hazard the sea travel to the neighboring islands. From Alto Mira and

114 Table 6.5 Period

Irrigated Agriculture Production of Major Irrigated Crops, 1971–1990 (in tons) Sugarcane

Sweet Potato

Cassava

Bananas

Vegetables

8,984 13,500 9,840 15,440

1,022 2,120 2,960 11,364

1,338 4,440 1,890 5,100

4,134 6,950 4,000 5,260

578 2,692 3,760 7,070

1971–1975 1976–1980 1981–1985 1986–1990 Source: MDRP 1992.

the other production spots on Santo Antão, vegetables flow primarily to São Vicente, the major market among the Barlavento Islands, but also to the islands of Sal, São Nicolau, and Boa Vista. Again, the interisland market activity is carried out by the small-volume rabidante traders.

Input Use in Irrigated Cropping Systems

Irrigated crops, in contrast to the rainfed staples, require significant levels of inputs and management. For example, labor needs per unit of land are several times greater, and many growers are active in the purchased inputs market for seeds, fertilizers, and pesticides. Capital investment is significant for some irrigated activities, such as the establishment of cane and banana plantations and the acquisition of irrigation equipment. Much irrigated land, particularly on Santo Antão, is steeply sloped and requires investment in the construction and maintenance of terraces. Labor

Irrigated crops are particularly more labor-intensive in terms of land preparation than are rainfed crops. Vegetables, cassava, sweet potatoes, and inhame are planted within or along the furrows of basins or along mounded rows that are constructed to concentrate and retain water in the root zone. Land is first cleared of residues from previous crops, then farmers use hoes to form the basins (or mounds), incorporating manure in the process. Men carry out the manual field activities, while women transport manure to the field in 20-kilogram sacks. Approximately 250 person-days are required to prepare 1 hectare of land for planting. After the land is sculpted into basins or mounded rows, the field is planted. This activity, by contrast, does not require much time, only 10–12 days per hectare for sweet potato, cassava, and inhame. Vegetables are started in small seedbeds, then transplanted to the fields in four to six

Irrigated Agriculture

115

weeks. Seeds for vegetables, including seed potatoes, are available for purchase at FAP warehouses or, increasingly, from private merchants, and farmers generally incur transportation costs. Sugarcane and bananas require even more extensive land preparation when new plantations are started. The old plants and root systems must be removed, the soil is deeply hoed and manured, then the basins are formed. The basins for sugarcane and bananas are larger than those for other irrigated crops, and new plantations of sugarcane and bananas require approximately 500 workdays per hectare to install. The seed stalks for bananas and cane can be acquired locally or taken from existing plants; however, some farmers actively seek out improved varieties, even from sources in other countries (e.g., sugarcane from Brazil). On Santo Antão, most irrigated fields are terraced, and the annual rainfall damage often requires substantial maintenance and repair of the terraces, such as rebuilding the walls and replenishing lost soil. Fields near the ribeira floors become filled with rocks and debris that must be removed, while, farther up the hillsides, soil must be transported up from the ribeira floors to fill in gaps behind reconstructed terraces. These activities vary from year to year according to the severity of damage. In our representative budget calculations, an average annual expenditure of 80 workdays per hectare has been attributed to the terrace maintenance on Santo Antão. During the agricultural cycle, labor demands are concentrated in weed and insect control, fertilization, and irrigation. In general, farmers weed most irrigated crops twice, with the first weeding occurring at about three weeks to one month after planting and the second about one month later. As in rainfed agriculture, weeding is a manual task, but not as time-consuming because of the smaller areas. Pest control varies according to the crop. Most vegetable growers attempt to control insects with periodic spraying (requiring the purchase of insecticides and the acquisition of back sprayers), but for the other crops, farmers primarily employ their local knowledge. Blight in banana (the dreaded mal do Panama) is present in Cape Verde but controlled through the eradication of affected plants; and, in the case of millipedes, farmers treat cassava stalks with ashes prior to planting to dissuade attack. Virtually all irrigated fields are manured, but many farmers also apply chemical fertilizers to bananas and vegetables and, less frequently, to sugarcane. Harvesting practices vary substantially by crop and depend on the destination of the output—home consumption or the market. Most vegetables require approximately 10 days per hectare to harvest, except for potatoes, which are more labor-demanding (60 days/ha). The harvesting of bananas, including transport to the edge of the field and weighing, requires approximately 25 days per hectare. The cutting of sugarcane is extremely labor-intensive, and

116

Irrigated Agriculture

we estimate that the activity to strip and cut the cane stalks requires about 60 person-days per hectare. The conversion of cane into grogue also involves significant amounts of labor, which for those without their own crushing mills (and stills) is assumed to be included in the rental fee. Products destined for household consumption are harvested and transported by family members, while marketed produce may either be sold to rabidantes in the field or transported to market centers for sale. For example, the vegetable growers in the isolated Alto Mira valley (on Santo Antão) prefer to transport their produce (on women’s heads or beasts of burden) out of the ribeira and into the port town of Porto-Novo, where rabidantes purchase agricultural commodities for shipment to the urban center of Mindelo on São Vicente and elsewhere. Commercial banana growers on Santiago have two marketing options. Bananas may be sold at the farm gate to domestic marketers who supply urban areas within the country, or growers may deliver their bananas to the alternative market outlets, including the FAP warehouses, for export. FAP pays growers only after bananas have been sold in Europe, which is often an exceedingly heavy constraint for small-scale producers. On Santo Antão, banana producers do not have access to export markets, and their commercial sales are destined primarily for the Mindello market. Growers may sell to rabidantes in the fields or at the port, or they can deliver the bananas to Mindello themselves. On Santiago, the small size of most irrigated plots means that family members can easily meet most of their labor requirements. Furthermore, unlike rainfed agriculture, in which all households within a given region are involved in a specific activity (like weeding) at the same time, irrigated tasks tend to be sequentially timed across neighboring fields due in part to the scheduling of water distributions. In contrast to rainfed agriculture, there is little demand for communal work groups, such as djunta-mon. Instead, there is a greater reliance on paid labor. The different island surveys revealed that 14 percent of farms with irrigated land use hired labor, as opposed to 5 percent of farms with only rainfed land. Thus, in contrast to rainfed farming, which is essentially subsistence oriented, irrigated farming displays the fundamental characteristics of a capitalist system of production. There exist significant differences with respect to management requirements across the irrigated crops. For example, sugarcane, bananas, and tubers require the least amount of oversight because the timing of activities is not so exigent. Irrigation and harvest can be delayed for several days or even weeks without significant effect, and pests do not create major problems requiring immediate control measures. Because of these less stringent management demands, sugarcane and bananas have advantages as extensive “plantation crops”; many field operations may be performed by hired labor with relatively little supervision, often only a hired

Irrigated Agriculture

117

foreman.5 Vegetables, however, must be closely monitored for diseases, pest infestations, or signs of water stress, and appropriate measures must be taken immediately. Moreover, the timing of harvest is crucial because of the perishable nature of ripened produce. As market prices are quite variable for these crops, decisions about the timing of planting and harvest can have great effects on revenues. As a result, family members either perform or closely supervise most of the labor activities, with the exception of land preparation. Purchased Inputs

Current technologies employed in irrigated agriculture require a more active presence in purchased input markets, as well as a certain level of working capital. The first category of cost is water. At the time of the surveys, all Santiago farmers receiving water from state-operated wells paid a fee of CVE 1 per cubic meter of water, as well as costs of the fuel to operate the pump. In contrast, on Santo Antão, there is no direct charge for water use, although farmers do pay the ditch master a minimal percentage of their output for his services in directing water to their fields. The other categories of purchased inputs include fertilizer, pesticides, and seeds. Most farmers apply organic fertilizer to their irrigated fields, and there is an active market for manure. In addition, some banana and vegetable growers also purchase chemical fertilizers, primarily from FAP, the government agricultural supply agency. The use of pesticides is quite limited in Cape Verde, although vegetable producers do tend to purchase back sprayers to control against insect infestations. Seeds for vegetables and potatoes are purchased at FAP or in local markets. In the latter case, the lack of adequate quantities of seed potatoes presents a significant constraint to the expansion of production. Cape Verde does have a program of domestic seed production in the higher elevations of Santiago; however, local supplies do not satisfy the demand. The government, through FAP, has traditionally imported a single annual shipment of seed potatoes, usually during the autumn. Since most farmers do not have adequate facilities to store their own stock, they tend to purchase and plant their seed potatoes at the same general time. As a result, potato farm gate prices exhibit a marked seasonal variation, even though growing conditions would permit potatoes to be harvested over a longer period of time. In fact, some farmers are beginning to experiment with on-farm storage of seed potatoes in order to provide more flexibility in planting and avoid the market glut and low prices at peak harvest time. Sugarcane producers have additional input costs. Most farmers do not own trapiches and must pay a fee for processing their cane at a neighbor’s mill. In general, a single ribeira may have one or two trapiches that serve

118

Irrigated Agriculture

several dozen farmers. Most commonly, a farmer pays a processing fee in grogue output, 20 percent of gross production, and is also expected to cover the operating expenses of the mill during the processing time, including fodder for the oxen (or, in rare cases of a diesel-operated mill, the cost of fuel) and the fuelwood burned in the furnaces. The government also assesses a production, or spirits, tax on each unit of area cultivated in sugarcane. Capital

The three major types of fixed capital in irrigated agriculture are the investments in establishing perennial crops, the construction of terraces and other erosion-control structures, and the acquisition of irrigation equipment including motors, pumps, and tubing. The patterns of capital expenditures differ by island and by crop. For example, on Santo Antão, terraces and check dams are much more prevalent due to erosion dangers created by the extreme nature of the topography. On the positive side, farmers on Santo Antão do not shoulder the cost of water-conveyance infrastructures, which for the most part are constructed and maintained by the state. In contrast, most irrigated fields on Santiago are located on flatter land, and terracing is generally unnecessary; however, all farmers are responsible for getting water from the well to their fields, either by acquiring their own plastic tubing or by maintaining earthen feeder canals. Moreover, on Santiago, a large portion of irrigation pumps are privately owned, and the hand-dug poços must be maintained individually. In the context of input requirements for irrigated agriculture, it is important to recognize that Cape Verdean rural society is cash poor to an extreme. In effect, many farmers do not have the liquidity necessary to operate irrigated lands. Furthermore, although irrigated agriculture demands both out-of-pocket expenditures and capital investment, credit availability is very limited from either formal sources, such as the banking system, or informal sources. In general, the government has shown reluctance to invest in high-risk agricultural enterprises; thus, most rural investments are financed out of savings or nonagricultural income, such as emigration remittances from family members.

Representative Budgets for Irrigated Crops

Consistent with our methodological framework, representative cropping activities were constructed for all the major irrigated crops grown on Santiago and Santo Antão. These representative activities attempt to provide a synthetic model of household decisionmaking patterns in irrigated agriculture.

Irrigated Agriculture

119

When prices are attributed to all of the inputs and outputs, we construct budgets that estimate the economic consequences of these representative allocative decisions. In this sense, the net revenue (calculated by subtracting the cost of purchased inputs) derived from an activity represents the income that accrues to the primary household resources—namely, owned land, capital, and household labor. The budget results therefore reflect the implicit income paid to a household for a particular allocation of its resources. Revenues may assume the form of cash, if the output is sold, or may represent the imputed value of products consumed by the household. In the economic sense, profits are calculated by subtracting from revenues all costs, including the imputed costs of inputs owned by or belonging to the household. When an activity exhibits positive profits, the use of household resources in this particular activity generates net income for the economy as a whole. Table 6.6 summarizes technological assumptions—specified as inputs and outputs—that underlie each of the representative irrigated activities on Santiago and Santo Antão. Some crops have several budgets, corresponding to different irrigation intervals, that is, alternative technologies. For example, there are two budgets for sugarcane on Santiago—one with an irrigation interval of 28 days and another with an interval of 21 days—and the yields, inputs, and labor requirements of these two activities are different due to the assumed disparities in access to water. More frequent irrigations generate greater yields but also increase water costs and require more labor time to undertake the irrigations and the harvest. The labor requirements for each cropping activity include land preparation, planting, watering, pesticide application, harvest, and transportation. The labor requirements for sweet potato, cassava, and vegetables are higher than for sugarcane and bananas, because the land-preparation costs for these two latter plantation crops are embodied in capital expenses extended over a period of several years. The labor requirements on Santo Antão are generally higher than those on Santiago for two reasons. First, more land is terraced and requires annual maintenance activities; second, much agricultural land on Santo Antão is inaccessible by road, so transportation by vehicle or mule is impossible. Transportation of inputs and outputs to and from fields requires significant and burdensome labor, usually performed by women. The representative technologies for vegetables and bananas include the use of chemical fertilizer in addition to manure. Fertilizer is not included in the activities with longer irrigation intervals because of limited plant response. On Santiago, only vegetable crops receive pesticide applications. The number of applications varies substantially, both spatially and temporally, depending on the level of insect populations. We assume in the representative budgets that farmers on Santiago and Santo Antão apply

Sugarcane/straw 28-day interval 21-day interval Sweet potato/cassava 28-day interval 21-day interval 14-day interval Banana 21-day interval 14-day interval Irish potato Cabbage Onion Tomato 150/30 200/35 30/500 500/800 900/1,500 3,000 5,000 1,700 1,700 1,700 3,400

39 41 45

29 31 51 43 42 47

Yield (kg/liter)

22 23

Labor (persondays/liter)

15 15 60 40 40 45

35/40 35/40 35/40

200/80 200/80

Price (CVE/kg)

45,000 75,000 102,000 68,000 68,000 153,000

30,500 49,500 91,500

32,400 42,800

Gross Revenue

Crop Budgets for Major Irrigated Crops (per liter of land)

Santiago Irrigated Crops

Table 6.6

10,200 11,300 13,500 5,800 2,900 17,500

3,900 4,500 5,500

11,100 15,800

Purchased Inputs/Capital

34,800 63,700 88,500 62,200 65,100 135,500

26,600 45,000 86,000

21,300 27,000

Net Revenue

Monetary Results (CVE)

8,100 8,600 13,600 11,500 11,200 12,800

10,500 10,900 11,800

5,600 5,800

Labor

(continues)

26,700 55,100 74,900 50,700 53,900 122,700

16,100 34,100 74,200

15,700 21,200

Profit

continued

Sugarcane/straw 28-day interval 21-day interval Sweet potato Cassava Inhame Pigeon pea Banana Irish potato Cabbage Carrots Onion Garlic Tomato

Santo Antão Irrigated Crops

Table 6.6

36 42 48 39 157 39 46 43 49 41 49 59 52

Labor (persondays/liter)

150/30 200/36 650 500 1,500 600 3,000 2,000 2,000 2,500 1,500 500 2,000

Yield (kg/liter)

200/80 200/80 30 35 40 50 18 40 35 30 60 300 45

Price (CVE/kg)

32,400 42,900 19,500 17,500 60,000 30,350 54,000 80,000 70,000 75,000 90,000 150,000 90,000

Gross Revenue 11,600 16,000 2,100 1,600 1,600 720 4,000 27,100 15,700 7,800 14,100 15,200 15,700

Purchased Inputs/Capital 20,800 26,900 17,400 15,900 58,400 29,630 50,000 52,900 54,300 67,200 75,900 134,800 74,300

Net Revenue

Monetary Results (CVE)

6,800 7,900 9,100 7,500 31,400 6,800 8,700 8,200 9,500 7,800 9,300 11,500 9,700

Labor

14,000 19,000 8,300 8,400 27,000 22,830 41,300 44,700 44,800 59,400 66,600 123,300 64,600

Profit

122

Irrigated Agriculture

pesticides at each irrigation. In fact, many farmers report that they are unable to acquire the appropriate types of pesticides at the necessary time. At the time of the surveys, pesticides were provided by FAP, and stocks were often limited. It is possible, however, that under current liberalization policies, private traders will handle a larger share of the market for all purchased inputs. Monetary accounts of the representative production activities estimate the “success” of these resource allocations (Table 6.6). Gross revenue per liter of land is calculated as the per liter yield multiplied by the price of the output. This is the imputed value of all production, and cash income may be smaller, depending on the amount sold. The primary cash crops—bananas and vegetables—generate much higher gross revenues than tuber food crops. With the exceptions of pure-stand cassava and sweet potatoes on Santo Antão, sugarcane exhibits the lowest gross revenues of all the irrigated crops. Revenues from bananas are approximately twice as great as those from sugarcane, while vegetable revenues range from two to five times greater than those from sugarcane. Purchased inputs include all current inputs used in the production process: seeds, fertilizer, pesticides, water charges, and spirits taxes in the case of sugarcane. Some inputs are not purchased (seeds may be saved from previous harvests, manure provided from household livestock, etc.), but these costs have all been valued at their imputed market value. The purchased input costs are relatively high for sugarcane for two reasons. First, we included in the budgets a processing and distillation fee equal to 20 percent of the grogue produced, which corresponds to the actual amount paid to trapiche owners. Second, growers of sugarcane must pay a tax of CVE 8,000 per hectare of land planted in sugarcane. Both of these items are included as purchased input costs in the sugarcane budgets. The purchased input costs for vegetables and potatoes are relatively high because they include fertilizers, pesticides, and purchased seeds. Tomato purchased input costs are also significantly greater than those for other crops because of numerous insecticide applications and the cost of materials needed to stake the plants. Water costs differ between Santo Antão and Santiago. On the latter island, the representative sugarcane, food crops, and banana activities are assumed to receive water from wells operated by the INGRH. It pays CVE 1 per cubic meter of water and also provides fuel to operate the pump. Vegetable producers are assumed to have their own pumps (included as a capital cost), so their sole purchased input cost is the fuel to operate the pump. Per liter costs vary by the number of months irrigated and the interval between irrigations—ranging from CVE 1,000 for 28-day interval sugarcane to over CVE 3,000 for 14-day interval bananas and sweet potato/cassava. On Santo Antão, all water is assumed to be provided from

Irrigated Agriculture

123

an INGRH-controlled network of captações. Here, the only charge to the farmer is for the services of the merim, calculated at the traditional fee of one-thirtieth of the total value of grogue production. Based on the 28-day interval cane budget, this cost amounts to approximately CVE 1,000 per liter of land. Farmers on Santo Antão must also pay higher charges for transporting inputs to their farms because road access is more restricted. Manure is often transported over considerable distances, since the cattle population is small and widely dispersed around the island. Because of larger applications of fertilizer and frequent market trips, vegetable producers also have higher transportation costs. Where vehicles do not pass (as is the case with many communities on Santo Antão), women are hired to transport this cargo on their heads. Capital expenses include those costs that are incurred over several years and interest charges on working capital during the growing season. They are generally lower in the Santo Antão budgets because farmers do not have to incur costs of irrigation tubing, motors, and pumps. Net revenue, as stated earlier, represents the total value of production minus purchased inputs and capital costs. For those households who use family labor, net revenue represents an informal wage payment for the use of that labor pool (in the given activity). If the output is sold, the budget result represents the net monetary receipts received by the household (assuming that purchased inputs and capital expenses are monetary costs). Profits, on the other hand, are obtained by subtracting the value of family (and nonfamily) labor, as well as the imputed value of land, from net revenue estimates. In these budgets, labor was valued at the existing agricultural wage rate of CVE 250, or the monetary cost that would be incurred if laborers were hired to carry out all operations. Wage rates for agricultural labor are higher on Santiago than on Santo Antão due to the influence of nearby urban labor markets in the capital city that provide greater alternative employment opportunities for rural workers. With regard to the imputed value of irrigated land (CVE 3,500/liter), estimates are drawn from official data on land rents and interviews with farmers (Varela, Langworthy, and Finan 1990). The results of the economic calculations given in Table 6.6 demonstrate that all irrigated crops generate positive net revenues that are significantly greater than those for rainfed crops. Even the irrigated crops with the lowest net revenues—sugarcane on Santiago and sweet potato on Santo Antão—provide 9 to 10 times more income than the various cornbeans combinations. Indisputably, access to irrigation water creates a much more attractive agricultural enterprise. A comparison of the irrigated crops also demonstrates the strong inverse correlation between irrigation interval and net revenue. The vegetable crops, which demand water most frequently,

124

Irrigated Agriculture

provide net revenues seven to eight times higher than those for the crops irrigated once per month. This analysis firmly asserts that farmers with frequent access to water face strong economic incentives to grow cash crops (vegetables and bananas). Yet the empirical reality belies these results, in the sense that only a small share of total irrigated area is cultivated in these high-value crops. In effect, farmers do recognize the economic attraction of vegetables and bananas but are constrained in their choice of crops by limited access to water. This analysis also points out the value to society of irrigated agriculture. All irrigated activities generate positive profits, indicating that there are strong economic incentives to invest in and to expand irrigated agriculture.

Representative Farm Systems

The synthetic crop budgets presented in this analysis model the economic returns from production of a single crop. In rural Cape Verde, however, virtually all irrigated farmers grow complicated combinations of crops, which we attempt to model here. In order to estimate returns from all irrigated activities (and not just a single crop), several representative crop systems were constructed. The concept of a system here borrows from the farming systems research paradigm that focuses on a set of interrelated cropping and livestock activities that households do not perceive as analytically distinct (since households probably don’t do representative budgeting). In an operational sense, the representative systems are constructed as model farms with specific plots of irrigated land allocated to particular combinations of irrigated crops. A number of factors influence crop mix decisions. As a first-level constraint, the conditions of access to irrigation water impose restrictions on the kinds of crops a farmer may feasibly grow. Individuals who expect to receive water on intervals of one month or longer have only one cropping alternative: sugarcane. If water is available every three or four weeks, sweet potatoes and cassava are also possible. Bananas require irrigation on three-week intervals or less, and farmers who receive water every 7–14 days may also grow vegetables. Water availability is not the only determinant of crop mix, because even those farmers who are able to grow bananas and vegetables also may include some sugarcane and tuber crops in their land allocations. Although bananas and vegetables can command higher net revenues, farmers also react to much higher levels of marketing and production risk. Vegetables, in particular, are more susceptible to pests and disease, their prices are much more variable, and their yields are more adversely affected by short

Irrigated Agriculture

125

interruptions in irrigation schedules. As a result, most farmers who have frequent access to water do not put all their land into bananas or vegetables, but also plant some of the lower-risk food crops, so that even in the event of low production or disappointing prices, they can still expect to obtain some return from the other crops (and also fulfill household consumption needs). Like their colleagues elsewhere in the world, Cape Verdean farmers do not allocate all irrigated land to, say, tomatoes despite high revenue expectations, because they recognize and appreciate the value of risk-averting diversification strategies. We have constructed five cropping systems as representative of the range of irrigated farms on Santiago. The first pattern, which characterizes the irrigated agriculture of 55 percent of the sampled households, includes farms that grow only sugarcane. This group suffers from infrequent access to water and very small amounts of available irrigated land. Seventy percent of all the households in this group have 0.5 liters or less of irrigated land, an area that precludes possibilities for multiple cropping. The second pattern is comprised of a crop mix of sugarcane, sweet potato, and cassava, and it is practiced by over one-quarter of the irrigated farms in our survey sample. These farms are also constrained by their small land size. The third system includes bananas in the crop mix, and this pattern accounts for 6 percent of all irrigated farms. These farms exhibit a wide range in terms of irrigated area, although on average they tend to be larger than the previous two systems. A fourth system represents a specialization in bananas found primarily among large-scale, capitalist growers. This system attributes 8 liters of land to bananas and 2 liters to tuber crops. While not common on Santiago, this system is nonetheless important because of its import potential. The last system is comprised of farms that include vegetables in their cropping patterns. Nine percent of irrigated farms fall into this category, and most vegetable growers have a relatively small amount of irrigated area—1 liter or less— but ready access to water. The representative cropping patterns are very different on Santo Antão. Approximately 30 percent of all farms grow only sugarcane, while another 11 percent grow tuber crops along with sugarcane. Many in this category also have a few banana plants, but only for home consumption. The third representative system is characterized by significant banana production destined in part for domestic markets. This group includes over 20 percent of the irrigated farms on the island. One-third of the irrigated farms on Santo Antão include vegetables in their crop mix. Overall, there is a significant percentage of farms with large irrigated areas—10 liters or more—that specialize in sugarcane production, although some areas are planted with other crops.

