Critical Debates in Public Finance 9783631810743, 9783631813492, 9783631813508, 9783631813515

Citation preview

Critical Debates in Public Finance

Adnan Gerçek and Metin Taş (eds.)

Critical Debates in Public Finance

Bibliographic Information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available online at http://dnb.d-nb.de. Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress.

Printed by CPI books GmbH, Leck ISBN 978-3-631-81074-3 (Print) E-ISBN 978-3-631-81349-2 (E-PDF) E-ISBN 978-3-631-81350-8 (EPUB) E-ISBN 978-3-631-81351-5 (MOBI) DOI 10.3726/b16603

© Peter Lang GmbH Internationaler Verlag der Wissenschaften Berlin 2019 All rights reserved.

Peter Lang – Berlin ∙ Bern ∙ Bruxelles ∙ New York ∙ Oxford ∙ Warszawa ∙ Wien All parts of this publication are protected by copyright. Any utilisation outside the strict limits of the copyright law, without the permission of the publisher, is forbidden and liable to prosecution. This applies in particular to reproductions, translations, microfilming, and storage and processing in electronic retrieval systems. This publication has been peer reviewed. www.peterlang.com

Preface This book examines the main issues discussed in the field of public finance today. These issues are perhaps identified among policy areas that will come to the agenda of many governments over the next decade. Topics covered in the book are as follows; revenue forecasting models, the taxation of sharing economy, tax incentives provided to green bonds in financing of energy efficiency, the importance of tax literacy in tax compliance, the concept of collective investment institutions, digitalization of tax administration and complexity of tax system, macro determinants of pharmaceutical spending, tax expenditures as internal tax bleedıng, the size of the public sector and the Armey Curve, Okun’s Law, subsidies granted to the private educational institutions, and taxation of artificial intelligence. The book consists of twelve chapters on “controversial issues in the public finance” mentioned above. The large part of chapters published in this volume was selected among the presented papers in the 34th International Public Finance Conference/Turkey in April 2019. They also went through a review process before publication. Erdoğdu and Yorulmaz compare the performance of three forecasting tax revenue models such as Random Walk, SARIMA, and BATS for Turkey throughout 2006:01 to 2018:12. They find that using the BATS model, rather than classical (SARIMA) in forecasting series of monthly tax revenues of Turkey, provides more accurate forecasts. The empirical findings of this study help the experts in the preparation process of the government’s budgets. Bozdoğanoğlu emphasizes that the sharing economy is a functioning economy through online platforms and makes it difficult to evaluate within the framework of tax and legal regulations, such as the traditional economy. In this study, taxes, which are the subject of sharing economy, which is a new economic model, and cooperation with platforms and determination of taxpayer awareness, are included. Yiğit Şakar discusses the financing of energy efficiency and argued that as an alternative to financing energy efficiency, green bonds are developing rapidly all over the World. Green bonds are financial instruments that provide opportunities for investors to participate in the financing of “green” projects that help reduce the negative impacts of climate change and adapt to the effects of climate change, reduce CO2 emissions, prevent environmental pollution, and improve social welfare. These structures have an essential impact on the realization of sustainable development.

6

Preface

Çetin Gerger, Bakar Türegün and Gerçek highlight the importance of tax literacy as one of the factors that determine tax compliance. They also examine arrangements and projects related to tax literacy in the OECD countries and the United States, along with the presentation of the projects and research related to tax literacy in Turkey. In the study, they conclude that the level of tax literacy in Turkey is not at the desired level. Thus they provide suggestions regarding the activities that could be conducted to increase tax literacy. Keskin evaluates the importance of collective investment institutions operating in the World under three legal structures, namely investment company, trust, and contractual model, to enable investors with low savings to work in the financial markets. Also, she analyses the advantages provided to these institutions and their investors in Turkish tax legislation. Giray argues that the digital tax paradigm would inevitably necessitate a change in countries’ tax systems. The digital tax administration can create an opportunity to raise tax-income without raising the tax burden. This study investigates the impacts of the digitalization of tax administration on the complexity of the tax system with the indicators of some OECD countries. Varol İyidoğan, Balıkçıoğlu and Yılmaz examine the effect of aging, chronic diseases, health care expenditures and social spending on pharmaceutical spending for 22 OECD countries by employing General Method of Moments (GMM) procedure of Arellano and Bond (1991) which utilizes the difference of dependent variable to eliminate the individual fixed effects. In this paper, they conclude that the rise in the elderly population leads to an increase in pharmaceutical spending, which is consistent with our expectations. Saygılıoğlu investigates the concept of tax expenditure and its meaning in theory. It is used as a concept that reduces the tax burden of taxpayers for various purposes and expresses regulations such as exemptions and exemptions in public. The study describes the theoretical framework and reasons for assets of tax expenditures, and discussing its size and results in Turkey to attract the ­attention of business and politics. Yüksel studies the relationship between economic growth and public spending as a percent of GDP (government size). One of the essential explanations of these debates is the Armey curve. The parabolic structure of the Armey curve is critical for estimating the optimal government size. This study aims to test the Armey curve using the ARDL bounds testing approach of time-series techniques between the years 1981–2018 in the Turkish economy. Mercan and Özpençe investigate the relationship between economic growth and unemployment via Okun’s Law. In this study, the relationship between economic growth and unemployment for the Turkish economy is calculated. In this

Preface

7

context, the growth policies determined by governments will contribute to minimizing this problem by encouraging employment. Özel Özer, Özer and Akın evaluate the subsidies granted to the private educational institutions within the framework of the Turkish tax system. This study elucidates and analyses the arrangements and recent developments concerning grants of space and location for investments and exceptions regarding the insurance and tax exceptions and exemption within a general framework in Turkey for educational institutions. Biyan and Yılmaz discuss the issues of how artificial intelligence can be taxed in accordance with the discussions going on about the same. The main point derived implies that it does not seem plausible that artificial intelligence could become a taxpayer as per the applicable legal system in force. We hope that the current volume would be very useful for both academics and policymakers not only in Turkey but also in many developing and developed countries alike. Adnan Gerçek Metin Taş

Contents List of Contributors ..................................................................................................  11 Hamza Erdoğdu and Recep Yorulmaz Comparison of Tax Revenue Forecasting Models for Turkey ............................  13 Burçin Bozdoğanoğlu Examination of Tax Administration by Digitalization: Taxation of Sharing Economy; Country Examples and Evaluation of Turkey .....................  25 Ayşe Yiğit Şakar Tax Incentives Provided to Green Bonds in Financing of Energy Efficiency and Its Importance for Turkey ..............................................................  37 Güneş Çetin Gerger, Feride Bakar Türegün, and Adnan Gerçek The Importance of Tax Literacy in Tax Compliance ...........................................  57 Filiz Keskin The Concept of Collective Investment Institution and Specific Tax Advantages Provided for These Institutions and Their Investors in Turkey ....  77 Filiz Giray The Impacts of Digitalization of Tax Administration on the Complexity of Tax System: OECD Countries Example ............................................................  95 Pelin Varol İyidoğan, Eda Balıkçıoğlu, and H. Hakan Yılmaz Empirical Findings on Macro Determinants of Pharmaceutical Spending in Selected OECD Countries ...............................................................  113 Nevzat Saygılıoğlu Example of Internal Tax Bleeding: “Tax Expenditures” ....................................  121 Cihan Yüksel The Size of the Public Sector and the Armey Curve: The Case of Turkey ......  137 Nedim Mercan and Özay Özpençe Okun’s Law: Turkey Case ......................................................................................  155

10

Contents

Aslıhan Özel Özer, Buğra Özer, and Sercan Akın An Evaluation of Subsidies Granted to the Private Educational Institutions within the Framework of Turkish Tax System ..............................  169 Özgür Biyan and Güneş Yılmaz Artificial Intelligence: If It’s Taxed, But How? ....................................................  183 List of Figures ..........................................................................................................  205 List of Tables ...........................................................................................................  207

List of Contributors Sercan Akın Manager for Mavişehir Science Private Educational Courses, Turkey, [email protected]

Adnan Gerçek Prof. PhD., Bursa Uludağ University, Department of Public Finance, Turkey, [email protected]

Eda Balıkçıoğlu Assoc. Prof., PhD., Kırıkkale University, Turkey, [email protected]

Filiz Giray Prof. PhD., Bursa Uludağ University, Department of Public Finance, Turkey, [email protected]

Özgür Biyan Assoc. Prof. PhD., Bandırma Onyedi Eylul University, Faculty of Economics and Administrative Sciences, Department of Public Finance, Turkey, ozgurbiyan@ hotmail.com Burçin Bozdoğanoğlu Assoc.Prof. PhD., Bandırma Onyedi Eylul University, Faculty of Economics and Administrative Sciences, Department of Public Finance, Turkey, [email protected] Hamza Erdoğdu Assist. Prof. PhD., Harran University, Faculty of Economics and Admini­strative Sciences, Department of Econometrics, Turkey, [email protected] Güneş Çetin Gerger Assoc. Prof. PhD., Manisa Celal Bayar University, Turkey, gunes.cetin@ hotmail.com, (corresponding author)

Pelin Varol İyidoğan Assoc. Prof., PhD., Hacettepe University, Turkey, pelinv@hacettepe. edu.tr (corresponding author) Filiz Keskin Prof. PhD, Istanbul Arel University, Faculty of Economics and Administrative Sciences, Department of Political Science and Public Administration, Turkey, [email protected], [email protected] Nedim Mercan PhD. Student, Pamukkale University, Turkey, [email protected] Aslıhan Özel Özer Assist. Prof., PhD., Manisa Celal Bayar University, Ahmetli Vocational College, Tax and Accounting Applications Program, Turkey, [email protected]

12

List of Contributors

Buğra Özer Assoc. Prof., PhD., Manisa Celal Bayar University, Faculty of Economics and Administrative Sciences, Department of Political Science and International Relations, Turkey, [email protected] Özay Özpençe Assoc. Prof. PhD., Pamukkale University, Turkey, oozpence@pau. edu.tr Ayşe Yiğit Şakar Prof. PhD., Istanbul Arel University, Faculty of Economics and Administrative Sciences, Department of Business Administration, Turkey, [email protected], ayseyigitsakar@ gmail.com Nevzat Saygılıoğlu Prof. PhD., Atılım University, Faculty of Business Administration, Turkey, [email protected] Feride Bakar Türegün Assist. Prof. PhD., Bursa Uludağ University, Department of Public Finance, Turkey, feridebakar@uludag. edu.tr

H. Hakan Yılmaz Prof., PhD., Ankara University, Turkey, [email protected]. edu.tr Recep Yorulmaz Assist. Prof. PhD., Ankara Yıldırım Beyazıt University, Faculty of Political Sciences, Department of Public Finance, Turkey, ryorulmaz@ybu. edu.tr Cihan Yüksel Assist. Prof. PhD., Mersin University, Department of Public Finance, Turkey, [email protected] Güneş Yılmaz Assoc. Prof. PhD., Alanya Alaaaddin Keykubat University, Faculty of Business Administration, Department of International Trade, Turkey, gunes. [email protected]

Hamza Erdoğdu and Recep Yorulmaz

Comparison of Tax Revenue Forecasting Models for Turkey Abstract The objective of this study is to compare the performance of three forecasting tax revenue models for Turkey throughout 2006:01 to 2018:12. Three different time series forecasting techniques such as Random Walk, SARIMA (Seasonal Autoregressive Integrated Moving Average), and BATS (Exponential Smoothing State Space Model with Box-Cox Transformation, ARMA Errors, Trend and Seasonal Components) are used in the study. At the beginning of the analysis, the data set was apportioned into two parts: training and testing. The training period is from 2006:01 to 2014:12, and the testing part is from 2015:01 to 2018:12. Based on different evaluation criteria, forecast points of 36 months are obtained for each forecasting model. We find that using the BATS model, rather than classical (SARIMA) in forecasting series of monthly tax revenues of Turkey, provides more accurate forecasts. The empirical findings of this study help the experts in the preparation process of the government’s budgets. Keywords: Forecasting, Tax Revenue, BATS, SARIMA, Turkey JEL Codes: C1, C5, H20

1 Introduction Tax revenues are considered amongst the fundamental sources of government budget planning. Governments collect taxes not only to finance their expenses but also aiming of stabilization, distribution, and allocation in the economy. They use taxes to stabilize the employment levels, the balance of payments, and/ or prizes. They might try to intervene with the income and wealth distribution by playing with the tax structure. Further, they might want to use taxes to the allocation of resources in the economy by using their allocative effects on certain goods (Brown and Jackson, 1986, p. 297). There are three fundamental classifications in the Turkish tax system. These are income taxes, taxes on expenditure, and taxes on wealth. The relative importance of these taxes in the Turkish tax system is presented in Tab. 1. Income taxes are classified as individual income and corporate income taxes. Income taxes yield about 30% of total revenues in the Turkish tax system. Taxes on expenditures, on the other hand, contain approximately 68% of total revenues in the Turkish tax system. Taxing expenditures are considered the standard and easy way to collect taxes for governments. Hence, that significant

14

Erdoğdu and Yorulmaz

Tab. 1:  Percentage Distribution of Tax Revenues in Turkey

amount of taxes is comprised of expenditures in Turkey. Finally, taxes on wealth only yield approximately 2% of total revenues. Tax analysis and forecasting of tax revenues for governments are crucial to ensure stability in tax and expenditure policies (Jenkins et al., 2000). Budgetary uncertainties directed governments to rely heavily on economic analysis in recent decades. Because of the extent of these fiscal problems forecasting tax revenues is essential for governments to manage their budget planning process. Recent fiscal problems of governments created reliability ­issues on economic and revenue forecasting. Hence, there are plenty of methods that are used to forecast tax revenues by policymakers (Fullerton, 1989). Transparency and accuracy are the critical components while determining the method for forecasting. Potential manipulation of forecasts might create government problems. Furthermore, inaccurate forecasts might hinder the abilities of policymakers to make accurate budget planning and harm levels of productivity in the economy (Kyobe and Danninger, 2005; Cirincione et al., 1999). It is considered that countries with high-income levels and relatively small central government tend to have high formality, accuracy, and transparency forecasts (Kyobe and Danninger, 2005). Government revenue forecasting studies for Turkey are rare in the literature; hence, this study aims to fill this gap. The rest of the study is organized as follows. Section II outlines the significant studies that make forecasting analysis

Comparison of Tax Revenue Forecasting Models

15

in the literature. Furthermore, Section III describes the methodologies of the forecasting techniques applied and the data that are used in the study. Section IV provides the outcomes of selected forecasting methods in the study. Finally, Section V contains the conclusion and discussions.

2 Literature Majority of forecasting studies focused on the private sector in the literature so far, hence the studies focused on government revenue are relatively less than private-sector studies. For instance, Gajewar and Bansal (2016) conducted a forecasting analysis for the private sector using machine-learning algorithms. Accurately, they performed ARIMA, ETS (Exponential Smoothing), STL (Seasonal and Trend Decomposition using Loess), and Random forest machinelearning algorithms to obtain revenue forecast for Microsoft. They suggested that using machine-learning algorithms methods would increase the accuracy of quarterly revenue forecasting. Many researchers also focused on state and/or municipal revenue forecasting analysis so far. Fullerton (1989) analyzed sales tax revenues using a composite forecasting model for Idaho. Using a time series model and econometric models, he examined the capability of the composite forecasting model. He found that the composite forecast model is more effective than baseline forecasts. The combined model was also found more accurate than previous forecast attempts for Idaho. Hambor et al. (1974) used an econometric forecasting method using a simple revenue structure for Hawaii. They forecasted state revenues, including; excise, personal income, corporate income, and other state tax revenues, for a single fiscal year of Hawaii. Furthermore, Kyobe and Danninger (2005) ­analyzed the revenue forecasting practices in 34 low-income countries, focusing primarily on institutional prospects. They claimed that there are three critical factors on forecasting practices, such as “formality, organizational simplicity, and ­transparency”. They empirically found that countries’ levels of corruption are associated with f­ormality and transparency of forecasting. Accordingly, they found that high levels of corruption are related to less formal and transparent forecasts. Cirincione et  al. (1999) examined the impact of using time series models, the length, and the frequency of the data on non-tax general fund revenue forecasting for the municipalities of Connecticut. They found that exponential smoothing models are most effective on bimonthly data in which they claim local governments should rely on rather than monthly or quarterly data.

16

Erdoğdu and Yorulmaz

As we pointed above, there are plenty of methods that were used to forecast the private sector or government/state/municipal revenues in the literature. It is also essential to analyze the methods used in these studies. In the case of the Box-Jenkins Auto-Regressive Integrated Moving Average Model (ARIMA), researchers found different results in the effectiveness of the ARIMA model. For instance, Makridakis and Hibon (1995) claimed that the ARIMA model performs relatively weak than other models. In doing so, Makridakis et al. (1979) found the reason for this poor performance of the ARIMA model as the usage of differencing in order to find stationary in the mean of the series. Similarly, in a series of studies that focused on local government revenue forecasting for the municipalities of Florida, researchers found similar results. They claimed that the Box-Jenkins ARIMA model performs poorly than other methods such as time series models, which produce lower forecast errors. Furthermore, they found that trend fitting by regression generated more forecast errors than its counterpart methods (Frank and Gianakis, 1990; Gianakis and Frank, 1993). It is important to point out that the studies that found poor performance for the ARIMA method mainly focused on municipal government revenue forecasting. Differently, Downs and Rock (1983) found evidence that the multivariate Auto Regressive Moving Average (ARMA) method is more effective than univariate techniques using the ARMA model for municipal government revenue forecasting. While most of the forecasting studies examine the relative performance of various methods so far, only a few researchers tested the impact of data quality on the performance of forecasting methods. Gianakis and Frank (1993), which is one of these studies, claimed that the length of the data does not have any impact on the accuracy of forecasting techniques. However, some scholars suggested that at least fifty observations are necessary to implement the Box-Jenkins ARIMA method. On the other hand, scholars have kept using this method with fewer numbers so far (Lorek et al., 1976). Lorek and McKeown (1978) analyzed the association between observation numbers and the performance of the Box-Jenkins method on quarterly market income data. They found that the forecast error is not significantly different in models based on fifty observations and models based on fewer observations. They suggested that if the number of observations of the BoxJenkins method decreases, forecast error increases. However, the performance of the model does not occur until at least twenty-four observations are made. Similarly, Lusk and Neves (1984) found a result consistent with the previous cases. They suggested that the performance of the Box-Jenkins

17

Comparison of Tax Revenue Forecasting Models

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis ADF Jarque-Bera Probability Observations

Tax Revenues 26127815 23373839 67930091 9591739 13173330 0.838831 3.031952 5.52* 18.30122 0.000106 156

Turkey Tax Revenues (thousand tl), 2006M01-2018M12 70,000,000 60,000,000 50,000,000 40,000,000 30,000,000 20,000,000 10,000,000 0

06

07

08

09

10

11

12

13

14

15

16

17

18

Fig. 1:  Turkey Tax Revenues, 2006M01 – 2018M12. Source: General Directorate of Budget and Fiscal Control and General Directorate of Budget and Fiscal Control. * the t – statistic value of the Augmented Dickey-Fuller test, indicating nonstationarity of the series at level 0.05.

model does not associate with the length of data or the frequency in their private-sector study.

3 Data and Methodology 3.1 Data In the analysis, the series of monthly tax revenues in the central government budget realizations, from January 2006 to December 2018 is used. The data is obtained from the web site of the General Directorate of Budget and Fiscal Control (BÜMKO). The series is plotted and shown in Fig. 1, also provides some descriptive statistics about the data to help better understanding the structure of the series. From the plot, it is clear that the series has a trend and seasonal component.

3.1 Methodology In this section, we provide the fundamentals of the forecasting methods used in the study, such as Random Walk, SARIMA, and BATS.

3.1.1  Random Walk-RW The random walk model is widely used in econometric forecasting studies as a benchmark.

18

Erdoğdu and Yorulmaz

A time series is said to follow a random walk process if the first differences are random. For a time seriesYt , a random walk  can be written as differences changes from one period to the next, Yt = Yt−1 + εt



where Yt−1 is the value in time period t − 1 and εt is a discrete white noise in time period t .

3.1.2  SARIMA Introduced by Box and Jenkins (1970), ARIMA (Auto-Regressive Integrated Moving Average) models are a broad category of univariate models. In forecasting a time series, these models bring together three components:  the auto-regressive (AR), the moving average (MA) part, and the integrated (I) part. The AR part indicates that linear models can describe individual values in a variable of interest based on its own lagged values. The MA part assumes that regression error is a long combination error terms. The integrated part (I) shows the degree of difference. The ARMA (AutoRegressive, MovingAverage) model is defined as follows: Yt = φ1 Yt−1 + φ2 Yt−2 + ... + φp Yt−p + αt − ψ1 αt−1 − ψ2 αt−2 − ... − ψq αt−q (1) where the Yt  s is the original time series, they φ s are the unknown autoregressive parameters, they ψ  s are the unknown moving average parameters, and they α s are the white noise error terms. A modification for nonstationary series, known as ARIMA( p, d, q) is;

d

φp (B)(1 − B) Yt = ψq (B)αt (2)

where B is the backshift operator, thus BYt = Yt−1and B2 Yt = Yt−2and the  d parameter indicates the order of differencing? For seasonal series, a more general form of the above equation, and known as the multiplicative seasonal ARIMA  –SARIMA ( p, d, q)(P, D, Q)s process is given by;

d

D

φp (B)ΦP (B)(1 − B) (1 − Bs ) Yt = ψq (B)ΘQ (Bs )αt (3)

(1 − Bs )Yt = Yt − Yt−s s  Where and is the number of seasons per year, (1 − Bs ) d Dand are the orders of differencing. Also φp ΦP ψq, and ΘQ are the polynomial functions of orders p, P, q and Q, respectively.

Comparison of Tax Revenue Forecasting Models

19

3.1.3  BATS To model series having only one seasonal pattern, Winters (1960) introduces the standard Holt-Winters method. However, Taylor(2003) extends the standard method to a double seasonal Holt-Winters method, following mainly De Livera et al. (2011);

(1)

Yt = Lt−1 + Bt−1 + St

(2)

+ St

+ Dt , (1A)



Lt = Lt−1 + Bt−1 + αDt , (1B)



Bt = Bt−1 + βDt , (1C) (1)

= St−k1 + λ1 Dt , (1D)

(2)

= St−k2 + λ2 Dt , (1E)



St



St

(1)

(1)

where Lt represents the level component of the series Yt at time t , Bt represents the trend component of the series Yt  at time t , (i) St  represents the ith seasonal component at time t , k1 and k2 are the periods of the seasonal cycles, Dt  is the disturbance (or prediction error), α, β, λ1 and λ2 are the smoothing parameters, Proposed by De Livera et al. (2011), the BATS models (Exponential Smoothing State Space Model with Box-Cox Transformation, ARMA Errors, Trend and Seasonal Components) are designed in the exponential smoothing framework. The models are constructed to handle more than one seasonality as well as complex seasonalities, for example, non-nested, non-integer, and substantial period seasonality. De Livera et al. (2011) extends the second seasonal Holt-Winters method by adding a Box-Cox transformation, ARMA errors, and T seasonal patterns: ® (w) Yt

(w)



Yt

=

Ytw −1 w ,w

 0, = log Yt , w = 0,

= Lt−1 + φBt−1 +

T  i=1

(2A)

(i)

St−ki + Dt , (2B)

Lt = Lt−1 + φBt−1 + αDt , (2C)

20

Erdoğdu and Yorulmaz

Bt = (1 − φ)B + φBt−1 + βDt , (2D) (i)



(i)

St = St−ki + λi Dt ,

Dt =

p 

ψi Dt−1 +

i=1

q  i=1

(2E)

ζi ηt−i + ηt , (2F)

where; Lt represents the local level in period t , Bt  represents the short-run trend in period t , B represents the long-run trend in period t , φ represents the damping parameter, (i) St represents the iseasonal component at time t , k1 , k2 , ..., kT are the seasonal periods, Dt is an ARMA( p, q ) process, ηt is a Gaussian process with zero mean and constant variance σ 2 , α, β, λ1and λ2 are the smoothing parameters,

4 Analysis and Empirical Results At the beginning of the analysis, we split the data into a training and testing set. The training set covers the period from 2006:01 to 2014:12, and the testing part covers the period from 2015:01 to 2018:12. The training data set is used only to estimate unknown model parameters. Once the model coefficients are estimated, forecasts for each model are made for the testing part. To evaluate the forecast accuracy of each model, the testing data is used. The results of the best ARIMA(0,1,2)(0,1,1)12 model for the tax revenues series is given in the following Tab. 2. The automatic ARIMA selection option was used in the forecast package in R, and the details can be found in Hyndman and Khandahar (2008). The results of the BATS(0.377, {0,0}, 1, {12}) model for the series is provided in Tab. 3. The forecast package in R is used to get the results. Tab. 2:  The Results of the ARIMA(0,1,2)(0,1,1)12 Model MA MA SMA

Lagged Length 1 2

Coefficients -1.0682    0.3932 -0.4693

Standard Error 0.0963 0.0969 0.0908

Comparison of Tax Revenue Forecasting Models

21

Tab. 3:  The Results of the BATS (0.377, {0,0}, 1, {12}) Model Parameters Lambda Alpha Beta Damping Gamma Values

Coefficients 0.376905 0.2432854 0.01566971 1 -0.1082185

After fitting three-time series models:  random walk, SARIMA, and BATS, forecasts for the testing period, 2015:01 to 2018:12, are obtained in Tab. 4. Finally, the accuracy of each model is measured on the testing set. Tab. 5 provides statistical measures of accuracy of each method based on ­various forecast evaluation criteria:  ME, RMSE, MAE, MPE, MAPE, MASE, and Theil’s U. Among the three generated models, the accuracies of the models are tested based on seven forecast evaluation criteria. From the Tab. 5 results, BATS is preferred as the best forecasting model for the tax revenues series of Turkey, since it provides lesser values of seven evaluation criteria:  ME, RMSE, MAE, MPE, MAPE, MASE and Theil’s U.

5 Conclusion and Discussion This study aims to evaluate the performance of three forecasting tax revenue models for Turkey throughout 2006:01 to 2018:12. Three different time series forecasting techniques such as Random Walk, SARIMA (Seasonal Autoregressive Integrated Moving Average), and BATS (Exponential Smoothing State Space Model with Box-Cox Transformation, ARMA Errors, Trend and Seasonal Components) are used in the study. At the beginning of the analysis, the data set was apportioned into two parts: training and testing. The training period is from 2006:01 to 2014:12, and the testing part is from 2015:01 to 2018:12. Based on different evaluation criteria, forecast points of 36 months are obtained for each forecasting model. The BATS model outperforms the benchmark RW and SARIMA models based on all evaluation criteria. We find that using the BATS model, rather than seasonal ARIMA Yılmaz (2019) in forecasting series of monthly tax revenues of Turkey, provide more accurate forecasts. The empirical findings of this study help the experts in the preparation process of the government’s budgets. For further forecasting of other tax types such as Corporate Income Tax, Value

22

Erdoğdu and Yorulmaz

Tab. 4:  Point Forecasts of the Methods for the Testing Data (2016M01–2018M12) Forecast Horizon

ACTUAL

Jan 2016 Feb 2016 Mar 2016 Apr 2016 May 2016 Jun 2016 Jul 2016 Aug 2016 Sep 2016 Oct 2016 Nov 2016 Dec 2016 Jan 2017 Feb 2017 Mar 2017 Apr 2017 May 2017 Jun 2017 Jul 2017 Aug 2017 Sep 2017 Oct 2017 Nov 2017 Dec 2017 Jan 2018 Feb 2018 Mar 2018 Apr 2018 May 2018 Jun 2018 Jul 2018 Aug 2018 Sep 2018 Oct 2018 Nov 2018 Dec 2018

39685212 38361380 30496694 32446011 42368600 33195345 36111701 45425215 30883849 36060795 54060129 39906810 48420673 39994384 33201256 37082457 50949456 36422643 46062984 51377479 41837993 45559415 58372034 46766884 51995609 52558220 41249512 45049034 61218542 42749559 54360053 60934207 49235735 48504135 67930091 45525901

Forecasting Methods Random Walk SARIMA 34729587 39551109 34729587 38414335 34729587 31463837 34729587 35440194 34729587 40607697 34729587 34699758 34729587 39678120 34729587 41889999 34729587 34262776 34729587 38123593 34729587 43576351 34729587 38712079 34729587 43405126 34729587 42309290 34729587 35358791 34729587 39335148 34729587 44502651 34729587 38594712 34729587 43573075 34729587 45784953 34729587 38157730 34729587 42018548 34729587 47471306 34729587 42607034 34729587 47300081 34729587 46204244 34729587 39253745 34729587 43230102 34729587 48397605 34729587 42489666 34729587 47468029 34729587 49679907 34729587 42052684 34729587 45913502 34729587 51366260 34729587 46501988

BATS 38989627 37396828 30878845 34802877 41212156 35611013 38694089 43894982 34663442 38056660 44574433 38555889 43683557 41972917 34951174 39182860 46067487 40052674 43366295 48940777 39032728 42681794 49667752 43217915 48714507 46882727 39342516 43890960 51264277 44824230 48374916 54333154 43729821 47642028 55108914 48216072

23

Comparison of Tax Revenue Forecasting Models Tab. 5:  Measures Accuracy of the Methods for Testing Set (2016M01–2018M12) Methods Random Walk SARIMA BATS

ME RMSE MAE MPE 10159746.6 13580773 10905568 19.50

MAPE 21.87

MASE 3.93

Theil’s U 1.27

2961666.3 2042864.3

8.73 7.43

1.56 1.29

0.55 0.44

5761357 4578144

4317229 4.84 3583768 3.13

The measures calculated are ME: Mean Error, RMSE: Root Mean Squared Error, MAE: Mean Absolute Error, MPE: Mean Percentage Error, MAPE: Mean Absolute Percentage Error, MASE: Mean Absolute Scaled Error, Theil’s U: Theil Inequality Coefficient.

Added Tax, and Total Tax, the BATS model may provide better performance. Also, the empirical results of the current study will be used to develop combined forecasting models.

References Box, G. E. P., & Jenkins, G. (1970). Time Series Analysis, Forecasting, and Control. Holden-Day, San Francisco, CA, USA. Brown, C. V., & Jackson, P. M. (1986). Public Sector Economics. Oxford, Blackwell, Basil. Cirincione, C., Gurrieri, G. A., & Sande, B. (1999). Municipal Government Revenue Forecasting: Issues of Method and Data. Public Budgeting and Finance. 19, 26–46. De Livera, A. M., Hyndman, R. J., & Snyder, R. D. (2011). Forecasting Time Series with Complex Seasonal Patterns Using Exponential Smoothing. Journal of the American Statistical Association. 106(496), 1513–1527. Downs, G. W., & Rocke, D. M. (1983). Municipal Budget Forecasting with Multivariate ARMA Models. Journal of Forecasting. 2, 377–387. Frank, H. A., & Gianakis, G. A. (1990). Raising the Bridge Using Time Series Forecasting Models. Public Productivity & Management Review. 14, 171–188. Fullerton, T. M. (1989). A Composite Approach to Forecasting State Government Revenues: Case Study of the Idaho Sales Tax. International Journal of Forecasting. 5, 373–380. Gajawar, A., & Bansal, G. (2016). Revenue Forecasting for Enterprise Products. International Symposium on Forecasting, https://arxiv.org/ftp/arxiv/ papers/1701/1701.06624.pdf, (22.08.2019). Gianakis, G. A., & Frank, H. A. (1993). Implementing Time Series Forecasting Models: Considerations for Local Governments. State & Local Government Review. 25, 130–144.

24

Erdoğdu and Yorulmaz

Hambor, J. C., Norman, M. R., & Russell, R. R. (1974). A Tax Revenue Forecasting Model for the State of Hawaii. Public Finance Quarterly. 2, 432–450. Hyndman, R. J., & Khandakar, Y. (2008). Automatic Time Series Forecasting: The Forecast Package for R. Journal of Statistical Software. 27, 1–22. Jenkins, G., Kuo, C. Y., & Shukla, G. (2000). Tax Analysis and Revenue Forecasting. Development Discussion Papers 2000–05, JDI Executive Programs, https://cri-world.com/publications/qed_dp_169.pdf, (12.08.2019). Kyobe, A., & Danninger, S. (2005). Revenue Forecasting: How Is It Done? Results from a Survey of Low-Income Countries. IMF Working Papers. 05, 1. Lorek, K. S., Mcdonald, C. L., & Patz, D. H. (1976). A Comparative Examination of Management Forecasts and Box-Jenkins Forecasts of Earnings. The Accounting Review. 51, 321–330. Lorek, K. S., & Mckeown, J. C. (1978). The Effect on Predictive Ability of Reducing the Number of Observations on a Time-Series Analysis of Quarterly Earnings Data. Journal of Accounting Research. 16, 204–214. Lusk, E. J., & Neves, J. S. (1984). A Comparative ARIMA Analysis of the 111 Series of the Makridakis Competition. Journal of Forecasting. 3, 329–332. Makridakis, S., & Hibon, M. (1997). ARMA Models and the Box-Jenkins Methodology. Journal of Forecasting. 16, 147–163. Makridakis, S., Hibon, M., & Moser, C. (1979). Accuracy of Forecasting: An Empirical Investigation. Journal of the Royal Statistical Society. Series A (General). 142, 97–145. Saracoglu, F., Engin, I., & Ejder, H. L. (2014). Maliye Ders Notlari. Gazi Kitabevi, Ankara. Taylor, J. W. (2003). Short-Term Electricity Demand Forecasting Using Double Seasonal Exponential Smoothing. Journal of the Operational Research Society. 54, 799–805. Winters, P. R. (1960). Forecasting Sales by Exponentially Weighted Moving Averages. Management Science. 6, 324–342. Yilmaz, E. (2019). Vergi Gelirlerinin Tahminlenmesine Yönelik Ekonometrik Model. Vergi Dünyası. 38 (449), 38–47.

Burçin Bozdoğanoğlu

Examination of Tax Administration by Digitalization: Taxation of Sharing Economy; Country Examples and Evaluation of Turkey Abstract: The sharing economy or collaborative economy is a new economic model that uses online platforms to share individuals’ assets, resources, time, and capabilities on a scale that was not previously possible. There is no clear consensus among international economic institutions on its definition. Service providers, users, and platforms that bring them together are the sides of this economy. Airbnb, Uber, Taskrabbit are the platforms created by this new economic model.The fact that the sharing economy is a functioning economy through online platforms makes it difficult to evaluate within the framework of tax and legal regulations, such as the traditional economy. It should adopt a ‘one-size-fitsall’ approach to determine whether the revenue from the sharing economy is the primary or an auxiliary source of income, to clarify the status of the parties involved in the sharing economy transactions, to clarify tax obligations and to ensure efficiency in taxation.In this study, taxes, which are the subject of sharing economy, which is a new economic model, and cooperation with platforms and determination of taxpayer awareness, will be included. Models and practices for increasing awareness of the taxpayers implemented in collaboration with the platforms of the sharing economy taxation in the EU countries will be examined further assessment will be made regarding the size of the economy and taxation in Turkey. Keywords: Sharing Economy, Taxation, Digital platforms, Tax compliance JEL Codes: H2, H24, K34

1 Introduction The sharing economy is technology applications that allow individuals to share goods and services through information and communication technologies, internet platforms, and are also known to the user as “peer to peer economy” or “collaborative economy”. The sharing economy, which has begun to be seen in the last few years, is an economic model that enables the emergence of new platforms that allow the production and consumption of goods and services. Uber, Airbnb, TaskRabbit platforms within the sharing economy provide the possibility of renting a variety of services on a daily or hourly basis, including driving, renting, or using personal skills from other users through a personal computer or mobile application for their payment to consumers.

26

Burçin Bozdoğanoğlu

At this point, these economic models, which are based on individuals making agreements on the internet to share the assets or skills they possess, significantly affect the enterprises operating in traditional commercial structures such as taxis, limousine services, and hotels. For this reason, the concept of sharing economy should be made in terms of new legal arrangements, whether the flexibility or changes in existing regulations should be ensured, the new rules or modifications will be made, how these will be done and propose the agenda for how can be the proper relationship between technological arrangements and technological reforms. First of all, the concept of sharing economy will be tried to be defined primarily in the context of international literature. In the second part of the study, the parties operating in the sharing economy will be explained, and the cooperation platforms, which are the most important actors in terms of taxation, will be included on a sectoral level. In the third part of the study, the size of the sharing economy, the taxes it is the subject, and the taxpayers of these taxes will be mentioned. In the fourth chapter, the importance of cooperation with platforms in the taxation of the sharing economy and the related country practices are included. In the last part of the study, data and evaluations concerning the taxation of the economic model of the size of the sharing economy in Turkey are evaluated.

2 The Concept of Sharing Economy The phenomenon of sharing economy, especially in the media and in the academic literature, which has shown significant growth since the crisis of 2008, attracts much attention as an umbrella concept, whose borders are still blurred. Therefore, it is not easy to provide an integrated and formal definition. In the literature, it is seen that the concept of ‘sharing economy’ is interpreted under different labels: like “collaborative consumption”,”cooperation economy”, “economy on demand “, “peer-to-peer economy”, “zero marginal cost economy” and “crowd-based capitalism”, terms are considered in connection with the concept of sharing economy (Selloni, 2017: 15). It should be noted that there is no consensus on the definition of the sharing economy. This uncertainty about meaning is one of the reasons for the problems of sharing economy. It is seen that the term “collaborative economy” and “sharing economy” are used interchangeably primarily in European Union based sources. In its decisions of September 2015 and October 2015, the European Parliament announced the sharing economy as follows: “the sharing economy, or collaborative consumption, is a new socio-economic model that has taken off thanks to

Examination of Tax Administration by Digitalization

27

the technological revolution, with the internet connecting people through online platforms on which transactions involving goods and services can be conducted securely and transparently” (EPRS, 2016b:3). The term “collaborative economy” refers to “business models where activities are facilitated by collaborative platforms that create an open marketplace for the temporary usage of goods or services often provided by private individuals” (EC, 2016:5).

3 Parties of Sharing Economy According to the European Commission, the actors involved in the sharing economy can be classified into three categories. (EPRS, 2016a:6): – Service providers sharing assets, resources, time, and skills:  these may be private individuals serving at regular periods (peers) or professional service providers – Users of services offered by service providers, – collaborative economy platforms that connect providers with users and facilitate transactions between them, also ensuring the quality of these transactions, e.g., through after-sales services (handling complaints), insurance services, etc.

4 Evaluation of the Value Created by the Sharing Economy and Its Taxation 4.1 Services Offered in the Sharing Economy and Estimated Economic Value The rapid change in technology, the economic crises experienced, the change in consumer models in the business world, changing the consumption models, and acting with environmental concerns, led many people to search for different sources of finance and shopping and changing the way they made holidays. From private renters, office space, renters of vehicles and other goods, to trade services, there are plenty of new platforms to suit the demand and supply of goods and services. The five sectors with the most activity in the sharing economy and the platforms in these sectors (with high economic value) are listed as follows. (EPRS, 2016b:4): – transportation (Uber, BlaBlaCar), – retail (Etsy, eBay), – accommodation (Airbnb, ShareDesk),

28

Burçin Bozdoğanoğlu

– service and labour (TaskRabbit, Shareyourmeal, Elance), – finance (Kickstarter, Kiva, Indiegogo), The annual growth rate of the sharing economy exceeds 25% (EPRS, 2016b:4). In 2016, the size of the sharing economy was 26.5 billion Euros. While service providers earn € 22.7 billion, it is stated that platforms generate revenue of € 3.8 billion due to intermediation. The largest share of revenues was in the financial sector with 9.6 billion euros, followed by the accommodation sector with 7.3 billion euros, the online service sector with 5.6 billion euros, and the transportation sector with 4 billion euros (EC, 2018:9).