126

Irrigated Agriculture

The individual crop budgets thus provided the basic information to model 10 representative systems on the two islands. These systems effectively provide estimates of the economic performance of the principal types of irrigated farms, as summarized in Table 6.7. Among the Santiago systems, small cane alternatives provide very meager returns, particularly when revenues are adjusted for the case of sharecroppers. The second small-scale system, which includes tuber crops along with sugarcane, provides a significantly higher level of returns, but one that is still very small in absolute terms. Together, these two systems represent the vast majority of households with irrigated land on Santiago. Although certainly superior to the rainfed alternative discussed in Chapter 5, the agricultural revenue generated by these irrigated farms does not provide an adequate level of livelihood for the households. In constrast, the two banana-based systems and the vegetable system generate significant net returns from irrigated activities. The large-scale banana system produces an annual income of almost CVE 600,000 (approximately U.S. $7,500), an extremely large income for rural Cape Verde. While the small-scale banana and vegetable systems earn less due to their reduced land area, their returns exceed those of the cane and tuber system by 4 to 10 times. When these results are compared on a per liter of land basis (in order to adjust for scale), they demonstrate the direct correlation between profitability per unit of land and access to irrigation water. Table 6.7 also presents estimates of returns per unit of labor, a measure of the implicit wage derived from labor allocated to agriculture. All labor employed in the irrigated systems, including that of sharecroppers, earns returns well above the existing agricultural wage rates (see Chapter 4). The labor productivity of the small-scale banana and vegetable systems is particularly high even with the large amounts of household labor absorbed by the vegetable system. As important is that fact that the irrigated systems employ much larger amounts of labor than do the rainfed systems, demonstrating the former’s potential to absorb excess labor supplies. On Santo Antão, irrigated activities earn higher total incomes due to the larger size of the farm. The sugarcane system has 3 liters of land that generated net revenues of CVE 61,100 for landowners and CVE 34,100 for sharecroppers. Both the vegetable and the large sugarcane systems earn incomes above CVE 400,000, but the sugarcane system has 20 liters of land as opposed to 4 liters for the vegetable system. The patterns of net returns per unit of land are quite similar to the Santiago systems, except in the case of the larger farms. The per liter returns of the large sugarcane system on Santo Antão are one-third those of the large banana system. Returns per unit of labor are generally lower on Santo Antão, because of the additional labor requirements in transportation and maintenance of the terraces; however, they are again well above the opportunity costs for family labor and significantly higher than the returns from rainfed agriculture.

127

Irrigated Agriculture Table 6.7

Characteristics of Representative Farm Systems

Santiago Systems Irrigated area (l) Irrigation interval (days) Crops (l) Cane Sweet potato/cassava Banana Onion Potato Cabbage Tomato Labor input (days) Per farm Per liter of land Net revenue per farm (CVE) Owner Sharecropper Net revenue (CVE/l) Owner Sharecropper Net revenue (CVE/person-day) Owner Sharecropper Santo Antão Systems Irrigated area (l) Irrigation interval (days) Crops (l) Cane Cassava Pigeon pea Banana Inhame Onion Potato Cabbage Tomato Garlic Carrot Fava bean Sweet potato Labor input (days) Per farm Per liter of land Net revenue per farm (CVE) Owner Sharecropper Net revenue (CVE/l) Owner Sharecropper Net revenue (CVE/work day) Owner Sharecropper

Small Cane

Cane/ Tubers

Small Banana

Large Banana

Vegetable

0.50 30

0.50 30

1.00 21

10.00 15

1.00 7

0.50 n.a. n.a. n.a. n.a. n.a. n.a.

0.25 0.25 n.a. n.a. n.a. n.a. n.a.

0.25 0.25 0.50 n.a. n.a. n.a. n.a.

n.a. 2.00 8.00 n.a. n.a. n.a. n.a.

0.25 0.25 n.a. 0.25 0.25 0.25 0.25

11 22

15 31

30 30

334 33

9,500 5,000

10,800 4,750

33,300 12,800

59,300 108,500 n.a. 45,200

19,000 10,000

21,600 9,500

33,300 12,800

59,300 108,500 n.a. 45,200

864 455

720 317

1,110 427

Small Cane

Large Cane

n.a. n.a.

Cane and Cane and Banana Cassava

63 63

1,722 718 Vegetable

3.00 30

2.00 15

4.00 21

2.00 30

4.00 7

3.0 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

15.00 1.25 0.25 1.25 1.25 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

1.50 0.25 0.25 1.50 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

1.25 0.25 0.25 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

n.a. n.a. n.a. n.a. n.a. 1.00 2.00 1.00 0.25 0.25 1.00 1.50 1.00

109 36 61,100 34,100

993 50

145 36

77 39

363 91

407,400 116,700 n.a. 56,600

38,300 18,300

433,300 161,300

20,400 11,400

20,400 n.a.

29,200 14,100

19,200 9,200

108,300 40,300

561 313

n.a. n.a.

805 390

497 238

1,194 444

128

Irrigated Agriculture Issues in Irrigated Agriculture

The analysis presented above offers several insights into the agrarian ecology of Cape Verde. First of all, we must again recognize that irrigation water is extremely scarce in Cape Verde and that only 5 percent of agricultural land is presently irrigated. Moreover, access to irrigated land by rural households has an extremely skewed distribution. Only one-quarter of all farm households have any irrigated land at all, and within this limited subgroup, the vast majority of families cultivate 1 liter of irrigated land or less. Nevertheless, the economic benefits of irrigation are substantial. Sugarcane, which has the lowest net returns of all irrigated crops, provides net revenues per unit of land 8 to 10 times greater than the returns generated from rainfed corn and beans. The comparison with the highervalued cash crops is even more dramatic. Thus, irrigation offers farmers tremendous opportunities to increase their agricultural incomes, but access is only for the privileged few. Second, the analysis helps explain certain aspects of farmer decisionmaking. It can be readily argued that the current prevalence of sugarcane is not due to economic incentives but, rather, to physical constraints that restrict the widespread production of more profitable crops. In fact, net returns per unit of land for bananas are two to three times greater than those for sugarcane, and those for vegetables are three to eight times greater; however, these crops have more stringent water requirements, and most farmers are not able to irrigate their fields often enough to be able to grow the more profitable crops. When the individual crop activities are combined into representative crop systems, the overall impact of irrigation on household income becomes apparent. On both islands, those systems with greater access to land and regular availability of water can generate very substantial incomes from agriculture. Again, the overall scarcity and the poor distribution of water supplies severely constrain the cropping choices of farmers. Overall, irrigation affords great opportunities to increase returns to rural households in Cape Verde, and the benefits to any investments or policies that increase the effective availability of irrigation water to rural households are substantial. These actions could take the form of measures to increase the physical irrigation capacity, or of changes in management of irrigation systems that would provide individuals with more flexibility in their water-utilization patterns. The following chapters will discuss particular types of public management strategies that could increase the effective availability of water to agricultural users. At the same time, there remain daunting issues with regard to the sustainability of irrigated agriculture in the context of an expanding population. The primary challenge is to determine the sustainable level of groundwater

Irrigated Agriculture

129

use within the reality of Cape Verdean climatic constraints. During the current prolonged drought, the potential conflict between water for human consumption and water for irrigation has begun to appear more imminent, as tanker trucks deliver water to rural communities all over Santiago. The expansion of irrigated agriculture is integrally related to the increasing nonagricultural demand for water. At the same time, the analysis raises the issue of equity in rural society. Farmers with more flexible access to water can generate more income; however, it is not known if all irrigated farmers can move toward the higher valued crops. In fact, in some localized irrigation networks, fewer users might generate greater total income by managing the water supplies more effectively. But even the poorest of irrigated farmers have superior levels of welfare over rainfed households. These issues, here raised but unresolved, will be readdressed in the following chapters.

Notes 1. One-half of the sample has irrigated land area less than or equal to the median. 2. On Santiago Island, the capital city of Praia has experienced a construction boom with subsequent increases in the demand for sand. Sand is mostly mined at beaches where ribeiras open to the sea. As a result of widespread sand extraction, seawater has invaded deeper into the ribeira, exacerbating the salinization problem already there. 3. A very minor amount of sugarcane is marketed fresh in the urban centers. 4. Cape Verde has not escaped current neoliberalist prescriptions and has begun to explore ways of privatizing the Justino Lopes property. 5. Most landowners are present during the cane harvesting and processing activities, during which time the labor demands are more pressing and, obviously, when much of the marketing occurs.

7 Household Management Strategies

Within the overall reality of resource scarcity in Cape Verde, households pursue livelihood strategies based on their particular opportunities and constraints. They allocate their available family labor, rainfed and irrigated land, livestock, and other types of capital among the different agricultural and nonagricultural activities that make up their individual horizons of opportunity. In Cape Verde, most households face constraints that severely restrict their use of resources in more profitable ways. Many of these constraints are inherent in the ecological characteristics and the overall paucity of resources relative to the rural population, but other constraints are institutional or technological in nature and therefore subject to change through government or local community actions. In previous chapters, we have examined the agricultural and livestock alternatives that define the nature of the Cape Verdean agrarian ecology. We have determined that rainfed agriculture (and livestock) rewards its family labor at a competitive level, but most households simply can’t fully employ all of their working-age members with their current endowments of rainfed land. In comparison, irrigated agriculture is much more productive and remunerative, but water is scarce and unevenly distributed. In this chapter, we shift our focus to the dynamics of household decisionmaking and creative strategies of resource allocation that households adopt and adapt. The rural household in Cape Verde allocates resources to agricultural activities, and we have provided estimates of the economic returns from both rainfed and irrigated farming. Many rural families, however, attain part of their livelihood from nonagricultural activities, such as public work fronts, petty trading, and emigration remittances. Now we will calculate the total household income derived from alternative patterns of resource allocation. Based on the surveyed households on Santiago and Santo Antão, we can examine the distribution of income across families and the relative contribution of the different income sources, thus modeling the 131

132

Household Management Strategies

actual allocation strategies of these families. In addition, we evaluate the economic incentives of alternative strategies for utilizing productive resources and identify the constraints that restrict households from adopting strategies that could provide greater economic benefits.

Resource-Management Strategies

Analytically, household resource-management decisions may be conveniently divided into two stages, with the first stage comprising the selection of the overall combination of resources to be utilized by the household, and the second stage concerning the application of these available resources in specific productive activities. Clearly, these two levels of decisionmaking are closely related, but there exist a number of factors that constrain the ability to rapidly adjust the overall resource base. So within the short term (a single agricultural production cycle), a household may take its available land, labor, and capital resources as given and choose the production activities that most effectively utilize its given resource base. Over time, however, households may pursue strategies to alter their resource endowments. Household labor availability depends on the dynamics of household demographics. Household labor supply increases as children attain the age at which they begin to assist in agriculture and domestic activities, and it decreases as family members leave to form their own households. At any given point in time, however, the pool of available household labor is fixed by its number of able-bodied individuals. At the same time, alternative labor allocations are determined by the level of “complementary” household resources, such as land. For example, if a household has more land than can be properly exploited with available family members, the household must either contract nonfamily labor or reduce the amount of land farmed directly by the household by transferring it to others through sale, indirect tenure contracts, or bequest. Conversely, a household with more available household labor than can be gainfully employed on its farmland will strive to acquire more land, or find alternative employment for household members. The demand for agricultural land depends on its economic importance within household management strategies. Two distinct alternatives can be identified. On the one hand, rainfed land may be considered an agricultural input that is highly complementary with labor. Access to rainfed land provides employment to available household labor, so households with more workers will have a greater demand for agricultural land. On the other hand, if the imputed cost of labor is subtracted from the expected value of rainfed production, the remaining profits attributed to the land are very

Household Management Strategies

133

low. Thus, any capitalist wishing to invest in only rainfed land and paying all of the production costs including hired labor would not realize any significant returns to such an investment. Land is thus more valuable to farm households with available labor that can be usefully employed on the land. Such households benefit not from the returns to the land per se, but from the returns to family labor employed on the land. In this case, we would expect to observe a positive correlation between the available household labor supply and the quantity of rainfed land farmed. In contrast, irrigated land, in addition to being an agricultural input, has significant value as an asset. Ownership of irrigated land provides substantial economic returns, even if labor costs must be paid in cash. A household with sufficient money would have incentives to buy irrigated land, even if labor must be hired to work the land. Thus, the ownership of irrigated land can be expected to be more directly related to a household’s existing wealth but not to the number of workers in the household. Patterns of bequeathing agricultural land reflect these differences in the demands for rainfed and irrigated land. On the one hand, when adult children leave the household, they are usually given a plot of rainfed land. This institutional form tends to maintain the relationship between available household labor and land: the amount of land farmed in the parents’ household is reduced in conjunction with the loss of household labor services, but at the same time land is provided for the labor services in the newly formed household. On the other hand, irrigated land tends to be held by the head of the household until his or her death, and the land will usually be bequeathed to a single child who takes over control of the homestead. Furthermore, irrigated land is more likely to be sold, rented, or sharecropped to individuals outside the family. As we saw in the previous chapter, on Santiago, less than 30 percent of irrigated land is owned by the households that actually farm the land.1 Our empirical results tend to support the hypothesized relationships between household labor supplies and the two types of agricultural land. The arguments outlined above suggest that the amount of rainfed land should be positively related to the household supply of available labor, whereas irrigated land should not show a strong relationship with household labor. Table 7.1 shows that the amount of rainfed land per household increases with the number of adults per household on both Santiago and Santo Antão. However, the increase in land is not proportional to the increase in number of adults, so the larger households have less land per adult. In the case of irrigated land, there is no clear pattern of the amount of land farmed in relation to the available labor in the household. Table 7.2 (p. 135) reports results of regression models relating the number of adults and other household characteristics with the amount of rainfed and irrigated land farmed per household. Three variables are

134

Household Management Strategies

Table 7.1

Distribution of Rainfed and Irrigated Land by Number of Adults in Household

Santiago Rainfed No. of Adults 1 2 3 4 >4 Total

% of Sample 13.5 27.4 14.8 19.9 24.4 100.0

Irrigated

Per Household

Per Adult

7.2 11.6 12.5 13.5 16.4 12.7

7.2 5.8 4.2 3.4 2.8 4.5

Per Household

Per Adult

0.2 0.9 1.3 0.6 1.0 1.0

0.2 0.5 0.4 0.2 0.2 0.3

Santo Antão Rainfed

Irrigated

No. of Adults

% of Sample

Per Household

Per Adult

Per Household

Per Adult

1 2 3 4 >4

10.8 19.3 17.0 19.9 33.0 100.0

8.3 13.2 11.6 18.9 22.3 16.5

8.3 6.6 3.9 4.7 3.7 4.7

3.3 6.4 8.2 11.5 8.1 8.1

3.3 3.2 2.7 2.9 1.3 2.4

Total

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

included to explain variations in rainfed and irrigated land farmed across surveyed households. The number of adults represents in gross terms the available stock of labor services within the household. In addition to the number of adults, the age of the head of the household was included to capture the possibility of time lags for households to acquire their desired quantity of farmland. For a given number of adults, an older household may have more land simply because it has had more time to acquire the land. The sex of the household head was also included as an explanatory variable to see if female-headed households have less access to land, after accounting for number of adults and age of the household head. The regressions for rainfed land are highly significant for both Santiago and Santo Antão, although the proportion of variation in the land area explained by the independent variables (the adjusted R2) is not very high. In both cases, the number of adults is the only independent variable that is significantly different from zero; the age and sex of the head of household do not explain variation in the amount of rainfed land farmed. While the coefficients on number of adults are significant, they are also quite small in absolute magnitude. On Santiago, an additional adult in the household corresponds to slightly more than 1 additional liter of rainfed land farmed. On Santo Antão, the figure is higher at about 3.6 liters of

135

Household Management Strategies Table 7.2

Rainfed and Irrigated Landholdings

Dependent Variable (liters) Santiago Rainfed Land Irrigated Land Santo Antão Rainfed Land Irrigated Land

Constant

Number of Adults

Age of Sex of Adjusted Household Household R2 Head Head

3.30 (1.73) 1.38 (1.80)

1.33 (4.30***) –0.02 (–0.19)

0.06 (1.73) –0.01 (–0.66)

2.29 (0.25) 21.83 (2.60*)

3.59 (4.17***) 0.93 (1.50)

0.10 (0.79) –0.13 (–1.24)

2.57 (2.25**) 0.09 (0.20) –4.56 (–0.79) –10.93 (–2.42*)

0.16

F Value

15.70***

0.00

0.18

0.08

5.82***

0.06

2.87*

Source: authors’ calculations. Notes: * Coefficient significant at the .10 level; ** coefficient significant at the .05 level; ***coefficient significant at the .01 level.

rainfed land per additional adult in the household, which may reflect the tendency to employ more hired labor on that island. Referring back to the crop budgets, 1 hectare of corn and beans requires approximately 20 person-days per month during the period of peak labor demand (planting and weeding), so one worker could potentially cultivate 1 to 1.5 hectares (10 to 15 liters) of rainfed land. Thus, both the empirical values are quite low in relation to the potential amount of additional land that one more worker could farm, which may reflect some “stickiness” due to the overall scarcity of land. In the case of irrigated land on Santiago, there is no relation between household demographic characteristics and the amount of land farmed. The equation is not significantly different from zero. The results differ somewhat on Santo Antão, where the sex of the head of the household is slightly significant, and show that female-headed households have more land. This unexpected result may be explained by the common phenomenon of households in which the husbands have emigrated. Their incomes often provide significant wealth that in turn provides means to acquire irrigated land, which the wife manages. As hypothesized, the number of adults is not significantly correlated with irrigated land on either island. Age of the head of the household is not significant in any of the equations, so there appears to be no standard dynamic of household acquisition of land over time, other than the correlation with changes in household size. The coefficients on the sex of the head of the household are not statistically significant for either type of land except, as noted above, for irrigated land on Santo Antão. These results suggest that the much smaller landholdings among female-headed households is not intrinsically related

136

Household Management Strategies

to the gender of the household head but, rather, is explained by the fact that they have fewer working-age family members than do male-headed households. With respect to the first, or longer-run, resource-management objective of acquiring access to the appropriate combination of agricultural resources to maximize returns from household labor, existing institutions do provide the flexibility for households to adjust rainfed landholdings to their available labor supply to some extent. However, the absolute shortage of land prohibits most households from acquiring sufficient land to fully utilize their available labor services. Irrigated land, in contrast, is even scarcer. Since such land is an extremely valuable asset, households generally do not give it up, even in the case of insufficient household labor supplies. In fact, many owners of irrigated land do not farm the land directly themselves at all, but either lease it out or employ hired labor. This analysis of landholdings and family size suggests strongly that family labor is underemployed in agriculture. Since most households simply cannot fully utilize the labor services of the family on their small plots of land, they often seek alternative ways to gainfully utilize their labor in addition to farming the land available to them. Access to off-farm employment has become a critical factor for most families and seems to gain in importance as pressure on land increases. Rural households in Cape Verde can allocate household labor to three general categories of activities: rainfed agriculture, irrigated agriculture, and off-farm employment. Table 7.3 gives the breakdown of the average amount of time per family spent in each of these activities. In order to compare this with available labor supplies, some assumptions must be made with regard to the amount of time that household members are willing to devote to labor. Suppose that all individuals between the ages of 15 and 65 are willing to provide 2,500 hours (equivalent to 50 hours per week for 50 weeks) per year to labor services. This is a conservative estimate of available family labor services, because many individuals both younger and older than the reference ages perform agricultural tasks. Average quantities of available household labor based on these assumptions are given in Table 7.3. The amount of available labor is substantially greater than the aggregate of utilization in the three categories, revealing the degree of excess family-labor resources. This imbalance is the result of the small amounts of land available to families and the limitations of off-farm employment. Demand for labor in agricultural activities exhibits large seasonal fluctuations, with marked peaks at particular points in the growing cycle. This characteristic of agricultural-labor demand explains the observation in many parts of the world that labor supplies are fully utilized at the peak demand periods.2 In Cape Verde, peak demands occur in June and July for land preparation and seeding of rainfed crops. These activities must be

137

Household Management Strategies Table 7.3

Average Household Labor Requirements and Availability, Annual and Peak Labor Season (in hours)

Labor Use Characteristics Agricultural labor requirements Rainfed Irrigated Off-farm employment Available household labor Average surplus available labor % households with surplus available labor Peak labor requirements Rainfed Irrigated Available monthly household labor % households with surplus available labor: Peak labor months Rainfed Irrigated

Santiago

Santo Antão

1,310 1,260 50 1,165 8,321 5,846 97

2,876 1,653 1,223 2,374 9,091 3,841 79

302 22 699

397 317 764

90 100

85 80

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

completed before or immediately after the first rains in order to take full advantage of the available moisture. Weeding, which takes place approximately one month after planting, is another peak demand time. In irrigated agriculture, land preparation at the beginning of the irrigation cycle is the most prominent labor peak. However, labor can be more easily spread over time in irrigated crops by staggering planting dates. Peak labor demands for irrigated and rainfed crops were calculated for the farms in the Santiago and Santo Antão surveys on the basis of the crop budgets and representative systems presented in Chapters 5 and 6. As shown in Table 7.3, almost all households have excess available labor supplies even in months of peak labor demand. Only 10 percent of households on Santiago were estimated to have less household labor than needed in the peak demand months. Peak labor demands in irrigated agriculture did not exceed available family supplies for any households in the survey. These results help to explain the infrequent use of hired labor in rainfed agriculture. Statistical estimates based on an economic model of household labor allocation decisions support the conclusion of excess labor supplies in most rural households. The results are reported in Table 7.4, and the model is described more fully in the appendix to this chapter. Briefly, the regression model measures the relationships among factors that affect demand for household labor (amount of rainfed and irrigated land, local availability of off-farm work), as well as the overall supply of household labor (number of adults in the household) and how these factors influence household allocations among the three categories of labor—rainfed and

–820.02 (–1.43) –22.81 (–0.06) 388.36 (1.21)

Off-farm labor

Irrigated labor

Rainfed labor

Constant

Dependent Variable

Rainfed labor

–1231.25 (–2.81**) –618.24 (–2.08*)

Constant –4.16 (–0.27) 44.21 (4.27***)

Rainfed Area

1586.87 (2.54*) –569.41 (–1.36) 87.37 (0.25)

Local Employment Rate 15.05 (0.75) 43.90 (3.24**) –0.23 (–0.02)

Rainfed Area –211.36 (–1.95) –112.15 (–1.54) 167.11 (2.76)

Irrigated Area

283.21 (3.02**) 167.41 (2.65**) –14.84 (–0.28)

Adults in Family

191.28 (2.63**) 129.05 (2.62**)

Adults in Family

Households with Access to Rainfed and Irrigated Land

2944.06 (4.88***) 868.53 (2.13*)

Local Employment Rate

Households with Access to Rainfed Land Only

Regression Models of Household Labor Allocation

Off-farm labor

Dependent Variable

Santiago

Table 7.4

0.05

0.22

0.23

Adjusted R2

0.27

0.18

Adjusted R2

(continues)

2.01

6.05***

6.49***

F Value

15.92***

9.97***

F Value

–1912.31 (–2.26*) –109.94 (–0.69)

Constant

–1686.10 (–2.26*) –71.98 (–0.54) 257.51 (2.03*)

Off-farm labor

Dependent Variable

Off-farm labor

–1.09 (–0.06) 8.43 (2.45*)

Rainfed Area

4161.86 (3.63***) 200.46 (0.98) –207.38 (–1.07)

Local Employment Rate 3.79 (0.34) 3.51 (1.78) 0.58 (0.31)

Rainfed Area –66.81 (–3.38**) –6.46 (–1.83) 4.25 (1.27)

Irrigated Area

434.99 (3.19**) 67.89 (2.80**) 7.74 (0.33)

Adults in Family

572.66 (4.22***) 64.66 (2.53*)

Adults in Family

Households with Access to Rainfed and Irrigated Land

4025.64 (3.75***) 253.00 (1.25)

Local Employment Rate

Households with Access to Rainfed Land Only

0.00

0.15

0.37

Adjusted R2

0.23

0.32

Adjusted R2

Source: authors’ calculations. Note: *Coefficient significant at the .10 level; **coefficient significant at the .05 level; ***coefficient significant at the .01 level.

Irrigated labor

Rainfed labor

Rainfed labor

Constant

continued

Dependent Variable

Santo Antão

Table 7.4

0.88

4.30**

11.64***

F Value

8.33***

12.96***

F Value

140

Household Management Strategies

irrigated agriculture and off-farm work. The results indicate that labor allocations are strongly influenced by demand factors. For example, an increase of one liter of rainfed land causes an increase of labor in rainfed activities of 44 hours, both for households with and without irrigated land. An increase of 1 liter of irrigated area generates an increase of 167 hours of labor on irrigated land. An increase of 1 percentage point in the local employment rate leads to an increase of 2,430 hours of off-farm employment for families without irrigated land and 1,587 for those with irrigated land.3 Increases in the demand for one type of labor, however, do not lead to decreases in labor allocated to other activities; the coefficients of rainfed area on off-farm labor and irrigated labor are not significantly different from zero. Similarly, the off-farm employment variable has no significant coefficients on rainfed or irrigated labor. These results indicate that increases in demand for a particular type of labor are not met by offsetting reductions in household allocations to other types of labor but, rather, by increasing the total amount of time worked by family members. In particular, increasing access to off-farm work opportunities does not reduce the amount of time a household devotes to agricultural activities. Instead, the total amount of work time expands to meet the greater demands for household labor, which suggests that greater employment opportunities will tend to absorb the surplus labor in rural households. The model results also indicate that households with greater supplies of available labor increase their allocations of off-farm employment and rainfed agriculture, but not irrigated agriculture. The increased allocation to off-farm work is explained in part by the operation of the public work fronts. To meet social welfare goals, the limited number of work positions is spread among the largest number of families, and the number of individuals from one household that may work on a front is limited by the household size (one person per six household members). Since larger households are able to place more individuals in the work fronts, a positive correlation is expected between family size and off-farm employment. These results also demonstrate the impact of trade-offs between family and hired labor in irrigated agriculture. With an additional family member, a household can intensify labor concentrations in supplemental rainfed activities, for example, in a third weeding. Although it is not necessary to obtain a production of corn and beans, this activity can marginally increase yields, especially when rainfall continues later into the growing season. In irrigated agriculture, in contrast, the families with more labor available do not intensify allocations to supplemental activities, either because irrigated activities exhibit more fixed labor requirements per unit of land or because households prefer to hire in the additional labor. In effect, the regression results indicate that time spent in irrigated activities is not related to the amount of family labor available.

Household Management Strategies

141

Resource Management in Rainfed Agriculture

Within the domain of rainfed agriculture, the choices available to most farmers are very limited. The government has been long concerned to reduce the production of corn and beans on marginal lands, where these crops are thought to exacerbate soil loss due to erosion. In the context of this preoccupation, there is a necessity to understand the local motivations for continuing to grow these crops, even in the face of very unfavorable conditions and miserable yields in most years. Many officials and agricultural specialists have argued that family farmers in Cape Verde continue to plant corn and beans for reasons of tradition, even if these activities are not economically viable. Thus, it is the presumed symbolic importance of corn in rural society that supposedly motivates families to continue its cultivation even when economic returns are low or negative. The resource-management decisions of households are undoubtedly complex and affected by many factors, but there are strong economic incentives for growing rainfed corn and beans, based on three unarguable realities. First, working-age individuals within rural households have few alternative employment opportunities. Second, the outputs from rainfed agriculture comprise not only the grain for human consumption, but also residue feed for livestock; the value of this by-product—of vital interest to local farming systems—is often overlooked by researchers and policymakers. Finally, rainfed farmers (as well as researchers) have not yet found an alternative crop that provides them with similar returns and that can be successfully grown under the existing range of agroclimatic conditions found in Cape Verde. That most households have large amounts of available labor supplies with few employment opportunities suggests that many will be willing to utilize this labor in activities with low expected returns. If individuals are not able to secure employment with higher returns—such as off-farm wage work—and they do not have access to irrigated land, then rainfed agriculture provides the only means to utilize this critical household resource. In addition, the total amount of labor required over the course of a year to bring a crop to harvest is quite low, less than 130 workdays per hectare of land. Since most households have approximately 1 hectare of land, they need only devote a half person-year to the rainfed enterprise. For most households, this represents only a small portion of their available workforce. Rainfed production of corn and beans provides grain for human consumption and straw residue for animal feed. In the representative cornbean budgets, this straw by-product accounts for over one-half of the value of all output. The importance of livestock to the rainfed household cannot be underestimated, since they generate a significant share of total household income in rural Cape Verde, particularly on Santiago. Those households that

142

Household Management Strategies

do not fully utilize their feed for their own animals are still able to receive the feed value of their straw by selling it to other livestock owners. The additional benefits deriving from livestock production, as well as the low labor inputs, imply that, while the physical returns of grain per unit of land may be very low, the economic returns per unit of labor input may still be relatively attractive. Based on the representative budgets, the expected return to family labor of the production of corn, beans, and straw is CVE 150 on Santiago and CVE 185 on Santo Antão. These returns are comparable to the average daily salary of CVE 150 on public work fronts. Thus, even though the expected total returns to rainfed agriculture per unit of land are quite low, the returns to family labor employed in these activities are at least as great as what can be earned in alternative forms of employment. In fact, the expected economic benefits to employing family labor in rainfed agriculture are quite attractive, and many farmers in Cape Verde will continue to grow corn and beans as long as they do not have any better alternative uses for their family-labor resources. In terms of rainfed land, there are very few viable alternatives to corn and beans, except in limited areas with particular agroclimatic characteristics. As presented in Chapter 5, Cape Verdean farmers have experimented with the pigeon pea and have incorporated it into their farming system on the poorest, hilltop lands. Rainfed pigeon pea is now grown only in relatively arid areas, where expected corn yields are particularly low. The production of beans, straw and fuel from pigeon pea is relatively attractive in these regions, but in more humid areas and on better soils, corn and beans will produce more, while pigeon peas will enjoy greater vegetative growth but little significant increase in usable yields. Attempts to introduce sorghum and millet into Cape Verde have not met with success, mostly because the rural population has no tradition of small-grain consumption. Any programs intending to redirect agricultural activities on rainfed land must take into consideration the economic benefits that farmers accrue from current practices. Farmers are unlikely to be interested in substituting corn and beans for any alternative that does not afford economic returns at least as high as the direct and indirect (feed and livestock) returns from corn and beans. Resource Management in Irrigated Agriculture

Resource-management strategies are more varied on irrigated land, and the options available to individual farmers are much wider than on rainfed land. As indicated in Chapter 6, flexibility of access to irrigation water is the critical variable that defines the range of cropping strategies, and such access is highly dependent on the nature of the irrigation system, particularly the water source, and the overall size of the network. Wells provide

143

Household Management Strategies

a more constant flow of water (as long as groundwater is available) than catchment dams (captações), which depend on the rate of alluvial flow. Large systems with more users are generally less flexible because of the greater difficulties in forging new agreements among all the interested parties. Table 7.5 reports the average irrigation intervals for the three sources of irrigation water—hand-dug wells (poços), tubewells (furos), and springs (nascentes)—from the survey of Santiago farmers. The reported average interval between irrigations of 19 days for poços is significantly lower than that of the other two water sources. More than 20 percent of all irrigated farms on Santiago receive water from poços, and most are privately owned and managed. Some poços are cooperatively built and operated, but, again, the allocation of water among members is negotiated among themselves. The number of users in the poço systems is usually quite small (between 5 and 10 farmers), so negotiations are usually not difficult. In some cases, the members have close social ties as neighbors or relatives, which serves to minimize inequitable allocations. In other cases, however, sharecropping farmers have little bargaining power to counter the allocation decisions of the owner of the land and well. If cooperative members cannot reach agreement among themselves, they may appeal to the INGRH to intervene in the operation of the well. In effect, the poço systems appear to enjoy greater scheduling flexibility and less onerous management structures. Furos are a less important source of water on Santiago, and they generally are less flexible than poços with regard to the timing of irrigations for individual users. Although the flow rate from the wellhead is generally more stable than that from poços, most furos supply water to a larger number of users, ranging from 10 to over 60 members in a network. The networks are maintained and operated by the INGRH, which is responsible for the scheduling of water to network users. There are some variations in

Table 7.5

Characteristics of Irrigated Farms on Santiago by Source of Water Water Source

Characteristics Average irrigation interval (days) % of households with irrigated land Distribution of crop mix (%) Cane only Cane, cassava, sweet potato Banana Vegetable

Poço

Furo

Nascente

19 22

28 4

28 74

4 22 57 46

0 3 0 35

96 75 43 19

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

144

Household Management Strategies

procedures at different installations, but the general procedure is for the INGRH to establish a rotation in which every member receives a fixed number of hours within the allocation. The member may elect not to use his or her allocation and may make bilateral trades of his or her allocation with other members. The agency hires an individual, usually a network member, to operate the pump, to ensure that the rotation is followed, and to keep records of water deliveries to all network members. Water-allocation records received from a furo network on Santiago reveal that the actual water deliveries for the 1988–1989 growing season varied substantially from the nominal allocations defined by the INGRH. If the rotation had followed the nominal allocations, all members would have had the same number of irrigations over the season and would have received water at the same interval, approximately three weeks. In fact, the distribution of water to members did not follow this pattern at all. Table 7.6 divides the system users into categories on the basis of water deliveries over the years. As the table indicates, about 6 percent of the network members received water on average once per week, while almost 60 percent received water eight times or less over the course of the irrigation season and had an average interval between irrigations of more than six weeks. These variations in the irrigation patterns across members were due to two factors. First, some members did not actually use their allocations during part of the season, because they were either not present to connect their hose to the wellhead at their scheduled time, or they had no crops on their fields, or, in some cases, they were in arrears on their payments to the

Table 7.6

Distribution of Irrigation Intervals Among Farmers Within a Surveyed Irrigation Network, Santiago

Group

Irrigation Category (by no. of irrigations)

Average No. of Irrigations

Average Irrigation Interval (weeks)

No. of Farms

Total Irrigated Area (liters)

I

31

33

1

10

3

20 (57) 6 (17) 4 (11) 3 (9) 2 (6) 35

19.9 (32) 12.5 (20) 10.8 (17) 8.4 (13) 11.2 (18) 62.8

Total sample

Source: Langworthy 1989b. Note: Numbers in parentheses are percentages of total sample values.

Household Management Strategies

145

INGRH and therefore not eligible to receive water. Other network members may utilize these available hours. Second, individuals may split their allotment within a rotation and trade these portions with other members in order to increase the number of times they receive water within a single network rotation. These arrangements are negotiated between individual members as side exchanges. The pump operator complies with changes in the official rotation as long as there is agreement by both parties. Since both members in a trade increase their number of irrigations, the very skewed distribution of irrigations across network members implies that bilateral trades occurred only among a small portion of all members. There are several reasons to suppose that the possibilities and incentives for such trades are greater for individuals with larger water allocations. First, the trades provide access to water at different points in the rotation cycle, so an individual with a larger allocation has a greater stock with which to negotiate and, at the same time, faces a greater pool of other members interested in acquiring a portion of that individual’s allocation. Second, there are costs associated with dividing allotments into more irrigations per rotation, primarily the fixed downtime cost per irrigation. Within the time allotted to a member, the pump must be stopped, the previous user’s tubing detached from the wellhead, the current user’s tubing attached, fuel measured and added to the tank of the pump engine, and the tubing must be filled with water before the water actually reaches the field, which is often quite distant from the wellhead. The downtime cost is the difference between the nominal time allocation and the effective irrigation time during which the field actually receives water. The length of the downtime—usually between 10 and 20 minutes—varies according to the distance from the wellhead to the member’s field, but does not vary with the size of the allocation. Therefore, the proportional cost of downtime is less for members with larger time allocations. Increasing the number of irrigations per rotation, by trading with others, increases the total downtime per rotation. For example, with a 15-minute downtime, an individual with an allocation of four hours would only reduce his or her effective irrigation time by 7 percent if the turn is split into two separate irrigations (from three and three-quarter hours to three and one-half hours), whereas an individual with a one-hour allocation would reduce his or her effective irrigation time by one-third (from 45 minutes to 30 minutes) in a similar transaction. The combination of these fixed downtime costs and the costs of identifying other trading partners leads farmers to minimize the number of transactions necessary to meet their desired reallocations. In this situation, individuals with larger blocks of time enjoy a distinct advantage in scale. Someone wishing to trade two hours will prefer to trade with a single individual for the entire amount rather than negotiate multiple trades to

146

Household Management Strategies

make up the desired reallocation. Thus, members with larger hourly allocations have greater incentives to trade irrigation time because they face lower proportional downtime costs, and they are more likely to trade with other individuals with large allocations. The most important source of irrigation water on Santiago is natural springs (nascentes), which supply almost three-quarters of all farmers with irrigated land. The flow of water in these systems varies significantly over the course of the year and between years, and the storage capacity is very limited, so these variations translate directly into fluctuations of water available to farmers. These limitations are inherent in the nature of the networks, and they restrict the possibilities for farmers to alter their access to water through trades. The extreme variability in system flow rates imposes significant risks on making agreements for future deliveries. The greater variability and generally longer intervals severely restrict cropping choices to the more drought-resistant crops. More than 95 percent of these farmers grow only sugarcane or a combination of cane, sweet potatoes, and cassava. Almost all water on Santo Antão comes from nascentes or subterranean captações and is subject to variations in the alluvial flow of water, although some irrigation networks draw water from permanent aquifers that supply a more predictable flow. Many of the irrigation systems on Santo Antão are quite large, with several kilometers of cement or earthen canals providing water to a large number of users. In some of these largescale networks with relatively abundant and steady water flows, users are able to coordinate irrigation schedules through bilateral trades of access time. The two most important areas in which these agreements take place are Ribeira das Patas/Alto Mira and the upper reaches of Ribeira Paúl. As a result, these are also the areas where vegetable production is concentrated. The other areas of irrigated agriculture on the island have experienced secular reductions in water availability and extended intervals between irrigations as a result of low rainfall. The average interval between irrigations on the island at the time of the farm survey was five weeks. In sum, most farmers receiving irrigation water in Cape Verde are very constrained in their management strategies and must adapt their cropping choices accordingly. In contrast, a small percentage has access to more flexible water sources that permit the manipulation of allocation patterns to shorten watering interval times. Through these intranetwork negotiations, farmers can shift to high-valued crops, although such strategies tend to favor the farmers with more land and larger initial allocations.

Household Income Patterns

In order to evaluate the impacts of household resource-management strategies, we estimated agricultural incomes for each of the households in the

Household Management Strategies

147

two island surveys using the representative crop budgets described in Chapters 5 and 6. The surveys provided information on the amount of rainfed and irrigated land to which each household claims use rights through ownership, rental, or sharecropping and on the size and composition of the family livestock herd. Net revenues per unit of land and per animal were multiplied by the total amount of land area and number of animals, then adjusted for the type of land tenure. No explicit land costs were attributed to owned land, so net revenues may be interpreted as family income from agricultural activities. For sharecropped land, 50 percent of the value of the output was attributed as land cost,4 and rented land was attributed a rental rate that falls within the official range. More productive land was charged a higher rental rate. All rainfed land was attributed the net revenues of the corn and beans crop budgets presented in Chapter 5. Although there are other crops grown on rainfed land, corn and beans are by far the most widespread rainfed crops, and other rainfed crops are limited to specific geographic areas with special agroclimatic conditions. Households with irrigated land were categorized into 1 of the 10 representative cropping systems described in Chapter 6, and the net revenues of the appropriate systems were applied to the irrigated land area of the households. Because the net revenues are based on the representative crop budgets, they reflect expected incomes under “normal” growing conditions and price levels. The crop budgets are based on expected yields, and the irrigated systems incorporate the additional assumption that water is provided on adequate intervals to generate the productivity levels specified in the systems. These results, therefore, do not reflect the actual conditions of the years covered by the surveys—years with specific rainfall patterns, pest and disease problems, and so forth. The surveys did not provide reliable information about the productivity of irrigated crops, but it is possible to compare actual reported productivity levels for corn and beans with those defined in the representative crop budgets. On Santiago, medians of the reported corn and bean yields were 40 and 20 percent, respectively, of the levels in the representative crop budget. Reported yields were slightly higher on Santo Antão, with the median corn yield equal to that used in the crop budget and the bean yield at 60 percent of the level used in the crop budget. The crop years covered by the two surveys were generally considered to be particularly bad because of poor rainfall patterns, so these results are not surprising. For similar reasons, the actual net revenues from irrigated activities in the years of the survey were significantly below the levels of the representative systems. These differences serve to underscore the characteristic year-to-year risks faced by Cape Verdean farmers. We estimated income from livestock by attributing to each animal the annualized net revenues described in Chapter 5. All livestock feed produced on the farm (straw from corn, beans, and sugarcane) was valued as

148

Household Management Strategies

an output at the going market rate for the straw, while the total feed requirements of the animals were included as costs. Thus, the feed that was both produced and utilized on the farm was netted out, neither valued as straw, nor as a feed input. Those farms with excess straw production generated net household income by the amount of the excess. Conversely, net purchasers of animal feed had their costs increased by the amount of the feed deficit. Finally, off-farm labor was attributed a wage rate that corresponds to the going rate on the public work fronts. The total number of days of offfarm work by all household members was multiplied by a wage rate of CVE 150. This may somewhat underestimate off-farm earnings, as some household members were actually engaged in higher-paying agricultural labor or commercial activities. The results of the household income calculations are given in Table 7.7. The average total income per household on Santiago was CVE 67,900, while on Santo Antão the figure was CVE 166,700 per household, a difference of 145 percent. However, the distribution of income among households was highly skewed, particularly on Santo Antão. As perhaps more appropriate descriptors, the median levels of income per household were CVE 61,500 for Santiago and CVE 94,500 for Santo Antão. The samples from the two islands were divided into quintiles on the basis of total household income, which provides a more detailed perspective of the distribution of incomes across households. On Santiago, the poorest 20 percent of the sampled households received only 6 percent of the income of the entire sample, and the 20 percent of households with the highest incomes accounted for 42 percent of the total sample income. The average total income per household of the richest quintile, CVE 156,000, was seven times the average income of the lowest quintile. On Santo Antão, the share of the lowest quintile was only 2 percent of the total income, and the highest quintile received 70 percent of the total. The average income of the highest quintile was almost 30 times higher than that of the lowest quintile. Average incomes were higher on Santo Antão across all five groups. The income of the first quintile was 45 percent higher on Santo Antão than on Santiago, while the ratio of the incomes for the fifth quintile on Santo Antão compared to Santiago was almost six to one. The discrimination of total household income by source reveals significant differences across income groups and between the two islands. On Santiago, the poorest families derived most of their household income from rainfed land and livestock because of their lack of access to irrigated land and off-farm employment. Only 9 percent of the families in the poorest quintile had access to irrigated land. In this group, 17 percent of the families had off-farm employment, as compared with 60 percent in the sample as a whole. Average off-farm income per household was only 6 percent of the average of the entire sample.

149

Household Management Strategies Table 7.7

Household Income on Santiago and Santo Antão by Quintile

Santiago % of total income Total income per household (CVE) % shares Rainfed Livestock Irrigated Off-farm Household size (no. of members) Total income per capita (CVE) % households below poverty level % households with emigration Santo Antão % of total income Total income per household (CVE) % shares Rainfed Livestock Irrigated Off-farm Household size (no. of members) Total income per capita (CVE) % households below poverty level % households with emigration

Quintiles Entire Sample

1

2

3

4

5

100

6

12

18

24

38

67,908

22,109

41,710

62,576

82,824

131,246

37 31 5 27

53 37 1 9

42 36 2 20

33 30 4 33

32 27 6 35

24 26 12 38

5.7

4.2

4.7

5.6

6.3

7.6

14,062

7,852

12,049

13,490

17,434

19,807

37

71

48

37

28

1

56

46

41

55

61

75

Quintiles Entire Sample

1

2

3

4

5

100

3

7

11

19

61

166,718

28,603

58,503

92,001

157,181

499,440

26 15 25 34

42 26 6 26

27 14 15 43

24 16 16 43

21 10 24 44

15 9 63 13

6.8

4.8

6.5

7.1

8.3

7.5

28,176

8,882

11,457

16,439

22,147

82,293

34

71

71

25

3

0

46

34

49

44

46

60

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

Even though rainfed and livestock activities were the predominant income sources in the lowest quintile, this group had significantly less access to these resources in comparison with the rest of the sample. The average amount of rainfed land per household was much smaller than in the higher groups, approximately one-half of the sample average. Although livestock represented a large share of total family income of the lowest

150

Household Management Strategies

quintile, only 20 percent of the households in this category had cattle; most of the households had only lower-valued pigs, goats, and chickens. The intermediate income groups (the second, third, and fourth quintiles) exhibit a much higher share of off-farm income as a proportion of total family income. The relative shares of incomes from rainfed and livestock activities were less than those of the first quintile, but remained significant. Rainfed agriculture accounted for 30 to 40 percent and livestock approximately 30 percent of total income in these groups. The amount of rainfed land per household was considerably greater than that in the first quintile, as was the share of households with cattle. However, the contribution of irrigated agriculture to total income remained relatively small in these intermediate groups, even though the share of households with irrigated land increased in the richer quintiles. Only in the highest quintile did irrigated agriculture provide a significant share of total household income. More than 40 percent of the households had irrigated land. However, off-farm income was an important source even for these richest families, accounting for over 30 percent of total household income for the group. The relative importance of rainfed land and livestock was much lower for this group: the two categories together represented 40 percent of household income. Thus, on Santiago, there is a definite shift in the relative importance of different income sources as total household income increases. The poorest households depended most heavily on subsistence-oriented activities, namely, rainfed agriculture and small livestock. The intermediate group of families exhibited an increasing reliance on off-farm employment, although the absolute incomes from agricultural activities also increased relative to the poorest quintile. Irrigated agriculture only contributed significantly to the income of the richest 20 percent of the surveyed families. However, this group also received a significant portion of total income from off-farm endeavors. The breakdown of household income by source exhibits quite different patterns on Santo Antão. First, the relative importance of rainfed agriculture and livestock was less than that on Santiago for all income groups, although the general pattern of declining reliance on these subsistence-oriented sources at higher income levels is also evident here. Irrigated agriculture was more important in all income groups, but especially in the three highest quintiles. Off-farm income was an important income source for the poorest households, in sharp contrast with the pattern on Santiago. Finally, the highest income group depended predominantly on income from irrigated activities. The share of off-farm income was negligible for this group, suggesting the presence of a small, but high-income, class of irrigated farmers. Because household income provides the means to acquire material goods for all household members, the welfare of the individual household

Household Management Strategies

151

members depends not on the total income that the household earns, but on the amount of income available for each member. Thus, a more meaningful measure of individual welfare is per capita income, which is calculated by dividing total household income by the number of family members. This figure provides an indication of the resources available to provide for each family member. As with total household income, per capita income was very unequally distributed in the samples of surveyed families on Santiago and Santo Antão. Available income per person can be compared with an estimate of the minimum level of expenditures necessary to sustain an individual over a year. This minimum expenditure, or absolute poverty level, was calculated for Cape Verde by estimating the financial cost of a normal annual diet of the main food staple of corn and beans and deriving an estimate of minimum expenditures for other necessities, such as clothing, fuel, and health products. The estimated per capita consumption of corn and beans was obtained from household calculations of the necessary portions needed to make one serving of cachupa. Minimum required expenditures for other food and nonfood items were calculated from a survey of household expenditures administered to 167 families on Santiago in 1989. Household per capita expenditures were broken down into the following expenditure categories: food items other than corn and beans, clothing, health products, fuel, transportation, and miscellaneous. The average of the lowest 20 percent of per capita expenditures for each of these categories was added to the estimate of minimum staple food requirements to obtain a basic needs per capita expenditure level. The resulting figure of CVE 10,000 (U.S.$125) per year therefore represents the amount of money necessary to meet one person’s minimum expenditure requirements in rural Cape Verde. A large portion of the households on both Santiago and Santo Antão had per capita incomes below this measure of absolute poverty. On Santiago, 37 percent of the households did not have sufficient incomes to meet the minimum expenditure requirements, while 34 percent of the Santo Antão households did not meet this standard. Moreover, a large portion of the households on both islands had per capita incomes that hovered just above this level of absolute poverty. Approximately one-third of all residents in rural areas simply do not have access to adequate incomes to be able to provide all of their subsistence needs, and they must make therefore difficult trade-offs to meet their most pressing material requirements. These individuals are in situations of chronic food insecurity. The extreme degree of variation in household incomes, both at the household and the per capita level, is explained by differences in access to resources and employment opportunities. Table 7.8 reports the amount of rainfed and irrigated land, number of livestock, and hours of off-farm work by quintile of per capita household income on the two islands. The

152 Table 7.8

Household Management Strategies Characteristics of Surveyed Households on Santiago and Santo Antão by Per Capita Income Quintile

Santiago Rainfed land (l) % households with irrigated land Irrigated land (l) Goats Pigs Cattle % households with cattle % households with off-farm work Off-farm work (hours) Santo Antão Rainfed land (l) % households with irrigated land Irrigated land (l) Goats Pigs Cattle % households with cattle % households with off-farm work Off-farm work (hours)

Quintiles Entire Sample

1

2

3

4

5

12.7

6.8

10.2

11.5

16.4

18.7

24 0.24 1.8 1.3 0.9 52

6 0.02 0.7 0.7 0.3 22

6 0.06 1.4 1.1 0.7 42

34 0.17 1.9 1.3 0.8 46

32 0.25 2.1 1.3 1.3 69

45 0.73 2.7 2.2 1.6 82

61 1,165

19 69

47 478

71 1,153

77 1,456

88 2,724

3

4

5

Quintiles Entire Sample

1

16.5

7.0

9.4

12.8

19.5

34.1

52 4.2 2.6 1.3 0.4 11

29 0.5 1.3 0.7 0.0 3

46 1.4 2.1 1.1 0.0 3

39 1.3 2.2 1.3 0.1 8

57 3.6 3.0 1.7 0.2 20

91 14.3 4.3 1.5 1.6 22

64 2,374

37 581

2

77 1,673

67 2,656

86 4,527

54 2,425

Sources: Finan and Belknap 1985; Langworthy, Finan, Varela, and Rodrígues 1986.

data reveal that households with higher per capita incomes had significantly greater access to both rainfed and irrigated land and had more cows, although the difference in smaller livestock was not very great between the income groups. On Santiago, access to off-farm work was highly correlated with per capita household income. An important reason for the low income level of the poorest households was their lack of off-farm employment. The richest quintile, although relying less in relative terms on off-farm income, had the greatest access to jobs away from the farm, both in terms of the share of families who had employment and of the average number of hours per family. On Santo Antão, the same general pattern can be seen in the first four categories of per capita income, with increasing access to off-farm work in the higher income categories. Only in the highest category of per capita income does this trend reverse, with a fall in average number of hours worked and the percentage of households with off-farm

Household Management Strategies

153

employment in comparison with the previous category. Thus, those households with greater levels of income available per member have more agricultural and livestock resources and more access to off-farm work. Because most households have more workers than can be fully employed on their available land, off-farm work is not a substitute for farming, but provides another means to generate additional per capita income. In summary, the distribution of incomes is very unequal on both Santo Antão and Santiago, with the majority of households having low incomes in both relative and absolute terms. The variation in rural household incomes is explained by the very skewed access to agricultural resources, most importantly irrigated land, but also rainfed land. In addition, having more land permits families to maintain more livestock, especially cattle. Access to off-farm work is also correlated with total household income, with the exception of the richest families on Santo Antão.