4.2 Determination of Taxes and Taxpayers on the Sharing Economy Since several transactions are conducted in tripartite relations through a platform, each taxpayer and each tax case must be considered separately. This situation will result in an obligation related to income tax and VAT if the revenue obtained exceeds a certain threshold. The challenge is to collect data on individuals using the platforms and raise their awareness of relevant tax obligations. However, since transactions are usually paid electronically, it may be possible to use the data to contribute to better identification of taxpayers and revenue tracking, especially if platforms report data they hold in transactions (EPRS, 2018:17). Platforms can act as intermediaries and undertake legal responsibility for the collection of income in this activity. They can also contribute to this process while offering the service offered through the platform. Therefore, as an employer, the said tax liability can be given to the platforms. This may lead to concerns similar to the concerns of B2C platforms in the sharing economy, their size, and tax implications caused by other multinational companies. (EPRS, 2018:18). As can be seen here, taxes on the sharing economy include personal income tax, corporate tax, VAT, and other taxes. In this context, personal income tax is to be considered as a tax that must be taken into consideration for the services provided by service providers. If the supplier conducts its activities within the framework of its capacity, the income earned by the individual supplier is included in the income tax base. (EPRS, 2018:19). When the supplier operates within a business, it no longer falls under the subject of income tax. Both the service providers and online platforms in the sharing economy are, in principle, subject to indirect taxes starting with VAT obligations. In the EU, the VAT framework is being reviewed to better understand the taxation of digital economic models, including the sharing economy. The EU VAT

Examination of Tax Administration by Digitalization

29

regime for online sales has recently been updated, and the one-stop-shop will be expanded until December 31, 2020 (EPRS, 2018:19).

5 The Role of Cooperation with Platforms in the Taxation of Sharing Economy and Country Applications 5.1 Role of Cooperation with Platforms The sharing economy is considered to be at risk of expanding the informal economy, as existing and new activities are not reported. International studies on the fight against the informal economy, especially the OECD, address this (OECD, 2017:48). The fact that all transactions on a platform where data is recorded, leaving digital traces allows the platforms to report these operations objectively. This provides the possibility to benefit from technology to improve tax management and frees sharing actors from monitoring a large number of small-scale P2P participants, especially small providers and self-employed. (EPRS, 2018:22). This also prevents a potential tax mismatch on behalf of small businesses and avoids the administrative burden of taxation (Aqib&Shah, 2017:26). Service provider’s tax compliance can be improved by raising awareness (EPRS, 2018:20). For this reason, the first step regarding tax obligations is to increase awareness (P2P) for individuals acting as service providers, especially on sharing economy platforms. P2P providers serving as failures are not naturally experienced in required tax record-keeping and filing obligations (Rahim et  al., 2017:453). If the platform collects tax and if record keeping is supported by information technologies reporting from the sharing-based economic platform, some of the administrative burdens will be alleviated. At this point, the most fundamental requirement is to determine the threshold at which the service provider determines the taxpayer. Collaboration with sharing economy platforms is of critical importance to validating the revenue reported by P2P vendors because platforms keep records of information about transactions that they facilitate or mediate. In other words, there is the possibility to use the large data generated by the platforms themselves to ‘pre-filled’ most of the tax declarations of taxpayers involved in the collaborative economy. This is particularly important in P2P transactions to be declared by service providers. A  system that allows tax administrations to provide direct access to information held by platforms, or that the platforms allow tax deductions from P2P users through withholding and then transfer them to the tax administration will reduce tax loss. Collaboration between sharing economy platforms and tax administration

30

Burçin Bozdoğanoğlu

can begin by sharing information about transactions carried out by service providers and may develop to automatically report relevant revenue information to tax authorities by the platform. Indeed, some country practices show that this is possible.

5.2 Country Applications In the EU, some countries have taken various measures to promote tax compliance in the sharing economy. Ireland created a “sharing economy tax center” in August 2016. The website of this center provides information and resources on the taxation of the sharing economy. Thus, individuals who earn income through platforms can fulfill their tax obligations (EPRS, 2018: 21). In France, the tax return form, known as ‘automatic reporting of revenues for online platforms’, and a number of information elements for each user responsible for taxation, are attached directly to the tax statement (total gross revenue generated by the user during the calendar year, online e-mail address, personal or professional status, or total gross income paid over the platform and the category in which gross revenue will be deducted for activities on the online platform). In this tax declaration, a copy of the information per user is sent to the relevant user online (OECD, 2018:98). A system established in Estonia, in September 2015 and since February 2016, allows drivers to register with a system where the transaction between service providers and the user is recorded by the platform. The platform then sends the information to the tax office about the income generated by the drivers involved in the ride-sharing system, which are automatically added to the tax returns based on the advanced online tax system (OECD, 2018:99). In the Netherlands, an agreement was signed between the Amsterdam city administration and Airbnb in 2014, requiring the platform to collect the city’s tourism tax on behalf of service providers (EPRS, 2016a: 161). Airbnb has been the pioneer of this model by transferring the tourist taxes that constitute 5% of the accommodation fee and transferring them to the state, in this model, where the homeowners also pay tourist tax (determined as %5 of the accommodation price) as well as the tax they will pay for the short-term lease. In Italy, revenues up to € 1000, which are included in the sharing economy, will be taxed at a rate of 10%, and income above € 1000 will be subject to the rate applicable to the professional income of the service provider. Platforms collect the taxes collected through deductions and transfer them to the state (OECD, 2018:97).

Examination of Tax Administration by Digitalization

31

6 The Value Created by the Sharing Economy in Turkey and Its Evaluation in Terms of Taxation 6.1 The Size of Sharing Economy in Turkey There is no research in the European Commission or the OECD level about the size of the sharing economy in Turkey. This new economic model, most current data about Turkey, contained in the report by PWC carried out the market research institute Faktenkontor GmbH between June and August 2017 and based on a survey which represents on over 4500 consumers from six countries including Turkey and Austria, Belgium, Germany, the Netherlands, Switzerland (PWC, 2017:5). The report explores the sharing economy in seven sub-headings such as; Media and Entertainment, Hotel and Hospitality, Automotive and Transportation, Retail and Consumer Goods, Services, Finance, and Industry (PWC, 2017:5). The sharing economy throughout the year among the countries surveyed, with expenditure amounting to 1,031 euros per person to use Turkey, realized the highest average spending (PWC, 2017:6). The estimated size of the market share in Turkey’s economy is 38.3 billion euros. The high market size can be explained by the presence of a well-developed sharing economy finance sector and in general high acceptance rate in Turkey. The size of this sector is approximately 11.2 billion Euros, followed by the Retail and Consumer Goods and Hotels and Accommodation sectors, with a turnover of 6.5 billion Euros and 6.3 billion Euros. Total revenues for Automotive and Transport are 4.6 billion Euros, 4.2 billion Euros for Services, 3.3 billion Euros for Machinery, and 2.2 billion Euros for Media and Entertainment. The estimated market size of the sharing economy in Turkey is 71.5 billion euros for the coming period (PWC, 2017:11).

6.2 The Taxes Which Are Sharing Economy Subject to in Turkey In Turkey, service providers and platforms as a player of the sharing economy actions are evaluated in the framework of income tax and value added tax and corporation tax. Income Tax Law Article 37 deals with the income generated from commercial activities without the definition of commercial activity and is considered to be commercial gain. Given that business activities are carried out within a commercial organization based on continuity, based on the combined use of labor and capital, and activities in the sharing economy are evaluated within this framework, it is clear that if the event includes a “continuity” component for the service providers, it will be considered as commercial gain.

32

Burçin Bozdoğanoğlu

At this point, the “continuity” element enables the activity to become profitable. However, there is no specific regulation as to which activity should be considered for how long. This situation makes it difficult to determine what gain component will be affected by the activity. If the activity does not include continuity, the profit in question will be considered as failure, and the exemption from tax will be exempted for 2019 ₺33.000. Concerning the income tax law, real estate capital income; Article 70 provides that the income obtained from the leasing of the goods and rights shall be considered as the real estate capital income. When the provisions of the income tax related to income from the immovable property are evaluated in the context of sharing economy, “for the benefit of the economic assets left for use by others”, the income tax law under Article 70 can be accepted as income from immovable property. For example, it is the subject of the income tax if an individual rent his house in a summer location at specific periods of the year or if a person is living in the metropolitan area rents for his own housing needs or shares a dwelling that is owned and owned by a tourist. If the activities are evaluated within the scope of commercial earnings, VAT will be born as delivery/service will be deemed realized. However, delivery and services shall not be subject to VAT in case of incidental commercial activity. The real estate capital income is subject to VAT, provided that these goods are included in an economic enterprise. Therefore, VAT will not be included in the rental transactions for natural persons. However, this situation is specific to real estate. In other words, VAT will be subject to the rental of goods other than real estate. The parties must be members of the relevant platform to access the share. Platforms receive commissions for this intermediation service. Although the legal nature of the activity is commercial, it is realized in an electronic environment, and users can use the platforms by using mobile applications. In this context, since the intermediation is a commercial activity, income/corporate tax liability will be established according to the platform is a real/legal person. In addition, intermediation services are subject to VAT within the scope of the commercial activity. On value-added tax against the state of the platform that mediates the sharing economy in Turkey, Law No. 7061, an arrangement was made within the framework of law VAT added to the stipulation of Article 9. The provision in question as follows; “In so far residence in Turkey, the workplace, the legal center, and business center value-added tax on real people who are not payers of value-added tax by not regarding the services offered electronically paid shall be declared by

Examination of Tax Administration by Digitalization

33

those who offer this service. Ministry of Finance is authorized to determine the scope and principles and procedures of the services provided by electronically”. Subsequently, 31.08.2018 date and 30318 numbered Official Gazette within 17 Serial No. VAT Application, determined that; the residence, the workplace, the legal center, and business center are not found in Turkey at a price to the real people who are not VAT payers electronically, It is stated that service providers will declare the VAT related to these services by establishing “Special VAT Liability for Electronic Service Providers “. Provisions added to Article 9 of the Act and regulations made by Communique Serial No. 17 along, taking place within the legal and business center of Turkey’s borders and encompasses organizations and institutions in the performance of services through electronic platforms. According to the communiqué mentioned above, service providers declare the Value Added Tax related to these transactions electronically with the VAT declaration no. 3. At this point, if the service provided by the said platforms is provided only on the internet, and the transactions are realized in a virtual environment without coming face to face with the beneficiaries as buyers or sellers on the internet, there is no obligation to use payment recording devices for the services provided and therefore each customer invoices must be issued within a maximum of seven days from the date on which the individual facility or transaction takes place (Özdemir, 2019). However, before this declaration, the form prepared by the tax administration shall be filled out electronically, and accordingly, a special VAT liability shall be provided to the Electronic Service Providers on behalf of the service provider in the Big Taxpayers Tax Office (Kara, 2018). Considering the regulations made in terms of VAT with both the law and the communiqué is understood that legislation applications may face problems in the face of the digital economy. Namely, these institutions shall issue a VAT Declaration No. 3 within the framework of the said regulations; in the event, they regulate commission invoices for the transactions they mediate. First of all, invoice arrangements will not be possible for service providers that do not have VAT liability. In addition, considering the fact that the platforms which are in the limited taxpayer category and which mediate the services have calculated the commission price on the basis of the transactions they mediate, the amount in the commission invoice for the intermediary will be at the discretion of the platform since there is no invoice in accordance with the legislation. Therefore, although the target capture system to digitalize said, this economic model cooperation arising from the taxable value through processing platform that can point to is of critical importance for Turkey because the provision and

34

Burçin Bozdoğanoğlu

the applicability of the VAT declaration and the provisions added in Article 9 of the VAT Law depend on this cooperation to a large extent. At this point, it is thought that it is possible to develop a system suitable for our country by taking the model of tax compliance and cooperation with the platforms applied in EU countries and to understand the tax base by using the knowledge of the platforms.

7 Conclusion The sharing economy, which is a relatively new phenomenon, has recently become widespread in some areas of activity, such as ride-sharing or short-term rentals. Although it is now a well-known term - from sharing to the peer-to-peer economy, it refers to cheap access to information on an enormous scale through a digital platform that meets supply and demands beyond the central feature to which potential consumers and providers are connected. The platforms that play an essential role in the development of this economy make it possible for nonprofessional service providers to offer goods and services in a wide variety of areas, to generate value and to develop more development potential. The sharing economy covering these situations is a rapidly developing phenomenon. However, the names used to describe this phenomenon, in particular, blur the lines with the use of the new three concepts: traditional consumer, business, and ‘user-provider-platforms’ that do not match the agent concepts. Moreover, the regulatory frameworks on which the legal provisions to be used for their implementation are based do not contain clarity. Three-sided transactions cover a wide range of applications, from non-monetary sharing to real person businesses, and in particular from business to consumer (B2C) business models. In this context, this new digital economic model raises the need to make some market arrangements, including taxation. In order to determine the tax requirements that arise in this new digital economic model, an approach that is appropriate to each event and a tax-based plan should be adopted, not a strategy that fits all, in order to clarify the situation of the parties involved in the sharing economy transactions and to ensure efficiency in taxation. Based on the existing examples of national and local regulatory approaches adopted so far, it may be possible to create alternative and practical solutions that are specific to some areas in which the sharing economy is developed and which address the defined side effects. Collaborating with platforms to ensure tax compliance of participants is one of these alternatives. Platforms may play a role in cooperation with tax authorities to exchange information on tax obligations.

Examination of Tax Administration by Digitalization

35

They can help with the tax statement process and may even take on the part of collecting some taxes (such as local taxes on tourism) by simplifying the collection of tax authorities. It is the first step to clarify the definitions to address the tax challenges of the sharing economy, to comprehend the rapidly developing, multifaceted reality, and to understand exactly what the sharing economy means. Arrangements made for the electronic service providers for Turkey in the understanding of the economic value created by these new digital models include some difficulties in practice. It may be possible to overcome these difficulties in cooperation with platforms by taking into account the implementation examples in EU countries.

Acknowledgements This study was supported by Scientific Research Projects Coordination Unit of Bandırma Onyedi Eylül University. Project Number: BAP-19-1009-019.

References Aslam A. & Shah A. (2017), “Taxation and the Peer-to-Peer Economy”, IMF Working Paper, August 2017, https://www.imf.org/~/media/Files/ Publications/ WP/2017/wp17187.ashx, (02.07.2018) European Commission (2016). A European Agenda for the Collaborative Economy, Brussels, https://ec.europa.eu/docsroom/documents/16881/ attachments/2/translations/en/renditions/pdf European Commission (2018), “Study to Monitor the Economic Development of the Collaborative Economy in the EU”, http://www.technopolis-group. com/wp-content/uploads/2018/08/CE_Final-report_PartA_Final_230218. pdf, (20.09.2018) European Parliament Research Service (2016a). The Cost of Non-Europe in the Sharing Economy, European Added Value Unit of the Directorate for Impact Assessment and European Added Value, http://www.europarl.europa.eu/ RegData/etudes/STUD/2016/558777/EPRS_STU(2016)558777_EN.pdf, (01.06.2018) European Parliament Research Service (2016b). A European Agenda for the Collaborative Economy, Briefing, November 2016, http://www.europarl. europa.eu/RegData/etudes/BRIE/2016/593510/EPRS_BRI(2016)593510_ EN.pdf, (12.09.2019). European Parliament Research Service (2018), Collaborative Economy and Taxation, February 2018, http://www.europarl.europa.eu/RegData/etudes/ IDAN/2018/614718/ EPRS_IDA(2018)614718_EN.pdf, (07.08.2018)

36

Burçin Bozdoğanoğlu

Kara M.S. (2018), “Elektronik Ortamda Sunulan Hizmetlerde KDV (3 No’lu KDV Beyannamesi)”, Yaklaşım, Aralık 2018, Sayı:312, www.yaklasim.com OECD (2017), Shining Light on the Shadow Economy: Opportunities and Threats, http://www.oecd.org/tax/crime/shining-light-on-the-shadoweconomy-opportunities-and-threats.pdf, (08.08.2018) OECD (2018), Tax Challenges Arising from Digitalisation-Interim Report 2018 OECD/G20 Base Erosion and Profit Shifting Project, https://read.oecdilibrary.org/taxation/tax-challenges-arising-from-digitalisation-interimreport_9789264293083-en#page1, (02.01.2019) Özdemir M. (2019), “Elektronik Ticarette Aracılık Hizmetlerinin vergisel Açıdan Değerlendirilmesi”, E-Yaklaşım, Ocak 2019, www.yaklaşim.com PWC (2017), Share Economy 2017 the New Business Model, https://www. pwc.de/de/digitale-transformation/share-economy-report-2017.pdf, (07.08.2018) Rahim N. et al., Research on the Sharing Economy, Her Majesty’s Revenue and Customs (HRMC)’s Report, May 2017. Selloni D. (2017). “New Forms of Economies: Sharing Economy, Collaborative Consumption, Peer-to-Peer Economy”, CoDesign for Public-Interest Services, Springer International Publishing, 2017, pp. 15–26.

Ayşe Yiğit Şakar

Tax Incentives Provided to Green Bonds in Financing of Energy Efficiency and Its Importance for Turkey Abstract: Energy efficiency is of vital importance for Turkey as one of the elements of sustainable development. Turkey is faced with difficulties in providing sustainable development due to its dependence on imported energy. Besides, the impacts of climate change have a negative effect on Turkey’s environment and economy. Turkey is therefore involved in international efforts to combat global climate change and reduce greenhouse gas emissions. As for many countries, financing of energy efficiency is also a significant issue for Turkey. As an alternative to financing energy efficiency, green bonds are developing rapidly all over the world. Green bonds are financial instruments that provide opportunities for investors to participate in the financing of “green” projects that help reduce the negative impacts of climate change and adapt to the effects of climate change, reduce CO2 emissions, prevent environmental pollution, and improve social welfare. These structures have an essential impact on the realization of sustainable development. Turkey’s first and only green bond was issued by the Industrial Development Bank of Turkey in 2016 and attracted investors’ attention. Countries such as the US, China, and Chile apply tax incentives for green bonds to attract investors. However, the level of awareness of green bonds in Turkey is low, and there are no tax incentives yet. Necessary measures should be taken to facilitate the financing of energy efficiency in Turkey, and tax incentives should be implemented for green bonds. In this paper, the development and types of green bonds in the world and Turkey, tax incentives provided for green bonds in the financing of energy efficiency in the world and Turkey, and recommendations for Turkey were discussed. Keywords: Green bonds, green bond principles, tax incentives, energy efficiency, sustainable energy, sustainable development JEL Codes: H23, K34, P18, P48, Q01, Q28, Q42, Q48, Q54, Q56

1 Introduction Natural disasters stemmed from climate change in 2017 are estimated to cause US$320 billion in losses (Bahuet, October 8, 2018). Turkey is one of the most affected countries by global climate change, especially by desertification and deterioration of water resources. Therefore, efforts are made at the national and international levels to ensure a sustainable environment and development. Combating climate change is not solely an environmental problem. Within the

38

Ayşe Yiğit Şakar

scope of the combat, the global transition to the low carbon economy will determine the countries’ sustainable development, energy, health, agriculture and food security, water resources utilization policies and bring about a social and economic transformation. Sustainable development can only be successful if environmental, economic, and social policies are implemented in a coherent and balanced manner. Energy efficiency is one of the elements of sustainable development. Environmental, economic, political, and social concerns increase the importance of energy efficiency policies. Recently, green bonds have attracted the attention of investors as a sustainable financing instrument in the financing of energy efficiency investments. Stock markets such as Italy, Oslo, London, Mexico, Luxembourg, Shanghai, and Shenzhen have established a specific green bond market segment to strengthen the green bond market (Reboredo, 2018: 39). Incomes obtained from green bonds are provided with tax incentives in the US, China, India, Brazil, and Chile. However, there are no tax incentives in Turkey. In this paper, tax incentives that can be provided to green bonds in the financing of energy efficiency in Turkey and its importance will be discussed.

2 Methods of Financing Energy Efficiency in Turkey Financing energy efficiency in Turkey is of considerable importance in terms of its benefits and results. In Turkey, which is dependent on foreign energy and whose economy is adversely affected by global climate change, financing energy efficiency projects provide significant and environmental benefits. Energy efficiency projects contribute to the increase of efficiency in the use of energy resources and the reduction of greenhouse gas emissions. They support the reduction of external dependency and current account deficit by ensuring supply security in energy. Energy efficiency investments in Turkey are not at the desired level due to difficulties arising from access to finance, and for Turkey to reach its 2023 energy intensity goal, it is necessary to remove the obstacles to the financing of energy efficiency (Ata, 2013:99). Necessary supports in the financing of energy efficiency in Turkey can be listed as such (http://www.yegm.gov.tr/verimlilik/destekler.aspx.): • Support for Efficiency Enhancing Projects in industrial enterprises. • Benefiting from 5th Region incentives for Energy Efficiency Investment Projects. • Support for reducing energy intensity by making Voluntary Agreements for industrial enterprises.

Tax Incentives Provided to Green Bonds

39

• Support provided for energy efficiency projects by the industrial enterprises of the Technology Development Foundation of Turkey. • Support provided by KOSGEB to small and medium enterprises for surveys and energy efficiency consultancy services taken from energy efficiency consultancy companies. • Loans provided by international development agencies and financial institutions and by national banks and financial institutions under favorable conditions. Efficiency Enhancing Project (EEP) is a project designed to implement the necessary measures to eliminate energy wastes, losses, and inefficiencies in industrial enterprises. (http://www.yegm.gov.tr/verimlilik/d_VAP.aspx). EEP is submitted to the General Directorate of Energy Affairs by the industrial enterprises and approved by the Ministry of Energy and Natural Resources. These projects are supported by up to thirty percent of the application project costs. Application project costs cannot exceed five million Turkish Liras (Energy Efficiency Law No: 5627, art. 8/1 -a-1). Voluntary agreements for industrial enterprises include support for reducing energy densities. The Voluntary Agreement is concluded between the industrial enterprise and the General Directorate of Energy Affairs. This agreement takes an enterprise’s energy intensity for the past five years as a reference. The industrial enterprise commits to reduce the energy density by an average of 10% within three years of the agreement. 30% of the energy expenditure for the year in which the deal is made of the industrial enterprise fulfilling its commitment is covered. This amount cannot exceed one million Turkish Liras (Energy Efficiency Law No: 5627, art. 8/1 -b-1). In accordance with the Decision on State Aids in Investments No.2012/35, Energy Efficiency Investment Projects are evaluated within the scope of 5th Region Incentives and provided with incentives such as tax, insurance premium support, and interest support. Support is provided to Industrial organizations’ future Energy Efficiency Projects with three new support programs launched in August 2006 by the Technology Development Foundation of Turkey (http://www.yegm.gov.tr/ verimlilik/d_kobi_en_ver_destek.aspx). KOSGEB supports Small and Medium Enterprises with their studies, consultancy, and training services within the scope of energy efficiency. It includes the support given to energy efficiency studies and consultancy services from Energy Efficiency Consultancy Companies authorized under the Energy Efficiency Law No. 5627. (http://www.yegm.gov.tr/verimlilik/d_kobi_en_ver_destek.aspx.)

40

Ayşe Yiğit Şakar

In addition, loans are provided by international development agencies and financial institutions and by national banks and financial institutions under favorable conditions. Furthermore, green bonds began to be used more and more in the financing of energy efficiency in Turkey and worldwide.

3 Green Bonds in Financing Energy Efficiency 3.1 The Concept of Green Bond and Its Development A green bond is a debt security that is issued to raise capital specifically to support climate-related or environmental projects (IBRD, 2015:23; Berensmann vd., 2016:2; Kandır &Yakar, 2017a:161; Tang & Zhang, 2018: 4). This financial instrument is attractive as an alternative investment tool for investors focusing on integrating Environmental-Social-Governance issues into investment processes in international financial markets (ESCARUS, 2018:42). Green bonds are financial instruments that provide opportunities for investors to participate in the financing of “green” projects that help reduce the negative impacts of climate change and adapt to the effects of climate change (Reichelt, 2010: 2; Tang & Zhang, 2018: 4; Kandır &Yakar, 2017a:161). Green Bonds can be defined as “any bond instrument where the proceeds will be exclusively applied to finance or re-finance, in part or in full, new and/ or existing eligible Green Projects and which are aligned with the four core components of the GBP” (International Capital Market Association, 2018). A vital feature of these bonds valued by many investors is the due diligence process that the issuer of green bonds conducts to identify and monitor ‘green’ projects (Reichelt, 2010: 2). The difference of green bonds from other bonds is that the funds obtained from the bond issuance must be used in green projects (Kandır &Yakar, 2017 b: 92). As can be seen in Tab. 1, renewable energy and energy efficiency investments have the largest share. In 2017, the share of renewable energy decreased by 3.7%, while the share of energy efficiency increased by 2.4% compared to 2016. The private sector, international financial institutions, supranational institutions as well as national level states, regional governments and municipalities can also issue green bonds (Flaherty vd., 2017: 471–472). As seen from Tab. 2, while financial institutions had the largest share with 36.45% of green bond issuers in 2016, its share in 2017 decreased to 16.59%. The most drastic increase in green bond issuance in 2017 occurred in agency and sovereign bonds. The global green bond market started in 2007 with an issue from the European Investment Bank. In 2008, the World Bank issued the first fixed-rate bond carrying a green label (Chiang, 2017:8).

41

Tax Incentives Provided to Green Bonds

Tab. 1:  Comparison of the Use of Proceeds for 2016 and 2017. Reference:  Green Bonds:  Review of 2017, s.  11. https://www.environmentalfinance.com/assets/files/ Green%20Bonds%20Review%20of%202017.pdf (16.02.2019).   Renewable Energy Energy Efficiency Clean Transportation Sustainable Water Management Pollution Prevention and Control Terrestrial and Aquatic Biodiversity Conservation Eco-Efficient Products, Production Technologies, and Processes Sustainable Management of Living Natural Resources Climate Change Adaptation

2016 29.1% 19.4% 14.02% 11.2% 11.3%  2.1%  1.2%  2.1%  5.6%

2017 25.4% 21.8% 14.06%  9.95%  9.95%  7.8%  5.6%  5.24%  0.2%

Tab. 2:  Comparison of Issuer Types (2016–2017). Reference: Green Bonds: Review of 2017, s.  11. https://www.environmentalfinance.com/assets/files/Green%20Bonds%20 Review%20of%202017.pdf (16.02.2019). Agency Corporate Financial Institution Municipal Sovereign Supranational

2016 10.% 31.5% 36.45% 10.5% 0.85% 10.5%

2017 31.27% 29.73% 16.59% 8.9% 7.07% 6.41%

Poland was the first national government to issue a green bond valued US$800 million in December 2016. It was followed in January 2017 by France. Argentina, Chile, Fiji, Lithuania, Malaysia, Nigeria, Singapore, Switzerland, and the UAE joined the fray in 2017 (Green Bonds: Review of 2017, s. 3). In March 2018, the Government of Indonesia issued the first Green Islamic Bond (green sovereign Sukuk). The  five-year  issuance reached US$1.25  billion (Bahuet, October 8, 2018). The US, China, and France are the leading countries in the green bond market. China’s first green bonds were issued in 2015 (Boulle et al., 2017: 4). Whereas in the US and China, issuers were mainly agencies and financial institutions in 2017, French issuers were more diverse. The French sovereign bond accounts for 53% of the value of issuance from the country, but corporates account for

42

Ayşe Yiğit Şakar

32%, agencies 10%, and financial institutions and municipals the remaining 5% (Green Bonds: Review of 2017, s. 3). The first green municipality bond was issued in June 2013 in Massachusetts. In October 2013, Gothenburg issued its green city bond. There are different US states, Ontario state (Canada), Johannesburg city (South Africa), and La Rioja state (Argentina) among the large scale green bond issuers. The green bond issuance of regional governments is also ongoing (ESCARUS, 2018: 39). On May, 18th 2016, the Industrial Development Bank of Turkey was the first to issue a 5-year term $ 300 million green/sustainable bond. The bond issuance received approximately $ 4 billion in demand from 317 institutional investors in international markets. 44% of the need for the bond issue came from England, 39% from Continental Europe, 9% from US off-shore funds, 8% from Asia, and the Middle East (TSKB, 2017:13). Industrial Development Bank of Turkey financed 75 projects in green bonds, automotive, iron and steel, cement, chemical and plastics, and other sectors. These projects are expected to contribute 4.6 billion energy savings (kcal/year) and reduce 3.6 million CO₂ emissions (TSKB, 2017: 7).

3.2 The Green Bond Principles In 2014, the Green Bond Principles, which are in the form of advice and voluntary guidance, have been formed (Ceres, 2014:1). Green Bond Principles are designed for the development of the green bond market. The issuers are guided and supported at the key points when issuing green bonds. It proposes the transparency, sharing, and integrity of the information to be reported to the stakeholders by the issuers. (International Capital Market Association, 2018). Green Bond helps investors by ensuring the availability of information necessary to assess the environmental impact of their investments and helps the insurance companies by moving the market towards standard explanations to facilitate transactions (Ceres, 2014:1). The GBP has four core components: “1. Use of Proceeds 2. Process for Project Evaluation and Selection 3. Management of Proceeds 4. Reporting”

Use of Proceeds: The primary requirement for incomes; green bond income is used for Green Projects. They should be appropriately identified in the legal documents for safety. The main green projects, but not limited to the ones below, areas such (Ceres, 2014: 2; International Capital Market Association, 2018):

Tax Incentives Provided to Green Bonds

43

• Renewable energy (including generation, transmission, equipment, and products); • “Energy efficiency (such as new and renovated buildings, energy storage, central heating, smart grids, devices, and products); • Prevention of pollution and its control (including reduction of air emissions, greenhouse gas control, soil improvement, waste prevention, waste reduction, waste recycling and energy/emission-efficient waste-to-energy conversion); • Environmentally sustainable management of living natural resources and land use (including environmentally sustainable agriculture, environmentally ­sustainable livestock, climatic intelligent farm inputs such as biological plant protection or drip irrigation, environmentally sustainable fishing, and aquaculture, environmentally sustainable forestry, conservation or restoration of natural landscapes); • Protection of terrestrial and aquatic biodiversity (including protection of coastal, marine and watershed environments); • Clean transport (such as electric, hybrid, public, railway, non-motorized, multimodal transport, infrastructure for clean energy vehicles and reduction of harmful emissions); • Sustainable water and wastewater management (including fresh and/or drinking water, wastewater treatment, sustainable urban drainage systems and sustainable infrastructure for river reclamation and other forms of flood reduction); • Adaptation to climate change (including information support systems such as climate monitoring and early warning systems); • Products, production technologies, and processes adapted to the eco-efficient and/or cyclic economy (such as eco-label or environmental certification, resource-efficient packaging and distribution, and the development and promotion of sustainable products in an environmental context); • Green buildings are meeting regional, national, or internationally accepted standards or requirements of certificates.” Process for Project Evaluation and Selection: The Green Bond issuer must clearly inform investors of the relevant eligibility criteria, including e­ nvironmental sustainability objectives, the issuer’s method of determining which projects fit into the Green Projects category, exclusion criteria or other ­processes applied to identify and manage potential environmental and social risks associated with the projects (International Capital Market Association, 2018). Management of Proceeds: Net revenues of Green Bonds must be moved to a sub-portfolio or otherwise monitored by the issuer and proven by a formal

44

Ayşe Yiğit Şakar

internal transaction that will be linked to the lender and investment transactions for the donor’s projects. As long as Green Bonds are extraordinary, the balance of tracked revenues should be periodically reduced by amounts that match the investments made during that period (Ceres, 2014:4). In order to ensure transparency, it is recommended that the internal monitoring method and resource allocation and management of revenues from Green Bond revenues should be supported by an auditor or another third party (International Capital Market Association, 2018). Reporting: The issuers should keep up-to-date information on the use of the fund and update it annually or in the event of significant developments until the allocation is completed. The annual report must include a list of projects to which Green Bond revenues are allocated, as well as a brief description and distribution of projects and the estimated impact of the projects (International Capital Market Association, 2018).

3.3 Types of Green Bonds There are four types of Green Bonds (International Capital Market Association, 2018). • Standard Green Use of Proceeds Bond: It is the obligation of a standard recourse debt to the issuer aligned with the green bond principles. • Green Revenue Bond: It is the obligation of a conventional non-recourse debt to the issuer aligned with the green bond principles in which the credit risk in the bond depends on pledged cash flows of inflows, wages, taxes, etc. and where their income is used for the relevant or unrelated green project(s). • Green Project Bond: It is the project bond that complies with the principles of green bonds for one or more green projects where the investor will be directly exposed to the risk of project/projects with or without potential recourse to the issuer. • Green Securitised Bond: It is the bond secured by one or more specific green projects, such as covered bonds, ABS, MBS, and other structures, and aligned with the GBP. Green bonds are also classified as labeled and unlabeled green bonds. While labeled green bonds can be marketed as green bonds, unlabeled green bonds cannot be traded as green bonds in the capital market, although they are used to finance environmentally friendly projects as labeled bonds (Kandır & Yakar, 2017b: 94).

Tax Incentives Provided to Green Bonds

45

3.4 Tax Incentives Provided for Green Bonds in the World and Turkey Tax incentives are one of the public finance instruments used for the growth of the green bond market in both developed and developing economies. Tax incentives applicable to green bonds may differ between bond issuers and bond investors (Climate Bonds Initiative & IISD-International Institute for Sustainable Development, 2016:10). The following tax incentives may be provided to issuers: (Asian Development Bank, 2018: 77): • A credit, in addition to deduction of bond interest and issuance expenses from income, which can be used to lower the issuer’s other taxes; • Reduction or exemption from VAT on social taxes for green bond projects; • Tax credits with refund policy for some or all green bond issuance; and • Subsidies or refundable tax credits provided for interest costs. Investors may find green bonds tempting due to their better post-tax returns compared to other bonds or investments. The benefits of tax incentives for investors can be listed as tax credits, deductions from taxable income, appropriate tax rates, or tax exemptions. The following tax incentives may be provided to investors (Asian Development Bank, 2018:  80; Climate Bonds Initiative & IISD-International Institute for Sustainable Development, 2016:10): • Investment loans for green bonds that can be used to reduce income from eligible bonds or tax in a broader income. • Deduction of income from eligible bonds • Reduced tax rates on income from eligible bonds, • Exemption from stamp duty and capital gains tax. The incentive can be a goal, especially at green bonds, or it can be a part of other programs to encourage the financing of infrastructure or bond markets. Tax credits or subsidies may promote the issuance. However, incentives can have a striking effect by encouraging the use of some form of funding, for instance, project bonds, not to contribute to the financing of a green project, but only to benefit from the tax incentive. And that will shift the funding from unionized bank loans to green bonds (Asian Development Bank, 2018: 77).

3.4.1  European Union The European Commission states that urgent action is needed to adapt public policies to climate change due to climate change and disasters caused by resource

46

Ayşe Yiğit Şakar

depletion. The Commission also adds that the sustainable financial system will play a key role in ensuring the green economy, sustainable development, financial stability, and transparency in the economy. For these purposes, at the end of 2016, the Commission appointed a Senior Expert Group on sustainable financing. On January 31, 2018, the expert group published its final report. “The Report argues that sustainable finance is about two urgent imperatives: (1) improving the contribution of finance to sustainable and inclusive growth by funding society’s long-term needs; (2) strengthening financial stability by incorporating environmental, social and governance (ESG) factors into investment decision-making.

The Report proposes eight key recommendations, several cross-cutting recommendations, and actions targeted at specific sectors of the financial system”. Based on these recommendations, the European Commission published an Action Plan on Financing Sustainable Growth in March 2018 (European Commission, 2018). In Action 2 of the Action Plan, the European Commission commits to creating standards and labels for green financial products. This action will preserve the integrity and trust of sustainable financial markets and make it easier for investors looking for these products. Thus, access to green bonds will be more comfortable (European Commission, 2018). In June 2018, the European Commission set up a Technical Expert Group (TEG) on sustainable financing to assist in four critical areas of the Action Plan. In June 2019, TEG published ten recommendations in its report. One of these recommendations is to encourage green bond investors (EU Technical Expert Group on Sustainable Finance, 2019:  11):  “Recommendation 4:  Investors, in particular, institutional investors are encouraged to use the requirements of the EU-GBS when designing their green fixed-income investment strategies and to communicate their preference and expectations actively to green bond issuers as well as to underwriters”. In the report, TEG recommends that the European Commission and the EU Member States consider developing short and long-term financial incentives to support the development of the EU Green Bond Market in line with the EU-Green Bond Standard (EU Technical Expert Group on Sustainable Finance, 2019: 47): “Recommendation 7: Consider developing financial incentives to support the EU Green Bond Market alignment with the EU-GBS.” TEG recommends giving tax incentives at the issuer and investor level to the EU countries for the development of the green bond market. In this regard, TEG shows U.S. federal government Clean Renewable Energy Bonds (CREBs99) and Qualified Energy Conservation Bonds (QECBs100) programs and Accelerated

Tax Incentives Provided to Green Bonds

47

Depreciation Scheme as examples for incentives (EU Technical Expert Group on Sustainable Finance, 2019: 49–50).

3.4.2  United States of America Tax incentives applied in the United States set an example for many countries, and there are various tax incentives provided to investors or issuers for green bonds (https://www.climatebonds.net/policy/policy-areas/tax-incentives): • Tax credit bonds: Bond investors receive tax credits instead of interest payments. An example of this in the clean energy field is the US Federal Government’s Clean Renewable Energy Bonds (CREBs) and Qualified Energy Conservation Bonds (QECBs) program. In the program, the issuance of tax credit bonds of municipalities is allowed; The municipality or the federal government subsidize a tax credit of up to 70% of the coupon. • Direct subsidy bonds: Bond issuers receive cashback from the state to subsidize net interest payments. Also, this structure is used under the US federal government Clean Renewable Energy Bonds (CREBs) and Qualified Energy Conservation Bonds (QECBs) program. • Tax-exempt bonds: Bond investors do not have to pay income tax on interest from their green bonds (which may lower the issuer’s interest rate).

3.4.3  France Green labels were developed by the French Government to make green assets more visible. The first label, the TEEC, was granted to 18 funds with 2 billion euro assets under management (AUM), the second one, the ISR was awarded to 119 funds with EUR22bn of AUM, and the third label, which targets crowdfunding platforms and provides funding for green growth projects, was granted to 16 platforms. Tax incentives for SMEs and the adoption of broader collection platforms, including green securitization, are other methods to facilitate market access for small businesses (Filkova et al., 2018:7).

3.4.4  Brazil Brazil allows the issuance of tax-free bonds for significant infrastructure investments, construction companies, and wind power plants (Climate Bonds Initiative & IISD-International Institute for Sustainable Development, 2016:10). Brazil launched infrastructure debenture bonds in 2011, and investors were granted tax-free income from bonds that are issued to provide financial support for priority projects on infrastructure. In 2016, approximately one-third of 68

48

Ayşe Yiğit Şakar

issues totaling BRL18.3 billion ($5.6 billion) was for renewable energy, and individuals owned around %45 of them, with major retail investment being made via mutual funds investing in infrastructure bonds (Asian Development Bank, 2018: 81).”

3.4.5  China The Chinese government provides various incentives to banks and companies, such as reduced central bank loan costs and subsidized interest payments on green bonds. Up to 12 percent of the interest rate on environmentally friendly loans is subsidized by the state (Morris, 2019).

3.4.6  Malaysia Malaysia has provided a series of encouraging tax applications that can be implemented equally to green bonds and sukuks to stimulate the growth of the bond and Sukuk markets, for example, exemption for resident investors from income tax, and for foreign investors from withholding tax, on interest income from ringgit-denominated debt securities and Sukuk, and foreign-currencydenominated Sukuk issued in Malaysia. Moreover, there is no stamp duty on sales of securities and no tax on capital gains is imposed in Malaysia (Asian Development Bank, 2018: 81) The popularity of environmentally friendly Islamic financial products has increased with the launch of the Responsible Investment (SRI) framework in 2014. The Securities Commission (SC) offered SRI products compliant with Islamic rules and stated that it would further strengthen the country’s leading position in the Sukuk market. The Commission also said that Malaysia could improve its value proposition as a center for Islamic finance and sustainable investment. In addition, an incentive package was introduced by the commission under the SRI initiative to increase the green Sukuk trend further. The kit consists of tax reductions on issuance costs of any SRI Sukuk authorized by the SC before 2020, tax relief on the use of green technology in energy, transport, building, waste management, and associated services (https://oxfordbusinessgroup.com/ news/launch-sustainable-finance-sees-malaysia-go green). One of the nine special categories to be financed with green Sukuk in Malaysia is energy efficiency projects. The government announced an RM6 million Green SRI Sukuk Grant Scheme in its 2018 budget. With this program, issuers were enabled to balance the expenses of external review of applications obtained by the SC between January 2018 and December 31, 2020. Although significant incentives seem to be provided for the issuance of green Sukuk, it can be said

Tax Incentives Provided to Green Bonds

49

that there is not much awareness among market participants about these potential tax benefits. Issuers can benefit from a tax deduction for the issuance costs of SRI Sukuk approved by the SC until 2020. Furthermore, to boost renewable energy investments, there are two incentive programs: the Green Investment Tax Allowance and Green Income Tax Exemption. It is likely that these programs elevate green projects’ financial feasibility and thereby contribute to green Sukuk (Asian Development Bank, 2018:  118; https://www.climatebonds.net/files/ reports/cbi-policyroundup_2017_final_3.pdf).

3.4.7  India Investors can benefit from tax-saving infrastructure bonds for a tax deduction. Individual investors can deduct the amount up to ₹20,000 ($315) of their investments in qualified infrastructure bonds from their taxable income. Additionally, bond interest income is included in the taxable income of the bondholder and taxed at the appropriate marginal rate. Non-bank financial institutions and semi-independent entities designated by the Reserve Bank of India may issue eligible infrastructure bonds. Individual investors are motivated to direct investment in infrastructure savings, offering semi-dependent bodies with non-budgetary funding. The low limit minimizes fiscal costs; on the other hand, the administrative costs for issuers associated with the typical certificate value of ₹5,000 ($77) are substantial. The program may additionally contribute to the wider growth of the bond market by attracting the attention of people who  may  not  otherwise  contemplate investing in bonds (Asian Development Bank, 2018: 81).

3.4.8  Turkey In Turkey, there is not a special provision concerning the taxation of green bonds. In this case, it is necessary to look at the general rules. In Turkey, income obtained from government bonds and private bonds is subject to withholding tax under provisional Article 67 of Income Tax Law. However, interest yield and trading income from Eurobonds are not subject to withholding tax, regardless of the issuance date. Eurobond interest yield must be declared if it exceeds the annual income tax declaration limit (which is 40.000 TL for 2019). Limited taxpayers are not obliged to submit a declaration for their income from Eurobonds. Revenues from government bonds and treasury bills are subject to different taxation regimes depending on their issuance before and after 1/1/2006 (Gelir İdaresi Başkanlığı, 2019: 8):

50

Ayşe Yiğit Şakar

• In accordance with Decision no.2010/926 of Council of Ministers, 10% withholding shall be made after 1/10/2010 on full and limited taxpayer natural entities’ interest yield obtained from the government bonds and treasury bills issued after 1/1/2006 and their revenues from purchase and sale of these bonds. The annual declaration shall not be submitted for these revenues, which are ultimately taxed through withholding. • Interest yield from government bonds and treasury bills issued before 1/1/2006 and trading income shall not be subject to withholding tax within the scope of provisional Article 67. The taxation of these incomes shall be made in accordance with the provisions of the legislation in force on 31/12/2005. Therefore, a reduction rate shall be applied depending on the issue date of the bonds and bills in question and whether they are issued in Turkish Lira. The amount ­remaining after the cost value indexing and exception application for trading income shall be declared. Interest yield and trading income on private-sector bonds are subject to withholding tax within the scope of provisional article 67. The revenues to be obtained are subject to different taxation regimes according to their issuance before and after 1/1/2006 (Gelir İdaresi Başkanlığı, 2019: 9): • Domestic private sector bond/bill interest yields and trading income are ultimately taxed through withholding after 1/1/2006 within the scope of provisional article 67. • Private sector bond/bill interest yield and trading income issued before 1/1/2006 are taxed utilizing declarations such as Government bond/ Treasury bill revenues in the same situation. Bonds issued abroad by full taxpayer institutions are accepted as the securities income stated in the subparagraph (5) of the second paragraph of Article 75 of the Income Tax Law. These bonds are not subject to withholding tax within the scope of provisional article 67 since they are not included in provisional article 67. However, withholding tax shall be applied to interest yield obtained from bonds issued abroad by full taxpayer institutions within the scope of Article  94 of Income Tax Law according to maturity and rates determined by the Decision No. 2011/1854 of Council of Ministers (Gelir İdaresi Başkanlığı, 2019: 10). • Accordingly; • For interest obtained from bonds with a maturity up to 1  year, %10 of withholding tax • For interest earned from bonds with a maturity between 1 and 2 years, %7 of withholding tax

51

Tax Incentives Provided to Green Bonds

• For interest obtained from bonds with a maturity between 3 and 5 years, %3 of withholding tax • For interest earned from bonds with a maturity between 5 and more years, %0 of withholding tax is applied. There are no tax incentives for green bonds in Turkey. Providing tax incentives in financing energy efficiency can increase interest in green bonds (For the same point of view, see Kandır &Yakar, 2017a: 104). Tab. 3 shows the incentives provided to green bonds in some countries.

4 Conclusion In many countries around the world, interest in green bonds is increasing, and this market is growing to combat climate change, to get rid of the dependence on fossil fuels in energy, and to ensure sustainable development. In 2016, the Industrial Development Bank of Turkey (TSKB) became the first establishment in Turkey to issue green bonds. There are countries such as Poland and France that issue sovereign bonds; however, Turkey is not one of them. In terms of our country, which is dependent on external resources in energy and which has a current account deficit, it is vital to strengthen and encourage the green bond market as an option for financing energy efficiency. Issuing green bonds in Turkey does not provide any tax benefits for investors. For the reasons mentioned above, tax incentives should be used as tools to be able to benefit from green bonds in financing energy efficiency and to improve the green bond market (See Kandır & Yakar, 2017a: 104).

Tab. 3:  Examples of Tax Incentives Related to Green Bonds. Reference: Climate Bonds Initiative & IISD-International Institute for Sustainable Development 2016: 14; the Author Added France, China, and Turkey. Country Bond Type Chile

All bonds

Degree For Who of Tax Exemption Full Foreign Institutional Investors

Description

Relevance for Green

Foreign institutional investors are exempt from tax on the bond

The incentive can be replicated for foreign investment into green bonds in particular

(continued on next page)

52

Ayşe Yiğit Şakar

Tab. 3: (continued) Country Bond Type India

USA

Degree For Who of Tax Exemption Full Investors

Muni bonds and selected corporate bonds from public entities Muni bonds Full

Investors

USA

Muni Partial bonds with proceeds for renewables and energy efficiency

Investors

Brazil

Bonds with Full proceeds for infrastructure including construction and wind energy

Investors

Description

Relevance for Green

Tax-free bonds issued by public corporations and municipal government Over 80% of the US muni bond market is taxexempt, intending to increase funding for municipalities for infrastructure Qualified Energy Conservation Bonds (QECBs) and Clean Renewable Energy Bonds (CREBs) offer special tax incentives offered for muni bonds with proceeds clean energy and energy conservation Tax-free bonds can be issued for significant infrastructure investments, construction conglomerates, and wind farm developers

Examples of tax incentives used to encourage investment in a policy priority area. The incentive can be replicated to apply to all labeled green bonds with robust green credentials e.g., that comply with set standards

53

Tax Incentives Provided to Green Bonds Tab. 3: (continued) Country Bond Type Malaysia Corporate ABS bonds

Degree For Who of Tax Exemption Partial Investors

Description

Relevance for Green

Issuance expenses for asset-backed securities are tax-deductible

The incentive can be replicated to cover issuance costs for green aBS, in particular, making it cheaper for issuers of green, etc. non-green aBS  

France

Green Bond

 

Investors

China

Green Bond

 

Banks and businesses

Include tax incentives for SMEs Form of lower central bank borrowing costs and subsidized interest payments on green bonds

Turkey

Green Bond

None

None

None

For the most environmentally friendly loans, the government subsidizes up to 12 percent of the interest rate None

References Asian Development Bank (April 2018). Promoting Green Local Currency Bonds for Infrastructure Development in ASEAN+3, Manila, Philippines. https:// www.adb.org/sites/default/files/publication/410326/green-lcy-bondsinfrastructure-development-asean3.pdf (15.04.2019). Ata, S.U. (2013). “Sürdürülebilir Enerjinin Finansmanı”, Türkiye’de İklim Değişikliği ve Sürdürülebilir Enerji, (Ed.) Ediger, V. Ş., Istanbul, ENİVAEnerji ve İklim Değişikliği Vakfı, ss. 99–119. Bahuet, C. (October 8, 2018). Indonesia’s Green Sukuk, https://www.undp. org/content/ undp/en/home/blog/2018/Indonesias-green-sukuk.html (20.02.2019). Berensmann, K., Dafe F., Kautz M. & Lindenberg N. (2016). Green Bonds: Taking off the Rose-Colored Glasses, Briefing Paper, 24/2016,

54

Ayşe Yiğit Şakar

Department “World Economy and Development Financing”, German Development Institute/Deutsches Institut für Entwicklungspolitik (DIE). http://www.die-gdi.de/uploads/media/BP_24.2016.pdf (28.01.2019). Boulle, B., Dai, L. & Meng, A (2017). China Green Bond Market 2016, Climate Bonds Initiative & the China Central Depository & Clearing Company (CCDC), January. https://www.climatebonds.net/resources/reports/chinagreen-bond-market-2016-0 (22.02.2019). Ceres (2014). Green Bond Principles, 2014, Voluntary Process Guidelines for Issuing Green Bonds, Ceres Report, January 13, 2014. https://www.ceres. org/sites/default/files/reports/2017-05/Green%20Bond%20 Principles.pdf (18.02.2019). Chiang, J., (January 2017). Growing the U.S. Green Bond Market: Volume 1, California State Treasurers Office. http://treasurer.ca.gov/greenbonds/ publications/ reports/1.pdf (18.02.2019). Climate Bond Initiative. Green Bonds Policy: Highlights from 2017, https:// www.climatebonds.net/files/reports/cbi-policyroundup_ 2017_final_ 3.pdf (14/09/2019). Climate Bonds Initiative & IISD-International Institute for Sustainable Development (April 2016). Roadmap for China: Using Green Securitization, Tax Incentives, and Credit Enhancements to Scale Green Bonds. https:// www.iisd.org/sites/default/files/publications/greening-securitisation-taxincentives-credit-enhancements-green-bonds-en.pdf (03.04.2019). ESCARUS (2018). Dönüşen Dünyada Fırsatları Yakalamak: Sürdürülebilir Finans Görünümü, Eylül 2018. Istanbul. http://www.escarus.com/raporlarve-analizler (17.02.2019). European Commission (2018). Action Plan: Financing Sustainable Growth, COM(2018) 97 final, Brussels, 8.3.2018. https://eur-lex.europa.eu/legalcontent/EN/TXT/?uri=CELEX%3A52018DC0097 (14/09/2019). EU Technical Expert Group on Sustainable Finance (June 2019). Report on EU Green Bond Standart, TEG Report Proposal for an EU Green Bond Standard. https://ec.europa.eu/info/sites/info/files/business_economy_euro/ banking_and_finance/documents/190618-sustainable-finance-teg-reportgreen-bond-standard_en.pdf, (11.09.2019). Filkova, M., Frandon-Martinez, C., Meng, A., Rad, G. (2018). The Green Bond Market in Europe 2018, Climate Bonds Initiative, https://www.climatebonds. net/files/files/ The%20Green%20Bond%20Market%20in %20Europe.pdf (14/09/2019). Flaherty, M., Gevorkyan, A. Radpour, S. & Semmler, W. (2017). Financing Climate Policies through Climate Bonds – A Three-Stage Model and Empirics, Research in International Business and Finance, 42, pp. 468–479.

Tax Incentives Provided to Green Bonds

55

Gelir İdaresi Başkanlığı (2019). G.V.K. Geçici 67 nci Madde Uygulaması İle İlgili Olarak Gerçek Kişilere Yönelik Vergi Rehberi, Mükellef Hizmetleri Daire Başkanlığı Yayın No: 313 Şubat 2019, Ankara. Green Bonds: Review of 2017. https://www.environmentalfinance.com/assets/ files/Green%20Bonds%20Review%20of%202017.pdf (16.02.2019). https://www.climatebonds.net/policy/policy-areas/tax-incentives (15.01.2019). https://oxfordbusinessgroup.com/news/launch-sustainable-finance-seesmalaysia-go green (14/09/2019). http://www.yegm.gov.tr/verimlilik/destekler.aspx. (20.02.2019). IBRD (2015). What Are Green Bonds?, International Bank for Reconstruction and Development –The World Bank, Washington, http://documents. worldbank.org/curated/en/400251468187810398/pdf/99662-REVISED-WBGreen-Bond-Box393208B-PUBLIC.pdf (15.02.2019). International Capital Market Association (June 2018), Green Bond Principles Voluntary Process Guidelines for Issuing Green Bonds,https://www.icmagroup. org/green-social-and-sustainability-bonds/green-bond-principles-gbp/ (16.02.2019). Kandır, S.Y. & Yakar, S. (2017a). “Yeşil Tahvil Piyasaları: Türkiye’de Yeşil Tahvil Piyasasanın Geliştirilebilmesi İçin Öneriler”, Ç.Ü. Sosyal Bilimler Enstitüsü Dergisi, 26 (2), ss. 159–175. Kandır, S.Y. & Yakar, S. (2017b). “Yenilenebilir Enerji Yatırımları İçin Yeni Bir Finansal Araç: Yeşil Tahviller”, Maliye Dergisi, 172, ss. 85–110. Morris, H. (2019). “China emerges as key player in green bonds market”, China Daily Global. http://www.chinadaily.com.cn/a/201903/11/ WS5c855866a3106c65c34edc7f.html, (07.08.2019). Reboredo, J. C. (August 2018). “Green Bond and Financial Markets: Co-Movement, Diversification and Price Spillover Effects”, Energy Economics, 74, pp. 38–50, https://doi.org/10.1016/j.eneco.2018.05.030 (05.01.2019). Reichelt, H. (2010). Green Bonds: A Model to Mobilize Private Capital to Fund Climate Change Mitigation and Adaptation Projects (English). Euromoney Handbook. Washington, D.C.: World Bank Group. http://documents. worldbank.org/curated/en/680921507013408005/Green-bonds-a-modelto-mobilize-private-capital-to-fund-climate-change-mitigation-andadaptation-projects (22.02.2019). Tang, D.Y. &, Zhang, Y. (2018). “Do Shareholders Benefit from Green Bonds?”, Journal of Corporate Finance (Corfin), (November 22, 2018). Available at SSRN: https://ssrn.com/abstract=3259555 or http://dx.doi.org/10.2139/ ssrn.3259555, (24.02.2019).

56

Ayşe Yiğit Şakar

TSKB (2017). Green/Sustainable Bond, Allocation & Impact Reporting 2017. http://www.tskb.com.tr/web/101-3133-1-1/tskb-site-tr/tr-hakkimizda/ tskbden-haberler/yesil/surdurulebilir-tahvil-ile-yaratilan-degeri-anlatanallocation-impact-reporting-2017-yayinlandi (20.02.2019).

Güneş Çetin Gerger, Feride Bakar Türegün, and Adnan Gerçek

The Importance of Tax Literacy in Tax Compliance Abstract: Taxes are the economic values received by the governments and tax authorities under the enforcements according to the rules specified by the law. As tax is the most important source of income of the state, the approaches that strengthen tax awareness and increase tax compliance have become essential for the revenue administrations. Tax literacy is defined as the ability to know and understand the tax-related issues and to interpret them, staying updated regarding the developments and maintaining the personal budget most efficiently while considering the tax payments.This study highlights the importance of tax literacy as one of the factors that determine tax compliance. The study also examines arrangements and projects related to tax literacy in the OECD countries and the United States, along with the presentation of the projects and research related to tax literacy in Turkey. In the study, we conclude that the level of tax literacy in our country is not at the desired level, and thus we provide suggestions regarding the activities that could be conducted to increase tax literacy. Keywords: Tax literacy, tax compliance, tax awareness, tax psychology JEL Codes: H2, K34, D91, H26

1 Introduction Tax literacy is a new concept that emerged from developed countries. Tax literacy can be defined as the individual’s understanding of tax laws related to tax liability in his/her own financial environment, fulfilling his/her tax obligations, and evaluating the possible tax risks independently (Cvrlj, 2015: 158). Therefore, there is a close relationship between tax literacy and tax compliance. We can classify the factors affecting tax compliance as moral, psychological, economic, tax management factors, and demographic factors (Çetin Gerger, 2011: 9). There is a positive correlation between tax compliance and education level, a demographic factor. Tax awareness of taxpayers also develops with an increase in education level. Therefore, taxpayers’ compliance with tax compliance behavior also increases (Otto et al., 1987: 304). Today, advanced revenue administrations in the world moved from the “despite the taxpayer” approach to “with the taxpayer” approach (Gerçek et al., 2015:  25). In this sense, revenue administrations have an essential role in the

58

Güneş Çetin Gerger et al.

Tab. 1:  Actors and Determinants in the Formation of Tax Climate. Source: (Alm et al., 2012: 136). ACTORS Government

Governance and regulation, the image of taxpayers, tax law, the tax rate

DETERMINANTS

Tax Authorities

Images of government, tax accountants, taxpayers, audits, fines

Accountants

Images of government and tax authorities, taxpayers and their goals

Other Taxpayers

Taxpayers

Images of government and tax authorities, attitudes, tax morale, knowledge of tax law, norms (personal, social, societal), justice (distributive, procedural, retributive)

development of tax literacy. In this study, the concept of de “tax literacy”, which is a basic sub-factor affecting tax compliance, is examined, the applications in developed countries are evaluated, and suggestion is listed for future applications in our country.

2 Factors Affecting Tax Compliance and Tax Literacy 2.1 Factors Affecting Tax Compliance Paying taxes is not a voluntary action for taxpayers. For this reason, a behavioral economy has emerged in the 1970s in which the effects of psychological factors on tax compliance are investigated. This taxpayer centered approach has adopted the “trust is good; control is better” principle in which intrinsic motivation is developed through measures. In this case, all the actors in the system determine the tax climate (Alm et al., 2012: 134). The following Tab. 1 shows the actors determining tax compliance. A tax climate is formed as a result of the interactions between occupations, including government, tax authorities, financial advisors, accountants along with the taxpayers and interaction among taxpayers. The development of tax literacy ensures that the tax climate among these actors is positively affected. Increasing tax compliance has become one of the critical contemporary tasks of revenue administrations (Gerçek et al., 2015: 162). Successful revenue administrations have a management approach that includes optimizing tax compliance, risk-based verification programs, simple laws and procedures, and taxpayer training and assistance (Russel, 2010:  2). In this respect, programs and exemplary practices that strengthen the communication between the taxpayer and the administration, including tax literacy, is a common practice in all the developed countries.

The Importance of Tax Literacy in Tax Compliance

59

Tab. 2:  Cognitive-Affective-Psychomotor Scope of Tax Literacy. Source: (Yılar & Akdağ, 2017: 366). Levels

Cognitive Level

Affective Level

Psychomotor Level

Scope Being able to know the definition of a tax, types of tax, tax legislation, tax rates, important tax payment dates. Being able to have tax perception, belief in the necessity of tax payments, positive tax attitude, tax awareness, tax ethics. Being able to fill the tax forms related to tax return Being able to pay the tax debt by the relevant institutions or the internet

2.2 Importance and Scope of Tax Literacy A tax literate person is aware of the basic concepts of taxation, has knowledge about the basic functioning of the tax system, is updated regarding the developments, and knows the rights of taxpayers. Also, he can follow tax updates, has a positive attitude towards taxation, has a tax ethic, has a clear idea regarding the place of the tax in the personal budget, and can fulfill obligations related to the tax. In this sense, passing through cognitive and emotional dimensions, tax literacy is accomplished at the behavioral aspect (Yılar & Akdağ, 2017: 368). Empirical studies indicated that the taxpayers’ level of knowledge increases tax compliance. Tax literacy is a concept that exceeds the taxpayer’s knowledge level. Tax literacy includes elements such as (i) having tax awareness, (ii) having the conceptual knowledge and skills, and (III) the ability to rationalize information and to make decisions. The first element, tax awareness, refers to the understanding of individuals’ role in the financial exchange or social contract with the state. This awareness forms the basis of the framework as it is a prerequisite for tax literacy. The second element includes a legal component of conceptual knowledge and skills, as well as the procedure. The procedural content allows the use of knowledge and skills to regulate tax records by the taxpayer’s interacting with tax authorities. The legal scope refers to the understanding of how taxes are levied. Legal tax information

60

Güneş Çetin Gerger et al.

has two dimensions:  understanding the legal requirements and the legislation. The first dimension is the ability to know whether something is taxable, and the second dimension is the ability to apply legal knowledge to the particular situation to calculate the tax effect (Lai, Zalilawati, Amran and Choong, 2013). The third element is to rationalize the information and to make decisions. The fact that taxpayers participate in the fulfillment of their tax obligations is to realize their awareness and knowledge based on their perceptions within the social structure. In other words, awareness, knowledge, skills, and attitudes need to come together to make decisions to act in compliance with the tax (Bomman and Wasserman 2018: 1). Figure 1 presents the concept of tax literacy as a process. An individual’s tax awareness is linked to tax information. Understanding the tax information is the key to obtaining the desired result. In other words, the one with tax awareness can make an informed decision to act by the tax. In this process, other factors, such as individual and social norms, perceptions, attitudes, and tax ethics, are also considered to be effective on the individual (Bomman and Wasserman 2018: 1). Niemirowsaki, Baldwin & Wearing (2003) research shows that there is a relationship between tax compliance behavior and taxpayer awareness. Eriksen & Fallan (1996) determined that obtaining additional tax information increases

Fig. 1:  Tax Literacy Framework. Source: (Bomman & Wasserman 2018: 7)

The Importance of Tax Literacy in Tax Compliance

61

tax compliance and reduced tax evasion. Citizens’ tax information increases the confidence in the use of state taxes, while incomplete or misunderstanding results in insecurity (Kirchler et al., 2008: 216). Tax illiterate people may unintentionally have a higher tax incompliance (Chardon et al., 2016). In the “Building Tax Culture, Compliance and Citizenship” report of the OECD, the importance of the transformation in the state-citizen relations and the cultural change in the tax administration were emphasized. Tax authorities once had a culture of fear, but currently, they recognize citizens not only as “liability holders” but also stakeholders (OECD, 2015: 17). Therefore, this change requires an understanding of tax perception and increasing tax awareness (Yaltı, 2018:  65). Today, most OECD countries recognize the limitations of traditional enforcement-based techniques in tax compliance and emphasize training programs for taxpayers to develop tax compliance and tax morality.

2.3 Regulations Regarding Tax Literacy in OECD Countries The OECD is a leader in measuring global financial literacy, including comprehensive measurement tools for adults and youth. In 2016, it published its second international financial literacy assessment report covering 30 countries (OECD, 2017:  13). Tax literacy is a sub-discipline of financial literacy. In order to get a general impression on tax literacy, financial literacy data are included in the research for G20 countries. The Fig. 2 illustrates the data from this research. This report shows that, on average, only 52% of adults in G20 countries accomplished six levels out of the nine attitudes discussed. Although national surveys have been conducted in countries for tax literacy, only a report on financial literacy has been published by the OECD. In a report published by the OECD in 2015, the tax literacy practices in developing countries were mentioned, but a measure of tax literacy was not included (OECD, 2015). Turkey is a little below the average financial literacy level of the G20 countries (12.7) with a score of 12.5. Figure 3 indicates the level of financial literacy in the Netherlands, Switzerland, and Germany concerning their education level. A positive correlation was found between financial literacy and education level. Having sufficient knowledge regarding the basic concepts of financial literacy is notably less likely among non-university graduates (Lusardi and Mitchell 2007a, 2011c). In addition, financial literacy is particularly weak for those with lower quantitative education levels (Lusardi 2012). Lusardi, Mitchell, and Curto (2010) found a positive correlation between financial literacy and cognitive skills among participants of the NLSY questionnaire and stated that cognitive factors should be considered seriously.

62

Güneş Çetin Gerger et al.

Fig. 2:  Financial Information, Behaviors and Attitudes. Source: (OECD, 2017:8).

2.4 US Tax Literacy Project Implementation The general purpose of the “Tax Literacy Project” carried out at Tulane University in the USA is to inform the public about taxation and to ensure that US citizens have knowledge about tax and tax policies and to enable citizens to make rational decisions in the future US tax policies. The Tax Literacy Project, which first started in 2013 at Arizona State University (ASU) Sandra Day O’Connor College of Law and headed by Professor Marjorie E. Kornhauser, has emerged as a non-political effort to inform the public about taxes. The project has three different concentration areas: (a) Why we are taxed (the purpose of taxes; the link between tax and expenditure), (b) fairness of the taxation (how the tax burden will be distributed, including the tax base and the rate structure), and (c) basic taxation concepts. The project is intended for all ages between 12 and 80 and all educational levels (https://taxjazz.com/). Web-based games, other internet activities, and social networks are used in the scope of the project. The project was supported by the university, volunteer participants, and individual donations, and it was also structured to receive personal and foundation support to produce innovative materials.

Primary or lower secondary

26.6

Less than HS

12.6

Some college

Vocational

44.3

63.8

Upper secondary

Tertiary

68.9

College Post graduate

44.9

Switzerland

43.1

Highschool

19.2

31.3

USA

80 70 60 50 40 30 20 10 0

0

20

40

60

80

Lower secondary

21.7

Primary

28

Upper secondary

51.6

Lower secondary

35.1

Non-GDR

52.4

GDR

55

Upper secondary

Germany

Middle secondary

41.7

54.4

Netherlands

Fig. 3:  Financial Literacy Level by Education Level. Source: (Lusardi & Mitchell, 2014: 19).

80 70 60 50 40 30 20 10 0

0

10

20

30

40

50

60

70

Postsecondary

70.1

Higher vocational

55.4

Tertiary

72

University

69.8

The Importance of Tax Literacy in Tax Compliance

63

64

Güneş Çetin Gerger et al.

Tab. 3:  Taxjazz Project Processes. Source: http://search.ebscohost.com/login.aspx?direct =true&db=a9h&AN=48076039&lang=tr&site=ehost-live, The Tax Literacy Project, www .taxjazz.com (01.04.2109). 2009–2010 * Developing informationbased tax games (developed with the support of ASU Computer, Informatics, and Decision Systems Engineering Department).

2010–2012 * Activities through the website have been developed. These are games, competitions, videos, social networks, and so on.

* A two-day workshop was organized for the expert group to be trained by Sandra Day O’Connor Law School. (Taxation, media, education and game design).

* Game Development: This section is longterm and includes game development in taxation. (Theater-like games have been developed).

* The project was developed Generally, the design and to receive funding from the implementation of such a workshop. game can take up to two years.

2012–… * Taxpayers are supported by Tulane University law school students with programs that provide information and framework that enables taxpayers to analyze problems and make a sensible decision. * Instructor guides have been developed and presented on the web for individuals who can learn at home and want to lead group discussions such as teachers and community leaders. * Voluntary donations were opened to develop the project.

H&R Block office survey in the United States found that most Americans “failed in tax 101”. People were not aware of the general concepts of taxation, their tax obligations, and the implementation of laws. Many of the participants did not even know why taxes exist (Kornhauser, 2009: 9–10). Tax ignorance hurts individuals and society as a whole. It can prevent taxpayers from benefiting from the tax benefits they deserve and may cause excessive tax payments. In the long run, tax ignorance precludes the introduction of robust tax policies, which increase revenues most fairly and efficiently possible (Kornhauser, 2009:  9–10). The project applied is a social project. It helps taxpayers to fill the information gap. The project provides only necessary information to support any tax structure or rate. The project focuses on federal income tax, but most of the content has the potential to be applied to other federal, state, and local taxes.

2.5 Austrian Tax Literacy Project In a study by Cechovsky (2018) at the Vienna University School of Economics and Business, concepts and misconceptions about tax literacy were investigated,

The Importance of Tax Literacy in Tax Compliance

65

and detailed results were presented. The relationship between knowledge and tax compliance was investigated. The primary audience was students from business schools in Austria. A three-stage process was adopted for this study. The first one was carried out by the faculty members with 22 students, and the results were analyzed by coding and categorizing method. In the next stage, a pilot questionnaire was applied to 94 students. Finally, the pilot study was improved, and data was collected using a survey method with 688 students. The study was completed in about two years. Within the scope of the project, students’ knowledge and skills, interest, tax attitudes, social norms and behavioral controls, tax compliance intentions, and the effects of demographic factors were quantitatively evaluated. The result of the study revealed that having information about tax was an important factor in tax compliance (Cechovsky, 2018:  114). The students had the knowledge of concepts and misconceptions that was not a form of expert instruction (Möller, 2007). The lack of awareness of tax payment has led to the integration of new information into preexisting ones. Although students were in business school, there were information gaps. Their own experiences and the media influence their knowledge. Having tax information is positively correlated with tax compliance and negatively associated with tax evasion behavior (Cechovsky, 2018: 119).

2.6 Tax Literacy Project in the UK In an empirical study conducted and financed by the Chartered Taxation Institute (CIOT) in the UK, study socio-demographic impacts on financial and tax literacy (FTL), tax ethics, tax administration, and compliance perceptions of 377 young adults from two UK universities was obtained. The scope of this research consisted of students samples from modern universities* (*post 92 universities) in England. It deals with the relationships between financial literacy and tax literacy, tax ethics, and tax compliance. Examples include cohorts that did not receive any tax training and groups that take one or two modules on UK taxation. Student research was conducted at the beginning and end of the 2015/16 academic year. The result indicated that the students who take a taxrelated module at the university have high financial and tax literacy. It was also observed that people with employment experience have a higher tax ethic than those who do not (Alexander et al., 2018: 1–9). Based on these results, the UK Revenue Administration, HMRC launched a “Tax Facts” initiative for youth in the UK in 2015. This was followed by the “Tax Facts for Children” program for the benefit of primary school children in 2016 (Alexander et al., 2018: 1–9).

66

Güneş Çetin Gerger et al.

2.7 Australian Tax Literacy Survey The research empirically examined the relationship between Australian tax literacy and demographic factors. A  survey was conducted with more than 600 Australians to determine their TLS (Tax Literacy Score). The Australian tax literacy level was quite elevated. It was found that 81% of the respondents had a “basic” or higher tax literacy score. This means that only 19% of Australians have a score classified as “low.” Moreover, it was determined that the TLS scores of those who were employed were higher than that of unemployed (Chardon et al., 2016: 323). Also, the Australian Tax Office (ATO) generally tries to train taxpayers on various topics such as preventing tax base wear, providing information on potential tax benefits. Because ensuring financial literacy in general and tax literacy, in particular, are adopted by the state in the development of public finance policies. In this respect, the “Financial Literacy of Adults in Australia” survey was conducted by the ANZ bank in 2003, 2005, 2008, 2011, and 2015. The results of the survey were considered as financial literacy standards (Chardon et  al., 2016: 326). According to the 2005 ANZ study, 7% of the participants expressed that they needed more training in taxation. According to the 2005 ANZ survey report, three main problems arose in the field of financial literacy. These were (i)  the taxpayers are not aware of their financial situation, (ii) the events and information deficiencies are not in their control, and (iii) they are not able to decide when to take advice (ANZ, 2005).

3 Tax Literacy in Turkey It was determined that a small number of studies have directly measured the tax literacy levels in Turkey. Yardımcıoğlu, Akpınar, and Günay (2014) asked questions to taxpayers in Kahramanmaraş about tax literacy and tax awareness. They found that the respondents had 65% knowledge of the names of the taxes in the tax system and 42% knowledge of the tax procedures, principles, and rates. In the study by Teyyare et  al. (2018), a questionnaire was applied to students of the Faculty of Economics and Administrative Sciences at Abant İzzet Baysal University, and the authors determined that tax literacy levels of the participants were above average. However, some studies measured tax awareness, perception, and compliance in Turkey, and some of them have had some measurements to determine the tax-related knowledge of taxpayers. A table of empirical studies and target groups for these may be seen below.

67

The Importance of Tax Literacy in Tax Compliance

Tab. 4:  Some Studies Related to Tax Perceptions and Their Target Population in Turkey. Source: Own Elaboration Based on the Studies Below. Taxpayers and Citizens Muter, Sakınç Manisa ve Çelebi (1993) Yüce ve Bursa Gerçek (1998) Demir (1999) Afyon Bayraklı, Uşak Sağbaş ve Ural (2004) Çoban ve Denizli Sezgin (2004)

University Students Ömürbek, Çiçek Süleyman ve Çiçek (2007) Demirel Şahin (2010) İzgi ve Saruç (2011) Sağlam (2013) Gülten (2014)

Other Students Sağbaş ve Afyon Başoğlu (2005) Gaziosmanpaşa Karaot İzmir (2010) Sakarya Hitit

Taytak Uşak (2010) Zorlu (2012) Ankara

Afyon Kocatepe Çelik ve Eroğlu (2014) Tuay ve 18 City Teyyare ve Abant İzzet Özen, Güvenç Kumbaşlı (2016) Baysal, Bülent Altunoğlu (2007) Ecevit ve ve Öztornacı Osmangazi (2015) Çiçek (2006) İstanbul Başdağ (2017) Kilis 7 Aralık Karaca (2015) Cansız (2006) Afyon Çiçek ve Bitlisli Mehmet Akif Demir ve (2017) Ersoy Ciğerci (2016) Çelikkaya Eskişehir Gür ve Yıldız Bingöl   ve Gürbüz (2017) (2008) İpek ve Çanakkale Koban ve Bulu Gaziantep   Kaynar (2017) (2009) Altuğ, Çak, İstanbul Teyyare, Ayyıldız, Abant İzzet   Şeker ve Dirican, Zıvalı ve Baysal Bingöl (2010) Renkli (2018)

Zonguldak İzmir

Kütahya Afyon      

Studies in the table had limited questions directly related to tax literacy. However, the overall assessment of these studies showed that the tax knowledge of taxpayers in Turkey is low, and it is increased by education on taxes. For example, in the survey by Teyyare (2018), a questionnaire was applied to university students to measure the effectiveness of public financial education on levels

68

Güneş Çetin Gerger et al.

of tax literacy, and it was concluded that education contributes to increasing tax literacy levels. Additionally, Demir and Ciğerci (2016) conducted a questionnaire on primary school students in Afyon in Turkey, approximately 30 minutes of tax training was then provided with the help of videos about public service advertisements and similar visual media, and the same questionnaire was repeated. After the training, positive changes were observed in the tax awareness levels of the students, and it was found that education to be provided, especially at the primary school age, was significantly essential to increase tax awareness levels. There are no specialized tax courses on the primary, middle, and high school levels in Turkey. However, it is seen that the subject is included in the contents of other classes. The scope of some courses with the aim of raising excellent and responsible citizens such as “life sciences”, “human rights” and especially “social studies” includes “duty to pay taxes”. Especially with the changes made in the Social Studies Curriculum, tax issues were involved, though to a limited extent, in economics-related fields limited and in 2017, financial literacy issues were added. Thus, some course contents from the fourth grade to the seventh grade were made suitable for explaining tax i­ssues (Yılar & Akdağ, 2017: 370– 375). However, the fact that these courses are very comprehensive, containing many other subjects, the absence of a separate course, and the deficiency in the tax-related knowledge of teachers who teach this course, prevent the topic of taxes from being prepared sufficiently. Tax topics and activities in social studies textbooks are given in the table below. In 2007, the “Vergibilir  – Training Program for Developing Tax Awareness in Children” working protocol was signed between the Revenue Administration Tab. 5:  Tax Topics and Activities in Social Studies Education in Turkey. Source: (Yılar & Akdağ, 2017: 375). Publisher Ministry of National Education

Class Subjects 6 • Taxes I pay a return to me

Other

6

Activities • Taxes I pay returns to me (Preparing slogan)

• Recognize Invoices • Taxes for Better Turkey • Collecting Tax Evidence • Collecting Tax • Write a letter Evidence • You are safe with your tax (Preparing poster)

The Importance of Tax Literacy in Tax Compliance

69

and the Ministry of National Education to provide information about tax to children in the third, fourth, and fifth grades. The micro-teaching course was given to the teachers in charge, the training started in 2009, and the project was completed in 2012. In this context, it is seen that various books, brochures, and especially a website with games were established (GİB, 2008–2014 Annual Reports). Additionally, guidance and booklets in different areas prepared by the Revenue Administration for taxpayers have been significant in increasing tax literacy (http://www.gib.gov.tr/node/128637). Currently, the Revenue Administration has developed materials such as web-based videos, games, songs, and presentations to enable students to learn tax-related concepts on the primary and high school levels. Furthermore, textbooks were developed for teachers, and these contents were provided with open access. Until 2018, approximately five and a half million students were trained in this program (www.vergibilinci.gov.tr). The following examples may be given to demonstrate the importance of this training. In Turkey, Başoğlu and Sağbaş (2005), who conducted the first study on elementary school students, carried out a survey among students in Afyon and found that the students could not establish any relationship between taxes and public services. After training, Karaca (2015) conducted a study on elementary school students in Kütahya, which has a similar demographic structure. 52.4% of the participating students were educated; most of them perceived the connection between tax and public services. 59.5% of them stated that the training increased their habit of taking invoices, and 61% of them said continuous tax training should be provided.