Conclusion

Farmers in Cape Verde face many different kinds of constraints that restrict the ways in which they can use the resources at their disposal. Most important, the overall scarcity of agricultural resources, combined with the very unequal distribution of these resources among farm families, implies that most households operate with access to a very limited resource base. Some households, particularly on Santo Antão, have large amounts of land that provide them with substantial incomes, but most households have very restricted amounts of agricultural resources. The skewed distribution of these resources implies that the vast majority of the rural population has extremely limited access to agricultural land, both rainfed and irrigated. Moreover, the agricultural resources available to households are scarce in both an absolute and a relative sense. In absolute terms, the lack of access to agricultural resources means that most households fail to balance incomes with family consumption requirements. The extremely limited access to land also implies that the demand for agricultural labor is significantly less than the available household supplies. Off-farm work, both locally and overseas, is important as a means to more fully employ family labor and thereby provide additional income. However, access to off-farm employment is also very restricted in comparison with households’ available supply of labor in excess of the requirements for their own agricultural activities. This condition of surplus household stocks of labor is confirmed even when peak seasonal agricultural demands are taken into account. With regard to constraints on agricultural activities, the options available for rainfed land are very limited. So, even though the economic

154

Household Management Strategies

incentives to produce corn and beans are not large, the costs from the perspective of individual households are also low, and farmers prefer the meager returns to these activities, with the possibility of much greater returns if rainfall is adequate, to nothing at all. If the government objective is to reduce the production of corn and beans in order to slow erosion, policymakers must be aware of the benefits that farmers perceive from corn and beans and ensure that alternatives are more attractive. Most farmers who receive irrigation water are also severely constrained by restrictions in access to water over the course of the growing season. The limited and variable flow rates of most water sources severely weakens the flexibility with which individual farmers can manage the timing of their access to water. The vastly greater incomes generated by the few farmers who are able to irrigate more frequently highlight the potential gains that could be obtained by improved management of irrigation water resources. The extreme constraints that rural households face, and the resulting impact of very low incomes, accentuate the need for public policies to address these problems and try to remove some of the constraints, particularly for the poorest households. Chapter 8 examines the various public agencies that are involved in the agricultural sector and evaluates possible alternatives for reducing some of the constraints identified in this chapter. Most of these alternatives involve changes in management or utilization of agricultural resources, particularly land and water.

Appendix

Households must decide how to allocate their available labor services to rainfed and irrigated agricultural activities and to off-farm work. Economic theory suggests that individuals within a household allocate their time between work and leisure on the basis of the additional income received from an additional unit of work as compared with the utility derived from allocating that same time to leisure. Individuals will have different preferences, but theory predicts that the amount of time allocated to labor will increase as marginal returns to labor (the wage rate) increase. Also, for given marginal returns to labor and leisure, the quantity allocated to labor will increase as the marginal utility of income increases. Since the marginal utility of income decreases with the level of income (an additional dollar of income provides more additional utility to a poor family than to a rich one), poorer families are expected to allocate more of their available time to labor, all else being equal.

155

Household Management Strategies

Consider a model in which households allocate labor to (1) rainfed agriculture, (2) irrigated agriculture, (3) off-farm employment, and (4) leisure, according to the following equations: DR(r,AR) + DI(r,AI) + O(w,r,Z,N) = SW (r,w,N) SW (r,w,N) + DL(r,w,N) = TL(N)

(1) (2)

where DR is the demand for labor in rainfed agriculture, DI is the demand for labor in irrigated agriculture, O is the household allocation to off-farm jobs, SW is the household supply of labor for all types of work, DL is the household demand for leisure time, TL is total time available in the household, r is the implicit return to household labor in agricultural activities, w is the off-farm wage rate, and Z is a variable that measures the degree of access to off-farm employment. Figure 7.1 provides a graphical representation of these equations. A given household will have a total demand for labor function similar to those shown in panels 4a and 4b, which is the aggregation of the demands for labor in rainfed and irrigated agriculture and in off-farm employment. These demands, in turn, are determined Figure 7.1

Household Labor Demands and Supply

156

Household Management Strategies

respectively by the amount of rainfed land, the amount of irrigated land, and the off-farm wage rate and the access to off-farm employment available to the household. The supply of labor depends on the number of people in the household, and the slope of the labor supply function indicates the willingness to increase labor time at the expense of leisure time. A household with excess labor can be characterized by a very flat labor supply curve within the relevant region, as shown in panel 4a. This signifies that the household is willing to provide significantly more total labor, by reducing leisure time, for a small increase in the return to labor. If a household experiences an increase in demand for labor, such as a shift in access to off-farm employment from QO0 to QO', the corresponding outward shift in the total demand curve will elicit a small increase in the marginal returns to household labor, from r0 to r'; total household labor supply will increase from QT0 to QT'; and only small reductions in allocations to agricultural activities will be necessary to maintain balance between supply and demand. In contrast, a steeper labor supply function, as shown in panel 4b, implies that a significant increase in returns to labor, from r0 to r', and a smaller increase in total household labor supply, from QT0 to QT'', would result from an increase in access to off-farm work. In this case, larger reductions in labor allocations to agricultural activities would be required to meet the increase demand for off-farm work. The model also predicts differential responses depending on whether or not the quantitative constraints on off-farm labor are binding. If the household supply curves intersect the horizontal portions of the total demand curves in the fourth panels, the quantity constraints on off-farm employment are not binding, and the amount of household labor allocated to off-farm employment will be less than QO0. Outward shifts in the household labor curve (due to an increase in number of working-age members) will be allocated entirely to off-farm employment, and there will be no change in agricultural labor allocations. If the supply curve intersects to the right of the horizontal portion of the demand curve, as shown in Figure 7.1, the quantitative constraints on off-farm employment are binding, and an increase in labor supply would result in greater labor allocations in agricultural activities. This is true unless the household supply curve is perfectly elastic (horizontal) at the point of intersection with the demand curve. In this case, outward shifts in the household labor supply curve would not affect any labor allocations, and the additional time would be allocated to leisure. Econometric estimates of the reduced forms of equations (1) and (2) were estimated using ordinary least squares: DR = F1(AR,AI,Z,N) DI = G1(AR,AI,Z,N) O = H1(AR,AI,Z,N)

(3) (4) (5)

Household Management Strategies

157

Theoretically, the off-farm wage rate, w, should also be included, but we did not have this information for each household. Because wage rates are quite uniform throughout the islands and the information is available only in a cross section, this omission is not considered to be serious.

Notes 1. The proportion of irrigated land owned by the operator is significantly higher on Santo Antão. However, on this island, owners of larger properties tend to use hired labor rather than sharecrop or rent out their irrigated land. 2 . See Netting (1993:102–122) and Stevens and Jabara (1988) for descriptions of particular case studies of seasonality in labor demands. 3. All of these coefficients are significant at the .01 level. 4. In the case of sugarcane, processing costs are not included as sharecroppers’ costs, since they are paid by the landowner.

8 Agriculture and Policy Choice

We now shift attention away from the conditions and activities of rural households in Cape Verde to examine the role of government policies directed toward the agricultural sector and rural populations. To begin, however, it is perhaps useful to retrace our steps. We have first documented scarcity of natural resources for agriculture, both at a national level and from the household perspective. Under current population pressure, the existing resource base is utilized to maximum intensity levels, and productivity is not only very low, but also highly uncertain. By examining resource-management patterns, we have sought to uncover the inherent logic that drives household strategizing in both rainfed and irrigated agriculture, and we have arrived at the conclusion that there are economic incentives to farm—strong incentives in the case of irrigated agriculture, but also positive incentives for even the marginal rainfed lands. Local farming systems, though technologically limited, are nonetheless complex in the manner that households combine agriculture, livestock, and off-farm employment to secure a livelihood. At the same time, our investigation has revealed widespread poverty and severe inequalities in the rural zone, as well as a dynamic sector of irrigated farmers. Within this reality, we juxtapose the virtual impossibility of national self-sufficiency with the potential for sustainable growth in irrigated production. One key element that may direct the trajectory of Cape Verdean agriculture is the state. Through its policies, the state can invest in agriculture in ways that are beyond the means of private groups and individuals, and it can direct signals—both negative and positive—to rural households that influence their access to productive resources and allocative decisions. Other policies may promote the flow of trapped labor from agriculture to other sectors. As we argued in the introductory chapter, the nature of constraints in Cape Verdean agriculture requires a strong policy presence in order to reduce the ecological imbalance between the demand for and the availability of productive resources. It is this policy dimension that we now address. 159

160

Agriculture and Policy Choice

From independence until the defeat of the PAICV political party, the Cape Verdean government placed major emphasis on agricultural development and undertook a wide range of policies and investments to increase the productive capacity of the sector. However, the overall impacts of these government activities on the long-term growth of the agricultural sector have been very limited. Yields remain extremely low by international standards, and rainfall is still the critical factor determining the level of aggregate agricultural production in any given year. The lack of significant advances in agricultural productivity, combined with a growing population, has necessitated increasing reliance on foreign aid to meet national food requirements. At the same time, the public work fronts, also funded by foreign aid, have been transformed from temporary interventions designed to provide subsistence incomes in rural areas during times of crisis into seemingly permanent programs that are vital to the economic wellbeing of many rural households. The failure to meet the long-term goals for stimulating agricultural growth, along with the dependence on donor food aid and stopgap income support programs, indicates the need for a reevaluation of agricultural policies in Cape Verde. Past performance calls for a more realistic appraisal of what is feasible within the severe physical constraints imposed on agriculture and for the identification of policies more likely to be successful in achieving these national goals. We draw upon the analyses presented in the previous chapters to examine the possible impacts of alternative policies on rural households. First, however, we summarize the major policy goals of, as well as the principal policies currently undertaken by, the Cape Verdean government. Alternative policy options will be evaluated in terms of their potential to meet stated objectives.

Policy Objectives

Agricultural policies implemented in Cape Verde are intended to pursue three general objectives.1 The first major goal related to agriculture has been to reduce the domestic food deficit—that is, the imbalance between domestic production and consumption of basic foodstuffs. The rapidly growing population, combined with the current drought, has led to the dependence on imports to meet 80 percent of domestic food needs in recent years (Soares 1984). Presently, food aid accounts for the vast majority of all food imports, including the basic necessities—cereals, edible oils, and processed dairy products. The government is naturally very concerned about continued dependence on this ephemeral source of survival, since the food deliveries are determined by the donor countries and agencies and are often decided by political forces largely outside the control of Cape

Agriculture and Policy Choice

161

Verdean policymakers.2 In order to reduce this dependency, the government looks to the agricultural sector to expand production of domestically consumed foodstuffs or to generate export earnings that can be used to purchase commercial food imports. Because virtually all agricultural land is already under cultivation, further increases in food production can only be accomplished by raising yields per unit of land. The second goal has been to improve the economic well-being of the two-thirds of the national population that is involved in agriculture. As Chapter 7 clearly demonstrated, rural poverty is a serious problem on both Santiago and Santo Antão, with large portions of the population at or near the minimal subsistence level. In addition to concern about the welfare of these individuals, public officials are also worried that rural poverty will escalate the rate of rural-urban migration, which already places alarming pressures on the urban infrastructures and social services and exacerbates urban unemployment problems. Improving agricultural incomes is thus seen as a means to slow the flow of migration to the towns. In the years immediately after independence, the agricultural sector was singled out as a major area where efforts should be focused to increase the productive capacity. Large investment projects funded by foreign donors were initiated on the major agricultural islands of Santiago, Santo Antão, Fogo, and São Nicolau. These projects covered a range of activities, from constructing and improving irrigation and erosion-control infrastructures, to supporting applied research and extension efforts, to implementing integrated development schemes that combined all these efforts in more limited geographical areas. Although these activities continue, often based on the initiative of the donor organizations, the economic development strategies of the government have begun to place more emphasis on other sectors of the economy (particularly, tourism and export-oriented light industry and services) as planners have become increasingly aware of the extreme challenges in increasing the productivity of agriculture in Cape Verde. At the time of independence, there was also very great concern to redress the unequal access to agricultural land in the country. It was thought at the time that redistribution of land could increase productivity as well as provide a more equitable distribution of income among households. Landreform legislation defined conditions under which lands could be expropriated (normally from absentee owners) and attempted to control the sales and rental of land. However, in recent years, the government has not actively pursued land redistribution, although there is still a monitoring of land transactions, as well as lawsuits between tenants and landowners that continue to appear in the courts. The third agricultural policy objective has been to retard the rate of environmental degradation, particularly soil erosion caused by agricultural

162

Agriculture and Policy Choice

activities. In Cape Verde, the motivations for environmental concern are fundamentally to preserve the agricultural resource base for the future and to maintain water resources for agricultural use and human consumption. Maintaining biological diversity carries less importance at a national level, because there is not great diversity of native terrestrial flora and fauna. Furthermore, there are no significant alternative uses for land and water resources, such as forestry or hydroelectric generation that provide competing demands for these resources, either now or in the foreseeable future. Thus, the main motivation for the antierosion effort is to preserve, or even augment if possible, the land and water resources available for future agricultural production. The Third National Development Plan for 1992–1995 (Republic of Cape Verde 1992) was prepared after national elections ushered in a multiparty system, and a new government assumed power for the first time since independence. The plan reveals a significant shift in the perception of the government’s role in the process of economic development and presents a change of priorities with regard to the agricultural sector. With respect to overall economic policy goals, the plan identifies three broad problems that require government action: (1) economic inefficiencies and mismanagement resulting from excessive public intervention in economic activities; (2) unemployment and underemployment of the nation’s workforce; and (3) the degradation of the natural resource base of the country. Proposed government actions to address these issues directly affect the agricultural sector and the rural population. Of most direct relevance to agricultural policies, the Third National Development Plan reconfirms the need to protect the national resource endowments: water, soils, and terrestrial and marine flora and fauna. Specific proposals for actions to retard degradation of agricultural resources cited in the plan include improving recharge rates of subterranean water supplies, continuing investments in reforestation and erosion-control structures, modernizing irrigation techniques, and introducing alternative agricultural practices for rainfed lands. One major area of proposed government action largely motivated by environmental concerns is support for research in science and technology that will identify sustainable agricultural and nonagricultural technologies that provide greater economic benefits to rural households without environmental degradation and that also generate employment opportunities. The issue of unemployment is addressed by increasing the capabilities of the workforce through education and training and by providing support to private enterprises that generate additional jobs. Although not specifically directed toward the agricultural sector, these actions can potentially have far-reaching impacts on the economic well-being of the rural population. Increased access to education for rural households will greatly

Agriculture and Policy Choice

163

improve the ability of household members to find remunerative employment outside of agriculture. Rural households already have good access to primary education; the next major hurdle is to provide secondary education, as well as more specific vocational training opportunities, to rural populations. Finally, reflecting a fundamental philosophical reorientation, the government has charted a course to decrease the level of state intervention in economic activities through privatization, reduced regulation, and the promotion of private enterprises. These changes will directly affect agriculture less than other sectors, since the level of public participation and regulation has been much less pervasive in agriculture than elsewhere in the economy.3 The proposed reduction in government regulation likely to have the greatest impact on agriculture is the dismantling of the land-reform legislation. The provisions of the legislation were most actively enforced by the government in the years immediately after independence. Although the reform legislation is not so vigorously implemented under the new government, it has continued to provide a favorable legal environment for tenants to resolve grievances with landowners. Under current proposals, even this formal support for tenants’ rights faces elimination. The government of Cape Verde pursues these various agricultural objectives, as well as a wide range of nonagricultural policies, within very restrictive budgetary constraints. In fact, the main source of funding for agricultural investments comes from outside the country in the form of foreign aid, which often carries specific conditions as to how the monies can be spent. Once a project is accepted, the Cape Verdean government has only limited control over the allocation of these funds. Furthermore, the discretionary funds of the government budget are extremely restricted. These severe financial limitations preclude the implementation of policies with large financial costs, such as consumer or producer subsidies. In addition, to the extent possible, the government strives to coordinate external projects to ensure that they are consistent with national policy objectives and, perhaps as important, to avoid costly duplication of efforts. The extreme scarcity of resources relative to the existing population and low productivity of the available resources accentuates the tension between the short-run and the long-run objectives of agricultural policies in Cape Verde. This is the fundamental dilemma facing agricultural policymakers, as outlined in Chapter 1 and shown in Figure 1.1. The goals of increased food self-sufficiency, rural employment, and agricultural incomes argue for maximizing the current production from the available agricultural resources—in other words, to get the curve that represents long-term changes in agricultural production capacity to approach as closely as possible the rising population curve. This strategy, however, runs great risks of depleting the agricultural resource base over time, through increased

164

Agriculture and Policy Choice

rates of erosion, water runoff, and mining of subterranean water reserves, eventually leading to a widening gap between population and the productive capacity of the resource base. At the other extreme, adopting a strategy to protect long-run agricultural capacity would require restricting current production levels, which would widen the gap between agricultural production and the demands of a growing population in the short run. Development of sustainable production techniques offers the hope of finding an intermediate path that can provide a more modest, but sustainable, increase in the agricultural production capacity. The goal of agricultural policy, then, must be to promote a sustainable development path that best meets both current and long-term needs.

Problems of Resource Degradation

Land and water resources in Cape Verde are highly susceptible to degradation and depletion as a direct consequence of their use by households in agriculture and other activities. Two important conditions can encourage households to undertake actions that deplete the domestic agricultural resource base. The first is related to the intertemporal nature of decisions about the utilization of agricultural resources. Land, land-improvement infrastructures, and irrigation networks are capital stocks that provide flows of services over time. These capital stocks may be degraded by overuse or neglect. Alternatively, they may be renewed through reduced use rates or investments in maintenance. If the stock of productive land is depleted through erosion, then future possibilities for agricultural production are curtailed. In planning current resource utilization, farmers consider the future costs, which take the form of lower future productivity due to resource depletion. However, most rural households in Cape Verde live in extremely impoverished conditions, and they are likely to discount future income very heavily in comparison with current income. If they are not even able to meet their current subsistence needs, these families are unlikely to be concerned about the future productivity of their land. The appropriate policy strategy to address this problem is to increase the income levels of the rural households, so that they will place a higher value on future income streams from their agricultural resources relative to the income that can be obtained from overusing their resources today. The second condition that may contribute to the problem of resource depletion is the existence of external costs. 4 External costs are generated when individuals undertake actions that impose costs on others without bearing these costs themselves. Take pollution as an example. A factory owner generates pollution, but the cost of spoiled air is borne by those who must breathe it and suffer health problems. The factory owner may

Agriculture and Policy Choice

165

also suffer somewhat from having to breathe polluted air, but the overall cost to society is much greater, since many other people must also share this burden. Economic theory states that activities that generate external costs are undertaken to a point where total benefits (factory output) are less than the total costs to society (the cost to the owner of operating the factory plus the costs borne by people living around the factory who must breathe the polluted air). Since the factory owner receives the value of the output from the factory but does not incur the total social cost of the activity, the factory will operate until the total incremental costs, including the external costs, are greater than the incremental benefits derived from the resource. Within Cape Verdean agriculture, two important types of external costs can be identified. The first, associated with rainfed agriculture, comprises the negative impacts of erosion resulting from planting corn on hillsides. Farmers in the upper reaches of watersheds may plant corn on their land, even though this activity can impose several kinds of costs downstream in the watershed. Soil erosion can decrease the amount of land available for agriculture, and it can also decrease the productivity of cropland as more fertile topsoil is removed. Perhaps even more important, erosion increases the rate of rainfall runoff into the ocean, thereby decreasing the subterranean water supplies that can be utilized for irrigation. These costs may be borne at a much later time than the activity that generated them. The second major external problem is found in irrigation networks that provide water to several users. The external cost is the exclusion of other potential users when an irrigator draws water from the network. In traditional water-allocation schemes, users are allowed to draw water from the distribution canal (or hose) as long as they wish, even though this restricts access to other farmers in the irrigation network, who may benefit more from water at that time. The extent to which individual irrigation patterns decrease the overall economic returns of the entire irrigation network represents the external cost associated with the existing water-allocation regime. Resolution of external cost problems requires an institutional framework that curtails the activities generating the external cost or provides a means for compensating those who must bear the cost. Two general types of institutional response can be distinguished. In the first, the relevant parties, namely, the potential generators and the victims of external costs, agree among themselves to share or limit the cost of externalities through some specific institutional arrangement. Such internal, or local, solutions to external costs can be found in Cape Verdean agriculture. Water users in several different irrigation networks on Santo Antão have coordinated and reduced their individual water applications within each rotation, thereby

166

Agriculture and Policy Choice

decreasing the interval between rotations for the whole network. This agreement permits farmers within the network to grow vegetables they can sell very profitably in the urban market on São Vicente. Internal agreement among the affected parties is not always possible, since they must all reach a consensus on an institutional framework that either regulates or taxes activities associated with external costs. In order to reach a private agreement, all parties must agree on the magnitudes of the costs and how they are generated. They must also agree on feasible strategies to monitor activities that generate externalities, as well as on ways to punish noncompliance. General agreement on an institutional framework that successfully addresses all of these issues is more likely for a relatively small and homogeneous group. If the number of individuals affected by an external cost is very large, it will be very difficult to reach a consensus solution preferable to the status quo of unregulated external costs. In such cases, an outside institutional authority will be needed to address external cost problems by imposing and enforcing restrictions. Lack of information about the direct relationships between resource use and the resulting external costs reduces the possibilities of agreement among affected parties and also reduces the efficacy with which external agents can manage external costs. Consider the problems associated with identifying the external costs associated with rainfed agriculture. There is a general perception that planting corn on steep slopes causes erosion that in turn increases the rate of runoff and decreases agricultural production, but the exact quantitative relationships are unknown. There is no agreement on the specific rates of erosion that result from planting corn on particular plots of land. The calculation is complicated by the fact that impacts are cumulative and depend on land-use patterns not only at the given location, but also elsewhere in the watershed. The quantitative impacts of erosion on the loss of agricultural land, the reduction in yields from lower soil fertility, and reductions in aquifer recharge rates are simply not known. If all interested parties cannot agree on the magnitudes of external costs associated with particular land uses at specific locations, they will experience great difficulties in reaching mutual agreements to curtail activities that generate these costs. Without general agreement on cost levels, those who generate external costs will argue that the costs are low, while those who must bear the costs will claim that they are high. Similarly, a management agency without the necessary information cannot identify the appropriate types of restrictions on land-use patterns or verify the extent to which changes in land use do in fact reduce the external costs. Continuing with the example of corn, if the government were to attempt to restrict corn planting, what would be the most appropriate policy? If we could suppose, for the moment, that the government were able to quantify external costs, then one solution would be to impose a tax

Agriculture and Policy Choice

167

equal to the amount of the damage. This decision would have the effect of internalizing the cost into the decisions of corn producers. Without the necessary information, however, government must use a more inexact instrument to restrict corn production. At one extreme, all corn production could be prohibited, but no Cape Verdean policymaker would ever seriously consider this option because of extreme social and political costs. In addition, a large portion of land currently in corn is not prone to erosion, so prohibiting corn production on this cropland would serve no purpose. However, more selective regulations raise the problems of defining the target areas. Should corn be restricted on lands with slopes more than 10 percent, or only greater than 15 percent? Without an understanding of the erosion impacts of corn on land of different slopes, these decisions cannot be made effectively. Thus, the complexity of agricultural sustainability is revealed like the facets of a gem turned slowly to the light. On one side, population pressure has forced the maximum use of the land base. From a slightly different angle, we see that widespread poverty has—for many households—severely limited the ability to incorporate future consequences into current resource allocation decisions. To these households, food and livelihood security takes immediate precedence over future considerations. Yet another twist of the problem reveals the complex and poorly understood interrelatedness of individual decisions made within an integrated watershed, where an activity engaged at one parcel can bear negative consequences on another parcel kilometers downstream. Policymaking within this complexity and with imperfect information is reserved for the brave-hearted.