4 Recommendations for Turkey to Improve Tax Literacy The development of academic literature in Turkey is required to increase tax literacy. However, these studies will be useful if they are carried out with broad participation in the empirical sense. Applications that have a high impact and can contribute to voluntary compliance must be implemented with the contribution of the Revenue Administration. Taxation concepts are included in social studies courses in primary education. Although it is essential to start tax-related education from primary education to increase tax literacy, in addition to this, families should transfer their knowledge to children positively about tax-related topics, and citizens should be informed at every stage to spread this literacy to the society. This situation, which is necessary for the development of a conceptual infrastructure, should be ensured for the target group between the ages of 12 and 80 as in the USA in pursuit of lifelong learning.

70

Güneş Çetin Gerger et al.

Tax literacy training may be carried out at a low cost with the contribution of students who are studying public finance or law at universities. These practices may be considered as an internship. This has the potential to be a practice that will both improve citizenship awareness and increase tax literacy. Questionnaires may determine tax compliance levels before and after these training, and the Revenue Administration may develop a strategy. Furthermore, tax literacy training within the scope of digital literacy may be planned under Industry 4.0. Designing long-term activities to improve tax literacy will develop tax compliance, reducing the need for the state to allocate resources to prevent tax evasion and undocumented income. Therefore, it would be useful to implement the following recommendations for improvements in tax literacy: – Academic studies should be increased to emphasize the importance of tax literacy. – A dataset in terms of tax literacy may be created in Turkey, and tax policies may be developed using the data obtained as a result of surveys to be applied periodically. – Since tax literacy is an issue that may be improved with the increase of financial literacy and tax knowledge, only studies carried out by the Revenue Administration remain on the administrative level. The Tax Awareness project is a good practice that can improve tax literacy. However, to increase the widespread impact of this study, quantitative data should be collected before and after training, and these data should be analyzed, and recommendations appropriate to the Turkish Tax System and tax culture should be developed. The readiness of the society and the reaction of it are decisive for the development of tax literacy. – In the organization of tax training, tax administration authorities, taxpayers, professionals, state officials, and academicians should develop a training program within the scope of a workshop, and education should be provided not only for children but also for certain specific target groups. – Taxpayers should be informed about tax guidelines and brochures, and their usage should be increased. Increasing the digital literacy of taxpayers will facilitate access to these brochures.

5 Conclusion In today’s world, where global realities are changing rapidly, protection of the tax revenues of nation-states is only possible by optimizing tax compliance. The main actor in tax compliance is the taxpayers. Education is one of the main elements that can ensure the compliance of the taxpayer. Many academic and

The Importance of Tax Literacy in Tax Compliance

71

administrative activities are organized to increase and improve tax literacy, which may include many concepts such as citizenship awareness, tax awareness, perceptions of the Revenue Administration, tax perception, tax morality, and tax awareness. In developed countries, there are more academic activities related to tax literacy. Studies have shown that there is a parallelism between tax literacy and tax compliance. Tax literacy-related activities in Turkey are still seen in the context of a more administrative level. For this reason, increasing academic studies aimed at raising tax awareness and increasing tax literacy and conducting them in cooperation with organizational research will positively affect tax compliance in Turkey.

References Alexander P., Balavac M., Mukherjee, S., Lymer A., & Massey D. (2018). “Improving Tax Literacy and Tax Morale of Young Adults”, BAFES Working Papers BAFES23, Department of Accounting, Finance & Economic, Bournemouth University, https://ideas.repec.org/p/bam/wpaper/bafes23. html, (18.07.2019). Alm J., Kirchler E., & Muehlbacher S. (2012). “Combining Psychology and Economics in the Analysis of Compliance: From Enforcement to Cooperation”, Economic Analysis and Policy, Elsevier, 42(2), 133–152. Altuğ, F., Çak, M. Şeker, M. & Bingöl, Ö. (2010). Türkiye’de Vergi Bilinci: İstanbul Araştırması, İSMMMO: 134, Istanbul. ANZ, ANZ Survey of Adult Financial Literacy in Australia, (2005). Available at www.financialliteracy.gov.au/media/465174/anz-survey-of-financialliteracy-in-australia -nov-2005.pdf, (28.05.2019). Başdağ, G.C. (2017). Vergi Bilinci ve Vergiye Bakış Açısı Bağlamında Gönüllü Vergi Uyumu: Kilis 7 Aralık Üniversitesi İİBF Öğrencileri Üzerine Bir Araştırma, Master Thesis, Gaziosmanpaşa SBE, Tokat. Bayraklı, H.H. Sağbaş, İ. & Ural, L. (2004). “Vergi Kaçırma Eğilimini Etkileyen Faktörlerin İncelenmesi: Uşak İli Örneği”, Afyon Kocatepe Üniversitesi İktisadi Ve İdari Bilimler Fakültesi Dergisi, VI(1), 197–214. Bornman, M., & Wassermann, M. (2018). “Tax Literacy in the Digital Economy”, https://www.business.unsw.edu.au/About-Site/Schools-Site/ Taxation-Business-Law-Site/Documents/Wassermann_Bornman_eJTR_ Tax-literacy-in-the-digital-economy.pdf (24.03.2019). Cansız, H. (2006). “Vergi Mükelleflerinin Vergiyi Algılama Hakkındaki Görüşleri: Afyonkarahisar İli Örneği”, Afyon Kocatepe Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi, 8(2), 115–138.

72

Güneş Çetin Gerger et al.

Cechovsky, N. (2018). The Importance of Tax Knowledge for Tax Compliance: A Study on the Tax Literacy of Vocational Business Students. In C. Nägele & B. E. Stalder (Eds.), Trends in Vocational Education and Training Research. Proceedings of the European Conference on Educational Research (ECER), Vocational Education and Training Network (VETNET) (pp. 113– 121), https://zenodo.cern.ch/record/1319646#.Xb7Ufy0lxhA, (03.03.2019). Chang Otto H., Nichols Donald R., & Schultz Joseph J. (1987). “Taxpayer Attitudes toward Tax Audit Risk”, Journal of Economic Psychology, 8(3), 299–309. Chardon, T., Freudenberg, B., & Brimble, M. (2016). “Tax Literacy in Australia: Not Knowing Your Deduction from Your Offset”, Australian Tax Forum, 31(2), 321–362. https://search.informit.com.au/documentSummary; dn=167713818520655;res=IELBUS (15.02.2019). Cvrlje D. (2015). “Tax Literacy as an Instrument of Combating and Overcoming Tax System Complexity, Low Tax Morale and Tax NonCompliance”, The Macrotheme Review 4(3), 156–167. Çelik, K. & Eroğlu, O. (2014). “İlköğretim Çağında Vergi Algısının İncelenmesi İdari: Zonguldak İli Örneği”, Çankırı Karatekin Üniversitesi İktisadi ve Bilimler Fakültesi Dergisi, 4(2), 301–326. Çelikkaya, A. & Gürbüz, H. (2008). “Mükelleflerin Vergiye Karşı Tutum ve Davranışlarını Etkileyen Çeşitli Değişkenler Arasındaki İlişkinin Analizi”, Sosyoekonomi Dergisi, 8(8), 23–54. Çetin Gerger, G. (2011). Mükellef Hakları ve Vergiye Gönüllü Uyum, 1. Baskı, Legal Kitabevi, Istanbul. Çiçek H. (2006). Psikolojik ve Sosyal Yönden Yükümlülerin Vergiye Karşı Tutum ve Tepkileri (İstanbul İli Anket Çalışması), Master Thesis, İstanbul Üniversitesi SBE, Istanbul. Çiçek, U. & Bitlisli F. (2017). “Vergi Eğitiminin Üniversite Öğrencilerinin Vergi Bilinci Ve Farkındalığı Üzerindeki Etkisi: Bir Araştırma”, Muhasebe ve Vergi Uygulamaları Dergisi, 10(2), 261–277. Çoban, H. & Sezgin, S. (2004). Vergi Mükelleflerinin Davranışları Üzerine Ampirik Bir İnceleme: Denizli Örneği, Türkiye’de Vergi Kayıp ve Kaçakları Önlenmesi Yolları. Türkiye 19. Türkiye Maliye Sempozyumu, Yaklaşım Yayınları, Ankara, 777–792. Demir, İ. C. (1999). Mükelleflerin Vergi Karşısındaki Davranışları: Afyon İli Anket Çalışması, Master Thesis, Afyon Kocatepe Üniversitesi SBE, Afyonkarahisar. Demir, İ. C. & Ciğerci, İ. (2016). “Vergi Bilincinin Oluşumunda Eğitimin Rolü: İlköğretim Öğrencileriyle Ampirik Bir Çalışma”, Yönetim ve Ekonomi

The Importance of Tax Literacy in Tax Compliance

73

Celal Bayar Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi, 23(1), 127–143. Eriksen, K., & Fallan, L. (1996). “Tax Knowledge and Attitudes towards Taxation: A Report on a Quasi-Experiment”, Journal of Economic Psychology, 17(3), 387–402. Gerçek, A., Çetin Gerger, G., Taşkın, Ç., Bakar, F. & Güzel S. (2015). Mükellef Hakları: Türkiye Perspektifi ve Geliştirilmesi, Seçkin Yayıncılık, Ankara. GİB, 2008–2014 Faaliyet Raporları, http://www.gib.gov.tr/kurumsal/stratejikyonetim/faaliyet-raporlari, (20.07.2019). GİB, www.vergibilinci.gov.tr (20.08.2019). Gülten, Y. (2014). Vergiye Uyum Sağlama Konusunda Yapılan Teşviklerin Vergi Algısı Üzerindeki Etkisi: Deneysel Bir Araştırma, Master Thesis, Afyon Kocatepe Üniversitesi SBE, Afyonkarahisar. Gür, E. & Yıldız, İ. (2017). “Vergi Algısı Ve Bilinci Üzerine Bir Araştırma: Sosyal Bilimler Meslek Yüksekokulu Öğrencilerinde Vergi Bilinci”, Route Educational and Social Science Journal, 4(5), 81–97. İpek, S. & Kaynar, H. (2009). “Vergiye Gönüllü Uyum Konusunda Çanakkale İline Yönelik Ampirik Bir Çalışma”, Yönetim ve Ekonomi, 16(1), 173–190. İzgi, K & Saruç, N. T. (2011). “Sosyokültürel Faktörlerin Vergi Ahlakı Üzerindeki Etkisi- Üniversite Öğrencileri ile Yapılan Anket Çalışması”, Ekonomi Bilimleri Dergisi, 3(2), 133–142. Karaca, R. (2015). Vergi Bilinci ve Vergi Algısı: Kütahya İlinde İlköğretim Öğrencileri Üzerine Bir Araştırma, Master Thesis, Dumlupınar Üniversitesi SBE, Kütahya. Karaot, A. (2010). İlköğretim Çağındaki Öğrencilerin Vergi Bilinç Düzeyi: İzmir İli Aliağa İlçesi Örneği, Master Thesis, Çanakkale Onsekiz Mart Üniversitesi SBE, Çanakkale. Kirchler, E., Hoelzl, E. & Wahl, I. (2008). “Enforced Versus Voluntary Tax Compliance: The ‘Slippery Slope’ Framework”, Journal of Economic Psychology, 29, 210–225. Koban, E. & Bulu, J. (2017). “Yükseköğretim Öğrencilerinin Vergi Farkındalığı ve Vergiye Uyuma Yönelik Yaklaşımlarının Değerlendirilmesi- Gaziantep Üniversitesi Özelinde”, Sosyal Bilimler Dergisi, 12, 118–139. Kornhauser, M. E. (2019). “TaxJazz: The Tax Literacy Project”, https://taxjazz. com/ (18.05.2019). Lai, M., Zalilawati, Y., Amran, M., & Choong, K. (2013). “Quest for Tax Education in Non Accounting Curriculum: A Malaysian Study”, Asian Social Science, 9 (2), 154–162.

74

Güneş Çetin Gerger et al.

Lusardi, A., and Mitchell, O.S. (2007). “Baby Boomer Retirement Security: The Roles of Planning, Financial Literacy, and Housing Wealth”, Journal of Monetary Economics, 54 (1), 205–224. Lusardi, A., and Mitchell, O.S. (2011). “Financial Literacy around the World: An Overview”, Journal of Pension Economics and Finance, 10 (4), 497–508. Lusardi, A., and Mitchell, O.S. (2014). “The Economic Importance of Financial Literacy: Theory and Evidence”, Journal of Economic Literature, 52 (1), 5–44. Lusardi, A., Mitchell, O.S., & Curto V. (2010). “Financial Literacy among the Young”, Journal of Consumer Affairs, 44 (2), 358–380. Lusardi, A. (2012). “Numeracy, Financial Literacy, and Financial DecisionMaking”, Numeracy, 5 (1), Art. 2, 1–12. Möller, K. (2007). Genetisches Lernen und Conceptual Change. In J. Kahlert, M. Fölling-Albers, M. Götz, & A. Hartinger (Eds.), Handbuch Didaktik des Sachunterrichts, (pp. 258–266). Bad Heilbrunn, Klinkhardt. Muter, B. N., Sakınç, S. & Çelebi, K., (1993). Mükelleflerin Vergi Karşısındaki Tutum ve Davranışları Araştırması, Celal Bayar Üniversitesi, İİBF, Maliye Bölümü, Manisa. Niemirowsaki, P., Baldwin, S. & Wearing, A. J. (2003). “Tax Related Behaviours, Beliefs, Attitudes and Values and Taxpayer Compliance in Australia”, Journal of Australian Taxation, 6(1), 132–165. OECD. (2015). The International and Ibero-American Foundation for Administration and Public Policies (FIIAPP), Building Tax Culture, Compliance and Citizenship: A Global Source Book on Taxpayer Education, OECD Publishing, Paris. OECD. (2017). G20/OECD INFE Report on Adult Financial Literacy in G20 Countries, http://www.oecd.org/daf/fin/financial-education/G20-OECDINFE-report-adult-financial-literacy-in-G20-countries.pdf (16.02.2019) Ömürbek, N., Çiçek, H. G. & Çiçek, S. (2007), “Vergi Bilinci Üzerine Bir İnceleme: Üniversite Öğrencileri Üzerinde Yapılan Anketin Bulguları”, Maliye Dergisi, 153, 102–122. Özen, A., Altunoğlu, B. K. & Öztornacı, E. (2015). “Orta Öğretim Düzeyindeki Öğrencilerin Vergi Algılama Düzeylerine İlişkin Ampirik Bir Değerlendirme”, Journal of Management & Economics, 22(2), 270–290. Russell, B. (2010). Revenue Administration: Developing a Taxpayer Compliance Program, Technical Notes and Manuals, IMF, Washington DC, https://www. imf.org/external/pubs/ft/tnm/2010/tnm1017.pdf (14.02.2019). Sağbaş, İ & Başoğlu, A. (2005). “İlköğretim Çağındaki Öğrencilerin Vergileri Algılaması: Afyonkarahisar İli Örneği”, Afyon Kocatepe Üniversitesi, İktisadi ve İdari Bilimler Fakültesi Dergisi, 7(2), 123–144.

The Importance of Tax Literacy in Tax Compliance

75

Sağlam, M. (2013). “Vergi Algısı ve Vergi Bilinci Üzerine Bir Araştırma: İktisadi ve İdari Bilimler Fakültesi Öğrencilerinde Vergi Algısı ve Vergi Bilinci”, Sosyoekonomi Dergisi, 1, 315–334. Şahin, A. (2010). Yüksek Öğrenim Öğrencilerinin Vergiyi Algılaması: Gaziosmanpaşa Üniversitesi Örneği, Master Thesis, Gaziosmanpaşa Üniversitesi SBE, Tokat. Taytak, M. (2010) “İlköğretim II. Kademe Öğrencilerinde Vergi Bilincinin Tespiti: Ampirik Bir Araştırma”, Maliye Dergisi, 158, 496–512. Teyyare, E. (2018). “Maliye Eğitiminin Vergi Okuryazarlığı Düzeyine Etkisi: Abant İzzet Baysal Üniversitesi Örneği”, Yönetim ve Ekonomi Araştırmaları Dergisi, 16 (4), 315–333. Teyyare, E. & Kumbaşlı, E. (2016). “Vergi Bilinci Ve Vergi Ahlakının Gelişmesinde Maliye Bölümü Eğitiminin Rolü”, AİBÜ Sosyal Bilimler Enstitüsü Dergisi, 16 (4), 1–29. Teyyare, E., Ayyıldız, B., Dirican, H., Zıvalı B.S. & Renkli, B. (2018), “İktisadi ve Mali Okuryazarlık Üzerine Bir Araştırma: Abant İzzet Baysal Üniversitesi İktisadi ve İdari Bilimler Fakültesi Örneği”, Bolu Abant İzzet Baysal Üniversitesi İİBF Ekonomik ve Sosyal Araştırmalar Dergisi, 14(1), 99–120. Tuay, E. & Güvenç, İ. (2007). Türkiye’de Mükelleflerin Vergiye Bakışı, Gelir İdaresi Başkanlığı, Yayın No:51. Yaltı B. (2018). Cross-Border Tax Challenges. In the 21st Century-21. Yüzyılda Sınır Ötesi Vergi Sorunları, 1.Baskı, Istanbul, Onikilevha Yayıncılık. Yardımcıoğlu, M., Akpınar, Y. & Günay, Y. (2014). “Vergi Okuryazarlığı ve Vergisel Farkındalık: Kahramanmaraş Araştırması”, Kahramanmaraş Sütçü İmam Üniversitesi İİBF Dergisi, 4(2), 95–122. Yılar M. B. & Akdağ H. (2017). “Sosyal Bilgiler Öğretiminde Vergi Okuryazarlığı”. In Turan R. & Akdağ H., (Eds.), Sosyal Bilgiler Eğitiminde Yeni Yaklaşımlar III, 1. Baskı, Pegem Akademi, Ankara, 356–388. Yüce, M. & Gerçek, A. (1998). Mükelleflerin Vergiye Yaklaşımı Açısından Türk Vergi Sisteminin Değerlendirilmesi, Bursa Ticaret ve Sanayi Odası, Bursa. Zorlu, Ö. (2012). İlköğretim Çağındaki Öğrencilerin Vergi Bilinci Düzeyi ve Vergi Bilinci Düzeyi İle İlgili Örnek Uygulama: Ankara İli Örneği, Master Thesis, Gazi Üniversitesi EBE, Ankara.

Filiz Keskin

The Concept of Collective Investment Institution and Specific Tax Advantages Provided for These Institutions and Their Investors in Turkey Abstract: The importance of collective investment institutions operating in the world under three legal structures, namely investment company, trust, and contractual model to enable investors with low savings to work in the financial markets, has gradually increased in the countries’ economies. In Turkey, according to regulations in Capital Markets Legislation, collective investment institutions operate in two legal structures, namely investment companies and investment funds. Advantages provided to these institutions and their investors in Turkish tax legislation are as follows: for investors; participation income exemption in Article 5 and deduction in venture capital fund in Article 10 of Corporate Income Tax Law; deduction in venture capital fund in Article 89 of Income Tax Law; non-declaration of incomes and %0 rate of withholding tax for income obtained from participation certificates and share of some funds and companies in Provisional Article of 67, for collective investment institutions; exception of portfolio management gains or corporate incomes in Article 5 and %0 rate of withholding tax for exempted gains in Article 15 of Corporate Income Tax Law; %0 rate of withholding tax for exempted gains and no withholding tax for some incomes in Provisional Article of 67 of Income Tax Law, exception of BITT for the money and capital market incomes of some institutions in Law on Taxes on Expenditure, exceptions of stamp tax for some papers in Stamp Tax Law. The relevant sections of the study and the conclusion section contain our suggestions and evaluation on the subject. Keywords: Collective, Investment, Institution, Fund, Company, Tax JEL Codes: G23, H20, K34

1 Introduction Collective investment institutions are organizations established to meet the needs of small savings owners and to have them in the economy. In other words, the purpose of collective investment institutions is to direct the savings of small savings holders to financial markets, who do not have the necessary knowledge to invest in financial markets, and hesitate to invest in these markets due to the risks, and do not have the opportunity to invest in various investment instruments due to their low savings. Therefore, the investment instruments in

78

Filiz Keskin

which big capital owners have the opportunity to invest also become available for small investors by minimizing the risks through professional institutions. Additionally, since trading in financial markets may bring risks, making these transactions by risk distribution principle through professionals can minimize the risks. Due to the benefits mentioned above, the shares and participation certificates of these institutions have become the most invested instruments in recent years. In order to encourage both these institutions and investors, some specific regulations can be seen in Turkish tax laws. The subject of this paper consists of the concept of collective investment institution and the general provisions of Capital Market Legislation for investment funds and companies, which are collective investment institutions operating in Turkey, explaining the regulations that provide tax advantages (exception, deduction, %0 tax rate, non-declaration) solely for collective investment institutions and their investors in Turkish tax legislation and evaluating the problems encountered in practice and giving suggestions about this issue.

2 Concept of Collective Investment Institutions (CII) With a general definition collective investment institution (CII) generally refers to incorporated companies or unincorporated undertakings that invest in financial assets (mainly marketable securities and bank deposits) and/or non-financial assets using the funds collected from investors by means of issuing shares/units (OECD, 2008: 192). The concept of collective investment institution is based on 3 basic principles: – Forming the portfolio with the capital collected from the public (collective capital), – Managing the portfolio according to the risk distribution principle1, – Managing the portfolio by experts (Nomer, 2013: 132). The first establishment in the world, similar to today’s CII, is The Foreign & Colonial Government Trust, founded in 1868 in the UK. The purpose of the fund in the certificate of formation was indicated as; “To provide the investor of moderate means the same advantage as the large capitals in diminishing risk… by spreading the investment over several stocks” (St. Giles, Alexeeva, Buxton, 2003: 14). 1 By the risk distribution, it is meant to minimize the investment risk by investing in businesses and markets operating in very different fields at the same time (Nomer, 2003: 5).

Collective Investment Institution and Tax Advantages

79

It is seen that collective investment institutions around the world can be e­ stablished under three different legal frameworks such as investment company2, trust3 and contractual structure (Hafeez, 2015: 152). Collective investment institutions established according to the partnership model are referred as “Yatırım Ortaklığı” in Turkish Law, “Investment Company” in American Law, whereas the ones with contractual structure are regarded as “yatırım fonları” in Turkish Law, “Unit Trust” in English Law and “Mutual Fund” in American Law (Yasaman, 1980: 2). Various legislative studies were carried out at the international level as the importance of collective investment institutions in financial markets increased and became one of the main instruments of investment. A directive on the collective investment models “Undertakings for Collective Investments in Transferable Securities – UCITS” issued by the European Community on 20.12.1985 can be given as an example4. The Directive has been updated and amended by various directives to this day, and it is now in force by the UCITS Directive 2009/65/EC of 13 July 20095. On the other hand, it is observed that studies on collective investment institutions are given importance in the eyes of OECD. For example, a study called “Governance System for Collective Investment Scheme in OECD Countries” was carried out in 2001 (OECD, 2001). Basic regulations relating to the definition and types of collective investment institutions in terms of legal aspects in Turkey can be found in Capital Markets Legislation. In article 3 of the Capital Markets Law no.  6362 titled “Abbreviations and Definitions”; “m) Collective investment institutions shall be defined as investment funds and investment companies.”

2 Investment companies developed mainly in British-American law and began to operate in the United States in 1889 and England in 1968 (Berzek, 1995: 7). 3 In a short description, the Trust is an arrangement enabling the property to be held by a person or persons (the trustees) for the benefit of some other person or persons (the beneficiaries) (Butler & Isaacs, 1997: 358; For more information on Trust, see Karayalçın, 1998: 650). 4 The full text of the directive can be found at http://europa.eu.int/comm/internal_ market/securities/ucits, (10.06.2019). Also See Sermaye Piyasası Kurulu, 1997. 5 For the full text of the Directive, see. https://eur-lex.europa.eu/legal-content/EN/ ALL/?uri=CELEX%3A32009L0065, (10.06.2019).

80

Filiz Keskin

In article 35/b of the same law titled “Capital market institutions,” “Collective investment institutions” are also counted among capital market institutions.

3 Tax Advantages Specific to the Collective Investment Institutions in Turkey When tax advantages specific to collective investment institutions in Turkey are analyzed in historical context, it is seen that in the tax legislation, regulations were made to support the capital markets especially after the period of 1980 and in private the investment funds and companies and tax advantages were provided according to the economic realities of the country and the market conditions (Aydın, 2000: 84). Tax advantages for these institutions in Turkish tax legislation will be explained under the following headings by the type of law.

3.1 Corporate Income Tax Law (CITL) 3.1.1  General Explanation In Article 2 of the Taxpayers Act of the Corporate Tax Law, it is stated that the funds subject to the regulation and supervision of the Capital Markets Board and foreign funds similar to these funds shall be considered as capital companies. According to the section titled “2.1.2 Funds” of the General Communiqué6 of CITL Serial No. 1 investment funds and pension investment funds are included among the main funds subject to the regulation and supervision of the Capital Markets Board. Additionally, according to the Communiqué; investment funds, provided in the bylaws, are named as bonds and bills fund, share fund, sector fund, participation fund, group fund, foreign securities fund, precious metal fund, gold fund, mixed fund, liquid fund, variable fund, index fund, special fund, free investment fund. Furthermore, venture capital investment funds and real estate investment funds shall also be considered as funds in the application of the CITL. On the other hand, investment companies established as joint stock companies are considered as CIT taxpayers due to the nature of their qualifications.

6 OG no. 26482, dated 03/04/2007.

Collective Investment Institution and Tax Advantages

81

3.1.2  Regulations in Article 5 of CITL: In accordance with article 5 of the CIT Law, “The dividends obtained from participation shares of full liable venture capital investment funds and stocks of full liable venture capital investment companies.”

are exempt from CIT. (CITL art. 5-1-a-3). Therefore, this exception is only for the income of participation share and dividends obtained from full-liable venture capital investment funds and companies, and the revenues received from other funds and companies are outside the scope of the exception. According to article 5 of CITL; – Incomes derived from portfolio management of securities investment funds or companies, – Incomes derived from portfolio management of investment funds or companies based on gold and precious metals traded on stock exchange whose portfolio is established in Turkey, – Incomes of venture capital investment funds and companies, – Incomes of real estate investment funds and companies, – incomes of pension investment funds

are exempt from CIT. (CITL art. 5-1-d) However, this exemption does not preclude the collection of withholding taxes from the incomes mentioned above of funds and companies. Additionally, if the aforementioned institutions have other incomes (such as management fees in securities investment funds, etc.) not qualifying for exemption, those incomes are subject to withholding tax. There are explanations in the section of “5.5. The exception regarding Income of Investment Funds and Companies located in Turkey” of General Communiqué7 of Corporate Income Tax Law concerning the application of these exceptions. Accordingly, for investment funds and companies based on gold and precious metals to qualify for the exception, portfolio structure of funds and companies that gain profits shall be taken into consideration and those that permanently invest at least %51 of their portfolio in; • gold and capital market instruments based on gold operating on the stock exchange in Turkey shall be considered as “gold fund or company.” • gold and precious metals, and capital market instruments based on these metals operating on the stock exchange in Turkey shall be considered as “precious metals fund or company.” 7 OG no. 26482, dated 03.04.2007.

82

Filiz Keskin

On the other hand, regulation of exceptions regarding real estate investment funds and companies mainly includes real estates, revenues of real estate investment funds and companies operating the portfolio that consists of the rights based on real estate and real estate projects. While creating the portfolios of these real estate investment funds or companies, the amount remaining from the ratio of obligation to invest in real estates, real estate projects and the rights based on real estate determined by CMB can be formed as deposit, participation ­account, repo, participation, and other rights and assets, and can benefit from the exception. Portfolios, including other rights and assets such as infrastructure investment and services, etc. cannot benefit from the exception. Even if these funds or companies are built according to capital markets legislation or have the titles of “Real Estate Investment Fund” or “Real Estate Investment Company”, they cannot benefit from the exception.

3.1.3  Regulation in Article 5/A of CITL In terms of foreign investment funds, tax advantages are provided in Article 5/A of CITL in case of certain conditions. Accordingly, the foreign funds referred to in the first paragraph of Article 2 of the CITL, due to their incomes obtained from transactions of: – – – – – – –

All kinds of securities and capital market instruments, Forward transaction and options contracts, Warrants, Foreign exchange, Forward transaction and options contracts based on commodity, Loan and similar financial assets and, Products in precious metal exchanges

whether traded on an organized market or not, through full-fledged taxpayer companies having the portfolio management authorization certificate given by the Capital Markets Board, those who manage portfolios, in case of fulfilling the following conditions, shall not be considered as permanent representative for the funds in question, their workplaces shall not be regarded as the workplace of these funds, no declaration shall be filed for these earnings and in the event that notification is filed for other earnings, these incomes shall not be included in the declaration: a) the transactions carried out on behalf of the fund must be among the regular activities of the portfolio management company.

Collective Investment Institution and Tax Advantages

83

b) the relationship between the foreign fund and company itself must be similar to the ones operating independently from each other when considering the commercial, legal, and financial characteristics of the portfolio management company. c) In return for the service provided by the portfolio management company, it is required to obtain the appropriate value for the peers, and the transfer pricing report must be submitted to the Revenue Administration within the period of the issuance of the corporate income tax declaration. ç) The portfolio management company and its related persons should not be entitled to more than 20% of the foreign fund’s gains directly or indirectly after deducting the estimated costs in return for the service they provide. The explanations on the applicable principles and procedures of the article are given in the General Communiqué Serial No. 78.

3.1.4  Regulation in Article 10 of CITL According to Article 10/g of the CITL, the portion (not exceeding %10 declared income) of the amount allocated as venture capital fund in Article 325/A of the Tax Procedure Law (TPL) can be deducted from the tax base. In article 325/A titled “Venture Capital Fund” of Tax Procedure Law (TPL) it is stated that;

– Venture capital funds can be allocated from the related period gains or the declared income to purchase venture capital investment fund shares or to capitalize venture capital investment companies established or to be established in Turkey subject to regulation and supervision of Capital Markets Board. – This fund cannot exceed 10% of the corporate income or the declared income, and 20% of the equity capital. – The amount allocated as a venture capital fund shall be kept in a temporary passive account. – If the investment is not made in venture capital investment companies or venture capital investment funds by the taxpayers until the end of the year, taxes that are not accrued in time shall be collected along with the default interest. – The transfer of this fund to any other account other than its purpose, withdrawing it from the business, distributing it to the partners, transferring it to the main center by the limited taxpayers or dissolving the work, the liquidation, transfer and division of the enterprise or in the event that it is not reused for the purpose specified in this article within six months following the disposal of the shares of venture capital

8 OG no. 28514 (4.rep.), dated 31.12.2012.

84

Filiz Keskin investment companies or venture capital investment fund participation shares, they shall be subject to taxation at the time of these transactions or when the term expires.

The explanations regarding the conditions and calculation method related to the application of deduction in question are included in the General Communiqué Serial no.7. Requirements for the deduction are as such; – The amount of funds allocated in the relevant year should not exceed 10% of the declared income, and the total amount of funds should not exceed 20% of the equity capital. (Two conditions must be fulfilled together.) – Investment should be made in venture capital investment funds or companies established in Turkey within the framework of regulation and supervision of the Capital Markets Board until the end of the year in which fund is allocated. – The amount of funds allocated should be included separately in the declaration of corporate income tax for the relevant year.

3.1.5  Regulation in Article 15 of CITL In article 15-(3) titled “Tax Deduction” of CITL, it is stated that; “barring gains of pension investment funds, a %15 withholding is made within the corporation from the gains (whether distributed or not) stated in subparagraph (d) of paragraph 1 of Article 5 of the Law.”

However, the withholding tax rate for investment funds and companies’ portfolio gains/corporate incomes mentioned above and exempted from ­ CITL is d ­ etermined as 0% with the Council of Ministers9 Decision (CMD) No. 2009/1459410.

3.2 Income Tax Law (ITL) 3.2.1  Regulation in Article 89 of ITL According to article 89/12 of ITL similar to the regulation in section 10/g of CITL, there is a provision that the venture capital fund in article 325A of TPL can also be deducted in terms of income tax.

9 OG no.27130, dated 03.02.2009 10 According to Provisional Article 67/8 of ITL, which is to be applied until 31.12.2020, incomes (whether distributed or not) of security investment funds (including exchange-traded funds) and companies established by CML that are exempt from CIT, are subject to %15 withholding tax. However, with CMD No. 2006/10731, this rate was determined as %0.

Collective Investment Institution and Tax Advantages

85

3.2.2  Regulations in Provisional Article 67 of ITL With the provisional Article 67 appendant to Income Tax Law to enter into force as of 01.01.2006 with Law No. 5281, it is aimed to simplify tax applications in terms of incomes obtained from money and capital market instruments. And in general preamble of the Law it is stated that with this application, in addition to simplicity, harmonization in taxation will also be provided for financial instruments (Sabuncu, Keskin, 2005: 227). On the other hand Provisional Article 67 of ITL whose validity is in force till 31.12.2020 requires taxation by means of withholding for security incomes, which are as follows:

– In paragraph 1, trading income of the stocks and bonds, bond interest, incomes of investment fund share and lending income, – In paragraph 2, excluding the payments to banks or intermediary institutions or to other real persons and legal entities through banks, security incomes (all kinds of bonds, treasury bill interests) stated in Article 75/5 of ITL, – In paragraph 3, banks and brokerage houses to acquire a security or other capital market instrument without being subject to withholding tax under paragraph (1), – In paragraph 4, security capital incomes stated in subparagraphs of 7 (deposit rates), 12 (for example, dividends paid to creditors not charging interest, and dividends paid for-profit and lost share certificates) and 14 (repo/reverse repo revenues)

are subject to %15 withholding tax. The tax advantages regarding collective investment institutions and their investors in the article will be explained under the following headings.

3.2.2.1 Tax Advantages Specific to Collective Investment Institutions In Provisional Article of 67/5 of ITL, it is stated that under the provisions of paragraphs (1)  and (4), no withholding tax will be made from incomes of exchange-traded funds and pension investment funds established according to CML. With CMD No. 2006/10731, withholding tax rate was determined as %0 for the incomes of some funds and companies stated in paragraphs (1), (2), (3) and (4) of Provisional Article 67. These are;

– exchange-traded funds – security investment funds and companies established by CML.

According to Provisional Article 67/8 of ITL, incomes (whether distributed or not) of security investment funds (including exchange-traded funds) and

86

Filiz Keskin

companies established following CML that are exempt from CIT, are subject to %15 withholding tax. However, with CMD No. 2006/10731, this rate was determined as %0.

3.2.2.2 Tax Advantages Specific to Investors Incomes from investment fund shares defined as capital market instruments following the Capital Markets Legislation can be classified as;

– Incomes arising from the return of the participation shares to the fund, – Incomes arising from the sale of the shares to the 3rd parties and – The periodical returns of the participation shares in the holding period.

While the incomes arising from the return of the participation shares to the fund and the sale to the 3rd parties are taxed through withholding under the ­provisional Article 67 (1) of ITL, it is not clear whether the periodical returns of these ­participation shares will be taxed in accordance with the provisional article 67/1 of ITL or accordance with the general provisions of the ITL regarding the taxation of dividends. In the tax ruling No. 32965 dated 10.04.2007 of Revenues Administration, it is stated that; “… According to paragraph 1 of provisional article 67 of Income Tax Law, withholding must be implemented on periodical payments made to investors within the scope of Protected Investment Funds committing with the framework of best effort to repay investor’s particular part or full amount of initial investment protected with Guaranteed Investment Fund in accordance with principles set forth in prospectus”.

Moreover, in the guide titled “Tax Applications in Investment Funds and Companies” prepared by Revenues Administration, yet not published, it is stated that periodical returns shall be subject to withholding tax under provisional article 67/1 of ITL. Therefore, periodical returns derived from investment fund participation certificates shall also be subject to withholding tax under provisional article 67/1 of ITL. On the other hand, types of income that can be obtained from investment companies come into question in two ways as dividend income and trading income. In this context, investment fund participation share incomes and trading incomes from investment company share are subject to withholding tax within the provisional article 67/1 of ITL11. 11 Share dividends of investment companies shall be taxed in accordance with the general provisions of the ITL and CITL (as they are not in the scope of Article 67 of the Income Tax Law).

Collective Investment Institution and Tax Advantages

87

After the amendment in Law12 No. 6009, however, the rate of withholding tax for fully and limited taxpayer companies’ incomes (also incomes of investment funds share and trading incomes of investment companies share) under the specified paragraph is determined as 0%. The withholding rate for full and limited taxpayer real persons’ incomes of investment funds shares and trading incomes of investment companies share is determined as %1013. The income derived from the disposal of the participation certificates of investment funds with at least 51% of their portfolio permanently consisting of shares traded on the BIST for more than one year is excluded from withholding tax (ITL Prov.67/1). In the section no.1.1.2. of General Communiqué of ITL Serial No. 258, it is stated that in order for income derived from the disposal of investment fund participation certificates to be excluded from withholding tax, during the 1 year period between the purchase and sale date of investment fund participation certificate, according to the fund by-law at least 51% of the fund participation certificate’s portfolio in question must permanently consist of shares traded on the Istanbul Stock Exchange. There is no regulation in the tax legislation regarding the content of the concept of permanency in the statement of “at least 51% of its portfolio permanently” included in aforementioned regulation. The stated rate is not the anticipated rate in Capital Market Legislation. Therefore, it is not clear that the 51% requirement will be sought every day. This issue needs to be clarified by legal regulation. Additionally, in the tax ruling no.  5515, dated 27.03.2017 of Revenues Administration; in the event that there is no statement in the fund by-law that at least 51% of its portfolio shall be composed of shares traded in BIST, it is concluded in the view that the actual provision of the condition in question will not lead to withholding tax exclusion of these incomes and this statement shall be included in the fund by-law. On the other hand, in accordance with CMD14 No. 2012/3141;

– Incomes of participation share of share intensive funds and – Incomes from the disposal of real estate and venture capital investment company  shares

1 2 OG no. 27659, dated 01.08.2010. 13 Per provisional article 67 of ITL, the types of funds within the withholding tax are not subject to declaration. 14 OG no. 28296, dated 18.05.2012.

88

Filiz Keskin

obtained by entirely and limited taxpayer real persons are subject to %0 withholding tax. It is seen that a similar statement stipulating the “permanency” criterion is included in the definition of the share intensive funds in the Capital Market Legislation. In Article 6/(2) of the Communiqué on Principles of Investment Funds, funds at least 80% of fund net asset value of which is permanently invested in issuer’s shares traded in BIAS, except for shares of securities investment companies, and units of which are issued under an umbrella fund as defined in subclause (2) of subparagraph (a) of the first paragraph of this Article are considered as “Share Intensive Funds”. In the same article, it is stated that “… Cash collaterals of futures contracts based on issuer’s shares and issuer’s share indices included in portfolios of share intensive funds, and premiums of option contracts based on issuer’s share, and covered warrants based on issuer’s shares traded in exchange are taken into account in calculation of 80%. Without prejudice to provisions of fifth paragraph of Article 24 of this Communiqué, in cases where a fund fails to meet the required conditions on daily basis for classification as a share intensive fund, for the relevant days, founder and portfolio custodian shall be jointly liable also for performance of all obligations of the fund, investors and/or institutions trading investment fund units.” If the limits specified in the information documents and this Communiqué are breached due to rights of option on newly issued shares or reasons beyond control of portfolio manager such as dividend distribution or price movements in value of portfolio assets, then it is required to re-establish compliance with the said limits within no later than 30 days. If it is impossible to re-establish compliance with the said limits within said period of time or it is determined that the re-establishment of compliance with the said limits shall lead to major losses, then this period may be extended by the Board. The Board may request liquidation or transformation of the funds which fail to apply to the Board by the end of this period of time or the funds which are not deemed appropriate by the Board for granting a time extension. Therefore, in case of remaining below the limit of 80% due to the rights of purchasing new shares or paying dividends and price movements of value of ­assets in the portfolio of share intensive funds that are beyond executive control, it is considered that tax exception may continue to be implemented in case of compliance with these limits within 30 days. The Turkish Council decided on State in a dispute over similar issues in the past years. In accordance with the legislation in force on the date of the case, because securities in portfolio of “the funds which shall be subject to %0 withholding tax

Collective Investment Institution and Tax Advantages

89

and whose at least 25% of the portfolio value invested permanently into shares of companies established in Turkey including State Economic Enterprises included in the scope of privatization according to the legislation” may decrease in value as a result of stock market fluctuations, it is stated that increasing the fund to its old rate requires a certain period of time, this period is determined as 10 working days in the Capital Market Legislation, and there is no regulation in the way that the 25% rate determined in the relevant articles of the income tax law and corporate income tax law is valid for each day of the year. Therefore, it is concluded that it is sufficient for the share rate in the portfolio to be above 25% on a monthly average basis15. On the other hand, in the opinion of the Revenue Administration No. 186 dated 14/08/2012; incomes of investors that purchased investment fund participation certificates prior to the amendment of the bylaw and disposed of the related participation certificates after the funds have been transformed into “share intensive funds” upon CMB approval, since investment funds were shared intensive funds on the date of the income shall be subject to %0 withholding tax according to article 1 of the Council of Ministers Decision No: 2012/3141.

3.3 Law on Taxes on Expenditure (BITT) In article 30 of the Law on Taxes on Expenditure No. 6802, taxpayers of BITT are considered as banks, bankers and insurance companies; in paragraph 2 of article 28, it is stated that; “The money that bankers received in cash or not in their own favor regardless of their titles due to their transaction and services (including the money received as commissions, fees, service over their benefit and income obtained from the money collected by those whose job is to collect money to charge deposit interest or to take advantage under other names and those who pledge to purchase and sell securities on account of themselves or others; and to act as mediator on sales and purchases or to pay off debts against the securities they purchase and sell) are subject to bank and insurance transactions tax (BITT).”

In paragraph 3, it is stipulated that those who carry out any of the transactions and services mentioned in paragraph 2 as the main field of business shall be considered as bankers in the implementation of this Law. Regulations on the rate of BITT are made in Article 33 of the Law. In tax rulings issued to date by Revenues Administration (for example; tax rulings no.23486 dated 05.06.1997, and no.  074581 dated 27.09.2006, and 15 The decision of the 4th Chamber of the Council of State dated 12.12.2000, numbered E. 2000/898 K. 2000/5234.

90

Filiz Keskin

no.150376 dated 31.05.2017), in the event that the investment funds and companies perform any of the transactions (such as buying and selling shares and acting as a mediator on purchase and sales) specified in paragraph 2 of article 28 of Law no.6802, they shall be considered as bankers and all kind of money received in favor of them as a result of these transactions shall be subject to BITT. However, under subparagraph 29/t of Law on Taxes on Expenditure; the money obtained by – – – – –

Pension investment funds, Securities investment funds, Securities investment companies, Venture capital investment funds and Venture capital investment companies

due to their transactions in money and capital markets shall be exempted from BITT. On the other hand, transactions of investment funds and companies which are outside of the scope of paragraph 2 of article 28 of Law No.6802 are subject to VAT16. Since the main field of business of real estate investment funds and companies is not trading securities or capital market instruments, these institutions are not subject to BITT17.

3.4 Stamp Tax Law (STL) The regulations related to stamp tax are included in the Stamp Tax Law no. 488. Whereas papers subject to stamp tax are counted in Table (1) attendant to the Law together with relative and lump sum stamp taxes, documents exempted from stamp tax are counted in Table II. The term “paper” which is a subject of stamp tax, refers to documents which are issued by means of being written and signed, and marked with a sign to replace the signature or which can be submitted to prove or indicate any matter and documents created as electronic data in magnetic medium by means of using electronic signature (STL art. 1). In accordance with article 3 of the Law, taxpayers of stamp tax are those who sign the papers. 16 In article 17/4-e of VATL, the transactions within the scope of banking and insurance transactions tax are exempted from VAT. 17 Real estate investment funds and companies are VAT taxpayers.

Collective Investment Institution and Tax Advantages

91

In section “IV-Commercial and Civil Affairs Paper” of Table II attendant to STL; “16. Papers issued concerning joint-stock companies, partnerships limited by shares and limited liability companies, and investment funds, share transfers, capital increases, and extensions of time.” “21. Trading contracts of real estate investment companies and real estate investment funds exclusively for real estate portfolios and preliminary contract for real estate sales.” “50. Contracts issued in relation to venture capital investments exclusively of venture capital investment companies and venture capital investment funds, and other papers issued regarding these contracts” are exempt from stamp tax.

Additionally, in section “V- Corporate Papers” of Table II attendant to STL; “21. Stamp tax of papers that are issued in all kinds of the transaction of insurance and retirement companies and pension investment fund companies, and these companies or funds should pay that” are exempt from stamp tax.

In the tax rulings regarding regulations of exception for real estate investment companies issued by Revenues Administration (tax ruling no.104570 dated 01.08.2016 and tax ruling no.580 dated 17.04.2016 of Revenues Administration), it is stated that;





– preliminary contracts for sale and sales contracts related to the immovables in the portfolio of the Real Estate Investment Company (REIC) can be exempted from the stamp tax, – stamp tax exception can be applied only for the part of the REIC share in preliminary contracts for real estate sales issued by sellers as joint venture including investment companies, – the revenue sharing contracts to be questioned regarding real estate (in the portfolio of REIC) sales shall not benefit from the stamp tax exception.

On the other hand, in the tax ruling no.  55872 dated 22.05.2015 of Revenues Administration, it is concluded that the “total stamp tax of the “KYD Indicies Agreement” aiming to have Indicies of …. Association to which Retirement company is party used in return of a specific price should be paid by the company that is a party to the agreement”. Therefore, solely the papers including pension investment funds of stamp tax taxpayers shall be excluded from stamp tax and in the event that the paper has another taxpayer, that taxpayer shall be obligated to pay total tax.

4 Conclusion When Turkish Tax Legislation is analyzed, in article 5 of CITL, income derived from venture capital investment funds and companies, and gains obtained from

92

Filiz Keskin

portfolio management of some investment funds and companies based in Turkey and corporate incomes of some of the others are exempt from CIT. In terms of foreign investment funds, tax advantages are also provided in the event that certain conditions in article 5/A of CITL exist. Moreover, in article 15/3 of CITL, the withholding rate for investment funds and companies’ gains exempted from CIT is determined as 0%. In Provisional Article 67 of ITL, tax advantages (such as exclusion from withholding tax, nondeclarative right, and 0% tax rate) are provided to corporate incomes of investment funds and companies and some incomes and revenues within the article some of their income and to investors of these institutions. By articles 89 of ITL and 10 of CITL, the amount allocated as venture capital funds in article 325/A of TPL can be deducted at certain rates. In Article 29/t of the Law on Taxes on Expenditure, the money obtained as a result of the transactions made in money and capital markets by pension investment funds, security investment funds and companies, and venture capital investment funds and companies are exempt from BITT. Exceptions regarding the papers related to various collective investment institutions are regulated in subparagraphs of 16,21 and 50 of Table II and subparagraph 21 of Table V attendant to Stamp Tax Law. In cases where the provisions providing tax advantages for CII are not sufficiently clear (such as, the portfolio structure of funds to be exempted from tax and the nature of the papers to be exempted), as the tax administration seems to have a narrow interpretation of their scope, the legislator needs to explicitly write these provisions. In withholding tax exemption, %0 tax rate or withholding tax exclusion applications, the statement “permanently” mentioned in the related regulations indicating that:

– To apply the exception to investment funds and companies based on gold and precious metal, at least %51 of their portfolio must permanently, – To apply withholding exemption to incomes derived from the disposal of participation certificates held more than a year, at least %51 of their portfolio must permanently, – To apply %0 withholding rate to incomes from participation share of share intensive funds, at least %80 of their portfolio must permanently,

consist of the assets mentioned above, and values should be clarified whether these rates are required every day or not. Also in subparagraph IV-50 of Table (II) attendant to STL, there are some hesitations in the application about the scope of the statement “Contracts issued in relation to venture capital investments exclusively of venture capital investment

Collective Investment Institution and Tax Advantages

93

companies and venture capital investment funds, and other papers issued regarding these contracts” in the provision of exception for Venture Capital Funds and Companies. Since it is seen in the application that there are provisions regarding revenue sharing in preliminary contracts for real estate sales and sales contracts regarding the portfolios of real estate investment funds and companies, the fact that the stamp tax exception within the scope of paragraph IV-21 of table (2) attendant to STL covers the revenue sharing transactions arranged together with the preliminary contracts for real estate sales and sales contracts is seen as an case to be evaluated. On the other hand, in accordance with Provisional Article of 67/5 of ITL, pension and exchange-traded funds and securities funds and companies’ incomes stated in paragraph 1,2,3 and 4 of the Law shall be subject to %0 withholding tax, but deposit and repo revenues of real estate and venture capital investment funds are subject to withholding tax. In this case, there is no difference between the investor opening a direct deposit account or buying a repurchase agreement and obtaining a real estate investment fund participation certificate in terms of the tax cost of deposit and repo accounts. Therefore, a regulation, which stipulates a withholding tax exclusion for the assets and the transaction in the portfolios of these funds, is thought to be advisable. Barring securities investment companies, a similar recommendation can also be made in terms of investors of investment companies. Again, especially in terms of real estate investment funds and companies, it is essential to make regulations for exceptions on VAT, title deeds, and real estate tax and thereby reduce the high tax and charging costs.

References Aydın, E. (2000). “Gelir Ve Kurumlar Vergisi Uygulamaları Açısından Yatırım Ortaklıkları ve Yatırım Fonları”, Vergi Dünyası Dergisi, Sayı 224, 84–99. Berzek, A.N. (1995). Yatırım Ortaklıkları, Vergi Sorunları Dergisi, Sayı 80. Butler, B. & Isaacs A. (1997),A Dictionary of Finance and Banking, Second Edition, Oxford Paperback Reference, Great Britain. Hafeez, M.M. (2015). Corporate Governance and Institutional Investment, Universal-Publishers, Florida. Karayalçın, Y. (1998). “İngiliz Hukukunda Trust ve Avrupa Hukuku”, Prof. Dr. Ali Bozer’e Armağan, Banka ve Ticaret Hukuku Araştırma Enstitüsü, Ankara. Nomer, F. (2003). Yatırım Ortaklıkları, Beta Yayınları, Istanbul.

94

Filiz Keskin

Nomer E.F. (2013). “6362 Sayılı Sermaye Piyasası Kanunu’nda Kolektif Yatırım Kuruluşları ve Özellikle Değişken Sermayeli Yatırım Ortaklığı”, İstanbul Üniversitesi Hukuk Fakültesi Mecmuası, Cilt 71, Sayı 2. OECD (2001). Governance System for Collective Investment Scheme in OECD Countries, John K. Thompson and Sang-Mok Choi, OECD Publishing, Paris, http://www.oecd.org/finance/financial-markets/1918211.pdf, (10.08.2019). OECD (2008). Benchmark Definition of Foreign Direct Investment, Fourth Edition, OECD Publishing, Paris, https://www.oecd.org/daf/inv/investments tatisticsandanalysis/40193734.pdf, (12.08.2019). Sabuncu & Keskin (2005). Gerçek Kişilerde Para ve Sermaye Piyasası Araçlarından Elde Edilen Gelirlerin Vergilendirilmesi, Beta Yayınları, Istanbul. Sermaye Piyasası Kurulu (1997). Menkul Kıymetler ve Borsalarla İlgili Avrupa Topluluğu Düzenlemeleri (Direktifler-Direktif Teklifleri-Ortak Tavırlar), Redaksiyon: Bahşayış Temir, Yayın No: 85. St. Giles & Alexeeva & Buxton (2003). Managing Collective Investment Funds, Second Edition, John Wiley & Sons Ltd., England. Yasaman, H.(1980). İsviçre ve Fransız Hukukunda Yatırım Fonları ve Türk Hukukunda Uygulanma İmkanları, Fakülteler Matbaası, Istanbul. www.spk.gov.tr www.gib.gov.tr http://europa.eu

Filiz Giray

The Impacts of Digitalization of Tax Administration on the Complexity of Tax System: OECD Countries Example Abstract: The traditional tax system may not support a new business model. The improvement of information and communication technology has changed the business model. The new business models are based on digital technologies and transactions. The traditional tax administration and system fail to tax on profits of digital companies. The digital tax paradigm will inevitably necessitate a change in countries’ tax systems. Today, the digitalization of business has played an essential role in the increased tax evasion and loss. Tax administrations have to go to digitization to adapt to these changing business structures. The digital tax administration can create an opportunity to raise tax-income without raising the tax burden. Also, the digitalization of tax administration affects the complexity of the tax system, which is a significant problem for many countries. The complicated tax system causes issues such as loss of tax revenues, injustice. The aim of this study investigates the impacts of the digitalization of tax administration on the complexity of the tax system with the indicators of some OECD countries. Keywords: Digitalization, Digitalization of Tax, Epistemic Tax Policy, Electronic Fiscal Device JEL Codes: H20, H26, H83, K34

1 Introduction The modern world is experiencing a digital age. The process of digitalization has been spread since the late 1990s. The data can be a more valuable asset than gold or oil in the current world. The digitalization is not only internet usage, but also it changes persons’ think the way (Naughton, 2010). The development of technology has disrupted the classical business model and created a new business type. These businesses are called platform business model or platform firms (Stallkamp and Schotter, 2019). The features of new business models can be listed as follows: (1) It is not necessary to create a company physically because of the digital products sold by these companies. This case invalidates classical national taxation based on local assessment. (2) The majority of these companies are multinational corporations to monopolize due to having a potential multiside platform, network, and scale economy effects and restrictions of use.

96

Filiz Giray

Although digital sectors still account for less than 10% of most economies based on the added value, income, and employment they created, this sector has been developing rapidly (IMF, 2018:1). For example, e-commerce in the Netherlands has increased as a share of total company revenue from 3.4% in 1999 to 14.1% in 2009. Similarly, between 2004 and 2011, this share increased from 2.7% to 18.5% in Norway and from 2.8% to 11% in Poland (OECD, 2015:56). The administration of tax, which was designed by business models before the revaluation of technology is far from insufficient needs. The development of e-commerce has made it difficult to reach taxable income. The OECD’s Shadow Economy Report for 2017 indicates that new technology and development in the digital economy will lead to informal economic activities (OECD, 2017). The digital economy creates opportunities for tax base erosion and profit shifting (Dover, 2016:47). New technology challenged both direct taxes and indirect taxes. The final product can be produced at the premises of the purchaser even if the design is made elsewhere, and the value created is determined (Hadzhieva, 2019:91). The tax system should follow these changes to reach the goals of tax authorities such as low tax evasion, tax erosion, more efficiently tax c­ ollection. These changes should include two parallel directions as the digitization of tax administration and tax policy changes. However, the digitalization of tax a­ dministration has not yet been fully realized in all countries. The effectiveness of the tax administration’s digitization depends on many factors. One of them is the epistemic tax policy. The performance of the digitalization of tax depends on epistemic tax policy, providing whether used tools in taxation have the test of reality (Campbell and Hanschitz, 2018:1). In the future, artificial intelligence will be used in the taxation area. Although digitalism has increased due to the aggressive tax planning of multinational corporations by moving their earnings to low-tax countries in particular (Hadzhieva, 2019:10), it is believed that digitalization in tax will be beneficial. Digitalization in tax administration will positively affect investments as a macroeconomic variable. Taxes are an important determinant of individuals’ entrepreneurial decisions. This is the most popular assumption, explained by Hundsdoerfer and Sichtmann’s irrational decide-making behavior (Hundsdoerfer and Sichtmann, 2008:19). Therefore digital tax system would positively affect entrepreneurial decisions. Digital taxation is an opportunity for countries that want to bring technology into their tax systems and to add more value to their businesses while it is a necessity for the new digital world (EY, 2017:26). In this study, it is expanded that the digitalization of the tax administration could also affect the complexity of the tax system, which is an important problem for many countries. Firstly, the digitization of the tax administration will be explained in

The Impacts of Digitalization of Tax Administration

97

this study. Then, the digital transformation of tax administration will be given with OECD country’s implementations. In the following chapter, it is assessed practices of digitalization in tax administration from the view of the complexity of the tax system. The findings will be evaluated in the conclusion.

2 The Digitalization of Tax Administration Although there is no generally agreed definition of the digital economy, the definition made by the International Monetary Fund (IMF) is the most comprehensive. The IMF has defined the digital economy in two ways: broad and narrow. The digital economy is defined as a narrowly online platform and liability activities on this platform. In broadly meaning, the digital economy can be defined as all activities that are using digitized data such as the internet into production processes and products, new forms of household and government consumption, and fixed-capital formation. But there is no compromise on the definitions yet (IMF, 2018). For this reason, Fortanier and Matei (2017) have been determined by three groups of digital transactions, rather than defining them. Fortanier and Matei (2017) determined three criteria for distinguishing digital transactions (Fig. 1): How the transaction includes the process of digitally ordered, enabled and delivered of trade, what the subject is goods, services and data, and who represents a corporation, household (customer) and government. Tax administrations can be seen as an essential component of e-government strategies (Marcus, Baron, 2018: 27). Digitalization in tax administration occurs with electronic fiscal devices. The use of electronic financial devices refers to a wide variety of technology instruments for the tax administration to help control business’ operations. In the modern global economy, the taxation of the profits of a business is more complicated and complex. There is not a link between where the value to be taxed and where taxes are paid. In other words, the source of income obtained can be uncertain (Fabregas, 2018:3). E-commerce can offer goods and services to many potential buyers by websites. Such a trade would raise various tax and government-level tax issues (Nellen, 2012:3). Digital taxation provides some benefits to both taxpayers and tax administration. In terms of the taxpayer: – Digitalization in taxation will not only enable individuals and companies to declare more easily and cheaper taxes, but will also allow unreported income to remain.

98

Filiz Giray

Nature (‘how’)

Digitally ordered and/or Platform enabled and/or Digitally delivered

Product (‘what’)

Actors (‘who’)

Goods

Corporations

Services

Households Government

Information/ data

Non-profit institutions serving households

Fig. 1:  The Dimensions of Digital Transactions. Source: Fortanier and Matei (2017): 10.

– Digital methods are less expensive than classical methods. For example, documenting certain transactions require written documents for taxpayers. The electronic record can substitute for the written record (Nellen, 2012:7). – Digitalization reduces the risk, which protects to company’s reputation (EY, 2017:18). – Individuals may not need expensive tax consultants (Campbell and Hanschitz, 2019:9). From the view of tax administration: – Digitalization may reduce tax evasion and tax evasion utilizing tools to collect tax information more efficiently and capture deficiencies (ICAEW, 2019:6). – It also reduces the time of data collection. – The digitalization of tax is more functional than the conventional tax process, especially in determining taxation’s data collection and tax liabilities. Innovative solutions can make the functions of tax administrations much more effective. Because it can provide more secure and full data and control of tax. The tax administration will be able to have structured and unstructured information about taxpayers from several various sources (Vuković, 2018; Volvach and Solovyev, 2018). The data can be able to give a picture of all networks of taxpayers who consist of stakeholders, consumers, and suppliers. This process can take place in a short time (about 45 minutes) (Gascon, 2018:15). – The use of electronic fiscal devices ensures better audit results with the same number or even fewer auditors (Casey and Castro, 2015:24).

The Impacts of Digitalization of Tax Administration

99

– Thanks to the online system, the exchange of financial data between the tax administration and companies will increase transparency in taxation, taxes will be collected more efficiently, and more tax fairness will be possible (Campbell and Hanschitz, 2019:2). – By analyzing the Data, the risk and the behavior, needs, and issues of the taxpayers’ behavior will be seen more clearly. – The human can behave differently. However, the operations in electronic devices are the same for everyone. This situation eliminates complaints that equal treatment is not taken. The tax administration’s gains on the result of these counted benefits will be (Strømme, 2018:50–51): – – – –

Increase in Efficiency. Increase Compliance. Provide Better Service. Equal Treatment.

All of the features require pressure on governments for digitalization. Digitalization is not easy for administration. It needs to be integrated into the business strategy. Holte emphasized digitization today as the lock for the tax administration to achieve the mission expected of him (Holte, 2018:9–10). Small - and medium-sized companies provide huge advantages from digitalization. Electronic applications would give opportunities to small - and medium-sized companies to reduce administration tasks in the calculation of their taxes. Notably, the online tax system means a less bureaucratic transaction for a self-employed person (Campbell and Hanschitz, 2019:2). The digitalization of the tax process is observed in five areas. On other words, the digitalization of the tax administration takes place in five steps: – E-filing: Filling the tax return with the standard electronic form. – E-accounting: Recording data in an electronic format by electronic invoices – E-matching: Cross-referencing with accounting, bank, and source data. – E-auditing: Electronic audit assessments. – E-assessment: Assessments by use of Blockchain technologies, etc. Most countries’ tax authorities began using e-filing tax returns to adopt digital technologies. E-accounting followed it. The most radical change is e-match. E-match involves matching with data to other sources such as banks in real-time.

100

Filiz Giray

Also, e-match analyzes across the taxpayers and jurisdiction to see hypothetical cases (EY, 2017:5). E-assessment is a significant step, too. Obtaining the data is not enough. The tax authority needs to analyze and assess them for purposes. Hwangbo (2004) called them e-tax technologies. It is stated that these technologies serve the implementation of a reliable, effective, transparent, and secure tax system. However, there are very few countries that implement these five steps (Hadzhieva, 2019:87). In many countries, still electronic tax statements, tax payments, and the integrated tax administration data system are optional (Casey and Castro, 2015:11). In order to solve problems in the process of digitalization, it is not only necessary to bring new taxes, but also digital conversion is required in the tax administration. The highest degree of digitalization is that tax administration use the submitted data to assess tax without tax forms (tax without tax forms (EY, 2017:5) If electronic invoicing is arrayed to meet the needs of all companies, it can decease administrative costs for business. Electronic invoicing should be mandatory for all enterprises in over the World. Another advantage of the electronic invoice is that it can improve the business process. The performance of the digitalization of tax administration depends on some factors. Technologies, people, managing tax risks, financial resources, and communication as five elements are required to digitalize tax administration (Vuković, 2018). In digitalization, technology is an essential element. If information technology experts only make the design of the digitalization of tax administration without taking account of tax authority and taxpayers, it leads to some mistakes (Vuković, 2018). However, human factors should not be ignored. Tax administration helps to taxpayers about their tax transactions. With digitization, the number of stakeholders in taxation has increased to three:  taxpayer, tax administration, and software. Software vendors are a new stakeholder in taxation. Software vendors take place in the market as supporters of taxpayers. Taxpayers want security software when submitting their information in data format (ICAEW, 2019:5). The issues of digitalization of taxation include data security, reliance on data, lack of nexus, taxable income determined, and expansion of e-commerce and new business model (Hadzhieva, 2019:13). Data security is one of most critical of tax digitalization applications (Campbell and Hanschitz, 2018:7). Additionally, the epistemic tax policy is needed to test the validity of tools and data used in taxation (Campbell and Hanschitz, 2019:2).

The Impacts of Digitalization of Tax Administration

101

Also, the digitalization of a country’s tax administration is not enough. It is necessary to use different digital systems and similar models with other countries in the process of globalization (EY, 2017:5).

3 The Implementation of Digital Transformation of Tax Administration in Some OECD Countries OECD developed the first “Base Erosion and Profit Shifting Action Plan (BEPS)” in 2015. Later, these plans were revised. BEPS Action Plans require to drive transparency on the part of taxpayers. OECD countries have signed some agreements over the years. For example, in 2016, only 84 of 111 OECD countries signed the Multilateral Competent Authority Agreement for Automatic Exchange of Information (EY, 2017:14). There are different views among 117 OECD members against BEBs measures. However, BEPS is a crucial instrument that draws attention to the issue of digital tax as an international tax issue (Hadzhieva, 2019:26). Tax administrations go digital. Although there is no consensus among OECD countries, the EU appears to have stepped up faster compared to OECD countries towards digital taxation. Governments in EU countries are under pressure from digital companies to create unfair tax pay. In the EU, the main aims of the digitalization of tax administration are to balance the tax obligations of member countries and to reduce the bureaucratic of the tax burden (Campbell and Hanschitz, 2019:5). Globally, 11 EU and some other OECD countries took unilateral measures. In direct taxes (Hadzhieva, 2019:39–40): - France’s New GAFA (Google Apple, Facebook, Amazon) and YouTube Taxes In 2018, a 2% tax was levied on advertising revenues by the resident, and nonresident platforms broadcast online videos such as YouTube, Netflix in France. This tax is called “Netflix tax”, which includes video given “for free” by web sites such as YouTube. However, some problems have been seen in the application of this tax. Because the persons to be taxed are not in France (Baron, 2018:5). The other charge, called GAFA, is levied on the income of the activities of global companies in Europe. This tax has been applied since 1 January 2019. – Italy’s Web Tax Italy has implemented a 3% web tax on internet services used by Italian resident and non-resident companies. This tax will be paid by the buyer of intangible goods, such as online advertising and sponsored links, but not to online retail.

102

Filiz Giray

– Austria’s Online Advertisement Tax In Australia, the digitalization of tax administration was begun by the digital transformation office within the government in 2013. Online advertising tax enacted in Austria in 2018. – Slovakia’s Intermediation Tax A tax on intermediation income from the website was introduced in Slovakia. – Belgium’s Fairness Tax Belgium has brought fairness tax at the rate of 5.5% over the distributed profits. This tax is different from corporate and income taxes. – Hungary’s Advertisement Tax Hungary introduced advertising tax in 2014 to tax advertising revenue for companies. However, the tax was revised in 2015 and 2016 because the European Commission does not comply with EU rules. Hungary amended it by reorganizing it in 2017 to harmonize the laws of the EU. In this respect, the threshold of the progressive tax rate to 7,5% was raised for advertising sales revenue of companies which have over Hungary forint 100  million advertising sales revenue. Although this tax is not directly for the digital sector, digital markets will be significantly affected. – UK’s Diverted Profits Tax The diverted profit tax aims to establish a link between the place where income originates and the business that provides income. This tax qualifies an upfront tax at 25%. Also, in 2018, it is proposed a withholding tax for the IP royalty payable by companies not established in the UK. This tax focuses on taxing the digital economy, since aiming at creating a nexus between consumer and user base rather than physical presence. – Australia’s Multinational Anti-Avoidance Law Australia introduced the multinational anti-avoidance law, which is similar to the structure of the UK’s Diverted Profits Tax. The taxpayer is a company that does not settle in Australia. – New Zealand’s Digital Services Tax Although New Zealand was reluctant to levy digital services tax, it has brought it as anti-tax avoidance measures with the effects of other countries.

The Impacts of Digitalization of Tax Administration

103

– India’s New Nexus and Equalization Levy In India, in 2018, a new income tax was legalized by the rate of 40% for foreign companies providing digital goods and services. – Turkey’s Withholding Tax on E-payments In Turkey in 2016, the monthly mandatory reporting obligation for transactions involving digital sales of service providers and advertising activities was enacted. Turkey has started to Apply 18% VAT for digital services offered by foreign companies abroad since January 2018. Some countries (the UK, US, China, Saudi Arabia, Kuwait, Israel, Taiwan, Turkey, Australia, and Japan) have introduced value-added tax as indirect taxes to capture the taxes of services provided via the internet (Hadzhieva, 2019:89). However, some criticism against unilateral measures has been put forward. These taxes carry some risks like negative impacts on investment, innovation, economic growth, the incidence of taxation on consumers (Hadzhieva, 2019:11). In addition, the unilateral measures can lead to the fragility of the single market (Fabregas, 2018:3). The unilateral actions should not be accepted in the international arena. In some countries, such as Italy and India, the equalization levy distorts the principle of ability to pay in taxation, because this may require tax payment from damaged businesses (Hadzhieva, 2019:43). Nexus-based Approaches include many issues. Latter, the European Commission proposed new legal rules for taxation of the activities of digital companies in 2018. The European Commission’s first proposal was the comprehensive solution, which includes changing the taxable nexus to account for the absence of physical presence. This proposal requires a long term. For this reason, an interim tax for certain income derived from digital 5,5% activities is proposed. Thus, the European Commission proposed “Digital  Services Tax” (DST) called an interim solution that harmonizes the  European Union. This tax is an indirect tax. Therefore, this tax will help certain European Union countries avoid unilateral taxes on digital activities. The DST includes revenues from i) Selling online advertising space, (ii) Digital intermediary actions, and (iii) The sale of data. The tax rate is 3% of gross revenue as a proxy of value created. The DST is a temporary implementation until the comprehensive solution (Fabregas, 2018:3). Lee and Hwangbo (1999) suggested a tax called the Consumer-Delivered Sales Tax for use in the electronic commerce order to prevent tax evasion and economic distortion. This tax is based on the direct payment of the consumer

104

Filiz Giray

without the intervention of the supplier. It is claimed that this tax is the best solution to a cyber-consumption taxation system to reach OEC’s seven criteria (Equitable, simple, confidence, effective, fairly, and adapting) (OECD, 1997) for the development of global electronic commerce. But the tax couldn’t solve the problem worldwide. Similarly, Hwangbo (2004) suggested a consumption tax system called the Global Electronic Tax Invoice System for cyber-taxation. Jin (2003), in his work, examines the effects of non-taxation of electronic commerce on state and local tax revenues. While electronic commerce is overgrowing, it is noted that the reflection of this on tax revenues is small.

4 The Assessment of Implementation of Digitalization in Tax Administration: The Complexity of the Tax System Tax is essential for both person and government. Tax is the most direct economic activity of connecting between citizens and the state (Sharman, 2012:18). Tax policy or tax principles should be changed to solve the problems of e-commerce related to taxation. For this, it is necessary first to determine the difference between tax policy and tax principles. Tanzi and Zee described tax policy in 2000 study for the IMF as designing a tax system to finance public spending most effectively and equally. The design of the tax includes its subject, rate, base, management, and compliance rules of the tax. Tax principles are tools that enable the creation of appropriate tax policies in a tax system. It is important to set principles of good tax policy. The problem is that applying current tax principles to e-commerce damages the tax policy of countries. While it is the work of some institutions (The American Institute of Certified Public Accountants and the National Conference of State Legislatures) to identify tax principles that can address this issue, the OECD’s study is the most general (Nellen, 2012:9). OCED’s five tax principles are as follows (OECD, 1998): – Neutrality: The tax system should be neutral in both commerce and e-commerce. – Efficiency: Administrative and compliance costs must be minimal. – Certainty and simplicity: Tax rules should be at ease, confidence, and simplicity. – Effectiveness and fairness: Taxes should be taken in amounts and fairly, which would not create tax losses. – Flexibility: The tax system must be flexible to keep pace with economic and technological changes. The simplicity in the tax system is one of the tax principles required for successful tax policy for countries. Tax complexity is an area that leads to a reaction against tax. The right tax statement is quite difficult in a complex tax system

The Impacts of Digitalization of Tax Administration

105

(Andreoni et al., 1998:852). The complexity of the tax system in many countries requires taxpayers to rely more on the tax administration and/or seek help from consultants in meeting their tax obligation (Casey and Castro, 2015:10). Countries have a more complex tax system that spends more time, waste, and effort on tax compliance, which leads to tax losses. One assessment of the application of the digitalization of tax administration is the degree of complexity in the tax system. Tax administration has several critical opportunities for digital conversion. These are centralization, data, and automation. Centralization will facilitate tax compliance, improve quality, greater transparency, and lower cost. Advanced data provides to see gaps in available data. Automation creates the development of tax processing, planning, and reporting. For this reason, tax administrations around the world are rapidly developing digitalization in the tax system. According to “The Financial Complexity-Index 2018 which was found by the “Netherlands-headquartered TMF group for 94 jurisdictions worldwide (Europe, the Middle East and Africa (EMEA) (50), the Americas (25) and Asia Pacific (APAC) (19)). They used the survey to measure four complexity parameters:  Compliance (Company representative, cross border transaction, data storage), Tax (Tax registration, type of taxes and compliance requirement), Reporting (Reporting process), and Bookkeeping (Accounting regulations, authorities, and technology). Global complexity rates for 2018 are as follows by reporting 57% (2017:  55%), tax 49% (2017:  48%) and bookkeeping 46% (2017: 51%) (The Financial Complexity-Index 2018, 2018:5). The top sixty-one countries in the index ranking showed in Tab. 1. Turkey ranged the most complex in 94 countries for accounting and tax compliance in 2017. In the same year, the Cayman Islands took place as the least complex. According to the Financial Complexity-Index 2017, most complex jurisdiction area focuses on Europe, the Middle East, and Africa. The main reasons for the increasing complexity in Turkey are as follows: Firstly, Turkey’s tax code is changed frequently. This case prevents both not easy follow-up and increases complexity. Secondly, an attempt to harmonize Turkish tax legislation was insufficient. Similarly, although taxes have been deceased in Italy, the fact that many specific requirements have risen the complexity of account and tax. In 2018, the index ranking was changed (Tab. 1). China became the most complex in the world for accounting and tax compliance. Also, the Cayman Islands are the least complex. Brazil is among countries that have the most complex for accounting and tax. An essential reason for Brazil is digital transactions. Especially, e-social (social

106

Filiz Giray

Tab. 1:  Top 10 Most Complex Jurisdictions for Accounting and Tax Compliance (2017–2018). Source: The Financial Complexity-Index, 2017, 2018. Global Ranking (2017) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

Jurisdiction Turkey Brazil Italy Greece Vietnam Colombia China Belgium Argentina India Russia France Bolivia Albania Kazakhstan Mexico Belarus Israel Spain Pakistan Croatia Austria Ecuador Honduras Luxembourg Philippines Uruguay Thailand El Salvador Guatemala Venezuela Angola Malta Chile Paraguay Latvia

Global Ranking (2018)  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

Jurisdiction China Brazil Turkey Italy Argentina France Bolivia Colombia Mexico Russia Vietnam Croatia India Albania Belarus Philippines Romania Venezuela Ukraine Belgium Germany Greece Slovakia Israel Guatemala Kazakhstan Luxembourg Ecuador Uruguay Portugal Thailand El Salvador Paraguay Spain Chile

The Impacts of Digitalization of Tax Administration

107

Tab. 1: (continued) Global Ranking (2017) 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61

Jurisdiction Indonesia Azerbaıjan Egypt Portugal Malta Dominican Republic Finland Montenegro Panama Jamaica Mauritius Poland Taiwan Canada South Korea Nicaragua Peru Russia Romania The Netherlands Costa Rica Lebanon Ireland Malaysia Czech Republic Estonia

Global Ranking (2018) 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61

Jurisdiction South Panama Peru Slovenia Bosnia Dominican Republic Egypt Honduras Lithuania Moldova Indonesıa Serbia Hungary Latvia Nicaragua Poland Cyprus Jamaica Algeria Australia Cambodıa Lebanon Estonia Austria Sweden Czech Republic

security and labor obligations) require increase data shared with authorities for companies, which causes some problems and complexity. Besides, another reason is that taxes on income transfers to foreign banks in Brazil have been increased. According to the 2018 Financial Complexity Index, Turkey ranked third. Some regulations provide simplification. For example, e-voice in export operations accelerates the customs process by reducing the administration burden on taxes and customs offices. Also, e-notification provided to companies by tax authorities has increased communication and effectiveness between parties.

108

Filiz Giray

Tab. 2:  Complexity Issues as Globally and Regional (2017, %). Source: The Financial Complexity-Index 2017: 26. Complexity Risk of non-compliance with local regulation Tax compliance Future impact of technology BEPS and transfer pricing Accounting complexity Cyber security/data privacy Other

GLOBAL 94 19

EMEA 50 14

AMERICAS 24 APAC 20 25 24

17 17 16 11 13  7

13 18 20  9 16 10

21 18 14  7 11  4

20 14 12 16  9  5

However, from Europe, the Middle East, and Africa, Turkey was still the most complex country in 2018. The factors affecting the complexity of accounting and tax compliance of TMF group are listed as follows: Risk of non-compliance with local regulation, Tax compliance (possibility of tax audits), Future impact of technology, Base Erosion and Profit Shifting (BEPS) and transfer pricing, Cyber security/data privacy. Tab. 2 indicates the weight of the factors by region with a global comparison rate. Globally, the first three factors impacting complexity were ranked ‘risk of non-compliance with local regulation’, ‘tax compliance’, and ‘the ‘future impact of technology’. This situation varies by region. The technology is an important factor in all areas except APAC. The digitization of the tax system will have a positive effect on three areas. For example, Estonia is a country that records significant distance in digitalization. Estonia has advanced digital tax administration. In Estonia, 29% of companies are online, 99% of bank transactions, and 99% of tax returns are filed online (Campbell and Hanschitz, 2019:4). Also, Estonia is among the lowest in OECD countries from the point of the cost of collecting taxes due to deliver the saving through digitalization (Laid, 2018:8). Parallel to this, Estonia, in the 2018 Financial Complexity Index, ranked 58. With this score, Estonia is the country with the least complex tax system.

5 Conclusion The change in information and communication techniques has changed the business forms and structures. E-commerce can offer goods and services to many potential buyers by websites. Such a trade would raise various tax and

The Impacts of Digitalization of Tax Administration

109

government-level tax issues. Digitalization is fundamentally changing the way businesses and governments interact. Digitalization is fundamentally changing the way businesses and governments interact. Digital tools provide significant opportunities through faster, easy, and low cost for businesses and consumers. The digital sector has been developing rapidly over the World. However, the growth of this sector brings with it several problems. Taxation is one of the areas where these problems occur. The digital economy creates opportunities for tax base erosion and profit shifting. Traditional tax administration has been unable to tax changing business gains. At the same time, the current tax principles which are used in a country occurred for tax administration without the internet. As a result of this situation, tax authorities face to loses tax revenues. As a solution, some authors have proposed separate taxes for e-commerce. The EU takes faster steps due to pressures in digital taxation than OECD. The EU has taken unilateral measures. In the unilateral actions, some taxes brought by member countries in solving the tax evasion and tax loss problems brought by digitalization have not been resolved. These measures have been even criticized in several respects. For example, these types of taxes negatively affect investment, innovation, economic growth, the incidence of taxation on consumers, and lead to the fragility of the single market. Whereas, digitalization creates an opportunity to tax authorities in collecting the right tax at the right time. The tax administration should be transferred to digitalization. A good tax system depends mainly on tax principles. One of the tax principles set out by the OECD for both traditional, and e-commerce is the simplicity of the tax system. This study searches the impacts of the digitalization of tax administration on the complexity of the tax system with the indicators of some OECD countries. Theoretically, by implementing digital services, the tax administration will lose Complexity. According to “The Financial Complexity-Index for 94 jurisdictions worldwide, Turkey ranged the most complex in 2017. China became the most complex in the world for accounting and tax compliance in 2018. The factors affecting the complexity of accounting and tax compliance of TMF group are listed as follows: Risk of non-compliance with local regulation, Tax compliance (possibility of tax audits), Future impact of technology, Base Erosion and Profit Shifting (BEPS) and transfer pricing, Cyber security/data privacy. According to the survey, technology is seen as an essential factor for all regions except the Asia Pacific. This result shows that the digitization of tax administration provides a positive effect. The stages of digitalization of the tax administration include e-filing, e-accounting, e-matching, e-auditing, and e-assessment. Very few countries

110

Filiz Giray

have completed these stages. Countries with complex tax system problems, such as Turkey, have to use new technologies in tax administration and adapt to the new economy and business conditions. In addition, the digitalization of tax administration should be parallel among countries.