Agricultural Policies in Cape Verde

To reconcile the complex demands of enhancing food security while promoting sustainable agricultural practices, the government of Cape Verde has pursued three institutional strategies of public intervention. First, the government administers the legal and bureaucratic framework that defines how individuals may use agricultural resources and the mechanisms for transferring use rights of resources among individuals. The second major area of public action is research to develop new production techniques that provide greater agricultural production from the given resource base under the specific environmental conditions found in Cape Verde. Finally, public work fronts provide direct income support to rural households while simultaneously undertaking public infrastructure investments, most of which benefit the agricultural sector. Public institutions regulate the utilization of the two most important agricultural resources in the country: land and water. The present landownership

168

Agriculture and Policy Choice

and use-right patterns in Cape Verde evolved from colonial mechanisms for allocating land to individuals, as described in Chapter 4. As expected in a situation of extreme scarcity of good-quality land, claims to land are clearly defined.5 At the time of independence, the General Directorate for Land Reform (Direcção Geral de Reforma Agrária [DGRA]) was formed within the Ministry of Agriculture to oversee issues of landownership.6 Legislation was enacted with the objective of transferring land from large absentee owners, including former colonials, to former tenants. Criteria for expropriation of land were established, the most important being nonresidence and lack of active participation by the owner in land management and utilization. The land-reform legislation also abolished sharecropping. Presently, the only legally sanctioned forms of transfer of use rights to land are by inheritance, sale, and rental contract. The DGRA was empowered to monitor all land sales and rental contracts. Maximum and minimum land prices and rental rates were established. In the years immediately after independence, active interventions to enforce land-reform laws provoked bitter conflicts in some areas, but in more recent years, the DGRA has been primarily involved in regulating prices and rental rates for registered land transactions. In fact, the DGRA does not greatly influence land transactions in the country; many agreements continue to be made privately among the interested parties without involvement of public authorities. The DGRA usually becomes involved only when there are disputes between landlords and tenants, but such disputes are not common. For example, sharecropping continues to be widely practiced, presumably because landlords and tenants are able to arrive at mutually acceptable agreements. However, the land-reform legislation does provide tenants with greater bargaining power in relation to landlords, because they can now threaten to bring the case before the DGRA. While the land-reform legislation presently affects, to some extent, the processes of land transfers, public policies do not place any restrictions on the ways in which agricultural land may be used. Farmers are free to decide how to best use their available land. As we mentioned in Chapter 6, all underground water resources in the country fall within the jurisdiction of the INGRH. The goal of this agency is to regulate the use of water on the islands in order to prevent depletion of available supplies. The agency drills and maintains wells that provide water for household consumption and for irrigation. In addition, private individuals must get authorization from the INGRH in order to open new wells. Technicians in the agency estimate the available water supplies in the ribeiras, and on the basis of this information, determine the number of wells that may be placed in a ribeira, as well as their respective pumping rates. The INGRH is also responsible for administering irrigation wells. In theory, the Ministry of Agriculture is supposed to provide technical assistance

Agriculture and Policy Choice

169

to ensure maximally efficient use of water allocated to agricultural production. However, this support has not yet been provided, and the INGRH has operated on an ad hoc basis, usually accepting traditional mechanisms for allocating water among network members. In recent years, the INGRH has become increasingly concerned about the management inadequacies in irrigation networks, and the agency has introduced alternative procedures that provide more autonomy and flexibility to individual users within the networks. INIDA undertakes applied research in the following major areas: plant production and protection (including integrated biological control); soil composition, fertility, and erosion; hydrology; agrometeorology; agricultural engineering, with emphasis on irrigation technologies; and social and economic studies of agricultural activities undertaken by rural populations. With respect to rainfed agriculture, the major activities have been to identify and collect samples of all the local varieties of corn and beans grown on the islands. The intention is to create an inventory of local varieties for use in selection trials. Variety trials of cassava are also being undertaken using local stock and improved varieties from other parts of the world. Research on irrigated crops has focused on food crops, particularly vegetables. This research includes variety trials and yield responses to alternative irrigation regimes. More recently, INIDA has initiated research on the traditional cash crops of bananas and sugarcane, where a wider short-term impact is thought to be possible. The institute has a long-run goal to construct soils maps for the entire country; however, the current lack of available resources limits this undertaking to soil samples taken in conjunction with other local project activities. Some small-scale experiments to measure runoff and erosion rates in microwatersheds are also being conducted. In the past, the research agenda of INIDA had been heavily influenced by the external projects that provide both the capital and the critical operating funds. Now, however, the institute has created a significant staff of researchers with their own research agendas into which external projects must now be more actively integrated. It is the intention of INIDA to become self-sufficient by selling its research and training services both domestically and internationally. The training center at INIDA has recently been designated the nation’s first agricultural faculty of higher education. The Ministry of Agriculture also operates a national extension service called the General Directorate for Rural Extension (Direcção Geral de Animação Rural [DGAR]). Contrary to the reigning wisdom of the farming systems approach to agricultural development, there is no direct institutional linkage or systematic collaboration between the extension service and INIDA researchers. The activities of the extension service are defined in large part by externally funded projects, which generally are distinct from the projects that collaborate with INIDA. As a result, the extension

170

Agriculture and Policy Choice

service does not provide a flow of information to researchers that could help set priorities within the research agenda, nor do the on-shelf technologies adapted at INIDA get into the hands of extensionists (or farmers). This lack of communication, while acknowledged by most parties, has continued to defy institutional solution, with the highly undesirable consequence that farmers have little access to national support structures and little opportunity to voice their technical concerns. The government programs that have perhaps the greatest direct impact on the current economic well-being of rural populations are the public work fronts, the FAIMOs. This public work mechanism appears at first glance to be the ideal compromise between the need to increase incomes of rural households and the need to conserve the agricultural resource base. Public work fronts function through cash payments for labor services, and they execute public investment works that protect slopes, increase infiltration, and retard erosion. Public investment projects are designed and supervised by the Ministry of Public Works (Ministério de Obras Públicas [MOP]) for road construction, by the Ministry of Agriculture for water- and soil-conservation projects, including reforestation, and by municipal governments for local infrastructure investment. The actual execution of these projects has been mostly designated to the National Institute for Rural Engineering and Forestry (Instituto Nacional de Engenharia Rural e Florestas [INERF]), which hires the workers and implements the contracted works.7 The agricultural investment projects employ 3,000 to 5,000 workers in erosion-control and reforestation operations during any given year. The general operational procedures of the work fronts, in terms of employment conditions, salary levels, and operating calendars, are uniform throughout the country and stipulated by the government. Onsite supervision of the work fronts is conducted by government employees who are members of the local community. Since the mid-1990s, the government has attempted to form legally constituted local ribeira associations that could also execute the rural engineering projects that demand less technological sophistication and heavy equipment. This experiment in participatory development approaches has been very encouraging in several of the Santiago ribeiras. The work fronts have been decidedly multiobjective by design. On the one hand, they have been used as employment vehicles to provide a minimal income for local households, a dimension that the government recognizes as the “social” objective of the FAIMOs. On the other hand, the work fronts have executed public investment projects meant to meet a “conservation” objective. Reconciling these objectives is not a trivial challenge. To provide effective erosion control, work must be undertaken initially in the upper reaches of watersheds; however, the social objective would dictate that work be concentrated in areas of denser population, which tend

Agriculture and Policy Choice

171

not to be found in the lower elevations of watersheds. In the past, the latter goal has been given higher priority, and the FAIMOs have been organized to provide work opportunities to the largest possible number of families, usually in the lower reaches of the ribeiras, where population density is higher. As a result, the effectiveness of the work fronts in terms of erosion control has been more limited. Without investments to control runoff in the headlands of ribeiras, heavy rains regularly cause severe flooding that destroys structures in the downstream areas, and they must continually be repaired. Although perhaps not attaining maximum impact in terms of erosion control, this strategy does maintain a continuing demand for labor in the areas of highest population concentration. In years of normal rainfall, work fronts operate six to eight months of the year and do not function during the planting season for rainfed crops, the peak period of agricultural labor demand. Recently, the current government has become more concerned with the efficiency of the FAIMOs and has introduced the concept of participatory planning within the Santiago ribeiras. In this case, a watershed plan is prepared with the community, and work fronts are organized around the implementation of the plan. The goal of this participatory approach is to ensure that the investments undertaken by the FAIMOs actually address the long-term needs of the communities in which they are undertaken, as well as providing employment in the short run. The scope of work front activities in terms of types of investments undertaken and geographical location is defined by the donor agencies. Thus, for example, in the late 1980s, agricultural work fronts were operating under the auspices of projects financed by the United States on Santiago, the Netherlands on Santo Antão, Germany on Fogo, and France on São Nicolau. Italians implemented road construction projects on several islands. In this operational framework, each donor project defines to a certain extent the overall objectives of its respective work front operations. INERF is charged with coordinating and monitoring all agricultural work fronts, but in fact individual project donors have had significant influence in defining work front activities. Thus, individual projects place different emphasis on structural versus vegetative control measures and the level of irrigation development, in part depending on the specific needs of the geographic areas in which they operate. All projects, however, place a very high premium on maximizing the amount of employment generated by their operations. Normally, evaluation of public works projects is based on the physical dimensions of the investments—number of meters of retention walls or check dams constructed, number of trees planted, and so on. While these measures provide a relatively reliable indicator of employment generated, they do not measure the effectiveness of the investment with respect to reducing erosion. Whereas the labor requirements of an erosion-control structure depend primarily on the physical dimensions, its

172

Agriculture and Policy Choice

effectiveness in controlling erosion depends critically on the design and placement in the watershed. Although the FAIMOs and the newly instituted community-based associations appear to solve the dilemma of simultaneous poverty alleviation and resource conservation raised earlier in this chapter, in fact the public works strategy itself faces a problem of sustainability. Such public investment is financed through donor funding, mostly generated from monetized food aid. Cape Verde has an overwhelming dependence on the international community for its basic food security, and most of the grain, edible oil, and dairy products provided in the form of food aid is monetized into domestic markets. The receipts from monetization are held either in donorcontrolled bank accounts or in the National Development Fund (Fundo de Desenvolvimento Nacional [FDN]), and these funds finance the public investment plan. In this way, rural households are dependent upon external donors not only for their basic food supplies, but also for the income to purchase them. As neoliberalist influence has reduced the food-aid dollar, the acute vulnerability of the Cape Verdean population has been dramatically exposed. For these reasons, the government does sense a certain urgency to convert current funds into a sustainable increase in domestic food production.

Analysis of Alternative Policies

Large amounts of financial, physical, and human resources have been committed to agricultural policies in Cape Verde since the 1970s, but with what effect? More specifically, to what extent have agricultural programs and investments been successful in increasing agricultural production, improving the well-being of the rural population, and preserving or improving the long-run agricultural productive capacity within the country? To address these questions, we evaluate current agricultural policies and alternative policy scenarios by estimating their impacts on rural households and the degree to which these impacts meet stated policy objectives. Both quantitative information about household access to resources and technical information about current production practices provide the basis for such evaluations. Policies meant to increase current incomes of agricultural households will be considered first, followed by an assessment of the policies designed to maintain or improve the availability of agricultural resources over the long term. To date, the work fronts constitute the policy program that has had the greatest direct impact on the incomes of rural households in Cape Verde. On Santiago, off-farm income, which is generated primarily from work fronts, represents one-third of the total net income of the interviewed

Agriculture and Policy Choice

173

households. On Santo Antão, work front employment provided 14 percent of total net income. The FAIMOs have also been successful in providing job opportunities to a large number of households. Information from the 1988 census reveals that 14 percent of all adults in rural areas had been employed on public projects, including 12 percent of the Santiago adults and almost one-quarter of the Santo Antão adults. Based on our survey information, 54 percent of the interviewed households on Santiago and 61 percent on Santo Antão had at least one family member on public work fronts. The surveys also indicated that jobs on the FAIMOs go disproportionately to the poorest households (net of work front receipts). Of those households with FAIMO employment, more than 70 percent would be below the poverty level if they did not have jobs on the work fronts. For the sample as a whole, without work front receipts, 63 percent of all interviewed households on Santiago would have fallen below the poverty level, compared with the actual figure of 37 percent. Thus, the public work fronts represent an effective safety net for a large number of poor rural families in Cape Verde, and the implementation of work fronts has been very successful in distributing the benefits widely and targeting the poorest households, thus meeting the social goals of this policy. Another possible strategy to improve rural incomes—and at the same time increase national agricultural production—focuses on the improvement and expansion of irrigated agriculture. As detailed in Chapter 6, the returns to all irrigated crops are substantially greater than those to rainfed crops, and even the least productive irrigated crop, sugarcane, provides net returns per hectare seven times greater than estimated returns from rainfed corn and beans. Thus, the expansion of irrigated land has the potential to dramatically increase agricultural incomes. However, the potential for increasing irrigation capacity in the country is limited by subterranean water supplies, and it therefore depends critically on erosion-control efforts that will increase the rate of water infiltration into aquifers. However, significant potential exists for increasing the efficiency with which existing water resources are used in agriculture. We have already demonstrated that returns to irrigated agriculture rise sharply with the reduction of intervals between irrigations. More frequent access to irrigation water gives farmers the opportunity to grow the high-value crops that have more stringent water requirements. Yet, more intensive water application would imply a reduction in the total irrigated area. For example, suppose that an irrigation network with 60 members is supplied by a well that can pump for 60 hours per week. Let us assume further that it takes 15 minutes to transfer tubing from one field to another and that 2 hours and 45 minutes are required to apply water to 1 liter of land. Under these assumptions, with a 6-week interval between irrigations, 125 liters of land could be irrigated with the amount of water supplied by the system. Sugarcane

174

Agriculture and Policy Choice

irrigated at 6-week intervals provides a net return of CVE 7,600 per liter of land area, so the total income for the network would be CVE 950,000. Reducing the interval between irrigations would diminish the amount of land each farmer could irrigate, but the higher value of the crops would generate a higher total income for each farmer and for the network as a whole (see Table 8.1). Thus, there are strong incentives to decrease irrigation intervals, at least to two weeks. In fact, some networks have achieved this result through locally agreed strategies, notably, the farmers in Alto Mira and Ribeira Seca. However, the inflexible nature of the fixed rotation mechanism for allocating water among users can discourage such locallevel agreements, since all producers within a network must agree to the same irrigation interval. For reasons related to differential farm sizes, network members would likely not reach a unanimous agreement to reduce the allocation per irrigation. Such a strategy penalizes the smaller farmer and favors the larger, as we have seen. In addition, some farmers might not be interested in growing the high-value crops either for reasons of excessive risk or because of the more intensive management inputs necessary for these crops. Even if an agreement within a given network could be reached, a difficult policy trade-off emerges. An increase in irrigation frequency will,

Table 8.1

Cropping Patterns and Net Revenues for Irrigation Network Under Alternative Intervals

Irrigation Interval (days)

Crops Grown

42 35 28

Sugarcane Sugarcane Sugarcane/ Sweet Potato/ Cassava Banana/ Sweet Potato/ Cassava/ Sugarcane Banana/ Sweet Potato/ Cassava Potatoes/ Cabbage/ Onions/ Tomatoes/ Sweet Potato/ Cassava/ Sugarcane

21

14 7

Source: authors’ calculations.

Net Revenue (CVE/liter)

Total Irrigated Area (liters)

Total System Net Revenue (CVE)

7,600 12,100 22,800

125 104 82

950,000 1,258,400 1,869,600

33,300

60

1,998,000

68,900

38

2,618,200

108,500

16

1,736,000

Agriculture and Policy Choice

175

under the current distribution system, drastically reduce the total amount of irrigated land in cultivation, as Table 8.1 illustrates. It is true that the overall value of the remaining land compensates, in economic terms, for the loss in area, but efficiency may triumph at the expense of equity. Since land is not distributed equally, there is little expectation that the number of farmers in the system would remain constant, as some smallholder households might be forced to transfer their irrigated plots to others. This raises the question of the value of the trade-off: does policy meet its social objective if a smaller number of water users grow high-value crops or if a larger number of users have access to enough water to practice a lower-value irrigated agriculture, remembering the difference in returns between sugarcane and the best rainfed alternative? This decision is a very difficult one. One promising initiative has been the recent introduction of association-based microirrigation systems that use drip technologies. With donor support and extension assistance, several pilot programs have begun on Santiago and on the very arid Sal Island with highly encouraging initial results. Small groups of producers have organized themselves, acquired titled land, and obtained financing to purchase the drip irrigation equipment. The management demand is very high for these systems, in part because of the complexity of the technology and the farmers’ lack of experience with it. Nevertheless, drip irrigation allows farmers to reduce transaction costs associated with transfer time and to expand watering frequency without loss of irrigated area due to the much greater watering efficiency. A final water-management issue has to do with the pricing of water. Presently, farmers receive irrigation water for free, or only have to pay a portion of the actual delivery cost, though many must pay for the fuel to operate the pump. However, users do not pay the high external cost of excluding access to water for other potential users. As a result, farmers have little financial incentive to economize on water use. Given the extreme scarcity of this agricultural resource, and the very high benefits that can be derived with irrigation water, there is great potential for improved management of water use through more appropriate pricing. Agricultural research at INIDA is directed toward developing new technologies that can enable farmers to produce more agricultural output from their available resources. The absolute magnitudes of benefits from any research effort are impossible to predict with any degree of a priori certainty. However, some generalizations about the likelihood of attaining different types of improvements, as well as their potential impacts on Cape Verdean agriculture, may be outlined. First, improvements in productivity are much more likely for irrigated crops than for rainfed crops. Variety improvement programs follow a strategy of selecting and propagating plants that produce the highest yields under a specific set of growing conditions. The selected varieties, however, may exhibit very poor yields under unfavorable

176

Agriculture and Policy Choice

conditions. Local varieties are better adapted to environmental stress and provide better yields in unfavorable conditions, albeit with lower potential productivity. Irrigated crops, by definition, grow in more favorable and stable conditions, so selection for higher yields under a specific set of environmental conditions is feasible. Improvements in management of irrigation water provide closer control over the environment and thus offer opportunities for additional gains from improved varieties. On the other hand, rainfed crops are subject to extreme drought stress in Cape Verde, which imposes severe restrictions on the potential to improve yields through the introduction of green revolution technologies. Although the probabilities for significant yield increases are greater for irrigated crops, the total benefits obtained from such improvements are limited by the small total amount of irrigated land. Presently, only 7 percent of all agricultural land in the country is irrigated, and any productivity increases would benefit only this restricted area. Some technicians believe that the potential for increases in total irrigation capacity is significant—either through the reduction of water loss caused by inefficient technologies and management, or through the expansion of wells into other aquifers, which would simultaneously increase the aggregate potential benefits from improved varieties of irrigated crops. Every additional hectare brought under irrigated cultivation represents a major increase in income and in food-production capacity. Table 8.2 presents results of several scenarios of alternative measures to increase agricultural incomes. These scenarios are based on very general assumptions about the impacts of different public policies on agricultural production levels. These assumptions are not intended to be “best guesses” about the probable impacts of different policies or projections of the future course of agricultural changes in Cape Verde; rather, they are rough estimates of the maximum potential impact of alternative public actions. The numbers should therefore be interpreted as the likely upper bounds on the potential benefits of alternative policies. The table reports estimates of net returns from rainfed and irrigated activities, as well as total household income for all of the households in the Santiago and Santo Antão surveys. Net income from rainfed agriculture is obtained by multiplying the amount of rainfed land the families farm by the per hectare net revenues that could be expected under favorable growing conditions (see Chapter 5). Net income for irrigated agriculture is calculated in the same manner, except that the per unit net revenues are adjusted to account for the average interval for which households report receiving irrigation water. The per unit returns correspond to the representative systems described in Chapter 6 and summarized in Table 8.1. The first row in Table 8.2 represents the base case of estimated total incomes of the surveyed households under current conditions. Looking first at Santiago, rainfed, livestock, and off-farm work each account for 30 percent of household income, and the share of irrigated agriculture is less

177

Agriculture and Policy Choice Table 8.2

Simulations of Potential Increases in Agricultural Incomes of Surveyed Households

Santiago

Base income levels Improved irrigation management (all irrigations on 2-week intervals) Increased irrigation capacity (double amount of irrigated land) Improved irrigation productivity (increase net returns per hectare by 30%) Improved rainfed productivity (increase net returns per hectare by 50%) (increase net returns per hectare by 100%) All changes Santo Antão

Base income levels Improved irrigation management (all irrigations on 2-week intervals) Increased irrigation capacity (double amount of irrigated land) Improved irrigation productivity (increase net returns per hectare by 30%) Improved rainfed productivity (increase net returns per hectare by 50%) (increase net returns per hectare by 100%) All changes

Aggregate Income of Surveyed Households (CVE) Rainfed Activities

Irrigated Activities

Total Household Income

38,468,000 38,468,000

9,554,000 12,735,000 (33)

122,394,000 125,575,000 (3)

38,468,000

25,470,000 (167)

138,310,000 (13)

38,468,000

33,111,000 (247)

145,951,000 (19)

57,702,000

9,554,000

76,936,000

9,554,000

76,936,000

33,111,000

141,628,000 (16) 160,862,000 (31) 184,419,000 (51)

Aggregate Income of Surveyed Households (CVE) Rainfed Activities

Irrigated Activities

Total Household Income

4,895,000 4,895,000

15,286,000 27,826,000 (82)

29,343,000 41,883,000 (43)

4,895,000

55,652,000 (264)

69,709,000 (138)

4,895,000

72,348,000 (373)

86,405,000 (194)

7,343,000

15,286,000

9,790,000

15,286,000

9,790,000

72,348,000

31,791,000 (8) 34,238,000 (17) 91,300,000 (211)

Source: authors’ calculations. Note: Numbers in parentheses are percentage increases relative to the base results.

than 10 percent. On Santo Antão, irrigated agriculture is a much more important source of income, over half of the household total. Policies directed toward irrigated agriculture are considered first. These scenarios are sequenced to maximize the cumulative benefits of the policies.