References Andreoni, J, B. Erard, and J. S. Feinstein (1998). ‘Tax Compliance’, Journal of Economic Literature, 36(2), 818–860. Baron, A. L. (2018). ‘How to Tax Digital Business – Countries Experiences’, Lisbon Tax Submit, Tax Administrations and the Challenges of the Digital World, Summary Report, https://www.ciat.org/Biblioteca/ ConferenciasTecnicas/2018/2018_Summary_Report_Portugal.pdf, (18.06.2019). Campbell, D. F. J and G. Hanschitz (2018). ‘Digitalization of Tax: Epistemic Tax Policy’, Handbook of Cyber-Development, Cyber-Democracy, and CyberDefense, E.G. Carayannis et al. (eds.), Springer International Publishing AG, Cham, Switzerland. Campbell, D. F. J and G. Hanschitz (2019). ‘The Innovation of Tax: Epistemic Tax Policy and Online Tax Accounts (Artificial-Intelligence-Based Tax Accounts)’, Encyclopedia of Creativity, Invention, Innovation and Entrepreneurship, E. G. Carayannis (ed.), Springer-Verlag, New York, 1–5. Casey, P. and P. Castro (2015). ‘Electronic Fiscal Devices (EFDs). An Empirical Study of Their Impact on Taxpayer Compliance and Administrative Efficiency’, IMF Working Paper, WP/15/73, file:///C:/Users/Sony/ Downloads/_wp1573.pdf, (14.06.2019). Dover, R. (2016). ‘Fixing Financial Plumbing: Tax, Leaks and Base Erosion and Profit Shifting in Europe’, The International Spectator, 51(4), 40–50. EY (2017). Tax Technology and Transformation Tax Functions ‘Go Digital’, Ernst & Young LLP India, New Delhi, https://assets.ey.com/content/dam/ ey-sites/ey-com/en_gl/topics/digital/ey-tax-technology-transformation.pdf, (22.06.2019). Fabregas, M. T. (2018). ‘Fair Taxation of the Digital Economy’, Lisbon Tax Submit, Tax Administrations and the Challenges of the Digital World, Summary Report, https://www.ciat.org/Biblioteca/ ConferenciasTecnicas/2018/2018_Summary_Report_Portugal.pdf, (18.06.2019). Fortanier, F. and S. Matei (2017). Measuring Digital Trade: Results of the OECDIMF Stocktaking Survey. Presented at 30th Meeting of the IMF Committee on Balance of Payments Statistics, October, 24–26, Paris.

The Impacts of Digitalization of Tax Administration

111

Gascon, J. (2018). ‘Advanced Analytics for Compliance Control’, Lisbon Tax Submit, Tax Administrations and the Challenges of the Digital World, Summary Report, https://www.ciat.org/Biblioteca/ ConferenciasTecnicas/2018/2018_Summary_Report_Portugal.pdf, (18.06.2019). Hadzhieva, E. (2019). Impact of Digitalisation on International Tax Matters Challenges and Remedies, European Parliament, Luxembourg, https:// www.europarl.europa.eu/cmsdata/161104/ST%20Impact%20of%20 Digitalisation%20publication.pdf, (23.06.2019). Holte, H.-C. (2018). ‘Making Tax Administration Digital’, Lisbon Tax Submit, Tax Administrations and the Challenges of the Digital World, Summary Report, https://www.ciat.org/Biblioteca/ConferenciasTecnicas/2018/2018_ Summary_Report_Portugal.pdf, (18.06.2019). Hundsdoerfer, J. and C. Sichtmann (2008). ‘The Importance of Taxes in Entrepreneurial Decisions: An Analysis of Practicing Physicians’ Behavior’, Review of Managerial Science, 3(1), 19–40. Hwangbo, Y. (2004). ‘Establishing a Trusted Third Party for Taxing Global Electronic Commerce: System Architecture of Global Electronic Tax Invoice (GETI)’, International Review of Public Administration, 9(1), 33–40. Jin, D.Y. (2003). ‘E-Tax or E-Commerce: The Debate on Taxing Electronic Commerce Transactions’, Journal of Internet Commerce, 2(1), 65–87. Laid, V. (2018). ‘Think Digitally, Act Meaningfully – Digital Transformation in Estonia’, Impact of Digitalisation on the Transformation of Tax Administrations, Budapest: Iota, 8–9. Lee, J. K. and Yeoul Hwangbo (1999). ‘Cyberconsumption Taxes and Electronic Collection Systems: A Canonical Consumer-Delivered Sales Tax’, International Journal of Electronic Commerce, 4(2), 61–82. ICAEW (2019). Digitalisation of Tax: International Perspectives, https://www. icaew.com/-/media/corporate/files/technical/information-technology/ thought-leadership/digital-tax.ashx, (28.06.2019). IMF (2018). Measuring the Digital Economy, https://www.imf.org/~/media/ Files/Publications/PP/2018/022818MeasuringDigitalEconomy.ashx, (11.06.2019). Marcus, E., A. L. Baron (2018). ‘The French Tax Administration (Dgfip) at the Heart of the Government’s Digital Agenda’, Impact of Digitalisation on the Transformation of Tax Administrations, Budapest: Iota, 26–27. Naughton, J. (2010). ‘The Internet: Is It Changing the Way We Think?’, The Guardian, 15 August 2010, https://www.theguardian.com/technology/2010/ aug/15/internet-brain-neuroscience-debate, (15.07.2019).

112

Filiz Giray

Nellen, A. (2012). ‘Internet Taxation and Principles of Good Tax Policy’, Policy and Internet, 4 (1), 1–21. OECD (1997). Electronic Commerce: The Challenges to Tax Authorities and Taxpayers. Turku, Finland, 5–22. OECD (1998). Electronic Commerce: Taxation Framework Conditions. http:// www.oecd.org/dataoecd/46/3/1923256.pdf, (08 July 2019). OECD (2015). Addressing the Tax Challenges of the Digital Economy, Action 1 2015 Final Report, Paris, https://www.oecd.org/tax/beps/policy-notebeps-inclusive-framework-addressing-tax-challenges-digitalisation.pdf, (13 July 2019). OECD (2017). OECD Report (2017) Shining Light on the Shadow Economy: Opportunities and Threats. https://www.oecd.org/tax/ administration/shining-light-on-the-shadow-economy-opportunities-andthreats.htm, (23 June 2019). Sharman, J.C. (February 2012). ‘Seeing Like the OECD on Tax’, New Political Economy, 17(1), 17–33. Stallkamp, M. and A. P. J. Schotter (2019). ‘Platforms without Borders? The International Strategies of Digital Platform Firms’, Global Strategy Journal, 9 (1), 1–23. Strømme, Ø. (2018). ‘Increased Compliance and Efficiency with Machine Learning’, Impact of Digitalisation on the Transformation of Tax Administrations, Budapest: Iota, 50–51. Tanzi, V. and H. H. Zee (2000). Tax Policy for Emerging Markets: Developing Countries, IMF Working Paper WP/00/35, https://www.imf.org/external/ pubs/ft/wp/2000/wp0035.pdf, (02.07.2019). The Financial Complexity-Index 2017, TMF Group. https://www.tmf-group. com/en/news-insights/publications/2017/financial-complexity-index-2017/, (15 July 2019). The Financial Complexity-Index 2018, TMF Group. https://www.tmf-group. com/en/news-insights/publications/2017/financial-complexity-index-2017/, (15 July 2019). Volvach, D. and M. Solovyev (2018). ‘Tax Administration in the Digital Era: The FTS of Russia Approach’, Impact of Digitalisation on the Transformation of Tax Administrations, Budapest: Iota, 13–15. Vuković, M. (2018). ‘Towards the Digitization of Tax Administration’. https:// www.cef-see.org/files/Digitization_Tax_Administration.pdf, (30 May 2019).

Pelin Varol İyidoğan, Eda Balıkçıoğlu, and H. Hakan Yılmaz

Empirical Findings on Macro Determinants of Pharmaceutical Spending in Selected OECD Countries Abstract: Pharmaceutical expenditure as a major determinant of health care spending arises as an attractive area of research in terms of public policy. Policymakers inevitably face the control of pharmaceutical expenditure, which corresponds to more than a %16 of health expenditure on average to optimize health policy. In this context, besides the historical pattern of pharmaceutical expenditure, the investigation of macro determinants of pharmaceuticals has come into prominence in the health care literature. Within this scope, we aim to examine the effect of aging, chronic diseases, health care expenditures and social spending on pharmaceutical spending for 22 OECD countries by employing General Method of Moments (GMM) procedure of Arellano and Bond (1991) which utilizes the difference of dependent variable to eliminate the individual fixed effects. In our paper, we conclude that the rise in the elderly population leads to an increase in pharmaceutical spending, which is consistent with our expectations.On the other hand, we find no significant effect of male cancer incidence and the negative impact of adjusted health spending on pharmaceutical spending. However, the results point out that a rise in female cancer incidence rises drug expenditures. More clearly, pharmaceutical expenditure is exposed to gender sensitivity in terms of cancer. Furthermore, we find that social spending has a positive effect on pharmaceutical spending. Keywords: pharmaceutical spending, dynamic panel, health, the general method of moments (GMM) JEL Codes: C33, I18

1 Introduction Pharmaceutical expenditure has a significant share of overall health care spending across OECD countries, which corresponds to more than a %16 of health expenditure on average. Besides the retail pharmaceutical spending, when hospital use is also taken into account, on average, 20 percent of budget canalizes to pharmaceutical demand. Moreover, similar to other health care functions, the cost of pharmaceuticals is predominantly covered by government financing or compulsory insurance schemes, which about % 56 of total health is spending in the OECD average (Graph 1). In this regard, pharmaceutical spending contributes

114

Pelin Varol İyidoğanet al.

%

Government and compulsory schemes

Voluntary health insurance

Out-of-pocket

Other

100 16 13

80

7

25 27 1

60

40

20

17 29 31 31 33 34 38 41 41 42 39 44 40 44 12 48 47 48 49 45

1

1 4

84 80

5

24

29 33 51 51

1

65 66

58

26 1

4 6

75 72 71 71 69 68 67 65 62 59 59 58 57 55 55 55 52 52 51 51 51 50 48 44

34 31

38 36 36 35 34

0

Graph 1:  Expenditure on Retail Pharmaceuticals by Type of Financing (2017). Source: OECD (2017).

notably to determine public health spending dynamics, which is a significant component of fiscal balance (OECD 2017). As pointed out by OECD (2017), during the last decades, pharmaceutical spending has upsurged sharply as a consequence of the growth in the consumption of recently developed drugs. This case has led to a rise in the share of health spending from %7 in the 1980s to more than %9 in the early 2000s. However, since the mid-2000s, this pattern of pharmaceutical spending has generally slowed compared with other subsectors of health care expenditure, such as in the hospital and outpatient sectors. The global financial and economic crisis, which induced fiscal policy implementations based on reductions in government expenditure and the introduction of cost-containment policies, coinciding with patent losses of several top-selling drugs led to a decline in pharmaceutical spending. While the extent of the slowdown varies widely across OECD countries, nearly all have seen a reduction in pharmaceutical spending growth since the onset of the crisis, and several European countries have noticed more dramatic reductions. On the other hand, recently, the consumption of pharmaceuticals has increased substantially concerning rising in the elderly population, the growing prevalence of chronic diseases, and changes in clinical practices (Belloni et  al., 2016). In 2017, as presented in Graph 2, pharmaceutical expenditures have reached to almost %30 for several OECD countries.

Findings on Macro Determinants of Pharmaceutical Spending

35,00

115

Pharmaceuticals

30,00 25,00

20,00 15,00 10,00 5,00 0,00

Graph 2:  OECD Pharmaceutical Expenditures – 2017 (%of Total Health Spending). Source: OECD Health Statistics (2017).

To sum up the framework of OECD (2017), the main drivers of pharmaceutical expenditures can be assumed to be an aging population, growth performance of the economy, indicators of chronic diseases, and fiscal policy instruments. In our model, we estimate a model embodying those macro determinants to examine the dynamics of pharmaceutical expenditures across OECD countries by utilizing panel estimation techniques. In the following years, the development of health technology, finding new diseases, especially driven by the aging and health pricing policy, will be useful in pharmaceutical expenditures. Our paper is organized as follows. In the next section, we give a brief overview of the literature regarding pharmaceutical expenditure. We explain the data and model as follows. Finally, we present empirical findings and conclude.

2 Literature While research on health expenditures has a wide and expanding literature, it can be observed that there is limited recent empirical evidence on the key drivers of pharmaceutical expenditure. We present an overview of those previous studies in Tab. 1 below. Although the findings differentiate concerning sample and methodology, some common determinants explaining pharmaceutical expenditure come into prominences such as GDP, the structure of the population, and health system

116

Pelin Varol İyidoğanet al.

Tab. 1:  Review of Literature. Source: Composed by the Authors. Study Roy and Madhavan (2012)

Sample 48 US States (1998–2002)

Methodology Result Panel data The impact of access to primary methodology care, the severity of the disease, unemployment and education level on pharmaceutical expenditures of government in the coverage of State Medicaid Programmes Huh et al. (2008) US (2000) Cross-section Drug coverage as a determinant of probit analysis drug expenditure Lauridsen et al. Spain (50 SUR analysis -The effect of GDP, health system (2008) provinces) characteristics such as the number of hospital beds, medical doctors, pharmacists, the young and elderly population -Heterogeneous results with regard to time and provinces Shaikh and 136 developed Panel data -The strong positive impact of GDP Gandjour (2019) and developing methodology on pharmaceutical expenditure in countries low spending countries economies (1995–2006) with large economic freedom Jung and Kwon 22 OECD Panel data -The effect of level of protection for (2018) countries methodology property rights on pharmaceutical (1970–2009) expenditure rather than GDP and elderly population BlazquezSpain Cyclical -The positive relationship between Fernandez et al. (1995–2012) sensitivity pharmaceutical spending and (2016) analysis economic development Çınaroğlu (2017) European -Canonical -Reverse the relationship between countries correlation pharmaceutical expenditures and including analysis health outcomes Turkey (2015)

characteristics. We contribute to the literature by a broader point of perspective, which includes variables indicating the incidence rates of major chronic disease (cancer). Moreover, we use a dynamic approach to estimate pharmaceutical spending, which excludes the endogeneity problem.

3 Data, Model and Estimation Results We estimate the model for pharmaceutical spending by employing unbalanced panel data for selected 22 OECD countries concerning data availability over the

Findings on Macro Determinants of Pharmaceutical Spending

117

Tab. 2:  The Description of Data Variable Dependent variable pharmaceutical spending (pharma) Explanatory variables GDP per capita growth (gr) adjusted health spending (adj_health)

Description   Pharmaceutical spending (%of health spending)

GDP per capita (USD 2010 constant prices) The difference between total health spending and pharmaceutical spending (%of GDP) male cancer incidence rate A cancer incidence rate is the number (male_cancer) of new cancers in a population during female cancer incidence a year, expressed as the number of rate (female_cancer) cancers per 100,000 people at risk. Incidence rate = (New cancers/ Population) × 100,000 World age-standardized An age-standardized rate is a summary rate measure of the rate that a population would have if it had a standard age structure. The most frequently used standard population is the World Standard Population. elder population (age) The share of 65 and older population (%of the total population) social spending Social spending (%of GDP) (soc_spend)

Source   OECD Health Statistics (2017) OECD National Accounts OECD Health Statistics (2018) Ferlay et al. (2018)

Ferlay et al. (2018)

OECD Labour Force Statistics OECD Social Expenditure (Aggregate data)

Notes: i) We obtain the elderly population of the UK from the Office for National Statistics (ONS). ii) The authors calculate GDP per capita growth by using the GDP per head series in the OECD database.

period 2000–2014. The description and the source of the data used in our model are summarized in Tab. 2. As for the methodology, we employ dynamic panel data analysis, which eliminates both the problem of the unobservable factors correlated with the dependent variable and regressors and the endogeneity that suppress the consistency and biasness property of the estimators. We perform the Generalized Method of Moments (GMM) methodology of Arellano and Bond (1991) which is based on the estimation of,

118

Pelin Varol İyidoğanet al.

∆yit = β 0 ∆yit−1 + β1 ∆xit + ∆εit (1)

where yit   indicate the dependent variable, that is, pharmaceutical spending. Furthermore, xit   it shows the explanatory variables presented in Tab. 2. The model utilizes the difference of the dependent variable  (∆yit−1 ) to remove the individual fixed effects. Finally, as a part of the Arellano and Bond (1991) methodology, we evaluate the consistency of GMM estimators by employing specification tests. In this regard, we both apply the Arellano-Bond test of autocorrelation in the first differenced errors at order 2 to analyze the serial correlation of the series, and Sargan test to examine the suitability of the instruments. We present the empirical results in Tab. 3 below. According to the results from the GMM procedure, we conclude that the rise in the elderly population leads to an increase in pharmaceutical spending, which is consistent with our expectations. Contrarily, we find no significant effect of male cancer incidence and the negative impact of adjusted health spending on pharmaceutical spending. On the other hand, the results show that a rise in female cancer incidence puts upward pressure on drug expenditures. Moreover, we explore that social spending has a positive impact on pharmaceutical spending. Finally, we find the significance of the instrument, which is the 1st difference of the dependent variable. According to the specification procedure, Sargan test results indicate that we accept the null hypothesis, which implies the validity of over identifying restrictions. Likewise, we detect no autocorrelation problem concerning AR (2) test results. Tab. 3:  GMM Results Dependent Coefficient Variable: pharma pharma (-1) -0.012* age    2.817* male_cancer    0.197 female_cancer    0.915* soc_spend    4.394* adj_health -1.602** gr    0.078 AR(2) test (p value): 0.1332 Sargan test (p value): 0.2318

Standard Error 0.001 0.228 0.135 0.299 0.237 0.061 0.292

Notes: i) Authors’ estimation. ii) *, ** and *** indicate the significance at %1, %5 and %10 levels, respectively.

Findings on Macro Determinants of Pharmaceutical Spending

119

4 Conclusion One of the main issues discussed with the structural change in health expenditures is the increase in pharmaceutical expenditures. Working on the determinants of pharmaceutical spending from a macro perspective is so vital for the establishment of public policies to be implemented in the following period. The fact that the cancer incidence (female), which is an excellent example in terms of chronic diseases, is related to drug expenditures, has revealed the necessity of evaluating the policies related to chronic diseases. In addition, differentiation in public social programs of countries in our study was related to differentiation in pharmaceutical expenditures. This shows us that policies and programs related to pharmaceuticals within health policies are related to policies and programs for social programs. One of the indicators used in terms of the effectiveness of health expenditures is the share of pharmaceutical expenditures. Increased efficiency in pharmaceutical policies would also contribute to increasing efficiency in public health policies and programs.

References Arellano, M. & Bond, S. (1991). “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations,” Review of Economic Studies, 58, 277–297. Belloni, A., Morgan D., & Paris V. (2016). “Pharmaceutical Expenditure and Policies: Past Trends and Future Challenges”, OECD Health Working Papers, No. 87, OECD Publishing, Paris. Blazquez-Fernández, C., Cantarero-Prieto, D., & Pascual-Saez, M. (2016). “Is Pharmaceutical Expenditure Related to the Business Cycles?”, Applied Economics Letters, 23(10), 705–707. Çınaroğlu, S. (2017). “İlaç Harcamalarının Sağlık Sonuçları İle İlişkisi: Bir Kanonik Korelasyon Analizi Uygulaması”, Hacettepe Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi, 35(2), 23–47. Ferlay J, Colombet, M. & Bray F. (2018). “Cancer Incidence in Five Continents, CI5plus: IARC CancerBase No. 9 [Internet]. Lyon, France: International Agency for Research on Cancer”. Available from: http://ci5.iarc.fr, (17.09.2019). Huh, S., Rice, T., & Ettner, S. L. (2008). “Prescription Drug Coverage and Effects on Drug Expenditures among Elderly Medicare Beneficiaries”, Health Services Research, 43(3), 810–832.

120

Pelin Varol İyidoğanet al.

Jung, Y. & Kwon, S. (2015). “The Effects of Intellectual Property Rights on Access to Medicines and Catastrophic Expenditure”, International Journal of Health Services, 45(3), 507–529. Lauridsen, J., Bech, M., López Hernández, F. A., Sánchez-Val, M., & Luz, M. (2008). “Geographic and Temporal Heterogeneity in Public Prescription Pharmaceutical Expenditures in Spain”, Review of Regional Studies, 38(1), 89–103. OECD (2017). Health at a Glance 2017: OECD Indicators. OECD Publishing, Paris. Roy, S., & Madhavan, S. S. (2012). “An Explanatory Model for State Medicaid Per Capita Prescription Drug Expenditures”, Social Work in Public Health, 27(6), 537–553. Shaikh, M., & Gandjour, A. (2019). “Pharmaceutical Expenditure and Gross Domestic Product: Evidence of Simultaneous Effects Using a twostep Instrumental Variables Strategy”, Health Economics, 28(1), 101–122.

Nevzat Saygılıoğlu

Example of Internal Tax Bleeding: “Tax Expenditures” Abstract: The issue of tax expenditures has been well known and adopted by developed countries for nearly half a century. This concept is perceived as a contradiction in that it includes the price collected as “tax ile and the price spent as” expenditure. It also reminds us of any administrative and legal expenses related to the collection of the tax, but it is not used in this sense. The concept of tax expenditure has the same meaning in theory. It is used as a concept that reduces the tax burden of taxpayers for various purposes and expresses regulations such as exemptions and exemptions in public. But, although it is the same definition in some respects, it does not have a structure suitable for international comparisons since it imposes different meanings in terms of scope. This study aims to describe the theoretical framework and reasons for assets of tax expenditures, and discussing its size and results in Turkey to attract the attention of business and politics. Keywords: Tax Exemptions, Tax Incentives, Tax Expenditures, Tax Erosion, Tax Justice JEL Codes: H 20, H 25, H 32

1 Theoretical Framework on Tax Expenditures These special arrangements commonly referred to as tax incentives or tax subsidies, referred to as “tax expenditures”, imply the application of provisions different from the regular tax system to a particular industry, activity, or group of persons. It is known that tax expenditures have various processes and mechanisms in terms of tax techniques and practices. States give incentives for specific activities by granting subsidies or reducing taxes for desired practices. States also make an expenditure when they allow individuals and legal entities to deduct, and even return to carry out certain activities. In other words, instead of making some public expenditures, the state aims to realize public investments with specific tax incentives. The definition of tax expenditures varies according to the country in which it is enacted. For example, Austria defines tax expenditure as the loss of tax revenue, except general tax norms, to grant privileges to the activities of some natural and legal persons. Tax expenditure in the Netherlands is defined as the loss of tax revenues arising from the legal situation and which is not compatible with the basic tax

122

Nevzat Saygılıoğlu

system. As another example, tax expenditure in Finland is defined as separation from the basic taxation structure to support specific objectives (Kulu, 2000: 3). Therefore, a tax expenditure is defined as loss of income from provisions that allow for a specific exemption, exemption, or deduction. But, the exemptions and exemptions arising from the tax technique such as value-added tax refunds, minimum subsistence allowance in continuous exportation in standard tax systems are not accepted as a tax expenditure. Tax expenditures are based on economic, fiscal, and even political reasons, especially social reasons. Tax expenditures arise as a permanent or temporary, complete or partial reduction of the burden of tax liability depending on the person or subject. The principle of legality of taxes has been explained under the title of “duty to pay taxes” in Article 73 of the Turkish Constitution. In this article, the executive body is given broad authority, provided that the upper and lower limits of exemptions, exceptions, reductions, and rates are included in the relevant tax law. The frequent use of taxation authority in terms of scope and intensity has become a subject of serious debate both in theory and in political and practical terms. To the increase of tax expenditures in Turkey and also began to emerge and configure tax amnesties too often, they have significant and negative consequences of the resort. First of all, this situation causes a loss of tax revenues to a great extent and amount. It also creates a severe tax injustice for tax-compliant and honest liabilities. More importantly, it encourages those who do not pay taxes and tend to evade. Thus, Turkey is becoming too institutional structures responsible for a large group of tax havens and tax return also almost obliged hell for small groups. The theoretical framework for tax expenditures is detailed in the following section.

1.1 Conceptually Tax Expenditures Stanley S. Surrey first expressed the concept of tax expenditure. Beyond some direct practices of the state, indirectly applied discounts, such as investment loans, special discounts, discounts on certain types of consumption, or rate cuts on specific activities, are called tax expenditures (Surrey and McDaniel, 1979: 228). The concept of tax expenditure is an approach whereby the tax system and the budget system are used together, suggesting that public expenditures are made not only through the state budget but also through the tax system (Ferhatoğlu, 2015:  91). In the definition created by the IMF, the tax revenues that income administrations give up for social, financial, and economic reasons are called tax expenditure (IMF, 2011: 99).

Example of Internal Tax Bleeding: “Tax Expenditures”

123

Tax expenditures lead to loss of income for the state, while taxpayers are seen as a decrease in tax liability. Tax expenditures are generally defined as tax revenues that the government has given up. However, tax expenditures are not included in the scope of public spending while creating a budget (Buhur, 2019: 69). The most important reason for the waiver of tax collection through tax expenditures is the transfer of funds from the public economy to the private economy through these expenditures. (Özlem ve Gürçam, 2015: 140). Although there is a consensus on the definition and concept of tax expenditure, there are different practices regarding the inclusion of exemptions, exceptions, and tax reductions in the scope of tax expenditure (Sabuncu, 2011: 10). All kinds of exemptions, exceptions, deductions, refund credits, and postponements that include privileged provisions other than the general tax technique, which continuously or temporarily, conditionally or ­unconditionally reduce the tax burden, are accepted as a tax expenditure. Particularly in the personal income tax and corporate tax systems, due to the different structures of the countries, numerous tax expenditure applications have been included for specific individuals or organizations in the activities and industries. Therefore, although the primary objectives and approaches are the same, there have been differences in the definition, scope, and applications of tax expenditures by c­ ountries. Despite different definitions or procedures, the typical characteristics of tax expenditures are as follows. In this context, tax expenditures; – Directed to a particular sector, activity or group of liabilities, – In terms of purpose, it should be defined as the aim that can be achieved by public expenditures, – The scope should be wide enough to determine a tax structure, – Abolition must be administratively possible, – Another tax regulation should not eliminate the effect of tax expenditure.

1.2 Tax Expenditures as Deviation from Normative Tax Approach The normative tax approach did not take into account the concept of tax expenditures when seeking the answer to the question of how to distribute the tax burden. At this point, the definition of a normative tax approach by governments emerged. Considering that the general aim of the phenomenon of tax expenditures is to provide social optimum, the economic dimension of tax expenditures emerges. In this context, according to the fact that tax expenditures deviate from the standard tax system, first of all, the normative tax order in a country should be

124

Nevzat Saygılıoğlu

defined. However, in practice, difficulties are encountered in determining the normative tax structure and, thus, tax expenditures.

2 Reasons of Assets of Tax Expenditures There are economic and social reasons for turning to tax expenditures. The common goals of tax expenditures are the accurate measurement of income, the distribution of financial aid, the change of tax burden, and the promotion of desired social behavior. The state implements tax expenditures to facilitate taxpayers’ achievement of their economic, political, social, and administrative purposes. Tax expenditures cause erosion of tax base (tax erosion), decrease in tax revenues, and decrease in tax flexibility. Tax expenditures increase the privilege of taxpayers in various income groups and increase their disposable income (Öztürk, 2011: 11). Therefore, tax expenditures are used to increase investments, to direct investments to specific sectors and areas, to support exports, to promote savings, or to implement the principle of the social state, especially for the protection of weak, disabled, and older people. Common objectives of tax expenditures; accurate measurement of income, distribution of financial assistance, changing tax burden, and encouraging desired social behavior. Tax expenditures; to increase investments, to direct investments to specific sectors and areas, to support exports, to encourage savings. Also, the social state comes to the agenda for social purposes, especially for the protection of the weak, the disabled, and the elderly, to fulfill the requirements. Therefore, the reasons for the existence of tax expenditures; economic, social, and even political.

3 Calculation Methods of Tax Expenditures Tax expenditures; It is possible to calculate with three different methods as the following abandoned income method, earned income method, and equivalent expenditure method. Abandoned income method: The measurement of losses in tax revenues due to tax privileges. Earned income method: Estimates the tax revenue that can be obtained when the tax privileges are removed. Equivalent expenditure method:  Estimating the amount of direct public expenditures to be made to achieve the same level of benefit instead of tax expenditure.

Example of Internal Tax Bleeding: “Tax Expenditures”

125

Located above the method of calculation of tax expenditures “waived income method”, it is the method most practiced in the world, including Turkey. The earned income method is used only in Japan.

4 Benefits and Problems of Tax Expenditures Tax expenditures are a financial tool used to produce social policies. Although, by definition, it is an abandoned income and thus seen as a loss, such expenditure affects available income (Tekin and Akdağ, 2013: 277). Monitoring of tax expenditures in central government budget law and is included in legal control provides an advantage in terms of seeing the system as a whole and transparency (Kara, 2019: 93). In this context, the benefits of tax expenditures can be stated as follows. – The cost of tax incentives is bearable as long as the increase in investment is higher than the tax loss increase. – Tax incentives lead to new investments over time, and the tax increases provided by these new investments may compensate for the loss of taxes incurred in the short term. – The participation of the private sector in the economic and social programs undertaken by the government is encouraged. – Provides the opportunity for the private sector to make decisions rather than government decisions When tax expenditure is made instead of public expenditure, this expenditure is not explicitly included in the budget as an expense. – The need for strict state supervision is reduced. Tax expenditures, which are an element of public expenditures, affect macro budget elements such as resource allocation and productivity, in particular, budget balance such as regular public expenditures. Because public expenditures, which are a component of the tax structure, make the budget balance more difficult by reducing the tax revenues in the first place. Since tax expenditures are financed from the tax base, they also affect the financial allocation and reduce efficiency and efficiency in public resource allocation (Saraç, 2010: 276). On the other hand, the drawbacks of tax expenditures are also revealed. A financial burden for the state means complexity and bureaucracy for the tax system (Uçanok, 2019: 113). These drawbacks can be summarized as follows: – Some tax expenditures are insufficient to override the basic economic forces or to balance them with local or foreign tax provisions (Inactivity).

126

Nevzat Saygılıoğlu

– Many tax expenditures correspond to a wide range of interest groups rather than actual needs. (Inefficiency) – The tendency to change the tax burden of taxpayers both horizontally and vertically (Inequality). – It is an obstacle for taxpayers among the poorest groups of society to benefit from tax expenditures. – The tax revenue base leads to contraction.

5 World Application Examples of Tax Expenditures Although the concept of tax expenditures is used in the same sense in all countries, it differs in terms of its definition, scope, and application forms. A tax expenditure is a concept that emerged in the world in the 1960s, and Germany prepared the first tax expenditure report in 1959. After Germany, the USA began to publish tax expenditure reports in 1968, Spain 1979, England 1979, France 1980, Canada 1997, and the Netherlands 1997. Today, the tax expenditure report is published in almost all OECD countries. Australia, Austria, Belgium, France, Germany, Portugal, Spain, Turkey, and the tax expenditure report in OECD countries such as the US and/or estimated amounts for the preparation of the list is a legal requirement. In most of these countries, tax expenditure reports are closely linked to the budget process (General Directorate of Revenue Regulations, Ministry of Treasury and Finance, 2018: 1). It is seen that countries are not in the unity of practice in terms of measuring tax expenditures. For example, in the US and UK tax expenditures, an abandoned income method is applied. But in countries such as Italy, different evaluation criteria such as cost and objective measures are used. Meanwhile, in countries such as Italy, Austria, Australia, Germany, the Netherlands, and Spain, accelerated or high-rate depreciation is considered tax expenditure. There are some essential criteria for tax expenditures in the European Union. In this context, according to the decisions of the Court of Justice, member states are prohibited from providing state aid as a rule. However, over time, the implementation of some regional and sectoral state aids has started. Performed through tax; The prohibition does not cover Depreciation, valuation methods, loss transfer, research and development, environment, education, employment support, development of underdeveloped regions, and privileges for sensitive and priority sectors such as agriculture, fisheries, and shipbuilding. The guidance published in 2010 for OECD countries, of which EU countries are members, is used. Moreover, in the IMF Fiscal Transparency Guide and published by the IMF, the inclusion of the central government tax expenditures

Example of Internal Tax Bleeding: “Tax Expenditures”

127

as a report in the budget documents was defined as the basic requirement of fiscal transparency. Most OECD countries regularly publish annual reports on tax expenditures, but OECD countries’ practices vary widely in scope and method.

6 Tax Expenditures in Turkey Explained in detail below is the view of tax expenditures in Turkey.

6.1 Legal Regulation Article 2 of the Turkish Constitution introduces the state as a social law state, and Article 73 underlines the fair and balanced distribution of the tax burden as the social purpose of the fiscal policy. Therefore, tax exemptions, deductions, and exemptions, tax refunds, or loans are applied within the framework of the financial functions of the tax, especially in the context of economic and social functions, and this definition is defined as a tax expenditure. In addition to the social role of the tax, especially since the economic function has become more prominent today, tax regulations for economics purposes emerge as another form of tax expenditure. In this context, tax expenditures are held in Article 18 of the Act No. 5018 in Turkey, is defined as the financial transparency principle that should be included in the central government budget each year (Batırel, 2013: 20). With the regulation mentioned above, the tax amounts that are waived due to tax exemptions, exemptions, and deductions are included in the tables added to the central government budget law. Thus, although there is no public expenditure, the legal framework of the tax expenditures, such as public expenditure, is put forward.

6.2 Fiscal Dimension: Tax Expenditures in Budget Laws The provisions of Act No. 5018 on Public Financial Management and Control concerning tax expenditures entered into force on 1 January 2005. Thus the 2006 central government budget law, which should come into force before the completion of the 2005 calendar year, includes lists and figures for tax expenditures for the first time. Turkey has started to calculate the tax expenditure after 5018 Public Financial Management and Control Law entered into force. In this respect, the definition and numerical dimensions of tax expenditures are included in Table B annexed to the Central Government Budget Law of each year. The evaluation of the numerical quantities related to tax expenditures in the Central Government Budget Law is given in detail below.

128

Nevzat Saygılıoğlu

Tab. 1:  Tax Expenditures and Gross Domestic Product (2006–2019) (Million TL) Source:  The author prepared it according to the statistics of the Turkish Statistical Institute, The Ministry of Treasury and Finance, and annual central government budget figures Year

Tax Expenditures (Million TL)

GDP (Million TL)

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (Program)

 8.592  9.479  12.444  14.684  14.363  17.566  17.918  22.417  23.888  26.112  29.963 102.216 132.142 178.696

  789.228   880.461   994.783   999.192 1.160.014 1.394.477 1.569.672 1.809.703 2.044.466 2.338.647 2.608.526 3.106.537 3.700.989 4.450.278

Share of Tax Expenditures in GDP (%) 1.08 1.07 1.25 1.46 1.24 1.26 1.14 1.24 1.17 1.12 1.15 3.30 3.57 4.01

6.2.1  Tax Expenditures and Gross Domestic Product Tax expenditures in Turkey and Gross Domestic Product (GDP) and the interaction between the relationships is given in the table below. When the above table is examined, in the fourteen years between 2006 and 2019, GDP increased by 5,6 times in current prices, whereas tax expenditures increased by 20,8 times in the same period. While the share of tax expenditures in GDP was 1,08% in 2006, this share was 4,01% in 2019. On the other hand, tax expenditures increased by only 3,5 times in the last 11 years from 2006 to 2016, when tax expenditures began to be monitored, but 1,7 times in the last three years between 2016–2019. Accordingly, the increase in tax expenditures, especially in the last few years, was much higher than the GDP.

6.2.2  Tax Expenditures and Public Expenditures Relations and interaction between public spending and tax expenditures in Turkey are given in the table below.

Example of Internal Tax Bleeding: “Tax Expenditures”

129

Tab. 2:  Tax Expenditures and Public Expenditures (2006–2019) (Million TL) Source: The author prepared it according to the statistics of the Turkish Statistical Institute, The Ministry of Treasury and Finance, and annual central government budget figures Year

Tax Expenditures Public Expenditures (Million TL) (Million TL)

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (Program)

 8.592  9.479  12.444  14.684  14.363  17.566  17.918  22.417  23.888  26.112  29.963 102.216 132.142 178.696

178.126 204.068 227.031 268.219 294.359 314.607 361.887 408.225 448.752 506.305 584.071 678.279 830.450 960.975

Share of Tax Expenditures in Public Expenditures (%)  4.82  4.64  5.48  5.47  4.88  5.58  4.95  5.49  5.32  5.16  5.13 15.07 15.91 18.60

When the above table is examined, in the fourteen years between 2006 and 2019, public expenditures increased by 5,4 times in current prices, whereas tax expenditures increased by 20,8 times in the same period. While the share of tax expenditures according to public expenditures was 4,82% in 2006, the share of tax expenditures according to public expenditures was 18,60% in 2019. In this context, the increase in tax expenditures was much faster than the public expenditures and was 3,9 times higher.

6.2.3  Tax Expenditures and Tax Revenues The relationship between tax expenditures and tax revenues in Turkey and interaction is given in the table below. When the above table is examined, in the fourteen years between 2006 and 2019, tax revenues increased by 5,5 times with current prices, whereas tax expenditures increased by 20,8 times in the same period. While the share of tax expenditures in 2006 was 6,25% in total tax revenues, the share of tax expenditures in 2019 was 23,62%. In this context, the increase in tax expenditures was 3,8 times higher than the tax revenues.

130

Nevzat Saygılıoğlu

Tab. 3:  Tax Expenditures and Tax Revenues (2006–2019) (Million TL). Source: The Author Prepared It According to the Statistics of the Turkish Statistical Institute, The Ministry of Treasury and Finance, and Annual Central Government Budget Figures. Year

Tax Expenditures (Million TL) 2006  8.592 2007  9.479 2008  12.444 2009  14.684 2010  14.363 2011  17.566 2012  17.918 2013  22.417 2014  23.888 2015  26.112 2016  29.963 2017 102.216 2018 132.142 2019(Program) 178.696

Tax Revenues (Million TL) 137.480 152.835 168.109 172.440 210.560 253.809 278.781 326.169 352.514 407.818 459.002 536.617 621.311 756.494

Share of Tax Expenditures in Tax Revenues (%)  6.25  6.20  7.40  8.51  6.82  6.92  6.42  6.87  6.77  6.40  6.53 19.05 21.27 23.62

6.3 Numerical Development of Legislation on Tax Expenditures The numerical development of legal regulations relating to expenditure taxes between the years of 2006–2019 in Turkey is located in the following table. When the above table is examined, in the fourteen years between 2006 and 2019, a total of 3,031 legal arrangements were made on tax expenditures. While the number of legal regulations on tax expenditures was 72 in 2006, this number was 606 in 2019. In 2016, the number of rules related to tax expenditures was 132, which was 641 in 2017 with an extraordinary numerical increase. The increase continued in 2018 and reached 673. The distribution of the tax laws and the number of legal regulations relating to expenditure taxes between the years of 2006–2019 in Turkey is located in the table below. When the above table is examined, Among the 72 legal regulations related to tax expenditures in 2006, 29 were associated with Income Tax Law, and four were related to other tax laws except for Corporate Tax, Value Added Tax, and Special Consumption Tax Laws. In 2016, the same numerical distribution was made ­concerning the Income Tax Law, 59 of which were related to the Income Tax Law, and 15 of the tax laws other than the Corporate Tax, Value Added Tax, and Special

Example of Internal Tax Bleeding: “Tax Expenditures”

131

Tab. 4:  Numerical Development of Legislation on Tax Expenditures (2006–2019). Source: The Author Prepares the Annual Central Government Budget. Year 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Total

Number of Regulations  72  79  87  89 100  99  99 106 121 127 132 641 673 606 3.031

Consumption Tax Laws. In 2017, although the number of regulations related to the Income Tax Law was 57 in the total number of 641 regulations, the number of rules related to other tax laws other than Corporate Tax, Value Added Tax, and Special Consumption Tax Laws increased by 113 times to 641. It is seen that occurred.

7 Result As explained before, “Tax expenditure” is defined as loss of income from provisions that allow for a special exemption, exemption, or deduction. But; The exemptions and exemptions arising from the tax technique such as value-added tax refunds, minimum subsistence allowance in continuous exportation in standard tax systems are not accepted as a tax expenditure — tax expenditures; economic, fiscal and political reasons. The phenomenon of tax expenditures, which has been in practice for almost half a century, is used in nearly all countries of the world in the same sense; in terms of scope and application. Tax expenditures; It has implications and implications for tax justice and income for public finance, business, and taxpayers. In terms of public finance, there are effects and consequences for the decrease in tax revenues and,

Law Income Tax Law Corporate Tax Law Value Added Tax Law Special Consumption Tax Law Other Laws Total

2007 29

25

13

 8

 4 79

2006 29

20

13

 6

 4 72

 4 87

 8

10

19

2008 46

 6 89

 9

10

19

2009 45

 8 100

10

11

24

2010 47

 7 99

10

11

24

2011 47

 6 99

10

11

24

2012 48

 6 106

10

14

26

2013 50

10 121

11

16

26

2014 58

13 127

11

17

28

2015 58

15 132

11

19

28

2016 59

498 641

18

35

33

2017 57

526 673

19

36

34

2018 58

452 606

18

41

34

2019 61

1559 3031

159

257

364

Total 692

Tab. 5:  Number Distribution of Tax Expenditures Regarding Tax Laws (2006–2019). Source: The Author Prepares the Annual Central Government Budget.

132 Nevzat Saygılıoğlu

Example of Internal Tax Bleeding: “Tax Expenditures”

133

consequently, the deterioration of the budget balance, the elimination of effective distribution of public resources, and the decline of fiscal discipline. In terms of business, while some sectors or firms are unduly supported in the form of inefficient use of abandoned public resources. In terms of taxpayers, social supports does not reach to low-income subsidies, economic supports does not replace them, and thus the tax justice deteriorates. In order to provide a useful tax expenditure model, the tax administration should first define the normative tax structure and prepare the tax expenditure budget accordingly. In addition, it should analyze the tax expenditure amounts in the budget prepared and measure their impact on cost and tax revenues and include the analysis of the state budget. Tax exemptions, exceptions, and reductions included in the Turkish Tax System should be assessed by comparing the economic or social benefits, and the limits of tax expenditures should be determined in such a way that they do not exceed the objective (Öz, 2002: 29). The state’s application of tax expenditures for economic, social, and cultural purposes does not comply with the principle of equality and justice. But it does not contradict the principle of openness as it is determined when, how, and how much tax will be paid in contemporary practice. On the other hand, tax expenditures are separated from the principle of impartiality because they prevent the effects of taxes on investment and consumption decisions and affect economic attitude and resource use. It complies with the principles of transparency and ease of payment. When tax expenditures are evaluated together with the principle of justice, it is seen that taxpayers may create tax privileges for some taxpayers contrary to the general principle. For this reason, it is vital to correctly determine which taxpayer group is given advantage with the revenue given up when tax expenditures are imposed (Gül, 2018: 82). The number of adjustments related to tax expenditures in Turkey is too much. This issue has been put forward in the relevant sections in numerical dimensions. Instead of the main tax laws such as income, institutions, value-added, private consumption, the tax exemptions and exemptions introduced by other laws and tax laws are very striking in terms of number and variety. Especially since 2016, tax expenditure amounts have increased considerably in Turkey. There is no sustainable side to tax exemptions and exceptions that are lack of tax justice, ineffective, non-objective, and very complex. Therefore, the number of tax expenditures in the form of exemptions, exceptions, and reductions in the Turkish Tax System should be reduced considerably.

134

Nevzat Saygılıoğlu

References Batırel, Ö. F. (2013). “Vergi Harcamaları, Mali Saydamlık Ilkesi ve Anayasaya Uygunluk”, İstanbul Ticaret Üniversitesi Sosyal Bilimler Dergisi, Yıl 12, Sayı 24, 13–20. Buhur, S. (2019). “Vergi Harcaması Kavramı ve Türkiye’deki Durumunun Analizi”, Maliye Hesap Uzmanları Derneği, Vergi Dünyası Dergisi, Mayıs 2019, Yıl 38, Sayı 453, 68–75. Ferhatoğlu, E. (2015). “Bir Kamu Harcaması Türü Olarak Vergi Harcaması ve Türk Kurumlar Vergisi Açısından Değerlendirilmesi”, Eskişehir Osmangazi Üniversitesi Sosyal Bilimler Dergisi, Cilt 6, Sayı 2, 77–93. Gül, K. (2018). “Türkiye’de Vergi Harcamaları Gelir Dağılımına Etkileri”, Vergi Sorunları Dergisi, Sayı 355, 64–84. Hazine ve Maliye Bakanlığı Gelir Düzenlemeleri Genel Müdürlüğü (2018). Vergi Harcamaları Raporu, Ankara. IMF (2011). Shifting Gears Tackling Challenges on the Road to Fiscal Adjustment, Fiscal Monitor, file:///C:/Users/Sony/Downloads/_fm1101pdf. pdf, (03.09.2019). Kara, K. Ö. (2019). “Vergi Harcamalarının Izonomisi”, Vergi Sorunları Dergisi, Sayı 365, 85–95. Kulu, B. (2000). “Vergi Harcaması ve Uygulama Örnekleri”, Vergi Dünyası Dergisi, Sayı 228, Ağustos 2000, 24–29. Öz, E. (2002). “Türk Vergi Sistemindeki Bazı Bazı Vergi Harcamalarının Optimal Vergileme İlkeleri Açısından Analizi”, Dokuz Eylül Üniversitesi İİBF Dergisi, Cilt 17, Sayı 1, 11–33. Özlem, A. E. & Gürçam, S. (2015). “Vergi Harcamaları: Seçilmiş Bazı OECD Ülke Uygulamaları”, Journal of Economics and Administrative Sciences, Volume XVII, Year 1, 138–158. Öztürk, Z. (2011). Türkiye’de Vergi Harcamalarının Gelişimi ve Değerlendirilmesi, Yayınlanmamış Yüksek Lisans Tezi, Erciyes Üniversitesi Sosyal Bilimler Enstitüsü, Kayseri. Sabuncu, C. G. (2011). Türkiye’de Vergi Harcamalarının Analizi, Yayınlanmamış Yüksek Lisans Tezi, Hacettepe Üniversitesi Sosyal Bilimler Enstitüsü, Ankara. Saraç, Ö. (2010). “Vergi Harcamaları ve İktisadi Etkileri”, Maliye Dergisi, Sayı 159, 262–277. Surrey, S. S. & McDaniel, P. R. (1979). “The Tax Expenditure Concept: Current Development and Emerging Issues”, Boston College Law Review, Volume XX, Number 2, 226-355.

Example of Internal Tax Bleeding: “Tax Expenditures”

135

Tekin, A. & Akdağ, Y. (2013). “Belçika, Danimarka, Finlandiya ve Hollanda’nin Vergi Harcamalari Gelişimi”, Afyon Kocatepe Üniversitesi İİBF Dergisi, Cilt XV, Sayı II, 277–292. Uçanok, O. (2019). “Beyana Dayanan Gelir Vergisi”, Vergi Dünyası Dergisi, Nisan 2019, Yıl 38, Sayı 452, 38–44.

Cihan Yüksel

The Size of the Public Sector and the Armey Curve: The Case of Turkey Abstract: The relationship between economic growth and public spending as a percent of GDP (government size) is a quite widespread issue in the literature. One of the important explanations of these debates is the Armey curve. The Armey curve is defined as a geometric expression that public spending below an optimal threshold level has an expanding effect, but that public spending above the threshold level affects economic growth adversely. The parabolic structure of the Armey curve is essential for estimating the optimal government size. This study aims to test the Armey curve using the ARDL bounds testing approach of time-series techniques between the years 1981–2018 in the Turkish economy. According to the coefficient values obtained in our study, the optimal level of public expenditure that maximizes economic growth is 16% of GDP. Between the years of 1981–2018 in Turkey, the actual rate varies from 12.1% to 33.5%, and the average rate is 20%. Accordingly, while the level of public expenditure between 1981 and 1992 remained below the optimal level, the level of public expenditure between 1993 and 2018 remained above the optimal level. Keywords: Public Expenditures, Economic Growth, Armey Curve, Turkey JEL Codes: E62, H11, H50, O40

1 Introduction Economic growth is one of the objectives of fiscal policy. Just like in other purposes of fiscal policy, public expenditures stand out from the tools used in the provision of economic growth. This is because a fiscal economist must know at which level and within which components he or she must use public expenditures to reach the target for economic growth. The composition of public expenditures is important here. Expenditures that directly affect economic growth utilizing implementing the total demand are generally real expenditures. The multiplier effect that these expenditures will create in the economy is expected to influence economic growth positively. The share of public expenditures within the Gross Domestic Product (GDP) also expresses the size of the government in an economy. However, when the size of the government comes into question, the differences between the paradigms grow deeper. What kind of effect was in the short term along with the expansive effects of the public expenditures in the long term should be discussed.

138

Cihan Yüksel

Those who say that there is a positive relationship between public expenditures and economic growth claim that the expansion of the public sector provides the function of private property insurance. Based on this, public expenditures encourage private investments that will lead to economic growth and also allow for the production of public goods that will improve the investment environment. Those who assert that there is a negative relationship between public expenditures and economic growth claim that the expansion of the size of the government (public spending) has diminishing returns effect and that the oversize of the government has a crowding-out impact on private investments. Based on this, public expenditures can transform into inefficient expenditures that lead to deterioration in resource allocation along with corruption. At the same time, the government will need higher taxes as public expenditures widen, but increasing taxes will slowly lead to negative effects on the economy. The Armey curve became one of the critical contributions brought to the debated relationship of the size of the government with economic growth. The geometric explanation that public expenditures have a widespread influence when beneath the optimal threshold level but negative impact over economic growth above the threshold level is expressed as the Armey curve (Armey, 1995). But the relationship between government size and economic growth can differ between economies and between periods, even in the same economy. The examination of this relationship based on the temporal and spatial distinctions still preserves its importance in the literature. The purpose of our study regarding this importance was the testing of the Armey curve between the years 1981–2018 in the Turkish economy. Our study used the ARDL approach, a time-series technique, and aimed to create an Armey curve for Turkey and to determine the level of public expenditure that maximizes economic growth in Turkey. For this reason, primarily public expenditures and the economic growth relationship in our study were examined theoretically in the framework of the discussions on the Armey curve, and the empirical literature that tries to respond to these discussions was subsequently compiled. Finally, we tried to determine the Armey curve and the optimal government size for the Turkish economy.

2 Armey Curve and Optimal Public Sector Size 2.1 Theoretical Literature The relationship between the size of the public sector and economic growth is a much-discussed topic certainly before it was explained with a curve. Solow (1956) and Swan (1956) don’t see public expenditures as a determinant for

The Size of the Public Sector and the Armey Curve

139

economic growth in the neoclassic growth models, which explained output level with labor and capital as a production function and which accepted technology as an exogenous variable. Barro (1990) revealed that there was a relationship between public expenditures and economic growth in the explained endogenous growth model. However, this relationship was positive to a certain amount, while economic growth is negatively affected at levels of high public expenditure. The first person who studied this relationship with the help of a graphic in a nonlinear extent was Armey (1995). Although there are studies that call this curve the BARS curve, based on the acronym of Barro (1989), Armey (1995), Rahn and Fox (1996), and Scully (1994), the frequent use in the literature is for Armey curve. Armey referred to Arthur Laffer, who explained the quadratic relationship between tax rates and total tax revenue and tried to explain with similar logic the relationship between government size and general welfare. According to Armey, the government is certainly necessary to ensure peace, prevent anarchy, and provide public services. This dimension of the government is similar to the constitutional description, such as guaranteeing the protection of freedom and increasing general welfare. However, if the government starts to grow after some point, it starts to erode the general welfare and liberty (Armey, 1995: 91–92). The Armey curve intercedes at this point. The horizontal axis in Fig.  1 expresses the growth of the government and the decline of liberty. The vertical axis shows the general welfare of society. It is seen in the graphic; there is an upper boundary on the topic of being able to make something better in the economy. Economic progress takes place with the increase of this upper boundary over time. The capability of increasing this is tied to an optimal mixture of elements such as government, savings, and investment. There is no prosperity at the level in which the government is zero because there is chaos, no domestic or international security, no system of justice, and no contract law. There is no prosperity at the level in which the government is 100% because there is no reason to work if the government owns everything. As is seen from the graphic, the government serves the people and increases the prosperity up to a certain point. However, after this point, the government begins to reduce productivity and, concerning this, reduce prosperity. The “X” point in the figure shows an optimal mixture that includes the activities of the public and the private sector. And the attainable prosperity comes to the highest level at this point (Armey, 1995: 92–93). It is understood from here that the Armey curve is a parabolic curve that demonstrates that government activities have the effect of increasing welfare up to a certain point but that the growth of the government beyond this certain level

140

Cihan Yüksel

Attainable prosperity

The general welfare

X

0%

The growth of government and decline of liberty

100%

Fig. 1:  Original Armey Curve. Source: Armey, 1995: 92.

reduces welfare. However, contrary to what is frequently used in the literature, Armey (1995) expressed the vertical axis in the original graph with the general welfare of society. The concept of welfare here has many determinant criteria. Vedder and Gallaway (1998) related this societal welfare to economic growth, and the Armey curve was expressed afterward in the literature with economic growth. As is known in addition to this, the size of the government is generally measured with the share of public expenditures within GDP. For this reason, the axes of the Armey curve today are expressed as economic growth and shares of public expenditures within GDP (government size), differently than the original version. The graphic below shows the version of the Armey curve used today. In Fig. 2 as the size of the public sector, shown in the horizontal axis, increases from zero (from complete anarchy), the rate of economic growth, shown in the vertical axis, grows from the start. The curve has a concave form because of diminishing marginal return. In other words, a proportional increase in public expenditures is slower than a proportional increase in economic growth. Along with growing positive externalities, an additional percentage increase in the contributions of the government to economic activities still creates further economic productivity (meaning, a positive slope in the curve). At one point, however, the marginal benefit obtained from growing public expenditures is zero. When the contrary effect of the growth of the government concludes with a

141

Real gross domestic product

The Size of the Public Sector and the Armey Curve

Public spending as a percent of GDP

Fig. 2:  Armey Curve. Source: Vedder ve Gallaway, 1998: 2.

decrease in the increase of output, the growth-increasing properties of the government begin to decrease (Alimi, 2014: 7). The absence of the government will lead to a state of anarchy and a low level of output per capita because the lack of legal government rules and the failure to protect property rights presents an extraordinarily little incentive for savings and to make investments. Similarly, the output per capita will be low in situations when the government makes all input and output decisions. However, when the mixture of public-sector and private-sector decisions becomes relevant, the output is expected to be more significant. Based on this, outputincreasing features are dominant when the government is exceedingly small. In addition to this, the functions of the government that increase growth subside after a point and the greater expansion of the government does not lead to the expansion of outputs (Vedder and Gallaway, 1998: 1–2). In other words, as public expenditures increase, government-financed additional projects gradually become less productive, and the taxes and debts that increase about this bring further burdens. At this point, the marginal benefit obtained from growing government expenditures is zero (Pevcin, 2004: 4). The Armey curve does not mean that the government is entirely evil. It emphasizes that an excess of something accepted as a good thing may be harmful. For this reason, they assert that the government must be measured in the economy (Vedder and Gallaway, 1998: 2).

142

Cihan Yüksel

Facchini and Melki (2011) say that the positive effects of public expenditures can be explained with the benefits obtained from the correction of market failures, and their negative effects can be explained with the costs that the nature of the state failures create. For this reason, they express that the Armey curve is the combination of two different curves that show the shortcomings of the market and state. According to Schaltegger and Torgler (2006), although there is a large empirical literature that has researched the relationship between government size and economic growth, the empirical evidence obtained is still insufficient. This is because the concept of a small or large government is not hypothetically a determinant on its own. While a negative relationship is only valid for rich countries with an expansive public sector, the growth in the size of the government in underdeveloped countries can lead to more secure property rights and the implementation of agreements. Analyses were conducted for this reason, considering the levels of development of the countries. Many factors like countries’ levels of development, levels of productivity, transaction costs, rates of corruption, bureaucratic unwieldiness, strength of rent-seeking operations, length of lags occurring in the observance results and the implementation of policies, and power of fiscal policy to penetrate conjuncture may be determinant in the effect of public expenditures on economic growth.

2.2 Empirical Literature Numerous studies test the Armey curve for countries and periods. The objective of seeking answers to the question of what kind of relationship there is between economic growth and the size of the public sector constitutes the foundation of these studies. Guseh (1997) concluded in an analysis for 59 middle-income, underdeveloped countries for the 1960–1985 period, and Fölster and Henrekson (1999) also completed a study they performed for 23 OECD countries for the 1970–1995 period that there was a negative relationship between public expenditure and economic growth. However, Ram (1986), in an analysis for four different groups of countries and the periods of 1960–1970 and 1970–1980, and Kormendi and Meguire (1986), for the 1931–1983 period, concluded that there was a positive relationship between public expenditures and economic growth. Vedder and Gallaway (1998) testes the Armey curve for the U.S. economy for the period of 1947–1997 with the least-squares regression analysis and calculated the optimal size of the government as 17.45%. The Armey curve was tested further in five countries in the continuation of the study. Based on this, it was

The Size of the Public Sector and the Armey Curve

143

calculated that the optimal size of a government for Canada in the 1926–1988 period was 21.37%, for Denmark in the 1854–1988 period was 26.14%, for Italy in the 1862–1988 period was 22.23%, for Sweden in the 1881–1988 period was 19.43%, and for the United Kingdom in the 1830–1988 period was 20.97%. Pevcin (2004) tested the Armey curve in a panel data analysis that covered 12 industrialized Western European countries for the 1950–1996 period and determined that the optimal size of the public sector ranged between 36.56–42.12%. The author found this rate to be high and tested the Armey curve with the time series method separately using country data in the continuation of the study. The author concluded that the optimal public sector size for eight countries whose results were statistically significant was between 37.09–45.96%. When referring to the year 1996, it was seen that only the size of the public sector in Ireland, from among these countries, was below the calculated optimal levels. Chen and Lee (2005) concluded that there was a nonlinear Armey curve for the period of 1979–2003 in Taiwan. Based on this, the threshold regime for the total public expenditures was 22.83%, while the threshold regime for the public investment expenditures was 7.30%, and the threshold regime for public consumption expenditures was found to be 14.96%. Schaltegger and Torgler (2006) tried to test the effect of the size of a subfederal government for a rich country on economic growth using panel data for 26 Swiss cantons for the period of 1981–2001. The general finding was that there was a strong negative relationship between the size of a government and economic growth. Davies (2009) added a different dimension to the literature on the optimal size of the government and correlated the effect of government consumption expenditures on social welfare. Thus, using the United Nations’ Development Programme’s Human Development Index as an outcome variable, Davies shifted the criterion for optimal government size from productivity to social welfare. By conducting a panel data analysis for 154 countries for the 1975–2002 period, Davies concluded that the optimal size based on the humanitarian-development standards of the government was significantly greater than the optimal size. Matuşcu and Miloş (2009) found the optimal public sector size to be 27.46% and 30.42% in the analysis they conducted in 12 old EU member states and 15 EU member states in the 1999–2008 period. Samimi, Nademi, and Zbeiri (2010) tested a two-sector production model by measuring the threshold government size in eight Muslim countries for the 1980–2007 period. Based on this, a nonlinear relationship was found between the size of the government and economic growth. A significant, positive correlation between the two variables when the government is small and a meaningful

144

Cihan Yüksel

negative relationship when the government was large (except for Jordan and Turkey) were determined. Abounoori and Nademi (2010) used a two-sector production function to test the Armey curve for Iran with a threshold regression model. According to a study that found a nonlinear relationship between economic growth and public expenditures for the 1959–2005 period, the threshold value for total public spending was 34.7%, while public consumption expenditures were 23.6%, and public investment expenditures were 8%. Facchini and Melki (2011) studied a long period of 1871–2008 in France and attained strong findings that the Armey curve had a relationship with the time series. According to this, the optimal government size in France for this period was at a rate of 30% of GDP. Fallahi and Montazeri Shoorkchali (2012) tested the existence of the Armey curve using a smooth transition model for the 1961–2008 period in Greece. As a result of their analysis, they concluded that there was a nonlinear relationship between economic growth and public expenditures but that this relationship was positive. According to the study, which found that the threshold was 13.26% in Greece for this period, the existence of the Armey curve could not be verified. Herath (2012) asserted that the Armey curve can be valid not only for developed nations but also for underdeveloped countries and tested the Armey curve for the Sri Lankan economy. The researcher performed an analysis using the least-squares method for the 1959–2009 period and found the level of public expenditures, which corresponds to the peak of the threshold of the Armey curve, to be about 27% of GDP. Alimi (2014) tested the Armey curve in the Nigerian economy between 1970 and 2012 and acquired different optimal public sector sizes under different assumptions. Based on this, the optimal size of the public sector is 19.8% when there is a GDP-dependent variable, including the component of the government, while it is 12.58% when there is a GDP-dependent variable in which the government component is not included. Ahmad and Othman (2014) concluded that the Armey curve was valid using the ARDL bounds test approach for the 1970–2012 period in Malaysia and determined that the optimal level of public expenditure was 16.32%. This rate is above the level of public expenditures that occurred in the year 2012. Hok, Jariyapan, Buddhawongsa, and Tansuchat (2014) tested the Armey curve with the help of a panel data analysis in the 1995–2011 period for eight Asian countries (Brunei, Cambodia, Indonesia, Malaysia, Philippines, Singapore, Thailand, and Vietnam) and concluded that the optimal rate was 28.5%.

The Size of the Public Sector and the Armey Curve

145

Turan (2014) tested the validity of the Armey curve for different periods in Turkey and found significant results. Based on this, the optimal public expenditures ranged between 8.8–9.1% in the 1950–2012 period and between 15.4–17% in the 1970–2012 period. Considering non-interest public expenditures, the researcher found the optimal level to be 14.4% for the 1980–2012 period. The discovered values were below the realized values. De Mendonça and Cacicedo (2015) tested the Armey curve with monthly data in the period of 2000–2013 in the Brazilian economy and found the optimal government size to range based on the established models, between 20.88–23.05%. Pamuk and Dündar (2016) calculated the optimal public sector size to be 23.5% of GDP using the Scully time series method for the Turkish economy in the 1950–2006 period. Varol İyidoğan and Turan (2017) tested the Armey curve with the threshold regression model for the period of 1998:1-2015:1 in Turkey, found strong findings that there was nonlinear relationship, and calculated the threshold values as 16.5% for the total public expenditures, 12.6% for the public consumption expenditures, and 3.9% for the public investment expenditures. Tabaghua (2017) found the optimal government size in the Georgian economy in 2002–2014 to be at a rate of 21% and determined that the public expenditures were beneath the optimal level before 2006 and above the optimal level after 2006. Bozma, Başar, and Eren (2019) tested the Armey curve with the ARDL cointegration model in G7 countries by dealing with different periods between the years of 1981–2014. Based on this, it was determined that the Armey curve in the United States, France, and Canada are valid and are not valid in other countries. Optimal public consumption expenditures were calculated as 12.46% in the United States, 23.57% in France, and 18.93% in Canada.

3 The Armey Curve and Optimal Public Sector Size in Turkey The share of public expenditures within GDP in Turkey in the period of 1981– 2018 ranges between 12.1% and 33.5%. This rate, which shows the size of the government, changes either based on fiscal policies implemented against conjuncture or based on the change in the understanding of the state. The economic growth rates are, on average, 3.9% in this period. As is seen from Fig. 3, public expenditures and economic growth rate move in the same direction in this period, except for years of crisis. The fundamental purpose of our study was to create an Armey curve that shows the relationship between economic growth and public expenditures based

146

Cihan Yüksel

40% 35% 30% 25% 20% 15% 10% 5% 0% –5% –10% 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09 20 11 20 13 20 15 20 17

19

19

81

–15%

Public expenditures/GDP

Economic growth rate

Fig. 3:  The Size of Government and Economic Growth Rates in Turkey (1981–2018). Data Source: Ministry of Treasury and Finance, Republic of Turkey (2019).

on Turkish economic data from the period of 1981–2018 and to calculate the ratio of public expenditure level to GDP that maximizes economic growth in Turkey.

3.1 Model and Data Set Our study aims to calculate the size of optimal public expenditures by testing the Armey curve in Turkey. The analysis in which the time series techniques were used took place based on the following model.

Yt = β0 + β1 Gt + β2 Gt 2 + εt (1)

In the model, Yt expresses the rate of economic growth, Gt expresses the percentage of total public expenditures for GDP, and εt expresses the error term. The data set used in the model was obtained from the Republic of Turkey Ministry of Treasury and Finance and comprised data belonging to the period of 1981–2018. The dependent variable, Yt , were deflated data of nominal GDP based on the CPI (1987=100) procured from the Turkish Statistical Institute and were obtained with the logarithmic difference. All public expenditure values were prepared by dealing with consolidated budget data from 1981 to 2005 and central government budget data for the period of 2006–2018. By creating a quadratic equation, the presence of the Armey curve was accepted. The purpose here is to determine the optimal level of public expenditure for Turkey regarding the preliminary acceptance in which the Armey curve exists. For this to

The Size of the Public Sector and the Armey Curve

147

Tab. 1:  ADF Unit Root Test Results Variables

ADF-Test Statistic

Y G G2 ΔG ΔG2

-5.446*** -1.533 -1.890 -5.109*** -4.635***

MacKinnon Critical Value (5%) -2.943 -2.943 -2.943 -2.943 -2.943

Lag Length (k) 0 1 1 0 0

Notes: Δ expresses the first-degree difference processor. Results were obtained based on the Akaike Information Criterion in the unit root test. Maximum lag lengths were taken as 4. Only the model with constant was used. *** expresses the level of statistical significance at the level of 1%.

occur, the first independent variable coefficient is expected to be positive, and the second independent variable coefficient is expected to be negative.

3.2 Method and Findings The stationary of the variables was tested in the analysis through a unit root test, and the long-term coefficients were obtained afterward by testing the cointegration relationship with the ARDL bounds test approach.

3.2.1  Unit Root Analysis The Augmented Dickey-Fuller (ADF) (1981) unit root test was used to research the unit root properties of the variables discussed in our study. As is known, if the ADF test statistic is smaller than the specified critical value, the null hypothesis in which the series is not stationary is rejected. As is understood from Tab. 1, the Y series is stationary at level, while the other series include the unit root at the level. But when the difference is taken, these become stationary at a scale of 1% significance. Because of the differences in the degree of integration in the series, the ARDL bounds test approach that considers this situation was used.

3.2.2  Bounds Test (ARDL) Approach The ARDL (Autoregressive Distributed Lag) bound test approach was used to research whether there was a long-term cointegration relationship between the variables in our study. The use of the ARDL approach means to test whether the lags of the variables are statistically significant by estimating a dynamic limited VAR model. Our study estimated equation number (1)  to determine the

148

Cihan Yüksel

Tab. 2:  (2) Numbered Equation Bound Test Results Dependent Variable Lag Length (k)

FY (Y | G, G2 )

2

F-statistic

Critical Values (Lower Bound - Upper Bound) 3.10 - 3.87

14.377

Notes: The critical values were given based on a significance level of 5%. The maximum lag length was taken as 4, and the lag length was specified based on the Akaike Information Criterion.

long-term relationship with the ARDL unrestricted error correction model (UECM), which is expressed in equation number (2). ΔYt = β0 + β1 Yt−1 + β2 Gt−1 + β3 Gt−1 2 +

p 

+

p 

λ2i ΔGt−1 +

i=0

p 

(

λ1i ΔYt−1

i=1

(2) 2

λ3i ΔGt−1 + εt

i=0

By estimating equation number (2)  with the least-squares method, the null hypothesis was tested in which the coefficients of the lagged variables are equal to zero (there is no cointegration relationship between the variables), and the alternative hypothesis was tested in which the coefficients of the lagged variables are not equal to zero (there is a cointegration relationship between the variables). Accordingly, if the F-statistic value exceeds the upper critical value, it can be said that there is cointegration between variables. As is understood from Tab. 2, the F-statistic is found above the upper critical value at a significance level of 5%. In this situation, the null hypothesis, which expresses that there is no cointegration relationship between the variables, is rejected. In other words, there is a long-term relationship between public expenditures and economic growth in the period of 1981–2018. After finding a long-term relationship between the variables, the ARDL long-term model was estimated. Based on this, ARDL long-term estimation for equation number (1) is obtained with equation number (3).



∆Yt = β0 +

p  i=1

β1 ∆Yt−i +

r  i=0

β2 ∆Gt−i +

k  i=0

β3 ∆Gt−i 2 + εt

(3)

As a result of the estimation of equation number (3)  with the least-squares method, the long-term coefficient estimations belonging to the ARDL (3,4,2)

The Size of the Public Sector and the Armey Curve

149

model whose lengths of lag are specified based on the Akaike Information Criterion are shown in Tab. 3. Tab. 3:  ARDL (3,4,2) Model Long-Term Coefficient Estimations Variable Coefficient Standard Error t-statistic Prob. C  0.010 0.025  0.398 0.694 G  0.513 0.252  2.035 0.053* G2 -1.606 0.605 -2.653 0.014** Note: * and ** express the level of statistical significance at the level of 10% and 5%, respectively.

As is seen in Tab. 3, the variable of public expenditures is positive and significant at a level of 10% while the variable of the square of the public expenditures is negative and significant at a scale of 5% in the long-term. This situation is a result that is expected theoretically, and that supports the Armey curve. Finally, the error correction coefficient (η) and the short-term dynamic parameters were estimated. This situation is shown with equation number (4). ∆Yt = β0 +

p 

β1 ∆Yt−i +

r 

β2 ∆Gt−i +

k 

β3 ∆Gt−i 2 + ηecmt−1 + εt

i=1 i=0 i=0 (4)

Here, the ecmt−1 variable (error-correction term) expresses the one-period lagged value of a series of error terms found in equation number (3). The error-correction coefficient (η) shows the short-term imbalance that might be corrected in the long term and is expected to has a negative sign and be statistically significant. As is seen from Tab. 4, the error-correction coefficient was found to be negative-signed and statistically significant at a level of 1%. The F-statistic was found to be statistically significant at a scale of 1%. According to the Jarque-Bera test, the error terms are distributed normally. According to the Breusch-Pagan-Godfrey heteroscedasticity test, there is no heteroscedasticity problem. There is no autocorrelation problem based on the Breusch-Godfrey autocorrelation test. It is possible to calculate the optimal public expenditure level from the longterm coefficients that we obtained based on the ARDL approach. Based on î this, as ó dY =0, a result of equalizing the derivative of the equation number (1) to zero dG ó î β1 the level of optimal public expenditure can be found with formula − 2β . Thus, 2 the level of public expenditures corresponding to the peak of the Armey curve will have been found. According to the coefficient values that we obtained in

150

Cihan Yüksel

Tab. 4:  ARDL Model Error Correction Coefficient Estimations Variable ΔY(-1) ΔY(-2) ΔG ΔG(-1) ΔG(-2) ΔG(-3) ΔG2 ΔG2(-1)

ecmt−1

Coefficient    1.076    0.331    0.121 -4.591 -1.218 -0.887 -4.501    7.449 -2.619

Standard Error 0.231 0.131 0.990 0.942 0.291 0.278 1.955 1.913 0.325

ecmt−1 = Y − (0.513 ∗ G − 1.606 ∗ G2 + 0.010)

R2 AIC

   0.830 -4.113

Adjusted R2 F-statistic

DW-statistic

   2.078

χ2BG

Jarque-Bera

   2.069 [0.355]

χ2BPG

t-statistic    4.669    2.521    0.122 -4.873 -4.186 -3.192 -2.303    3.894 -8.063

Prob. 0.000*** 0.019** 0.904 0.000*** 0.000*** 0.004*** 0.031** 0.000*** 0.000***  0.748 10.182 [0.000]  2.042 [0.155] 10.540 [0.482]

Notes: *, **, and *** express the level of statistical significance at the level of 10%, 5%, and 1%, respectively. AIC: the Akaike Information Criterion, DW-statistic: the Durbin-Watson statistic, χ2BG the Breusch-Godfrey LM serial correlation test, and χ2BPG: the Breusch-Pagan-Godfrey heteroscedasticity test.

our study, the (optimal) level of public expenditures that maximizes economic growth constitutes 16% of GDP. This rate varies between 12.1% - 33.5% for the years 1981–2018 in Turkey, and its average value is 20%. Based on this, the level of public expenditure that occurred between 1981 and 1992 was under the optimal level, while the level of public expenditure that occurred between 1993 and 2018 was above the optimal level. According to the findings we obtained in our study, the Armey curve belonging to the period of 1981–2018 in Turkey was shown in Fig. 4. By placing the values of the coefficients in the model we estimated and the value of public expenditures that grow with certain intervals into the equation, we can geometrically demonstrate the relationship between public expenditure and economic growth. Indeed, as is to be understood from Fig. 4, the optimal public expenditure level for the relevant period in Turkey (the level of public expenditure that demonstrates the peak point for the Armey curve) is 16%, and the red

151

The Size of the Public Sector and the Armey Curve 6,00%

Economic growth

5,00% 4,00% 3,00% 2,00% 1,00%

32,0%

30,2%

28,5%

26,7%

24,9%

23,1%

21,4%

19,6%

17,8%

16,0%

14,2%

12,5%

10,7%

8,9%

7,1%

5,3%

3,6%

1,8%

0,0%

0,00%

Government size

Fig. 4:  The Armey Curve in Turkey

line shows it. The green line is the level of public expenditures that occurred in Turkey in 2018 (22.2%) and is to the right of the red line. In this situation, it is possible to say that the level of public expenditure in the year 2018 was above optimal.

4 Conclusion The Armey curve tries to explain the hypothesis that the rate of public expenditures to GDP, or in other words the size of the government, will contribute positively to economic growth up to a certain level but will negatively affect economic growth after this certain level, and it is an important topic that is discussed in the fiscal economics literature. Based on studies that provide different results in different periods, different countries, and various economic structures, it is not possible to express an optimal size of government that is de facto. Therefore, the Armey curve in the Turkish economy was tested for the 1981–2018 period in our study, and we attempted to determine the level of government that maximized economic growth. It was seen in the model for which the bound test (ARDL) approach, a time series technique, was used based on yearly data from these periods that the Armey curve provided statistically significant results and met theoretical expectations. Based on this, the rate of public expenditure to GDP that maximizes economic growth was calculated as 16%. While there is information that the levels of public expenditure that occurred in the Turkish economy in the 1981–2018 period were

152

Cihan Yüksel

in a range of 12.1 - 33.5% and the average value obtained was 20%, we can say that the actual level mostly remained above the optimal level. Reviewing based on year, we can say that the level of public expenditure that occurred between 1981 and 1992 was below the optimal level and between 1993 and 2018 was above the optimal level. Our study only aimed to test the Armey curve with the presupposition that accepted economic growth as a dependent variable. But economic growth is only one of the objectives of fiscal policy. Therefore, public expenditures are also expected to serve purposes like price stability, development, and equity in income distribution. For this reason, the determinant of optimal government size is not only economic growth. It is a fact that the size of the government that maximizes each objective of the fiscal policy may be different and that each type of public expenditure contributes to different objectives at different levels. For this reason, determining an optimal level of public expenditure (or types of public expenditure), that will maximize the overall set of fiscal policy objectives will advance the literature toward a wider discussion.

References Abounoori, E., & Nademi, Y. (2010). Government size threshold and economic growth in Iran. International Journal of Business and Development Studies, 2(1), 95–108. Ahmad, R., & Othman, N. (2014). Optimal size of government and economic growth in Malaysia: Empirical evidence. Prosiding Persidangan Kebangsaan Ekonomi Malaysia, 9, 41–48. Alimi, R. S. (2014). Does optimal government size exist for developing economies? The case of Nigeria. MPRA Paper No: 56073, https://mpra. ub.uni-muenchen.de/56073/1/MPRA_paper_56073.pdf, (27.05.2019). Armey, R. K. (1995). The freedom revolution: The new Republican house majority leader tells why big government failed, why freedom works, and how we will rebuild America. Regnery Publishing, Washington, D.C. Barro, R. J. (1989). A cross-country study of growth, saving and government. NBER Working Paper, No. 2855, https://www.nber.org/papers/w2855.pdf, (15.05.2019). Barro, R. J. (1990). Government spending in a simple model of endogenous growth. The Journal of Political Economy, 98(5), 103–125. Bozma, G., Başar, S., & Eren, M. (2019). Investigating validation of Armey curve hypothesis for G7 countries using ARDL model. Doğuş Üniversitesi Dergisi, 20(1), 49–59.