178

Agriculture and Policy Choice

The first policy scenario estimates the potential income impact of improving the effectiveness with which the existing water supplies are distributed among farmers. All households with irrigated land and irrigation intervals of more than two weeks are assumed to switch to two-week intervals and adopt the more profitable crop mix associated with this shorter interval. Bananas are used as the basis for estimating returns with twoweek intervals because the export market establishes an exogenous minimum price level, whereas the prices of the other irrigated crops are determined by domestic market conditions. Higher net returns for these other crops could be expected to diminish as increased production drives down domestic prices. The net returns of bananas thus represent the minimum potential returns for irrigated crops at two-week intervals, even with expanded output levels. The amount of land under irrigation is reduced to maintain the same total level of water application during the course of the irrigation season, thus ensuring that overall water use is constant. The potential for increasing returns from irrigated land through improved irrigation management is great. Net returns from irrigated agriculture increase by almost 30 percent over the base case in Santiago. However, because of the very small amount of irrigated land there, the overall impact on household incomes is very small—less than 3 percent. In fact, only one-quarter of the surveyed households would be affected at all by this action. On Santo Antão, the effects of improved irrigation management on aggregate household income are much greater because of the larger share of irrigated agriculture in total income, increasing total household income by more than 40 percent. These benefits of improved irrigation water management could be obtained with very little financial expenditure on the part of the government. The greatest obstacles are likely to be coordinating changes in watering schedules among irrigation network members and agreeing on how reductions in farmed area should be allocated among farmers. The second scenario examines the impact of increased irrigation capacity, specifically, doubling the amount of water available for irrigation. Several studies have indicated that at the national level, groundwater availability would permit this scale of increase in irrigation, as presented in Chapter 3. This scenario assumes all farmers irrigate at two-week intervals. On Santiago, this option increases the returns from irrigated agriculture by 167 percent and increases total household income by 13 percent. On Santo Antão, doubling irrigation capacity would more than double total household income. Increasing irrigation capacity has great potential for increasing agricultural incomes, particularly in conjunction with improved efficiency in the allocation of irrigation water. This strategy would require some direct investments in irrigation infrastructures, but more important, much more detailed management of activities within watersheds, including

Agriculture and Policy Choice

179

careful monitoring of rainfall patterns and of pumping rates from all wells, and possibly regulating the utilization of rainfed land, so as to maximize water deliveries without causing permanent overdraft from the watershed. The distributional impacts of this scenario depend in part on the physical constraints that circumscribe irrigable land, as well as on the process by which new irrigation capacity is allocated to rural households. For example, landholdings along the upper slopes of the watersheds would not benefit as easily from irrigation development as those along the ribeira bottoms. The next scenario examines the impacts of increased economic returns for irrigated crops resulting from improved technologies or varieties. Net returns are assumed to increase by 30 percent under the improved technologies. In conjunction with improved management and greater irrigation capacity, such yield increases can raise household incomes on Santiago by nearly 20 percent. Increased productivity alone, however, would expand total household incomes by less than 3 percent. On Santo Antão, similar productivity increases would raise total incomes by almost 200 percent. The potential impacts of improvements in the productivity of rainfed crops are evaluated in the fifth and sixth rows of Table 8.2. This scenario incorporates the assumption that productivity gains double net returns to rainfed activities. This is an extremely ambitious assumption; there are no results from experimental research on alternative crops or preliminary results from variety trials that suggest the possibility of such significant productivity increases in Cape Verde in the immediate future. The probability of attaining this scenario in the foreseeable future is thus very low and depends on major technological and biological breakthroughs. Because of the relatively small share of rainfed agriculture in total income, the overall impact of doubling returns to rainfed activities produces a much more modest relative increase in total household income. On Santiago, total household income increases by 31 percent and on Santo Antão, by 17 percent. It should be noted that this simulation does not incorporate any complementary increases in livestock production resulting from increased rainfed production. If new crops or growing techniques significantly increased the feed production on rainfed land, the spillover effects on livestock could significantly increase the impacts on total household income. Although this scenario does increase incomes for virtually all households in both island samples, there remains a significant number of households below the poverty level, 23 percent on Santiago and 21 percent on Santo Antão. Finally, if all of the proposed policies to increase or improve the productivity of agricultural resources were successfully implemented, total household incomes would increase by slightly more than 50 percent on Santiago, while they would triple on Santo Antão. This indicates the upper

180

Agriculture and Policy Choice

limit on the extent to which agricultural policies, under the most favorable assumptions, can increase incomes of rural households. On Santiago, the potential increase in incomes, while substantial, does not represent a basis for an expectation of major transformation in the economic condition of rural households. Against these potential benefits must be weighed the costs, especially the risks inherent in the proposed policies. Unfortunately, the policies most likely to be successful would provide benefits to only a small segment of the rural population, while the policies with more wide-ranging potential benefits have a much greater risk of failure. The increased efficiency of the use of irrigation water is technically feasible under current conditions, but the impacts on rural incomes would benefit only those households that currently receive irrigation water. Increasing total irrigation capacity would provide benefits to a larger, but still restricted, number of households. This strategy would also require successful programs to manage water resources in watersheds, with systemwide erosion-control investments. Improvements in returns to rainfed activities could potentially have the greatest aggregate impact on incomes and would provide benefits to all rural households; however, the probability of encountering technological solutions to the extremely adverse environmental conditions of Cape Verde is low at the present time. Furthermore, strategies to increase rainfed agricultural production over the short run must be carefully examined against potential impacts on erosion. Overall, the possibilities to substantially increase current returns to agriculture on Santiago are limited, and the benefits are likely to be concentrated among a small proportion of rural households. On Santo Antão, the potential for agricultural development is much greater, and agricultural programs directed toward increasing the scale and productivity of irrigated activities can dramatically increase rural incomes. This is so because the share of irrigated agriculture is much higher on Santo Antão than on Santiago, and there exist great opportunities to increase the efficiency of irrigation. Rainfed agriculture on this island faces the same difficulties found throughout the rest of the archipelago. Attempts to analyze the policies designed to reduce long-run environmental degradation are hobbled by the widespread lack of quantitative information about the factors that affect erosion processes, especially landuse practices. Without this information, it is very difficult to estimate potential benefits of investments to control erosion; however, some preliminary judgments can be made about the relative merits of alternative policy strategies to control erosion and runoff. One strategy implemented recently is the effort to improve the level of maintenance of the erosioncontrol investments undertaken by the work fronts. The Cape Verdean government, as well as donor agencies, has been concerned about the lack of

Agriculture and Policy Choice

181

repairs on erosion-control structures constructed by the FAIMOs. Heavy rains regularly damage or wash out retaining walls, but local farmers tend to not undertake repairs, even on the structures that protect their own parcels. Instead, they await the formation of FAIMOs to repair the annual damage. Similarly, local populations do not protect the trees planted in reforestation programs; in fact, the trees are sometimes removed for firewood. In order to address this perceived problem, the government and donor agencies have attempted to introduce participatory methods into the operation of the public work fronts. It is argued that local populations do not undertake maintenance operations under the existing system because they have no direct input into deciding the kinds of investments that the FAIMOs undertake within their communities, resulting in a lack of incentives to maintain the structures. The proposed participatory framework assigns local groups more direct control over the public work front activities as well as greater responsibility for the quality of the work. Under this participatory development scheme, groups within local communities decide on what kinds of activities should be undertaken by the public work fronts, and they also manage and execute the projects, while the government provides the necessary financial and technical support. Ideally, the resultant projects would reflect the needs of local communities, so the resident population would have direct interest in maintaining the investments. This strategy is likely to be successful for those investments in which the benefits of erosion control are in fact concentrated within a localized community and can be clearly identified. Irrigation construction and maintenance would fall into this category; however, many of the benefits of investments to control erosion are widely dispersed throughout a watershed and are thus perhaps less visible. Because of the highly dispersed nature of the external costs of erosion, local groups have few incentives to undertake these types of investments, much less to maintain them. For this reason, government might best consider erosion-control structures to be public investments that provide for the general good of the population, similar in this regard to roads, schools, hospitals, and the like. This line of argument could be extended to government responsibility for the maintenance of the structures (just as with roads). Note that this strategy also supports the employment objective of work fronts over time. Overall, the management of public works programs should take into consideration the differing natures of different types of investment. For small projects in which the benefits of the investments fall to a relatively small number of individuals, the appropriate strategy would be to let these individuals manage and undertake the work. In contrast, public investments, which provide widely dispersed benefits to the population, should continue to be managed and maintained by the government.

182

Agriculture and Policy Choice

In addition to construction of physical structures and reforestation, management of the use of rainfed land is an important area of agricultural resource policy. Land-use patterns affect both soil erosion and water runoff rates. One possible strategy to reorient land-use patterns on rainfed land is to regulate or tax activities that generate external costs. In particular, if planting corn and beans on steep slopes accelerates erosion, then restrictions on this activity might be appropriate policy measures to address this external cost. However, the consequent distress that this measure would impose on rural populations, as well as the difficulties in effective enforcement, render it an infeasible policy option. Farmers have historically had the right to grow corn on their land, and they place a very high value on this right. Any curtailment on individual rights to grow corn on their land would be met with extremely strong resistance on the part of farmers and landowners throughout the country. Furthermore, the land most prone to erosion is generally farmed by the poorest families, so sacrifices of such restrictive policies would be borne most heavily by the most disadvantaged strata of the rural population. Moreover, farmers would have strong incentives to violate any regulations, and effective monitoring over disperse areas, often remote and without convenient access, would be very difficult. Another possible policy would be to promote alternatives to corn and beans that provide better protection against erosion. The fact that the prohibition of corn and beans production is not a feasible policy option places a very high priority on research to identify alternative crops or management practices that will reduce erosion. However, for farmers to adopt any alternative crop, it must provide benefits to the farmer at least as great as those corn and beans currently do. As discussed in Chapter 5, corn and beans are attractive crops for farmers to grow, not only for their food production (grain and beans), but also for the forage for livestock and because of their low labor requirements. One technology currently being tested is to plant pigeon peas and other species along contour ridges within fields to act as “living fences” or vegetative barriers to water and soil runoff. Corn and beans are planted between the ridges. The reduction in area planted with these latter crops as a result of the introduction of the ridges is expected to be offset by increased yields due to the greater retention of moisture. Pigeon peas also fix nitrogen in the soil as a further advantage. Close examination of farming systems in Cape Verde reveals the importance of livestock in the total incomes of rural households and the high degree of complementarity between rainfed agricultural and livestock activities. Any proposal for alternative rainfed land-use patterns must take into consideration the potential effects on livestock assets. In fact, recent donor proposals have suggested increasing the use of rainfed lands for pasture. This conversion has clear advantages for reducing erosion by

Agriculture and Policy Choice

183

increasing vegetation cover, but it would also require management of herd stocking rates to avoid overgrazing.

Conclusion

The potential for increased agricultural development in Cape Verde is ultimately constrained by the ecological imbalance that characterizes the archipelago. Because of this imbalance, the rural population cannot expect to live solely on rainfed agricultural activities within the foreseeable future. Crop and livestock production is simply too dependent on the vagaries of the weather. Irrigated agriculture provides opportunities for greatly increased returns, but the limited subterranean water supplies presently restrict this option to a small portion of the total rural population. Given these realities, Cape Verdean policymakers must accept the fact that agriculture in and of itself does not provide the key to the economic well-being of the existing rural population. The problem will be exacerbated in the future as the population growth rate continues to diverge from the limited agricultural production capacity. However, this rather stark reality does not imply that agricultural policies are not an extremely important area of public action in Cape Verde. The majority of the country’s population live in poverty and depend at least in part on agriculture for their livelihood, so any increase in returns to agricultural activities provides significant increases in the overall welfare of the rural population. The more appropriate conclusion to be drawn is that policies should be directed to attain feasible goals and to maximize the potential impact given the limited resources available to implement programs. In terms of Figure 1.1 in Chapter 1, the goal of agricultural policy must be to pursue agricultural programs that will sustain long-term development of the agricultural production capacity, rather than to attempt to immediately increase agricultural production by intensification of activities that will exacerbate soil erosion and water runoff, which would in turn jeopardize the long-run productivity of agricultural resources. One implication of the low rainfed agricultural productivity is that the benefits attained from current exploitation of the agricultural resource base are relatively small in comparison with the possible long-term costs of erosion and reduced subterranean water supplies. Policy decisions must therefore take a long view of the potential impacts. Unfortunately, the necessary information to assess long-run environmental costs is not presently available. A primary research objective must be to identify and analyze the factors that affect soil erosion and water runoff and how various agricultural activities affect these processes. Within the small geographical scale and abrupt topographic relief of the major agricultural islands, these processes occur within the confines of quite small and well-defined watersheds.

184

Agriculture and Policy Choice

Quantitative understanding of the physical processes in the watersheds must be obtained through field observations. In the area of farm-level practices, the search for alternative uses of rainfed land that reduce erosion and runoff problems should continue. However, researchers must be aware of the full magnitude of the economic benefits that corn and beans provide to farmers, both for human and animal consumption. Any alternative crop must provide greater benefits to farmers if it is to be widely adopted. Management of water resources provides perhaps the greatest potential for increasing agricultural production and rural incomes. Important management issues at the watershed level include large-scale erosioncontrol investments: reforestation or structures, management of the use of rainfed land to minimize erosion, and control over the amount of water removed from subterranean supplies. Effective management requires information about the hydrologic characteristics of the watershed. In addition, the existence of significant external costs associated with erosion, water runoff, and water pumping implies that these management decisions should be made at the level of the watershed, where these costs are internalized. Institutionally, there is a much-needed role for the public management of both erosion-control investments and pumping of water so as to protect the interests of all households within a given watershed. Such management of rainfed land use is currently restricted by the lack of alternatives to corn and beans, which provide superior economic benefits to the farmers. High research priority should be given to finding alternative crops that reduce erosion and also provide benefits to farmers superior to the current farming system. On a smaller scale, there exists great potential for increasing agricultural returns by improving the management of water within irrigation networks, specifically, how often individual producers are able to put water on their fields. Under current management practices, most farmers are constrained to growing relatively low-value crops. A more flexible management regime could dramatically increase the overall income generated within existing irrigation networks, even with the same amount of available water. Technological improvements, such as drip irrigation, also hold significant promise for expanding this important activity. In sum, we have held the rural reality of Cape Verde to the policy candle and have seen that agriculture cannot be made to bear the public burden for total food security and complete poverty alleviation. It simply does not have the resources. Nonetheless, our analysis suggests that there is policy leverage and room for maneuverability in agriculture that can begin to narrow the gap between production capacity and the consumer demand associated with a growing population.

Agriculture and Policy Choice

185

Notes 1. These objectives are explicitly identified in the national Constitution. 2. In fact, the current fashion toward neoliberalism and privatization has threatened to reduce Cape Verdean access to international food aid, causing great consternation among policymakers. 3. The exceptions to this are the state-run farm and livestock operations on Santiago. These public enterprises accounted for only a tiny fraction of national agricultural production, and they are now in the process of being privatized. 4. For a comprehensive review of the concept of external costs, see Zilberman and Marra (1993). 5. Individual disputes occasionally occur in cases of transfer, but these are usually among family members or result from different interpretations of the transfer agreement by the parties involved. 6. It should be noted that the ministry responsible for agricultural development has changed names several times in recent years, depending on the emphasis placed on fisheries, forests, or rural extension. It is currently called the Ministry of Fisheries, Agriculture, and Rural Extension (MPAAR). As a convention, we refer simply to the Ministry of Agriculture. 7. INERF was once a division within the Ministry of Agriculture. It became a semiprivate, self-sustaining institute in order to reduce the potential for misuse of public investment funds.

9 Conclusions: The Future of Agriculture in Cape Verde

As elsewhere in the Sahel, agriculture on the Cape Verde Islands has had to adapt to a harsh and uncertain physical environment. Throughout the first 500 years after settlement in the second half of the fifteenth century, the dynamic interaction of nature and culture created a socioeconomic livelihood system, the fortunes of which ebbed and flowed with the vagaries of climate. At first, commerce flourished and farming occupied an ancillary position; but as Portugal increasingly constricted the commercial opportunities of its fledgling colony, agriculture came to be the primary means of livelihood for the majority of the population. In those years of adequate rainfall, domestic production could support a growing population, but the dreaded appearance of drought took heavy tolls on the island population with disease and famine. Throughout the centuries, the cyclical pattern of population growth then contraction directly followed the variations in rainfall. Then, in the 1960s, the international community—partly for reasons of humanitarianism and partly motivated by geopolitical maneuverings—decided to break this Malthusian interrelationship. With the arrival of food aid and the development of industry, population growth became disconnected from the wide fluctuations in agricultural production, a change that improved the quality of life for many archipelago inhabitants and wove an effective “poverty safety net” to eliminate starvation from the country. In response, the population began to grow steadily and continues to do so, but Cape Verde’s natural resource endowment that provides the productive base for agriculture cannot expand. On the contrary, the increasing population pressure on fragile agricultural lands threatens to reduce the productive capacity of the resource base. It is the current unchecked growth of population in the face of uncertain and inadequate food supplies that has added a new sense of urgency to developing an economic base that can support the nation’s population into the future. 187

188

The Future of Agriculture

Population growth is vigorous for several reasons. First, food aid has rendered famine a memory, albeit a not too distant one. Second, the large number of Cape Verdean émigrés who reside and work elsewhere in the world continue to support their extended families on the islands with remittances, the second largest source of foreign currency. Finally, improved health care and sanitation, also largely financed by the international community, have drastically decreased the mortality rate at both ends of the population pyramid. As a consequence of these factors, the national population now expands even as agricultural production experiences wide fluctuations, with only a very modest upward growth trend. Since most people still live in rural areas, an ever-larger population depends, at least in part, for its livelihood on the paltry and extremely variable production that island farming can provide. To be frank, there are simply too many people in agriculture. As we outlined in Chapter 1, the increasing imbalance between the size of the population and the limited agricultural resource base manifests itself in several important ways. Perhaps the most obvious is the perennial gap between local food production and consumption requirements of the population. In recent years, local production has provided a mere 20 percent of national food-consumption needs, while the remainder has been met with imports, primarily in the form of food aid. National leaders harbor grave doubts as to the long-term sustainability of this situation, since it implies such heavy dependence on the goodwill and largesse of other nations and organizations. Another bothersome dimension of the inherent imbalance between population and resources is that population pressure has encouraged the overutilization of scarce agricultural resources—including marginal lands that provide very low incomes to the rural population. Not only do households have inadequate access to land, but productivity levels oscillate widely with the rainfall outcomes. In addition to the overall scarcity of agricultural resources relative to the existing population, the distribution of these resources across households is very unequal. As a result, a small number of families live comfortably from their agricultural incomes, while the vast majority must turn to other activities besides agriculture just to meet minimal subsistence needs. Again, to state the case simply, not only are there too many people in agriculture, they are also too poor. One consequence of this widespread rural poverty is the increasing tendency for households to degrade the existing agricultural resource base by planting on highly erodible hillsides and drawing down subterranean water resources more rapidly than they can be replenished by rainfall infiltration. Thus, poverty and increasing population pressure lead to a vicious cycle of agricultural resource degradation that will further reduce both agricultural productivity and the incomes of the ever-growing rural

The Future of Agriculture

189

population. This seemingly irreconcilable dialectic between short-term welfare and ecological sustainability is clearly the major dilemma confronting those Sahelian countries—like Cape Verde—where local resources are fragile and scarce in relation to the population density. And it is a dilemma that policymakers have little time to resolve. Given these stark realities, what are feasible goals with regard to agricultural resource management in Cape Verde? Although current discussions of agricultural policy highlight the notion of sustainability within Cape Verdean agriculture, the appropriate definition of sustainability to use in guiding policy and resource management lacks sharp resolution. One definition, which formed the basis for agricultural policies immediately after independence, interprets sustainability as the ability of the Cape Verdean agricultural sector to meet all domestic food requirements; that is, sustainability was equated with national self-sufficiency. Under this view, continued reliance on foreign food supplies is considered unsustainable, since Cape Verde cannot depend indefinitely on international assistance and does not have the means to pay for commercial imports. In addition, since two-thirds of the population live in rural areas, increasing agricultural production to meet food needs will also provide increased incomes to a major portion of the population. The postindependence policy stance placed much emphasis on increasing agricultural production and productivity, and the agricultural sector became the primary recipient of both foreign aid and national budget expenditures. Unfortunately, the uncooperative growing conditions in the archipelago and the very small total amount of arable land thwarted these goals, and food self-sufficiency is now generally perceived to be infeasible in the foreseeable future. Furthermore, overzealous attempts to increase total agricultural production probably only hasten the damaging impacts of erosion and groundwater depletion. Our analysis of the factors affecting the resource-management strategies of rural households suggests that a different interpretation of sustainability is more appropriate for Cape Verde. This alternative view emphasizes the need to maintain, or even improve, the quality (the productive potential) of the agricultural resource base. This goal is very different from food self-sufficiency and perhaps even argues for a lower level of selfsufficiency in order to minimize depletion of resources. Adoption of this definition of sustainability has a number of important policy implications. First, and most generally, under this sustainability goal, agriculture cannot be seen as the main source of economic development providing attractive employment opportunities for the rural population of Cape Verde. Although substantial benefits can be obtained through increased efficiency in resource use, particularly irrigation water, and technical improvements are possible in the current farming systems, the agricultural potential in Cape Verde is ultimately and inexorably limited by the weather.

190

The Future of Agriculture

This notion of sustainability also accepts that there are strict limits to the extent to which agricultural production can be increased, given known production techniques. If, as we argue, food self-sufficiency is an infeasible goal, alternative solutions must be sought to acquire food supplies for the national population. In essence, the economy must generate the sources of foreign exchange necessary to purchase food on international markets and, at the same time, provide rural households the income to buy it. In effect, the nonagricultural sectors of the economy must be developed and expanded. Similarly, since the agricultural sector does not offer a sufficient source of employment opportunities, a successful development strategy must find alternative employment outside of agriculture. Ironically, then, the move toward increased sustainability may very likely reduce total employment in the agricultural sector. Furthermore, any increases in the economic returns to labor in agricultural activities will occur in large part through reductions in the total amount of labor employed in the sector. All these factors indicate the need to attempt to promote alternative employment opportunities outside of agriculture, even for the existing rural population. This strategy may involve large-scale population movements to urban areas, a process already under way but one that many policymakers would prefer to retard. Perhaps a better position would be to view this movement as inevitable, even beneficial, in the long run, and then to attempt to minimize the economic and social costs associated with this adjustment process by preparing and planning for it. One manifestation of the limited employment opportunities in agriculture has been the high rate of emigration from rural areas to foreign countries. In the near future, this process is likely to continue, but this strategy is fraught with uncertainties, depending as it does on the capricious immigration policies of receiving countries. Furthermore, this strategy imposes significant social and personal costs on rural populations. In order to successfully promote agricultural sustainability, public policies must be designed to induce private resource managers—households—to generate the highest possible level of output while simultaneously avoiding the long-run impacts of resource degradation. This goal has important implications for designing and executing policies and also points to crucial areas for research. With regard to research, more detailed information is needed in several key areas in order to design effective policy interventions to promote sustainable agricultural development. In Cape Verde, all physical and social processes crucial to agriculture occur in the watershed, or ribeira. Thus, a better theoretical and quantitative understanding of the physical, biological, and social interactions within watersheds is of fundamental importance to long-term sectoral planning in Cape Verde. In particular, this information would help to better determine the exploitable subterranean water supplies, the rate of soil erosion, and how

The Future of Agriculture

191

different kinds of land use affect these processes. To capture this complex and interactive set of variables, an integrated research approach—incorporating geological, chemical, biological, meteorological, agronomic, and social processes—is necessary. One of the main objectives of this research should be the identification and, to the extent possible, the quantification of the ways in which particular agricultural (and other land-use) activities affect the various physical processes in watersheds and result in the degradation of the agricultural resource base. Only when the environmental impacts of agricultural activities are understood can policies be designed to address these problems. In order to formulate effective public policies, it is necessary to identify the context of constraints and opportunities in which households make resource-management decisions. The continuing preference of farmers to plant corn and beans in the face of an aggressive policy favoring the erosion-mitigating pigeon pea can only be understood by the fact that these crops provide significant food and feed outputs, even in years of low rainfall. The limited land area of the islands prohibits extensive livestock systems, so there is a great premium on high productivity of feed per hectare, and corn meets this need in addition to providing grain for human consumption. Furthermore, the labor requirements of corn and beans do not impose great strain on the supplies available in most households, with intense demands falling only during a short period of the agricultural cycle. The corn and beans combination provides foods that are well incorporated into local diets, as well as significant amounts of livestock feed, even in years of poor rainfall. Most households cannot expect to produce sufficient production to meet household consumption needs during most years; however, something is better than nothing, and the production cost—family labor that generally has few other opportunities for gainful employment— is very low from the household perspective. The future of dryland agriculture in Cape Verde will respect the past. The introduction of new crops or techniques must compete with the current farming system and its particular food-livestock combination that has evolved through time. To be acceptable, any alternative technology must either address the system logic and present superior results or generate greater overall household welfare through better use of available inputs. For example, technicians have focused attentions on developing systems oriented toward livestock production in order to take advantage of the ecological advantages of permanent forages. However, without significant improvements in vegetative biomass production, existing holdings of 1 hectare or less will probably not sustain herds large enough to offer an economically viable alternative to present agricultural practices. It is possible that the current movement toward participatory watershed planning and integrated upstream/downstream strategies to control

192

The Future of Agriculture

erosion will improve moisture retention and increase dryland yields. However, the extent of potential benefits from this approach in terms of either farmers’ incomes or erosion control remains unknown; and in the absence of good information, it is difficult to assess the policy. While one can imagine the physical possibility—under an increasingly precarious assumption of continued foreign assistance—of entire watersheds terraced from top to bottom and from the highest peaks to the sea, the subsequent impacts on food security and poverty can be surmised only in the most hypothetical, almost romantic, terms. We know from our analysis only that yields would have to jump vertically to make a significant difference. With regard to irrigated agriculture, more information is needed about the institutional forms that govern farmer access to water. Our research in this area focused on a small number of pump-fed irrigation networks in Ribeira Seca, one of the principal watersheds of Santiago. Key informant and group interviews indicated that there is a high degree of variation in management procedures across irrigation networks, so a more detailed cataloging of these variations and evaluation of their effectiveness would create a baseline for comparing possible improvements in management of irrigation networks. The current movement toward small, private water associations and drip technologies presents one of the more optimistic scenarios for Cape Verdean food production. In addition to highlighting these important research issues, the notion of sustainability defined above points to several institutional aspects of policy implementation. First, because erosion and water infiltration occur over an entire watershed, policies to address these processes must also operate at the watershed level. The actions of farmers at the higher reaches of watersheds have very significant effects on individuals at lower elevations, so some public intervention may be necessary to monitor and even regulate the resource-management practices of farmers. The specific form of such interventions must be very carefully evaluated with respect to its distribution of burdens and benefits among individuals and groups in the community. At the same time, the financial and other costs to implement the interventions must be considered. For example, one possible public intervention would be to prohibit agricultural activities in certain portions of watersheds. This strategy would provide benefits to the downstream inhabitants at the expense of those upstream individuals whose activities would be curtailed. On the cost side, the government would have to allocate resources to monitoring activities to ensure that restrictions were obeyed. Alternatively, the government could pay restitution to the landowners, but this too would imply financial strains on the government budget. Agriculture occupies an important position in the economy of Cape Verde. The resource-management issues are particularly crucial in this environment because of the low and unstable productivity of agricultural

The Future of Agriculture

193

resources and the rapidly growing population that depends on these resources. There is widespread concern that the increased pressure on the agricultural resource base is accelerating the rates of degradation. While past policies and investments in agriculture attempted to pursue an overly optimistic goal of dramatic increases in overall production, more recent government priorities have downplayed the emphasis on investment in the agricultural sector. This overall reorientation of national priorities is probably correct in the abstract; however, poverty resides and thrives in the rural reality. Furthermore, there is much that can be done to improve the performance of the sector. Although aggregate growth in agricultural production is unlikely to provide a basis for growth of the whole economy, the sector can and should continue to be developed in the sense of promoting more efficient and productive use of available resources. Even though agriculture as a share of total economic activity can be expected to decline as growth occurs, this does not obviate the importance of absolute growth in agricultural production to provide increased welfare for those households employed in this sector. The fact that such a large portion of the population depends on agriculture and continues to live in poverty places a very high social value on any improvement in agricultural returns. Returning to the dilemma illustrated in Figure 1.1, the gap between the population and the agricultural resource base in Cape Verde will not disappear. Furthermore, overambitious attempts to increase agricultural production are likely to be unsustainable and to exacerbate degradation of the agricultural resource base through erosion and depletion of groundwater supplies. Sustainable agricultural development will be characterized by something like the intermediate path in the figure—with the quality of the agricultural resource stabilizing or perhaps increasing slightly over time. The agricultural output that can be realized under this scenario can be increased somewhat, although not dramatically, through improved production technologies and institutional innovations to utilize water more effectively. The gap between food requirements and production will likely increase, at least in the near future; however, this conclusion does not imply the abandonment of the sector. On the contrary, that the resource base is so limited in comparison with the rapidly growing rural population places a particularly high premium on generating the greatest possible returns without sacrificing the long-run production potential of the agricultural resources of the islands. The other policy area that holds significant potential for improving agriculture and alleviating poverty is found, once again, in the nonagricultural sector. Nature, with particularly vicious intent, controlled the historical population of Cape Verde; perhaps now policy (and culture) can play this role in a much less excruciating manner. The postindependence government placed a high premium on family planning, but the lack of education

194

The Future of Agriculture

among women, especially poor women, has reduced the impacts of this policy strategy. Since family planning is a sensitive issue in a deeply Catholic country, its success must be assessed over the long run when the impacts of widespread investment in education begin to yield fruits. It is evident from the arguments in this book that a downward bend in the population curve is part of the ultimate solution. In conclusion, we must point out that the government of Cape Verde has placed high priority on developing sustainable resource-management policies. The protection of the environment, particularly agricultural resources, is highlighted for public action in the Third National Development Plan. Nevertheless, while the plan discusses the need for development and promotion of new technologies to improve agricultural productivity, there is very little emphasis on increasing total national agricultural production or achieving food self-sufficiency. In these shifting political times, maintenance of the agricultural resource base has become more important than promoting current agricultural production levels. The plan rightly emphasizes the need to coordinate actions within watersheds, since the physical processes that affect soil erosion and water infiltration are all interdependent within these extensive geographical formations. Reforestation, investment in erosion-control structures, subterranean water development and management, and research and extension directed toward household resource-management strategies are highly interrelated activities, linked by their combined impacts on water, soils, and vegetation in watersheds. They should be coordinated so their interactions are complementary rather than contradictory. This commitment to sustaining the agricultural resource base in Cape Verde has a number of implications for public policy. First, the external costs of both rainfed and irrigated agriculture require that the government adopt a more interventionist role. This argument lies in direct contrast to the general philosophical orientation of the new government to reduce state involvement in economic spheres. Within the domain of public resource management, decisions must be made about the appropriate means to intervene in rainfed farming and irrigation water management (i.e., regulations, taxation, public ownership of resources, etc.) and the appropriate institutional forms to execute the programs (public agencies, legal system, user groups, etc.). Whatever final form the interventions take, however, there will be a need for public institutions, since local or private institutions alone cannot be relied upon to adequately resolve external cost problems. The Third National Development Plan cites past inefficiencies as its argument to minimize state intervention in the national economy. This concern remains valid, and the selection of the appropriate forms of public resource-management institutions and policies must be sensitive to past

The Future of Agriculture

195

problems. However, some form of public intervention will be necessary to successfully undertake a sustainable resource-management strategy. The implementation of sustainable resource-management strategies will require additional policies to promote alternative employment opportunities for existing rural populations. Reducing the rate of utilization of agricultural resources will also perforce diminish the demand for agricultural labor, and successful structural transformation within the economy requires a significant reduction in the rural workforce dependent on agriculture. In particular, if total agricultural output is stabilized as a consequence of attempts to reduce overutilization of the resource base, then the only way to increase the returns to agricultural labor is to decrease the amount of labor employed to generate that output. The government has already placed a high priority on education and training as a means of enhancing the employment opportunities for the national workforce. A specific effort to direct this training toward rural populations should help to facilitate the flow of labor into nonagricultural sectors. The major structural reality facing the Cape Verdean economy is that new employment opportunities will come outside the agricultural sector, mostly in urban areas, whereas the majority of the workforce is still in the countryside. The challenge will be to provide rural workers with the necessary skills to take advantage of all employment opportunities that become available. The level of primary education is already very high in rural areas, but secondary and vocational training is scarce in more isolated communities. Finally, the government of Cape Verde has not exhibited a great interest in addressing distributional issues in recent years. This may in part be explained by the social turmoil that resulted from early efforts to redistribute agricultural properties. Although agricultural assets and incomes continue to be very inequitably distributed across rural households (and islands) in Cape Verde, there does not appear to be strong popular pressure to alter the status quo, even with a more open political environment. However, the kinds of interventions that will be necessary to promote sustainable use of agricultural resources will have very significant distributional impacts. On the one hand, attempts to restrict agricultural activities on more marginal lands will fall disproportionately on poor households, those in the higher reaches of the ribeiras without access to irrigated land. On the other hand, increased utilization of water resources will benefit only those farmers in the bottoms of the ribeiras and will fall disproportionately on households already endowed with irrigated land. The strong regressive nature of such public resource strategies must be recognized, and supplemental polices to offset the deleterious distributional impacts should be contemplated. In the short term, focusing public work fronts in the upper ends of ribeiras, for example, can provide income to those families that

196

The Future of Agriculture

must bear most heavily the restrictions on agricultural activities. An added advantage of this strategy is that erosion-control interventions are most effective if they begin at the top of watersheds and work their way down over time. The inescapable fact, however, is that improved utilization of agricultural resources in Cape Verde will necessarily entail concentration of agricultural activities in specific regions and restrictions over wide areas. These changes cannot be evenly apportioned across all rural households; there will be some big winners and a larger number of losers. Policymakers must be aware of this challenge that their commitment to a sustainable resource-management strategy poses. Perhaps there is some uneasiness with the conclusions to our story. We apprehensively admit to no clear policy directives or prescriptions to solve the dilemma of rural Cape Verde, but such is the agricultural reality in much of Sahelian Africa. Instead of solutions, then, there are choices. Policymakers must reconcile poverty alleviation with ecological sustainability, social justice with efficiency, and long-term goals with short-term urgencies. Our analysis has attempted to identify the nature of those trade-offs and the reasonable expectations from informed policymaking. We have also attempted to faithfully relate the situation of rural Cape Verde and to shed outside light upon the daily struggle of rural households as they wait for rain. In the final accounting, it is in their hands that the future of Cape Verdean agriculture will be molded.

References

Adelaja, Adesoji O. 1991 “Price Changes, Supply Elasticities, Industry Organization, and Dairy Output Distribution.” American Journal of Agricultural Economics 73:89–102. Amaral, Ilídio do 1964 Santiago de Cabo Verde: A terra e os homens. Lisbon: Junta de Investigações do Ultramar. Antonovitz, Frances, and Richard Green 1990 “Alternative Estimates of Fed Beef Supply Response to Risk.” American Journal of Agricultural Economics 72:475–487. Bailey, F. G. 1969 Stratagems and Spoils: The Social Anthropology of Politics. New York: Schocken. Ball, Nicole 1976 “Understanding the Causes of African Famine.” Journal of Modern African Studies 14(3): 517–522. Barlett, Peggy F. 1980 “Cost-Benefit Analysis: A Test of Alternative Technologies.” In Agricultural Decision-Making, Peggy F. Barlett, ed., pp. 137–160. New York: Academic Press. Barnum, Howard N., and Lyn Squire A Model of an Agricultural Household: Theory and Evidence. Balti1979 more: Johns Hopkins University Press. Barth, Frederik 1966 Models of Social Organization. London: Royal Anthropological Institute. Becker, Gary 1965 “A Theory of Allocation of Time.” Economic Journal 75:493–517. 1981 A Treatise on the Family. Cambridge, Mass.: Harvard University Press. Beets, Willem C. 1990 Raising and Sustaining Productivity of Smallholder Farming Systems in the Tropics. Alkmaar, Holland: AgBé Publishing. Bellon, Mauricio R., and J. Edward Taylor 1993 “‘Folk’ Soil Taxonomy and the Partial Adoption of New Seed Varieties.” Economic Development and Cultural Change 41:763–786.

197

198

References

Berry R. A., and W. R. Cline 1979 Agrarian Structure and Productivity in Developing Countries. Baltimore: Johns Hopkins University Press. Binswanger, Hans P., and John McIntyre 1987 “Behavioral and Material Determinants of Production Relations in Land Abundant Tropical Agriculture.” Economic Development and Cultural Change 36:73–99. Binswanger, Hans P., and M. R. Rosenzwieg 1986 “Behavioural and Material Determinants of Production Relations in Agriculture.” Journal of Development Studies 22:503–539. Boserup, Esther 1965 The Conditions of Agricultural Growth: The Economics of Agrarian Change Under Population Pressure. Chicago: Aldine. Bromley, Daniel W. 1991 Economic Interests and Institutions: The Conceptual Foundations of Public Policy. Oxford: Basil Blackwell. Bromley, Daniel W., C. Taylor, and D. E. Parker 1980 “Water Reform and Economic Development: Institutional Aspects of Water Management in the Developing Countries.” Economic Development and Cultural Change 28:365–387. BURGEAP 1974 Etude et Mise en Valeur des Eaux Souterraines Dans L’Archipel du Cap Vert. Lisbon: Ministerio de Coordenacao, Brigada de Aguas Subterraneas. Carreira, Antonio 1982 The People of the Cape Verde Islands: Exploitation and Emigration. Translated by Christopher Fyfe. Hamden, Conn.: Archon Books. Cabo Verde: Formação e Extimção de uma Sociedade Escravocrata 1983 (1460–1878). Praia: Instituto Caboverdiano do Livro. 1984 Cabo Verde: Aspectos Sociais. Secas e Fomes do Século XX. Lisbon: Ulmeiro Press. Chayanov, A. V. 1966 The Theory of Peasant Economy. D. Thorner, B. Kerbly, and R. E. F. Smith, eds. Manchester: Manchester University Press. Clark, Gracis, ed. Traders Versus the State: Anthropological Approaches to Unofficial 1988 Economies. Boulder, Colo.: Westview Press. Dalton, George Studies in Economic Anthropology. Washington, D.C.: American An1971 thropological Association. Easter, K. W. 1977 “Improving Village Irrigation Systems: An Example from India.” Land Economics 53:56–66. Eggertsson, T. 1990 Economic Behavior of Institutions. Cambridge: Cambridge University Press. de Janvry, A., M. Fafchamps, and E. Sadoulet 1991 “Peasant Household Behavior with Missing Markets: Some Paradoxes Explained.” Working paper no. 578. Giannini Foundation of Agricultural Economics, University of California, Berkeley, California.

References

199

de Pratere, Marguerite 1986 “Milho/Feijão: Significação, simbolism, e sociedade.” Paper presented to the Social Science Research Conference on Contemporary Cape Verde, Praia, Santiago. Fernandopulle, Denis Aperçu sur les Resources en Eau et les Possibilites d’Amenagement 1977 Hidraulique, Iles du Cap Vert. Praia: United Nations Development Programme. Finan, Timothy J. 1988 “The Farm System Under Duress: Agricultural Adaptations on the Cape Verde Islands.” Human Organization 47:109–117. 1990 “Farmer Perspective on Pigeon Pea Cultivation in the Target Area of Santiago Island.” Working paper no. 9, report no. 0263-CV. Rome: International Fund for Agricultural Development (IFAD). 1993 Socio-Economic Characteristics of Santiago Households in the FIDA/PRODAP (Fundo International de Desenvolvimento Agrícola/ Programa de Desenvolvimento Agro-Pecuário) Intervention Zone. São Jorge, Cape Verde: Instituto de Investigação de Desenvolvimento Agrícola. 1996 “Anthropological Research Methods in a Changing World.” In Transforming Anthropology. Emilio F. Moran, ed. Ann Arbor: University of Michigan Press. Finan, Timothy J., and John Belknap Characteristics of Santiago Agriculture:1984 Report on Survey of 1985 Santiago Agriculture. Praia: USAID/Food Crops Research Project. Finan, Timothy J., and Camilla Harshbarger 1996 “Evaluation of the P.L. 480 Title II Monetization Program.” Report presented to Agricultural Cooperative Development International, Washington, D.C. Finan, Timothy J., and Helen K. Henderson 1988 “The Logic of Cape Verdean Female-Headed Households: Social Response to Economic Scarcity.” Urban Anthropology and Studies of Cultural Systems and World Economic Development 17:87–104. Foy, Colm Cape Verde: Politics, Economics, and Society. London: Pinter Publishers. 1988 Franke, Richard W., and Barbara H. Chasin Seeds of Famine. Montclair, N.J.: Allanheld, Osmun & Co. 1980 Freeman, Peter H., V. E. Green, R. B. Hickok, E. F. Moran, and M. D. Whittaker Cape Verde: Assessment of the Agricultural Sector. Maclean, Va.: 1978 General Research Corporation. Geertz, Clifford 1956 “Religious Belief and Economic Behavior in a Central Javanese Town: Some Preliminary Considerations.” Economic Development and Cultural Change 4:134–158. Gorse, Jean Eugene, and David R. Steeds 1987 “Desertification in the Sahelian and Sudanian Zones of West Africa.” World Bank Technical Paper no. 61, Washington, D.C. Hildebrand, Peter E. “Combining Disciplines in Rapid Appraisal: The Sondeo Approach.” 1981 Agricultural Administration 8(6):423–432.

200

References

IFAD (International Fund for Agricultural Development) 1990 “Community-Based Agricultural and Livestock Development Project. Appraisal Report.” Rome. INGRH (Instituto Naçional de Gestão de Recursos Hídricos) 1993 Esquema Director para os Recursos Hídricos. Praia: Republic of Cape Verde. Langworthy, Mark 1989a Income and Expenditure Patterns of Rural Households on Santiago Island, Cape Verde. Praia: USAID/Food Crops Research Project. 1989b “Temporal Water Allocation and Crop Choice in Small-Scale Irrigation Systems: The Case of Cape Verde.” Working paper no. 58. Department of Agricultural and Resource Economics, University of Arizona, Tucson. 1991 “Economic Viability in Cape Verde.” Journal of Farming Systems Research-Extension 2:109–124. Langworthy, Mark, and Timothy J. Finan 1990 Crop Budgets and Representative Farm Systems for Santo Antão Island, Cape Verde. Praia: USAID/Food Crops Research Project. Langworthy, Mark, Timothy J. Finan, Raul Varela, and Elísio Rodrígues Characteristics of Santo Antão Agriculture. Praia: USAID/Food Crops 1986 Research Project. Langworthy, Mark, Raul Varela, and Timothy J. Finan 1986 Characteristics of São Nicolau Agriculture. Praia: USAID/Food Crops Research Project. Lawrance, Emily C. 1991 “Poverty and the Rate of Time Preference: Evidence from Panel Data.” Journal of Political Economy 99:54–77. Lopes, Vicente L., and John Meyer 1993 “Watershed Management Program on Santiago Island, Cape Verde.” Environmental Management 17:51–57. Lopez, Ramon E. 1986 “Structural Models of the Farm Household That Allow for Interdependent Utility and Profit-Maximization Decisons.” In Agricultural Household Models: Extensions, Applications, and Policy, Inderjit Singh, Lyn Squire, and John Strauss, eds. Baltimore: Johns Hopkins University Press. McIntire, John, Daniel Bourzat, and Prabhu Pingali Crop-Livestock Interaction in Sub-Saharan Africa. Washington, D.C.: 1992 World Bank. MDRP (Ministério de Desenvolvimento Rural e Pescas) Recenseamento Agrícola. Praia: Gabinete de Planeamento, Ministério 1979 de Agricultura, República de Cabo Verde. 1986 Estatísticas Agrícolas. Praia: Ministério de Agricultura, República de Cabo Verde. 1988 Recenseamento Agrícola. Praia: Gabinete de Planeamento, Ministério de Agricultura, República de Cabo Verde. 1989 Estatísticas Agrícolas. Praia: Gabinete de Planeamento, Ministério de Agricultura, República de Cabo Verde. 1990a Recenseamento Agrícola, 1988. Volume 1: Dados Globais. Praia: Ministério de Agricultura, República de Cabo Verde. 1990b Estatísticas Agrícolas. Praia: Gabinete de Planeamento, Ministério de Agricultura, República de Cabo Verde.

References 1992

201

Estatísticas Agrícolas. Praia: Gabinete de Planeamento, Ministério de Agricultura, República de Cabo Verde. Meintel, Deirdre 1984 Race, Culture, and Portuguese Colonialism in Cabo Verde. Syracuse, N.Y.: Maxwell School of Citizenship and Public Affairs, Syracuse University. Monke, Eric A., and Scott R. Pearson The Policy Analysis Matrix for Agricultural Devlopment. Ithaca, N.Y.: 1989 Cornell University Press. Moran, Emilio 1982 “The Evolution of Cape Verde’s Agriculture.” African Economic History 11:63–86. Moran, Emilio F., ed. Transforming Societies, Transforming Anthropology. Ann Arbor: Uni1996 versity of Michigan Press. Netting, Robert McC. 1993 Smallholders, Householders: Farm Families and the Ecology of Intensive, Sustainable Agriculture. Stanford, Calif.: Stanford University Press. Netting, Robert McC., Richard R. Wilk, and Eric J. Arnould Households: Comparative and Historical Studies of the Domestic 1984 Group. Berkeley: University of California Press. Norgaard, Richard B. 1984 “Traditional Agricultural Knowledge: Past Performance, Future Prospects, and Institutional Implications.” American Journal of Agricultural Economics 66:874–878. North, Douglass, C. 1990 Institutions, Institutional Change, and Economic Performance. Cambridge: Cambridge University Press. Pollack, R. A. 1985 “A Transaction Cost Approach to Families and Households.” Journal of Economic Literature 23:581–608. Reidinger, R. B. 1974 “Institutional Rationing of Canal Water in Northern India: Conflict Between Traditional Patterns and Modern Needs.” Economic Development and Cultural Change 23:79–104. Ribeiro, Orlando 1954 A Ilha do Fogo e as suas Erupções. Lisbon: Junta de Investigações do Ultramar. Repetto, Robert 1986 World Enough and Time. New Haven, Conn.: Yale University Press. Republic of Cape Verde 1982 Food Strategy Study. “Volume 1: General Outline of the Strategy.” Praia. 1992 Ill Plano Nacional de Desenvolvimento 1992-1995. “1 Volume: Relatório Geral.” Praia: Ministério das Finanças e do Planeamento. Richard, Jacques 1983 Aspects de l’imigration rurale dans l’archipel du Cap Vert. Republique du Cap Vert, Secretariat d’Etat à la Cooperation et à la Planification. Richards, Paul Indigenous Agricultural Revolution: Ecology and Food Production in 1987 West Africa. London: Hutchinson.

202

References

Ruttan, Vernon 1988 “Sustainability Is Not Enough.” American Journal of Alternative Agriculture 3 (2 and 3):128–130. SCET-AGRI 1985a Ilha de Santiago: Monograph. Paris: Ministere Français des Relations Exterieures Cooperation et Developpement. 1985b Ilha de Santo Antão: Monograph. Paris: Ministere Français des Relations Exterieures Cooperation et Developpement. 1985c Ilha de São Nicolau: Monograph. Paris: Ministere Français des Relations Exterieures Cooperation et Developpement. 1985d Ilha do Fogo: Monograph. Paris: Ministere Français des Relations Exterieures Cooperation et Developpement. Segurado, Joaquim Bivar 1983 “Estratificação Social e Necessidades Básicas num Ecossistema Frágil.” Report presented to the Instituto Gulbenkian de Ciência, Oeiras, Portugal. Singh, Inderjit, Lyn Squire, and John Strauss, eds. 1986 Agricultural Household Models: Extensions, Applications, and Policy. Baltimore: Johns Hopkins University Press. Soares, Horácio da Silva 1984 Assessment of Agricultural Research Resources in the Sahel. Volume 3, National Report: Cape Verde. Bamako, Mali: Permanent Interstate Committee for Drought Control in the Sahel. Socio-Economics Studies Team 1987 Crop Budgets for Selected Irrigated Crops on Santiago and Santo Antão Islands. Praia: USAID/Food Crops Research Project. Stevens, Robert D., and Cathy L. Jabara 1988 Agricultural Development Principles: Economic Theory and Empirical Evidence. Baltimore: Johns Hopkins University Press. Traxler, Greg, and Derek Byerlee 1993 “Joint-Product Analysis of the Adoption of Modern Cereal Varieties in Developing Countries.” American Journal of Agricultural Economics 75:981–989. Trigo de Abreu, Armando, and Horácio Soares Desenvolvimento e Pesquisa no longo prazo em Cabo Verde: Mesa 1983 Redonda. Oeiras, Portugal: Fundação Calouste Gulbenkian. USAID (United States Agency for International Development) 1984 “Cape Verde Watershed Development,” project paper. Washington, D.C. van Willigen, John, and Timothy J. Finan 1990 “Soundings: Rapid and Reliable Research Methods for Practicing Anthropologists.” NAPA (National Association of Practicing Anthropologists) bulletin no. 10, American Anthropological Association. Varela, Raul, Tito Andrade, Mark Langworthy, and Timothy J. Finan 1988a Contas Culturais para as Principais Culturas em Santiago e Santo Antão. Praia: USAID/Food Crops Research Project. 1988b Descrição e Analise dos Sistemas Agrícolas Representativas em Santiago. Praia: USAID/Food Crops Research Project. Varela, Raul J., Mark Langworthy, and Timothy J. Finan 1990 “A Socio-Economic Survey of Selected Localities in the Island of Santiago, Cape Verde.” Report presented to the International Fund for Agricultural Development, Rome.

References

203

Veiga, Manuel 1982 Diskrison Strutural di Lingua Kabuverdianu. Praia: Praia Institutu Kabuverdianu Di Livru. Wade, R. A. 1988 Village Republics: Economic Conditions for Collective Action in South India. Cambridge: Cambridge University Press. Wharton, Clifford R. 1971 “Risk, Uncertainty, and the Subsistence Farmer: Technological Innovation and Resistance to Change in the Context of Survival.” In Studies in Economic Anthropology, George Dalton, ed. Washington, D.C.: American Anthropological Association. Wolf, Eric R. Europe and the People Without History. Berkeley: University of Cal1982 ifornia Press. World Bank 1993a African Development Indicators. Washington, D.C.: World Bank. 1993b Social Indicators of Development, 1993. Baltimore: Johns Hopkins University Press. 1993c World Bank Development Report, 1993. Oxford: Oxford University Press. Zilberman, David, and Michell Marra 1993 “Agricultural Externalities.” In Agricultural and Environmental Resources Economics, Gerald A. Carson, David Zilberman, and John A. Miranowski, eds. New York: Oxford University Press.

Index

Agricultural extension service, 161, 169–170 Agricultural policy, 2, 9–11, 159–164, 168, 183–184; alternatives, 172– 182 Agricultural production: early, 3, 58–59 Agricultural research, 169–170, 175–176, 184, 190–191 Agricultural resource base, 4, 14, 15(fig.), 20, 164–167, 188, 192, 193, 194. See also Household resource management; Resource management Agricultural resource degradation. See Environmental degradation Agricultural technologies, 3, 22, 23, 25, 29–30, 32, 71, 75, 82; alternative, 191; water-capturing, 40, 41, 104–105, 142–144. See also Indigenous knowledge Agriculture. See Irrigated agriculture; Rainfed agriculture; Subsistence agriculture; Sustainable agriculture Agriculture, Ministry of. See Fisheries, Agriculture, and Rural Extension, Ministry of Agroclimatic zones, 38, 39(table) Alluvial flows. See under Water sources Amaral, Ilídio do, 76 Aquifers. See under Water sources Arnould, Eric J., 14 Arrendamento, 68 Arretes, 48, 49 Assento de casa, 69

Ball, Nicole, 3 Bananas, 30, 109, 114(table), 116, 143(table); crop budget, 119, 120–121(table), 122; cultivation, 112–113, 115, 116, 117, 124, 143(table), 174(table); income, 174(table); in representative irrigated systems, 124, 125, 126, 127(table) Banquetas, 48 Barlavento islands, 1 Barter, 59 Beans, 58; crop budget, 90(table), 147; cultivation, 41, 67, 80, 81, 182, 191; economic returns, 93, 94, 184; research, 169; tradition of growing, 141, 142, 182, 191; yields, 83, 84(table), 87, 90(table), 147 Binswanger, Hans P., 71 Biodiversity, 51(n1), 162 Birthrate, 63 Boa Vista Island, 1, 35, 57, 61(table) Boserup, Esther, 3, 59 Brava Island, 1, 35, 60, 61, 67, 72(n2), 80(table) Bromley, Daniel W., 17 Budget analysis. See Crop budgets; Livestock, budget Cachupa, 76, 80, 81, 88, 99(nn 1, 2), 111 Caldeiras, 48, 50 Canals, 105, 107, 146 Capelas, 57, 58

204

Index Cape Verdean National Agricultural Research and Development Institute (Instituto Nacional de Investigação e Desenvolvimento Agrário [INIDA]), 26, 169–170, 175 Capitalist production, 116 Capitanias, 57 Captações, 46, 48, 105, 107, 108, 143, 146 Carreira, Antonio, 5, 54, 56, 57, 59 Cash crops, 3, 10, 57, 58, 101, 124. See also individual crops Cassava, 109, 111, 114(table), 143(table), 146, 169; crop budget, 119, 120–121(table), 122; cultivation, 111, 114, 124; income, 174(table); in representative systems, 124, 125, 127(table) Catchment dams, 48, 143 Catholic Church, 62, 65, 194 Cattle, 47, 84, 91, 92, 97(table), 152(table) Chasin, Barbara H., 3 Check dams, 48, 118 Chickens, 84, 92, 93(table), 94, 98(table) Children, 75, 84, 86(table), 88 Cidade Velha, 54 Class, 5, 55, 56, 61, 62, 77(nn 2, 3) Climate, 3, 4, 35, 36–38, 39(table), 82 Colonial land institutions, 57, 62 Commerce, 1–2, 19 Companhia do Grão Pará e Maranhão, 56 Conservation structures, 41–42, 48–49, 96; maintenance, 30, 49, 180–181 Conta própria, 67 Corn, 4, 9, 58, 169; in crop budget, 90(table), 147; cultivation, 67, 80, 81, 182; economic returns, 93, 94, 184; tradition of growing, 9, 141, 142, 182, 191; yields, 83, 84(table), 87, 89, 90(table), 147. See also Straw Cotton, 2, 57, 58, 59 Credit availability, 118 Crioulo, 27, 55, 56, 58, 78(n4) Crop budgets, 22–23, 29, 30, 31, 33(n7), 147. See also under Livestock; Irrigated agriculture;

205

Rainfed agriculture; Santiago Island; Santo Antão Island; individual crops Crop mixes, 124–125, 127(table). See also Intercropping Crop selection, 175–176, 179 Crop substitution, 9, 50, 142. See also Pigeon pea Dams, 46, 48, 105, 118, 143. See also Captações Darwin, Charles, 51(n1) Deforestation, 49. See also Reforestation DGAR. See General Directorate for Rural Extension DGRA. See General Directorate for Land Reform Diet, 76, 77, 80, 83–84, 86–87, 99(n2). See also Cachupa; Food sufficiency Diques de correcção torrencial, 48 Djuda, 72, 73 Djunta-mon, 72–73, 81, 86(table), 87, 88 Drought, 37, 38, 61(table); and population, 5–6, 59–60, 61(table) Economy, 3, 162 Education, 8, 19, 27, 162; primary/secondary, 66, 163, 195 Emigration, 6, 15(fig.), 16, 19, 22, 24, 27, 60, 62(table), 63, 78(n9), 96, 188, 190; in income calculation, 149(table); of women, 60–61, 62 EMPA. See Empresa Nacional de Abastecimento Employment: agricultural, 53; off-farm, 6, 27, 137(table), 140, 152–153; private sector, 162. See also Labor, household; Labor, wage; Public work fronts Empresa Nacional de Abastecimento (EMPA), 74 Environmental degradation, 2–4, 50–51, 162, 164–167, 188–189. See also Erosion Erosion, 45, 46; and cultivation practices, 9, 47–48, 96; external costs, 165, 166–167, 184. See also Flooding; Rainfall; Soil, loss Erosion control, 10, 30; policy, 96, 154, 161, 162, 180, 192, 196; in

206

Index

rainfed agriculture, 162, 182, 183; research, 169, 184. See also Conservation structures; Public work fronts Exports, 3, 57, 59, 112, 116, 161 External costs, 164–167, 182, 184, 194 FAIMOs (frentes da alta intensidade de mão-de-obra) 74, 170–171, 181. See also Public work fronts Fallow lands, 59, 107 Family, 14, 26, 27; extended, 65, 66, 73. See also Households Family planning, 193–194 FAP. See Fomento Agro-Pecuário Farm size, 33(n4), 58, 68–71, 85, 86, 102, 103 FDN. See National Development Fund Feitor, 57 Fernandopulle, Denis, 41 Ferta, 76 Fertilizers, 82, 115, 117, 119, 122 Fisheries, Agriculture, and Rural Extension, Ministry of (MPAAR), 104, 169, 170, 185(n6) Fishing, 75 Flooding, 37, 41, 47, 49; control, 48 Flood-recession agriculture, 104 Fogo Island, 1, 4, 35, 49, 50, 57, 60, 61, 64, 77(n2), 78(n9); agricultural land, 38, 39(table), 43, 45, 46(table), 51(n3), 58, 61(table), 67, 80(table) Fomento Agro-Pecuário (FAP), 112, 115, 116, 117, 122 Food aid, 6, 10, 74, 160–161, 185(n2), 187; monetized, 10, 74, 172 Food Crops Research Project (INIDA), 26, 169–170, 175 Food imports, 10, 59, 160, 161, 190 Food sharing, 76–77 Food sufficiency, 5, 8, 9, 151, 160, 161, 189, 191; and rainfed agriculture, 86–87, 91, 95. See also Diet Foreign aid. See Food aid; International aid Forests. See Reforestation Franke, Richard W., 3 Freeman, Peter H., 36, 44 Fuelwood, 27, 47, 49, 50, 81, 88, 94 Furos, 105, 106, 143–144

Galerias, 40, 46, 105 General Directorate for Land Reform (Direcção Geral de Reforma Agrária) (DGRA), 168 General Directorate for Rural Extension (Direcção Geral de Animação Rural) (DGAR), 169 Goats, 47, 57, 58, 84, 91, 92, 93(table), 94, 97(table), 152(table) Gomes, Diogo, 57 Gomes, Fernão, 55 Government: administrative structure, 64–65 Grogue, 10, 73, 74, 109–111, 116, 117, 122 Groundwater, 35, 40; jurisdiction, 17, 18, 41, 106–107, 143, 168–169; policy, 10, 162; potential, 42–43, 176, 178. See also Irrigation water; Rainfall; Watersheds; Water sources Guinea Coast (Africa), 54, 55 Harmattan, 38 Harvesting practices, 81–82, 115–116 Health care, 188 Hoe, 71, 75, 82 Household resource management, 14, 15(fig.), 17, 19, 20, 21, 31–32, 33(nn 2, 5), 132, 190, 192, 193, 194–195; data collection and analysis, 25–33; representative budget for, 21–25, 34(n8); resource access, 4, 5, 14, 24, 26, 27, 30–31, 32, 33; resource transfer, 16–17, 33(n4); and risk, 24, 34(n9) shortterm/long-term, 4, 20, 24–25, 132, 167. See also Employment; Income; Institutional environment; Labor, household; Labor, wage; Land Households: in community, 65, 76–77; family structure, 14, 62(table), 63, 65–66, 75, 78(n9); maintenance, 19, 27, 88, 94. See also Children; Family; Land; Women Imports, 59 Income, 4, 15(fig.), 16, 19, 20, 24, 30–31, 126; alternatives for increasing, 172–173, 174(table), 176, 177(table), 178–180; delayed returns on, 24–25, 34(n10); in irrigated

Index system, 120–121(table), 123–124, 126, 127(table), 128, 147, 148, 149(table), 150, 152(table); off-farm, 75, 148, 149(table), 150, 152–153; per capita, 149(table), 151; in rainfed system, 93–94, 148, 149(table), 150, 152(table). See also Crop budgets Independence, 5, 6 Indigenous knowledge, 4, 26, 29, 65, 81, 82 INERF. See National Institute for Rural Engineering and Forestry Inflation, 31 Infrastructure, 74, 170 INGRH. See National Institute for Water Resources Management Inhame, 109, 111, 112, 114, 121(table), 127(table) Inheritance, 69, 85, 168; partible, 14, 17, 67–68 INIDA. See Cape Verdean National Agricultural Research and Development Institute Insects. See Pests Institutional environment, 14, 15(fig.), 16, 17–19, 20, 25, 26, 33(n2); sharing institutions, 5, 76–77. See also Land-ownership; Land rental; Sharecropping Intercropping, 81, 82, 90(table), 93, 111, 182 International aid, 48, 50, 74, 161. See also Food aid Intertropical convergence (ITC), 36, 37 Irrigated agriculture, 10, 35, 42, 67; capital costs, 30, 118, 120–121(table), 123; crop budgets, 118–119, 120–121(table), 122–124; cropping systems, 124–125, 127(table), 128; cycle of, 42, 67; market activity, 67, 77; purchased inputs, 117–118. See also Irrigated land; Irrigation intervals; Irrigation water; under Income; Labor, household; Labor, hired; Santiago Island; Santo Antão Island Irrigated land, 42, 46–47, 51; access/distribution, 67, 69, 70–71, 72(table), 101–104, 128, 133–135, 136; cost, 104, 129(n2); plot size, 102–103, 129(n1); potential, 42, 43,

207

46(table), 47, 173, 176. See also Irrigated agriculture; under Landownership Irrigation intervals, 27, 107–108, 119, 120–121(table), 127(table), 143–146, 184; and income, 173–175, 177(table), 178 Irrigation networks: management, 17, 22, 30, 105–106, 165–166, 169, 174–175, 184, 192; scheduling, 22, 24, 105–107, 178 Irrigation water: capture of, 104–108, 142–146; costs, 22, 106, 117, 122–123, 175; external costs, 165–166, 184; potential, 43, 47, 173, 178; saltwater contamination, 44, 106, 107, 129(n3); seasonal access, 107; temporary, 41, 104, 106(table) See also Groundwater; Irrigation intervals; Watersheds; Water sources ITC. See Intertropical convergence Justino Lopes state farm, 112, 129(n5) Kin-based settlement, 64, 65 Labor, household, 14, 19, 24, 27, 57, 58; cyclic demands, 71, 94, 136–137; exchange, 18, 19, 72–73, 78(n11), 86(table), 87, 88; in irrigated system, 114–116, 119, 120–121(table), 123, 126, 127(table); and land relationship in rainfed and irrigated systems, 132–135, 136, 137, 138–139(table), 140, 142, 154–156, 157(fig.); off-farm, 27, 36, 137, 138–139(table), 140, 153, 154–156, 157(fig.); in rainfed system, 71, 86(table), 87, 88, 89–90; slave, 57; surplus, 87, 91, 95–96, 136, 137, 140. See also Labor, wage; Public work fronts; under Women Labor, wage, 19, 24, 27; in irrigated system, 71, 73–74, 87, 116, 123, 136, 140, 156(n1); in rainfed system, 86(table), 87. See also Employment; Labor, household Lançados, 55, 56 Land, 44–47, 51(n4), 59, 168; acquistion of, 67–69, 133–136; colonial institutions, 57, 62;

208

Index

expropriation, 85, 161, 168. See also Inheritance; Irrigated land; Landownership; Land rental; Rainfed land; Sharecropping; under Labor, household Land-ownership, 14, 17, 67–68, 71, 72(tables), 103–104; absentee, 85; of irrigated land, 103, 133, 156(n1); of rainfed land, 85; transfer, 17, 18, 19, 168, 185(n5) Land preparation, 47, 80, 115, 136, 137 Land reform, 18, 68, 69, 161, 163, 168, 195 Land Reform, Office of (MPAAR), 104 Land rental, 58, 68, 69, 85, 103–104, 161 Leisure, 154, 155 Levadas, 105, 107 Literacy, 66 Livestock, 2, 3, 16, 19, 47, 57, 58, 83–85, 88, 191; budget, 22–23, 31, 33(n7), 90–91; fodder, 81–82, 83, 110, 141; income, 27, 141–142, 147–148, 149–150, 151–152, 179, 182; manure, 82, 83, 92(table), 114, 115, 117; net values and outputs, 91, 92, 97–98(table). See also individual animals Maio Island, 1, 35, 57, 61(table) Maize. See Corn Market access, 75, 113, 114, 116–117; in household resource management, 14, 16, 22, 26, 33 Marriage, 65–66, 69 MDRP. See Fisheries, Agriculture, and Rural Extension, Ministry of Merim, 107, 123 Migration, 60. See also Emigration Millet, 9, 142 Mindelo, 57, 116 Miscegenation, 55, 61 MOP. See Public Works, Ministry of Moran, Emilio, 4 Morgadio, 57, 58, 85 Mortality rate, 5, 61(table), 63 MPAAR. See Fisheries, Agriculture, and Rural Extension, Ministry of Nascentes, 40, 46, 64, 104–105, 106(table), 107, 108, 143, 146 National Development Fund (Fundo de

Desenvolvimento Nacional) (FDN), 172 National Institute for Rural Engineering and Forestry (Instituto Nacional de Engenharia Rural e Florestas) (INERF), 170, 171, 185(n7) National Institute for Water Resources Management (Instituto Nacional de Gestão de Recursos Hídricos (INGRH), 41, 42, 106–107, 143, 168–169 Natural resource base. See Environmental degradation; Groundwater; Land; Rainfall; Water sources Negros forros, 55 Noli, António de, 54, 57 Opportunity costs, 25, 34(n8), 87 PAICV. See Partido Africano da Independência de Cabo Verde Panos, 2, 57, 59, 78(n5) Parceria, 68 Partido Africano da Independência de Cabo Verde (PAICV), 68, 85 Pastureland, 46(table), 47, 51, 182 Pensions, 24 Pests, 83, 111; control, 115, 117, 119, 122 Pigeon pea (Cajanus cajan), 50, 81, 82, 127(table), 142, 191; crop budget, 90(table), 121(table); as erosion control, 182; in representative systems, 93, 94 Pigs, 84, 92, 93(table), 94, 98(table), 152(table) Pilares, 46 Poços, 105, 106, 118, 143 Political ecology, 13–14, 15(fig.) Pombal, Marquês de, 56 Population, 1, 53; and agricultural capacity, 2, 6, 7(fig.), 8; dependency ratio, 63; and drought, 5–6, 59–60, 61(table) Population growth, 3, 4, 6, 59, 63, 183, 187–188 Portugal/Portuguese, 1, 4, 5–6, 9, 54, 55, 56, 60 Potatoes, 111, 117. See also Inhame; Sweet potatoes

Index Poverty, 149(table), 151, 161, 167, 173, 188 Praia, 37, 44, 56, 63, 113, 129(n3) Prices: in household resource management, 21, 22, 25, 31, 33(nn 6, 7) Príncipe, 60 Privatization, 163 Proprietários, 103 Public enterprises, 185(n3). See also State farms Public work fronts, 9, 16, 22, 48, 63, 74–75, 76, 96, 160, 195–196; funding, 171, 172; local decisionmaking, 180–182; operations, 170–172; social goals, 170, 173 Public Works, Ministry of (Ministério de Obras Públicas) (MOP), 170 Pumps, 106–107, 108, 118 Rabidantes, 75, 113, 114, 116 Race, 55, 61, 77(n2) Rainfall, 3, 4, 6, 14, 36–38, 39(table), 40, 41, 51(n2); runoff, 9, 41, 45, 183, 184. See also Flooding; Groundwater Rainfed agriculture, 4, 79–80, 87; crop alternatives, 9, 142, 179; crop budgets, 88–90, 92; growing cycle, 67, 73, 80–83; policy, 182, 183–184; technology, 71, 82. See also Erosion; Livestock; Pigeon Pea; Rainfed land; Subsistence agriculture; under Erosion control; Income; Labor, household; Labor, wage; Santiago Island; Santo Antão Island Rainfed land, 45–46; distribution, 69, 70(table), 72(table), 86; farm size, 86; household access to, 68–69, 85–87, 133, 136; management, 179, 181–182; rights to, 68. See also Rainfed agriculture; under Landownership Rangeland, 46, 47, 50 Rapid rural appraisal (RRA), 26, 30 Reforestation, 42, 47, 48, 49–50, 74, 96, 162, 170, 181 Regadio temporário, 41, 104, 106(table) Resource access. See under Household resource management

209

Resource base. See Agricultural resource base; Household resource management Resource conservation. See Conservation structures; Erosion control Resource management, 4, 5, 16, 17, 192, 193, 194–195. See also Household resource management Ribeira Grande, 54, 56 Ribeiras, 35, 41, 43, 49, 64, 104, 105; associations, 170, 171. See also Groundwater; Watersheds Ribeiro, Orlando, 4 Risk, 24, 34(n9), 147 Road construction, 170 Rosenzweig, M. R., 71 RRA. See Rapid rural appraisal Rural-urban migration, 8, 9, 10, 74, 96, 161, 190 Sahel, 1, 2, 3, 35, 36, 38 Sal Island, 1, 35, 61(table), 175 Saltwater intrusion, 44, 106, 107, 129(n3) Sand extraction, 129(n3) Sanitation, 188 Santa Luzia Island, 1 Santiago Island, 1, 4, 26, 27, 51(n1), 57, 64, 85, 129(n3), 141; crop budgets, 90(table), 118–119, 120(table), 122–123; crops, 50, 81, 82, 87–88, 89, 110, 111, 112, 113, 116, 117; emigration, 60, 62(table); household structure, 62(table), 63, 65, 66, 75, 78(n9); income, 74, 147, 148, 149(table), 150–151, 152, 153, 172, 173; income projections from alternative practices, 176, 177(table), 178, 179, 180; irrigated land, 42, 43, 44, 70–71, 72(table), 102–104; irrigated systems, representative, 125, 126, 127(table); irrigation access, 104–105, 106, 107, 108, 143–146, 175; labor and land relationships, 86(table), 87–88, 133, 134, 135, 137, 138–139(table); livestock, 57, 84, 141–142; natural characteristics, 35, 38, 39(table), 45, 46, 47, 49, 61(table), 67; rainfed system, 48, 69–71, 72(table),

210

Index

80(table), 86–88, 142; rainfed system, representative, 92, 93–94; watershed plans, 171 Santo Antão Island, 1, 26, 27, 58, 64, 85; crop budgets, 90(table), 118–119, 121(table), 122–123; crops, 50, 82, 89, 110, 111, 112, 113, 114, 116; emigration, 60, 62(table), 78(n9); household structure, 62(table), 63, 65, 75; income, 74, 147, 148, 149(table), 150–151, 152, 153, 173; income projections from alternative practices, 176, 177–178, 179, 180; irrigated land, 42, 43, 44, 49, 70–71, 72(table), 87, 102–104, 115, 116, 117, 156(n1); irrigated system, representative, 125, 126, 127(table); irrigation water access, 104–105, 106, 107, 108; labor and land relationships, 86(table), 87–88, 115, 116, 133, 134, 135, 137, 138–139(table); natural characteristics, 35, 39(table), 43(table), 45, 46, 47, 48, 61(table), 67; rainfed system, 48, 69–70, 80(table), 86, 142; rainfed system, representative, 92, 93(table), 94– 95 São Nicolau Island, 1, 26, 64, 82; emigration, 60, 62(table), 78(n9); household structure, 62(table), 63, 66; land characteristics, 35, 39(table), 43(table), 45, 46, 48, 58, 61(table), 67; rainfed land; 80(table) São Tomé, 60 São Vicente Island, 1, 35, 61(table), 114, 116 SCET-AGRI study, 42, 43, 83 Seca verde, 37 Seeding, 80–81, 99(n3) Seeds: costs, 115, 117, 122; selection, 82; sharing, 76 Segurado, Joaquim Bivar, 68 Sequential rural research, 25–26, 31, 32 Série Supérieure (geologic formation), 40, 44 Settlement, 1–2, 54, 55, 58, 64–65 Sharecropping, 58, 59, 69, 85, 103, 104, 127(table), 147; abolition, 18, 168; terms of, 68, 78(n11), 104,

156(n4) Sharing institutions, 5, 76–77 Slaves, as agricultural labor, 57 Slaves, runaway, 57, 58, 64, 78(n6) Slave trade, 2, 54–55, 56 Social institutions, 5 Society, 5, 53–54, 55, 77(n3) Soil: composition, 44, 169; conservation, 48, 74, 170; fertility, 6, 14, 44, 45, 46, 82; loss, 6, 9, 37, 41, 47, 115. See also Conservation structures; Erosion control Sorghum, 9, 142 Sotavento islands, 1 Springs. See Nascentes Starvation, 5, 6, 60 State farms, 112, 129(n5), 185(n3) Straw, 16, 81–82, 90(table), 110, 141–142, 147–148 Subsidies, 163 Subsistence agriculture, 2, 57, 58, 59, 67, 68, 78(n6), 86–87. See also Rainfed agriculture Sugarcane, 10, 57, 114(table); crop budget, 119, 120–121(table), 122, 123; cultivation, 73, 109–111, 115–116, 124, 143(table), 146; income, 174(table); processing, 73, 117–118, 129(nn 4, 6), 156(n4); in representative irrigated systems, 124, 125, 126, 127(table) Sustainable agriculture, 6, 7(fig.), 8–9, 162, 189–190, 193 Sweet potatoes, 94, 109, 114(table), 143(table), 146; crop budget, 90(table), 119, 120–121(table), 122, 123; cultivation, 93(table), 111–112, 114, 124; income, 93(table), 174(table); in representative irrigated systems, 124, 125, 127(table). See also Inhame Tangamaos, 55 Taxation, 118, 122, 182 Tchuba, 51(n2), 79 Technologies. See Agricultural technologies Tenants’ rights, 163, 168 Terracing, 45, 46, 48, 49, 110, 111, 115, 118 Third National Development Plan for

Index 1992–1995, 162, 194 Topography, 35, 37–38 Tornador-em-tornador (irrigation distribution system), 107, 108. See also Pumps Tourism, 161 Trade, 1, 54, 55–56, 57, 59 Transportation, 78(n8), 82, 83, 115, 119, 123 Tuber crops, 109, 111–112, 116. See also individual crops Unemployment, 162. See also Labor, surplus Urban access, 64, 78(n8). See also Rural-urban migration U.S. Agency for International Development (USAID), 26 Vegetable crops, 109; crop budget, 119, 120–121(table), 122, 123–124; cultivation, 113–114, 115, 116, 117, 143(table), 146; income, 174(table); in representative irrigated systems, 124, 125, 126, 127(table) Vegetation, 49–50 Vilas, 64, 66 Vocational training, 163, 195 Wage labor. See Labor, wage; Public work fronts Water conservation, 170. See also Conservation structures

211

Water gathering: household, 88, 94 Water resource policy. See Watersheds Watersheds, 35, 40; management, 41–42, 48, 184, 195; participatory planning, 170, 171, 191–192; research, 190–191; See also Groundwater; Irrigation water; Ribeiras Water sources: alluvial flows, 40, 41, 42, 46, 48, 67, 104; aquifers, 35, 40, 41, 42, 43, 48, 67, 104, 162; capturing technologies, 40, 41, 46, 104–105, 142–144; surface, 40, 42, 104. See also Groundwater; Irrigation water; Rainfall; Watersheds; individual technologies Water tunnels. See Galerias Wealth, 62 Weeding practices, 48, 78(n11), 81, 115, 137 Wells, 105, 106, 142, 168, 178. See also Furos; Poços Wilk, Richard R., 14 Wolf, Eric R., 53 Women, 53; emigration, 60, 78(n9); in households, 14, 60–61, 62(table), 65, 66, 78(n9); labor, 73, 75, 86(table), 88, 114, 119, 123; in population, 62(table), 63; produce marketing, 75, 113, 114, 116; single, 66 Yams. See Inhame

About the Book

This ethnographic study of Cape Verde tackles critical development issues: the struggle for self-sufficient food security, the tension between agricultural production and natural resource sustainability, and the appropriate role of government policy in food production and natural resource management. Cape Verde has moved into an ecological imbalance between the sustainable production capacity of the resource base and the size of the population seeking to derive a livelihood from agriculture. Within this context, the authors discuss the current agricultural practices and the forms of indigenous knowledge employed by Cape Verdean farmers; estimate the sources and magnitude of incomes among rural households; describe the natural resource base available to farmers, emphasizing the nature of constraints to production; and document the rules, both public and traditional, that define resource access and management. They also describe the past and present policies that Cape Verdean leaders have directed toward the problem of food production in a fragile ecosystem. All Sahelian countries face the challenge of balancing the demand for food (and income) with the capacity of the resource base to respond. Langworthy and Finan’s assessment of agriculture in Cape Verde informs a thoughtful discussion of policy alternatives that may help relieve the current plight and avoid even more serious consequences in the near future. Mark Langworthy is assistant research scientist in the Department of Agricultural Economics, University of Arizona. Timothy J. Finan is director of the university’s Bureau of Applied Research in Anthropology.

212