The Size of the Public Sector and the Armey Curve

153

Chen, S. T., & Lee, C. C. (2005). Government size and economic growth in Taiwan: A threshold regression approach. Journal of Policy Modeling, 27(2005), 1051–1066. Davies, A. (2009). Human development and the optimal size of government. The Journal of Socio-Economics, 38(2), 326–330. De Mendonça, H. F., & Cacicedo, T. (2015). Size of government and economic growth in the largest Latin American country. Applied Economics Letters, 22(11), 904–910. Dickey, D. A., & Fuller, W. A. (1981), Likelihood ratio statistics for autoregressive time series with a unit root. Econometrica, 49, 1057–1072. Facchini, F., & Melki, M. (2011). Optimal government size and economic growth in France (1871-2008): An explanation by the state and market failures. CES Working Papers, ISSN: 1955-611X, Paris, 1–37. Fallahi, F., & Montazeri Shoorkchali, J. (2012). Government size and economic growth in Greece: A smooth transition approach. MPRA Paper No: 74078, https://mpra.ub.uni-muenchen.de/74078/1/6a6ba1cf08f52069e825ac968de c3b06.pdf, (12.05.2019). Fölster, S., & Henrekson, M. (1999). Growth and the public sector: A critique of the critics. European Journal of Political Economy, 15(2), 337–358. Guseh, J. S. (1997). Government size and economic growth in developing countries: A political-economy framework. Journal of Macroeconomics, 19(1), 175–192. Herath, S. (2012). Size of government and economic growth: A nonlinear analysis. Economic Annals, 57(194), 7–30. Hok, L., Jariyapan, P., Buddhawongsa, P., & Tansuchat, R. (2014). Optimal size of government spending: Empirical evidence from eight countries in Southeast Asia. The Empirical Econometrics and Quantitative Economics Letters, 3(4), 31–44. Kormendi, R. C., & Meguire, P. (1986). Government debt, government spending, and private sector behavior: Reply. American Economic Review, 76(5), 1180–1187. Ministry of Treasury and Finance, Republic of Turkey (2019). Budget Sizes and Budget Realizations, https://ms.hmb.gov.tr/uploads/2019/04/ butcegiderlerixls.xls, (23.04.2019). Mutaşcu, M., & Miloş, M. (2009). Optimal size of government spending. The case of European Union member states. Annales Universitatis Apulensis Series Oeconomica, 11(1), 447–456. Pamuk, Y., & Dündar, U. (2016). Kamu harcamalarının optimal boyutu: Türkiye örneği. Hacettepe Üniversitesi İİBF Dergisi, 34(3), 23–50.

154

Cihan Yüksel

Pevcin, P. (2004). Does optimal size of government spending exist?. EGPA (European Group of Public Administration) 2004 Annual Conference, University of Ljubljana, September. Rahn, R., & Fox, H. (1996). What is the optimum size of government? Vernon K. Krieble Foundation, Denver. Ram, R. (1986). Government size and economic growth: A new framework and some evidence from cross section and time-series data. American Economic Review, 76(1), 191–203. Samimi, A., Nademi, Y., & Zobeiri, H. (2010). Government size and economic growth: A threshold regression approach in selected Islamic countries. Australian Journal of Basic and Applied Sciences, 4(8), 2247–2249. Schaltegger, C. A., & Torgler, B. (2006). Growth effects of public expenditure on the state and local level: Evidence from a sample of rich governments. Applied Economics, 38(10), 1181–1192. Scully, G. (1994). What is the optimal size of government? Policy report. No: 188. National Centre for Policy Analysis, Dallas. Solow, R. M. (1956). A contribution to the theory of economic growth. Quarterly Journal of Economics, 70(1), 65–94. Swan, T. W. (1956). Economic growth and capital accumulation. Economic Record, 32(2), 334–361. Tabaghua, S. (2017). Optimal size of government and economic growth: The case of Georgia. Actual Problems of Economics, 7(193), 58–69. Turan, T. (2014). Optimal size of government in Turkey. International Journal of Economics and Financial Issues, 4(2), 286–294. Turkish Statistical Institute (2019). Consumer Price Index (CPI), http://tuik.gov. tr/PreTablo.do?alt_id=1014, (23.04.2019). Varol İyidogan, P., & Turan, T. (2017). Government size and economic growth in Turkey: A threshold regression analysis. Prague Economic Papers, 2017(2), 142–154. Vedder, R. K., & Gallaway, L. E. (1998). Government size and economic growth. Paper prepared for the Joint Economic Committee of the US Congress, Washington.

Nedim Mercan and Özay Özpençe

Okun’s Law: Turkey Case1 Abstract: The concepts of unemployment and economic growth, which have an important position in macroeconomic issues, are dynamic and are issues that are always on the agenda in every economic point. In addition to economic growth and increasing production, there is a desire to create jobs and reduce unemployment. Today, however, unemployment is not decreasing despite economic growth. There is a difference between the literature and today’s economic situation. Long-term and more sustainable economic growth is needed rather than short-term to reduce unemployment or increase employment. Therefore, unemployment is seen to be a broader, more complex, and more critical issue in terms of its policies and effects.While economic growth can be achieved through structural policies such as investments, and demand-increasing real wage growth, unemployment is not a problem that can only be solved by structural policies that increase economic growth. This is because there is a socio-cultural dimension as well as the economic dimension of the unemployment problem.Economic growth and unemployment are always up-to-date in all economies. Especially after the Second World War, the importance of this relationship increased. This study, which later entered the literature as the Okun’s Law, examined data on the U.S. economy between 1948 and 1960. The study concluded that there was a negative correlation between economic growth and unemployment. In other words, an increase in real GDP reduces unemployment. In this study, the relationship between economic growth and unemployment for the Turkish economy is investigated. As a result of the analysis, which was based on annual data from 1980 to 2016, an increase of 1% in economic growth reduces unemployment by 0.11%. In other words, this is the result of the validity of the Law of The Okun in Turkey. However, the inability of growth to adequately reduce unemployment is the basis for unemployment problems. In this context, the growth policies determined by governments will contribute to minimizing this problem by encouraging employment. Keywords: Economic Growth, Unemployment, ARDL, Regression Test JEL Codes: F43, E24, R15, C22

1 This study bases on the unpublished master thesis with the title “The Analysis on Turkey in Framework of The Okun’s Law of The Relationship Between Economic Growth and Unemployment” of Mr. Nedim Mercan and consultancy with Assoc. Prof. Ozay Ozpence at Pamukkale University Social Sciences Institute on 19/07/2017.

156

Mercan and Özpençe

1 Introduction The first to examine the relationship between growth and unemployment was the American economist Arthur M. Okun in 1962. According to this approach, which is entered into the literature as the Okun Law, the existence of a negative relationship between the real growth rate and the unemployment rate is emphasized (Ceylan & Şahin, 2010:  158). Accordingly, it was determined that the unemployment rate was low in the years when the real growth rate was high, and the unemployment rate increased in the years when the actual growth rate was flat or negative (Mıhçı & Atılgan, 2010: 48, Ahmad vd., 2011: 293). Theoretically, as long as growth contributes to employment growth, it is possible to benefit from the positive effects of growth by providing better income to individuals. In countries with high economic growth rates, the employment rate is expected to be high. However, due to the complex and multifaceted nature of unemployment, it is observed that this expectation does not occur at the desired level. (Takım, 2010:  3). Okun Law is one of the most common methods describing the relationship between unemployment and economic growth (Göçer, 2015: 2). The empirical analyses of the validity of the law of the Okun to date focus on the assumption that the relationship, in general, is symmetry. Symmetry relationship, in the expansion and contraction phases that may occur during periods of cyclical fluctuation, it has been accepted that the effect of real output on unemployment is similar. However, today’s studies indicate that the impact in real production on unemployment may be different during periods of contraction and expansion. It has an effect that increases unemployment during contraction periods and reduces unemployment during periods of expansion (Ceylan & Sahin, 2010: 158). This study aims to investigate the relationship between economic growth and unemployment within the framework of the Okun Law and to examine whether the Okun Law is valid in Turkey.

2 Growth and Unemployment in the Turkish Economy The problems experienced before 1980 and the lateness of the measures to solve these problems have led to a deepening of the problems. Facing high inflation, difficulties in finding financing, inadequacies in oil and energy, difficulties in finding imported inputs are the main economic problems before 1980. The decisions of January 24, 1980, which included export incentives to address export shortages arising from their shortcomings in imports, played a crucial role in solving the problems. The reduction of investments due to insufficient savings, high foreign debt and foreign currency bottleneck problem has also been effective in taking these decisions (Parasız, 2003: 281 – 283; Kılıçbay, 1984: 176 – 177).

Okun’s Law: Turkey Case

157

Tab. 1:  Economic Growth and Unemployment Rates in Turkey (1980–2016). Source: IMF (International Monetary Fund) (http://www.imf.org, 22.01.2017) Years Growth Unemployment 1980  -0,77 7.20 1981  4,36 7.20 1982  3,42 7.60 1983  4,75 7.51 1984  6,82 7.40 1985  4,25 6.94 1986  6,94 7.71 1987 10,02 8.13 1988  2,12 8.70 1989  0,25 8.57 1990  9,25 7.99 1991  0,92 8.19 1992  5,98 8.48

Years Growth Unemployment 1993    8,04 8.93 1994 -5,45 8.55 1995    7,19 7.62 1996    7,00 6.62 1997    7,52 6.81 1998    3,09 6.37 1999 -3,36 7.15 2000    6,77 5.99 2001 -5,69 7.80 2002    6,16 9.76 2003    5,26 9.92 2004    9,36 9.68 2005    8,40 9.48

Years Growth Unemployment 2006    6,89  9.03 2007    4,66  9.18 2008    0,65 10.02 2009 -4,82 13.05 2010    9,15 11.12 2011    8,77  9.09 2012    2,12  8.43 2013    4,19  9.04 2014    2,91  9.91 2015    3,84 10.27 2016    3,80 10.79

Looking at the relationship in Turkey before 1980, although it appears to have a relationship between growth rates and unemployment and employment, the relationship decreased after 1980. This is because production is carried out for the foreign market rather than the domestic market with the policies followed. This has led to a cost increase in the labor market, and it is considered positive to produce more with less labor to reduce costs (Akkaya & Gürbüz 2012 5). In this period, GDP conditions were observed to be insufficient in terms of job creation (Mıhçı & Atılgan 2010: 38–39). In Turkey, the unemployment rate in the 1990s was generally observed to exceed international standards (Tab. 1). Compared to other countries, Turkey is not only in the group of countries with low employment rates but also among the countries with the highest unemployment rates. The most important reason for this is that the rate of population growth generally increases more than employment. It was also observed that GDP growths during this period remained incapable of creating jobs (Mıhçı & Atılgan, 2010: 38 – 39). Structural reforms from 2002 have allowed the period between 2002 and 2007 to achieve a high and sustained growth momentum. However, the global crisis, which began to be felt in 2008 and continued in 2009, has negatively affected the economy, growth rates have decreased considerably, and the economy has shrunk by about 5%. The crisis is not only financially sourced, but it also affects

158

Mercan and Özpençe

the real sector. During this period, unemployment rates continued to increase and reached about 14% in 2009 (Acar, 2013: 17). In Turkey, the economy, which began to recover during the 2009 crisis, achieved high growth in 2010, but has lost this momentum since 2011 and has continued to fall below 5%, which is considered a potential growth rate. In 2014, it was stated that the growth rate would be below 4% when the Medium Term Program was announced. There was no improvement in unemployment rates during this period and remained in the 9–10% band (Köse, 2016: 62 – 64).

3 Literature Review Many surpluses are examining the relationship between economic growth and unemployment. Some of the essential studies related to Okun’s Law, which are dealt with both globally and nationally, are given in Tab. 2.

Tab. 2:  Literature Review. Source: Authors’ elaboration No Author Name/Year/Reviewed Country(s)/Years and Model 1  Okun Arthur M. (1962)/USA/ 1947 - 1960 2

3

4

Results Obtained

In this study, which entered the literature as Okun’s law, it was stated that there was an inverse relationship between unemployment and economic growth. Harris & Silverstone (2000) As a result of the analysis, the existence of New Zealand/1978 – 1999/ a long-term relationship between real GDP Co-integration Analysis and unemployment was denied. However, real GDP is the cause of unemployment. In addition, the model stated that real GDP is weak external, and the short-term Okun coefficient is estimated at -0.103. Silvapulle vs. (2004)/USA/1947 The Okun’s Law, which is being calculated 1Q – 1999 4Q/Time Series using USA post-war data, states that results Analysis and Regression Analysis support the asymmetric relationship between unemployment and output were obtained. Sinclair (2007)/USA/1948 1Q – The coefficient of the Okun was found 2005 4Q/Kalman Filtering Model to be negative for both output and and OLS unemployment. It has been stated that there is a negative relationship between output and unemployment. It also concluded that a 1% decrease in the temporary unemployment rate would result in a 1.4% increase in GDP.

Okun’s Law: Turkey Case

159

Tab. 2: (continued) No Author Name/Year/Reviewed Country(s)/Years and Model 5 Moazzami & Dadgostar (2009)/13 OECD Countries/ 1988 1Q – 2007 4Q/ Regression Model 6

Ball vs. (2013)/USA/1948 – 2011/20 OECD Countries/1980 – 2011/Regression Analysis/OLS

7

Huang & Yeh (2013)/53 Countries (21 OECD Countries and 32 NonOECD)/1980 – 2005/1976 – 2006/ Panel ARDL/ Pooled Mean Group Model/ Co-integration Analysis Khaliq vs. (2014)/9 Arab Countries/1994 – 2010/Pooled EGLS (Cross – Section SUR) Model/Unit Root Test

8

9

Palombi vs. (2015)/United Kingdom/1985 – 2011/Spatial Panel Approach

10 Ayhan (2008)/Turkey/1970 – 2006/Co-integration Analysis/ Granger Causality Analysis 11 Uysal & Alptekin (2009)/ Turkey/1980 – 2007/VAR Analysis

Results Obtained The analysis will show a 1% decrease in the unemployment rate, growth between 2.6% and 4.7% in the countries examined. Employment in Canada, Finland, Norway and the United States is more sensitive to economic growth. In most of the countries examined in the analysis, it was determined that the Okun law had a strong and stable relationship. The relationship between unemployment and output varies between countries. Furthermore, it was stated that there was no significant change in the analysis during periods of the great recession As a result of the analysis, it was stated that the long-term coefficients between unemployment and output are identical between countries. It is also stated that the Okun’s Law applies among the countries examined. At the end of the analysis, unemployment has negatively affected economic growth. In addition, it has been stated that the 1% increase in economic growth will reduce unemployment rates by 0.16%. The analysis, which also referred to the population, stated that a growth rate of 1% at the rate of population growth would increase unemployment by 0.37%. As a result of the analysis, it was found that Okun’s Law was valid, and the negative relationship between output and unemployment was reported to be strong. In the analysis, a long-term positive relationship was found between unemployment and growth. Also, as a result of the analysis, a one-way causality relationship from GDP to unemployment is determined. According to the results obtained in receiving in Turkey, it is stated that there is a Granger causality relationship from unemployment to growth. (continued on next page)

160

Mercan and Özpençe

Tab. 2: (continued) No Author Name/Year/Reviewed Country(s)/Years and Model 12 Ceylan & Şahin (2010)/ Turkey/1950 – 2007/ Co-integration Analysis/ TAR and M-TAR

Results Obtained

As a result of the analysis, it was stated that the Okun’s law is valid for a long time in the Turkish economy. There was also an asymmetric relationship between unemployment and growth. 13 Tarı & Abasız (2010)/ In the analysis, the fluctuations in the Turkey/1968 – 2008/Two Regime contraction periods of the economy affect the Threshold Co-integration unemployment more than the fluctuations Analysis/ Threshold Error in the expansion periods of the economy. Correction Model Furthermore, the Okun coefficient was -0.48 in the long run. 14 Muratoğlu (2011)/Turkey/2000 – As a result of the analysis, there was no long2011/Engle-Grangerer term relationship between employment and Co-integration Analysis/ Granger GDP, but a short-term relationship. Causality Test 15 Altuntepe & Güner (2013)/ In the study, two results were obtained. First, the Turkey/1988 – 2011/OLS growth in the services sector leads to increased employment. Secondly, there is a negative relationship between employment growth and economic growth in the services sector. 16 Göçer (2015)/Turkey/2001 2Q – As a result of the regression analysis, it is 2015 1Q/Granger Causality Test/ stated that growth hurts unemployment. In Regression Analysis this respect, the conclusion that the Okun’s Law applies to the Turkish economy has been put forward. Granger causality test stated that growth affects unemployment, while unemployment does not affect growth. 17 Bulut (2016)/Turkey/2005 – 2015 As a result of the analysis, it was determined (Quarterly Data)/ Structural that there are asymmetric relations between Fractured Unit Root Test/ growth and unemployment in Turkey. Asymmetric Causality Test Moreover, while an acceleration in the growth rate does not reduce the unemployment rate, a slowdown in growth is negatively affecting unemployment.

161

Okun’s Law: Turkey Case

4 Data Set, Model, and Empirical Findings In this study, which examined the relationship between economic growth and unemployment, the validity of Okun law for Turkey was tested. It was analyzed using the annual data of the period between 1980 and 2016 with the ARDL Co-integration model. As a result of the analysis, it was seen that there was a correlation between economic growth and unemployment in the long term. However, Regression analysis was performed after the ARDL Co-integration test to produce a more concrete result as a coefficient. The time series for real GDP and unemployment rate variables in the data set were taken from the International Monetary Fund (IMF) electronic database. The reason for the data being made since 1980 is the liberalization policies that began in Turkey at that time and then the desensitization of unemployment towards growth and the lack of desired results in employment. Tab. 3:  Descriptive statistics of Economic Growth and Unemployment Data. Source: Authors’ elaboration Statistics/Variables Mean Median Maximum Minimum Standard Deviation Skewness Kurtosis Jarque Bera

GDP    4.5677    5.6080 11.1130 -5.9620    4.3552 -0.9486    3.2174    5.6229

Unemployment(U)    8.5509    8.4860 13.0530    5.9970    1.4720    0.7163    3.7463    4.0232

When Tab. 3 is examined, descriptive statistics of variables are shown. The Skewness coefficient shows a skew distribution to the right in the growth data, while the unemployment data shows a skew distribution to the left. When we look at the Kurtosis coefficient, it is seen that the variables show a basic distribution.

162

Mercan and Özpençe

Tab. 4:  ADF and PP Unit Root Test Results. Source: Authors’ elaboration GDP Critical Value ADF %1 -3.6267 %5 -2.9458 PP %1 -3.6267 %5 -2.9458 ΔGDP ADF %1 -3.6329 %5 -2.9484 PP %1 -3.6329 %5 -2.9484

Test Values -6.4813* -6.7198*

-10.1949* -21.1336*

Unemployment(U) Critical Value %1 -3.6267 %5 -2.9458 %1 -3.6267 %5 -2.9458 ΔUnemployment(U) %1 -3.6329 %5 -2.9484 %1 -3.6329 %5 -2.9484

Test Values -1.8595 -1.7213

-5.0630* -8.3488*

*: * It represents a level of significance at 5%

Looking at the results of the ADF and PP unit root test values in Tab. 4, it is seen that the economic growth level is stable, and the unemployment variable is stable at the first level. 20 15 10 5 0 –5 –10 –15 –20

86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 Cusum

5% significance

Fig. 1:  CUSUM Test. Source: Authors’ elaboration.

Okun’s Law: Turkey Case

163

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 –0.2 –0.4

86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 Cusum of squares

5% significance

Fig. 2:  CUSUMQ Test. Source: Authors’ elaboration

Figure 1 and Figure 2 examine the stability of the series. In CUSUM tests, the graph should be within the specified limits. The fact that the graph is within the boundaries means that there is no structural break. In the analysis, it is seen that the graph is within the determined limits. In other words, it is concluded that there is no structural break. Tab. 5:  Diagnostic Tests ARDL (1, 1) Model. Source: Authors’ elaboration Diagnostic Tests F - statistics Breusch – Godfrey LM Arch LM Ramsey Reset Jarquera Bera

Statistics 4.2682 (0.0121) 0.0267 (0.9736) 0.1246 (0.7263) 2.8957 (0.0988) 4.1058 (0.1283)

According to the results of the diagnostic tests in Tab. 5, 5% at the level of significance, the relationship between economic growth and unemployment is significant, and there is no autocorrelation and heteroscedasticity problem in the model. Furthermore, it is observed that there is no model setting error in the parameters to be used in the analysis and that the model has a normal distribution.

164

Mercan and Özpençe

Tab. 6:  Bounds Test. Source: Authors’ elaboration. K

F statistics

1

28.0454

Critical Value at 1% Significance Level Lower Upper Bound Bound 6.84 7.84

Critical Value at 5% Significance Level Lower Upper Bound Bound 4.94 5.73

Critical Value at 10% Significance Level Lower Upper Bound Bound 4.04 4.78

In the analysis, it is observed that the F statistical value is higher than all values, with a significance of 1%, 5%, and 10% compared to the critical value. If the F statistical value is smaller than the upper bound values of essential values, it means that there is no co-integration relationship between economic growth and unemployment. However, since the F statistical value is higher than the upper bound values of all critical values, it is determined that there is a co-integration relationship between economic growth and unemployment. Tab. 7:  Error Correction Model. Source: Authors’ elaboration. Variables ΔU ECM

Coefficient -2.1352 -1.0878

T statistics -3.0665 -7.4218

Prob 0.0044 0.0000

The ECM coefficient in Tab. 7 indicates how much of the short-term shocks will be eliminated in the long term. As expected, the ECM coefficient should be negative. The value of the error correction coefficient indicates that the shocks that occur have disappeared in less than one year. According to the findings obtained in the ARDL method, it is concluded that the variables are co-integrated in the short and long term. On the other hand, long-term coefficients cannot be interpreted statistically. In this context, Regression analysis was performed to reveal the effects of variables on each other more concretely. Two models were created within the scope of Regression analysis. In the first model, unemployment is the dependent variable, while in the second model, economic growth is chosen as the dependent variable (Tab. 8). Data and results for models are shown below: Model 1: ΔU = 0.646798537143 - 0.115973233152*GDP Model 2: GDP = 4.94447041473 - 2.28545768787*ΔU

(-3.5016) 1 (0.0013) (-3.5016) 2 (0.0013)

165

Okun’s Law: Turkey Case Tab. 8:  Diagnostic Test Results. Source: Authors’ elaboration. Model 1 F-stat. Model Data 12.2617 Breusch – Godfrey Serial Correlation LM Test    1.0090 Heteroskedasticity Test: White Test    1.1210 Ramsey Reset Test    0.0105

Prob 0.0013 0.3759 0.2972 0.9188

Model 2 F-stat. 12.2617 0.1162 0.0019 0.3475

Prob 0.0013 0.8906 0.9650 0.5595

The aim of this study is to investigate the validity of the Okun’s law with annual data for the Turkish economy between 1980 and 2016. In this study, it was appropriate to use the ARDL co-integration method because the time series was stable at different levels. As a result of the study, the relationship between economic growth and unemployment emerged as co-integrated. In other words, economic growth and unemployment are acting together in the long term. This result states that a change in economic growth will lead to a change in unemployment. However, the coefficient of this change could not be found because the long-term coefficient could not be interpreted statistically. Therefore, Regression analysis after ARDL co-integration analysis was deemed appropriate. The reason for the Regression analysis is to see the statistically more tangible results of the relationship between economic growth and unemployment. In this context, a 1% growth in economic growth by Regression analysis reduces the unemployment rate by 0.11%. Later, the impact of unemployment on economic growth was analyzed. According to this analysis, a 1% increase in unemployment reduces economic growth by 2.28%. From this point of view, the negative impact of unemployment on economic growth is more significant than the effects of economic growth on unemployment. According to the analyses, it is concluded that the Okun’s Law is partially valid in Turkey.

5 Result and Evaluations According to the analysis, although Turkey is a developing country, the growth ratios are not affected sufficiently by unemployment. In this case, it is seen that a solution to the problem of unemployment cannot be produced only by economic growth. Policymakers should resort to different policies when addressing the issue of unemployment. Because the economic growth in Turkey is seen to be more capital-intensive growth. It is known that such growth does not contribute much to unemployment. It is possible to see this during periods when

166

Mercan and Özpençe

high growth does not reduce unemployment to the desired level. In this respect, it is essential that Turkey increases its importance to the unemployment problem and produces more permanent solutions, not palliative. It is essential to increase employment to solve the problem of unemployment. Employment created in this way will gradually reduce unemployment. Vocational training courses to be given by local governments can also be useful in reducing unemployment. However, opening courses to specialize in various professions of local governments and operating effectively here is another important step to be taken. Although this practice is carried out activities within the İş-Kur today, the participation rate is not at the desired level because it is not announced much. Studies show that many people are not aware of these courses. Another important step towards reducing the unemployment rate is to remove the overburden on the minimum wage. The tax burden on the minimum wage is a high cost for the employer. By reducing this cost, the employer can stop layoffs and take in more workers. Recruitment in this way can lead to significant reductions in unemployment rates in the long run. Improvements in education could also yield positive results in unemployment. For example, the cooperation of vocational high schools with the industries allows them to be more qualified and learn the job better. Implementation of this in universities are relevant policies that must be implemented to give more positive results in order to reduce youth unemployment.

References Acar F. (2013). “Türkiye Ekonomisine Genel Bir Bakış (2001 – 2013)”, ÇSGB Çalışma Dünyası Dergisi, 1(2), 15–32. Ahmad K., Khalil S., Saeed A. D. (2011). “Does There Exist Okun’s Law in Pakistan?”, International Journal of Humanities and Social Science, 1(12), 293–299. Akkaya Y., Gürbüz R. (2012). “Ekonomik Büyüme ve İşsizlik Üzerine…”, Türkiye Ekonomi Kurumu Tartışma Metni 2012/79, (http://www.tek.org.tr) (24.05.2017). Altuntepe N., Güner T. (2013). “Türkiye’de İstihdam – Büyüme İlişkisinin Analizi (1988 – 2011)”, Uluslararası Alanya İşletme Fakültesi Dergisi, 5(1), 73–84. Ayhan F. (2008). “İşsizlik ve İktisadi Büyüme İlişkisinin Türkiye Üzerinde Analizine Yönelik Bir Uygulama”, Balıkesir Üniversitesi Sosyal Bilimler Enstitüsü, (Yayımlanmamış Yüksek Lisans Tezi), Balıkesir. Ball L., Leigh D., Loungani P. (2013). “Okun’s Law: Fit at 50?”, IMF Working Paper, 13(10), 1–39.

Okun’s Law: Turkey Case

167

Bulut Ü. (2016). “Ekonomik Büyüme ile İşsizlik Arasındaki Asimetrik İlişki: Türkiye Örneği”, https://www.academia.edu/29415055/ EkonomikBüyüme_ile_İşsizlikArasındaki_Asimetrik_İlişki_ Türkiye_Örneği, (07.06.2017). Ceylan S., Şahin B. Y. (2010). “İşsizlik ve Ekonomik Büyüme İlişkisinde Asimetri”, Doğuş Üniversitesi Dergisi, 11(2), 157–165. Göçer İ. (2015). “Okun Yasası: Türkiye Üzerine Bir Uygulama”, Uluslararası Ekonomik Araştırmalar Dergisi, 1(1), 1–12. Harris R., Silverstone B. (2000). “Asymmetric Adjustment of Unemployment and Output in New Zealand: Rediscovering Okun’s Law”, University of Waikato Department of Economics Working Paper in Economics, 2, 1–23. Huang H. C., Yeh C. C. (2013). “Okun’s Law in Panels of Countries and States”, Applied Economics, 45, 191–199. Khaliq S. A., Soufan T., Shihab R. A. (2014). “The Relationship between Unemployment and Economic Growth Rate in Arab Country”, Journal of Economics and Sustainable Development, 5(9), 56–59. Kılıçbay A. (1984). Türk Ekonomisi: Modeller, Politikalar, Stratejiler, Ankara: Türkiye İş Bankası Kültür Yayınları. Köse Z. (2016). “Türkiye Ekonomisinde 2003 – 2014 Döneminde Ekonomik Büyüme İşsizlik ve Enflasyon İlişkisi”, Türk Sosyal Bilimler Araştırmaları Dergisi, 1(1), 58–76. Mıhçı S., Atılgan E. (2010). “İşsizlik ve Büyüme: Türkiye Ekonomisi için Okun Katsayıları”, İktisat İşletme ve Finans Dergisi, 25(296), 33–54. Moazzami B., Dadgostar B. (2009). “Okun’s Law Revisited: Evidence from OECD Countries”, International Business & Economics Research Journal, 8(8), 21–24. Muratoğlu Y. (2011). “Büyüme ve İstihdam Arasındaki İlişki: Türkiye Örneği”, International Conference on Eurasian Economics, SESSION 2C, 167–173. Okun A. (1962). “Potential GNP: Its Measurement and Significance”, https://pamukkaleuniv.on.worldcat.org/search?databaseList= 2375 %2C2264 %2C2087 %2C1271 %2C2261 % 2C173 %2C2260 %2C1910 % 2C1855 %2C3218 %2C1953 % 2C2237 %2C2259 %2C2897 % 2C1653 %2C137 %2C3313 %2C1079 %2C239 %2C1708 %2C1609 %2C638 %2C2507 &queryString=ti%3APotential +Gnp%3A+Its+ Measurement+and +Significance#/oclc/66101485, (09.06.2017). Palombi S., Perman R., Tavera C. (2015). “Commuting Effects in Okun’s Law Among British Areas: Evidence from Spatial Panel Econometrics”, Papers in Regional Science, 96(1), 191–209. Parasız İ. (2003). Türkiye Ekonomisi, Bursa: Ezgi Kitabevi.

168

Mercan and Özpençe

Silvapulle P., Moosa I. A., Silvapulle M. J. (2004). “Asymmetry in Okun’s Law”, The Canadian Journal of Economics, 37(2), 353–374. Sinclair T. M. (2007). “The Relationships between Permanent and Transitory Movements in U. S. Output and the Unemployment Rate”, Journal of Money, Credit and Banking, 41(2–3), 529–542. Takım A. (2010). “Türkiye’de Ekonomik Büyüme ile İşsizlik Arasındaki İlişki: Granger Nedensellik Testi”, Dumlupınar Üniversitesi Sosyal Bilimler Dergisi, 27, 1–8. Tarı R., Abasız T. (2010). “Asimetrik Etkiler Altında Okun Yasasının Eşik Hata Düzeltme Modeli ile Sınanması: Türkiye Örneği”, İktisat İşletme ve Finans Dergisi, 25(291), 53–77. Uysal D., Alptekin V. (2009). “Türkiye Ekonomisinde Büyüme – İşsizlik İlişkisinin Var Modeli Yardımıyla Sınanması (1980 – 2007)”, Dumlupınar Üniversitesi Sosyal Bilimler Dergisi, 25, 69–78.

Aslıhan Özel Özer, Buğra Özer, and Sercan Akın

An Evaluation of Subsidies Granted to the Private Educational Institutions within the Framework of Turkish Tax System1 Abstract: One of the essential concerns that governments have to deal with has been the theme of economic growth and development. In order to be able to sustain the objectives mentioned above, states and governments instrumentalize fiscal policies in which subsidies occupy a substantial place. The primary rationale behind these fiscal policies, including grants, has been the allocation of resources to those fields with better and more efficient prospects within the general good of the economy. Despite convergences seen in terms of types and implementation of subsidies, the basic objective is to accomplish higher rates of economic growth and investment.Through the Decision of the Council of Ministers of the Turkish Republic dated June 19th, 2012, investments to be handled for primary, secondary, and high school educational institutions were evaluated within the framework of the fifth region with the labeling of priority investment. Along with the closure of private-mentoring facilities, the related facilities investors were foreseen to utilize the subsidies to convert these facilities to schools, thereby minimizing the costs of investments coupled with rises in investments. The effort of the study, given the given scope and framework, is to elucidate and to analyze arrangements and recent developments concerning grants of space and location for investments and exceptions regarding the insurance and tax exceptions and exemption within a general framework of aforementioned subsidy program in Turkey for educational institutions. Keywords: Incentives, Education, Tax Exemption JEL Codes: H71, H52, H26, I22

1 Introduction Governments’ grants and subsidies provide the private sector and related actors in the name of economic development and growth. As the concerned grants may vary from location to another, these grant schemes may differ based on a sectoral basis as well. 1 This study has been abridged and developed from the MA project titled “ A General Evaluation in the Field of Subsidies in Education Sector and The Subsidies in Turkish taxation system” by Sercan Akın supervised Assistant Professor Aslıhan Özel Özer PhD.

170

Özer and Özer

Education as a field has a quintessential role to play from contributing to the making of the qualified labor force as a component of the making of human capital. While the public sector is predominantly involved in the private sector, the private sector is also actively indulged in the related field. The subsidies and grants program oriented at the private sector actors in the education field strengthen the private sector partner, thereby contributing to a more effective and quality provision of services along with the realization of higher rates of economic growth and development. A review of implementations and applications in the education fief characterized with positive externalities in particular with emphasis on regulations directing at reinforcement of private investments and analysis of the aforementioned schemes shall contribute to the development of the private sector in education. This work has the objective to present a literature review on the implementation of the private education sector within a context of Turkish taxation system along with elaboration on different aspects of the scheme The effort shall be to provide information on the fundamental regulations regarding the subsidies granted to the private sector in Turkey while putting elaborations in the education fields with emphasis on private partners’ interactions and their future investments.

2 Concepts in Regards to Subsidies and Education System and Evaluations The concept “subsidies” refer to “material and/or non-material assistance and encouragements with the aim of a faster and a more abundant provision of certain economic activities compared to other provision frameworks” (İncekara, 1995: 9). Subsides do not only develop the provision of economic activities but also contribute to coping with regional economic in inequalities, henceforth contributing to the national economic development on an economic basis. Given the default perspective, subsidies may be defined to be “a summation of supports given by governments with the objective of creation of new employment opportunities in certain sectors and on specific regions and increasing the life quality” (Karabıçak, 2013: 265). The final objective of the related schemes has been to back economic growth. There exists a direct relation between subsidies and economic growth. By means of economic growth, governments seek to achieve a higher rate of ­employment, create a broader basis of taxation, and to increase the welfare level (Eser, 2011: 12).

Subsidies Granted to the Private Educational Institutions

171

The primary characteristics of subsides are respectively: a) They are granted by governments, b) they may also be given to state enterprises as well as the private counterparts, c) they impose a cost on government leading to decreases in public funds through cash grants, low interest credits and grants and causing the level of public revenues to roll back due to the granting of state revenues whose source is the taxation, d) the subsidies may be a benefit for the private counterparts while they may be attributed as receding public revenues and funds, e) subsidies are used in accordance with the content, region, sector, magnitude and timing factors, f) subsidies and grants may be direct or indirect, g) they may be hidden or open (Duran, 2003:6). This being asserted, among the different subsidies classification, tax-based subsidies have been the most popular one with examples of implementation in many international examples (Giray, 2008:95). The term subsidies may be able to be explained in terms of their objectives. The term may be defined as ad economic-aimed, producer-based subsidies, transfer payments, premiums, credit with favorable conditions (Küçüktürkman, 2007:  61). Succinctly speaking, to be able to realize economic growth and to increase employment sources, to maintain economic stability, all measures taken by governments in economic, social, legal, and financial terms may be summarized as subsidies (Tuncer, 2008: 191). Subsidies in the form of in-kind, cash, and tax-related have served for different purposes throughout history with various usages. Many governments have aimed to encourage investments using subsidies, tax-rebate, and direct guarantees programs (Thomas, 2007: 1). By the second mid 980s, subsidies have tremendously increased along with the fact that both developing and developed countries have taken advantage of the related schemes (OECD, 2002: 169). The fact that markets do not operate automatically, thus leading to market failures, is also valid for many countries where income distribution is less than just. When market dynamics left alone in countries where the necessary economic allocation of production, consumption, and redistribution mechanism are left to the mercy of price mechanism and market conditions, the redistribution schemes are expected to be just either. Governments attempt to carry out a more just economic redistribution through fiscal policies by influencing social befit of public expenditures and distribution of tax burden (Aksoy, 2011:45–1). Several criteria are taken under consideration whist classifying the subsidies being respectively (Duran 2011:19): 1. Objective-Based Subsidies:  they may be in the form of influencing of increasing investment and production, supporting exports, gaining international competitiveness, decreasing regional disparities, attracting

172

Özer and Özer

foreign direct investment, increasing Research and Development Activities, reinforcing Small and Medium-sized Enterprises, finishing the unfinished investments. 2. Content-Based Subsidies:  General  –Aimed and Particular-Based Subsidies are to be mentioned. While the initial one refers to the implementation identical level of subsidies being applied to all sector with the very same rates including tax exemptions and value-added-tax exemptions, the latter one refers to the exemptions granted to the particular regional and firm based privileges with the concrete example of exemptions given to the investment on the research and development activities based on certain conditions. 3. Objective-Based subsidies correspond to those subsidies that based on investment and enterprise based along with pre-investment subsidies. 4. Instrument-Based Subsidies are those kinds that refer to subsidies in kind, in cash taxation one and state guarantees. 5. Resource-Based subsidies are those aimed at profit/revenue-based, capital investment based, labor-based, value-added bases and import and export-based ones. Most contemporary public expenditures are realized through taxation. Tax systems provide necessary investment for public expenditures. Taxes play quintessential roles in the redistribution of income and the contribution to increases in savings and capital accumulation strategies’. While trading off the ­tax-based income, the objective has been to realize economic and fiscal policies (Bıyık, 2001:4) Tax based subsidies with different implementation at different international settings have contributed to increases in savings via increasing disposable income through reductions income rates, and these savings have been diverted to the savings via tax-based subsidies and added up to the expansion of savings volume. Indeed the impact of these schemes had been positive upon to the increases of investments thereby with high growth and development rates in the less developed and the developing countries (Siverekli, 2003: 105) The aims of taxation based subsidies may be juxtaposed as follows a) Increasing savings and investments b) diverting investments to specific fields and regions c) alleviating the negative impact of taxation on economic decisions d) maintaining economic stability and decreasing the economic impact of inflationist pressures e) supporting employment e) maintaining a just economic redistribution f)sustaining justice in the making of tax burdens f) contributing to a higher level of international competitiveness rate for foreign currency rates (Giray, Koban and Gerçek, 1998:10).

Subsidies Granted to the Private Educational Institutions

173

The forms of taxation based subsidies may be in the form of reduction in tax rate, investment bonuses, investment credits, tax holiday, tax postponement, accelerated depreciation, exceptions and exemptions, reduced import and tax applications, VAT support, loss cuts, preferential treatment for long term capital gains, definite discount rules, reduced withholding tax, employment-based cuts, tax credits, reduction in property taxes.

2.1 Education as a Concept and Educational Expenditures Many definitions and explanations have been put forth by different scholars throughout history to locate the term education. While Plato place emphasizes the education’s quintessential role in the making of Excellency for the mind and body, Cicero stresses the process of how education leads to the making of humans. One of the related definitions regarding the term is to create a process of change in human behavior along with the desired direction with a given objective (Ertürk, 1982: 2). The enlightenment Age witnesses the usage of the term derived:  from the Latin word “educere” for the cultivation of animals and plants. The term has come to be applied for humans due to the cognitive traits inherent in the concept. While the intelligentsia has come to be identified with the term, the term has come to denote a mature level of moral, physical and character development. Accordingly, education as a concept has a connotation of intellectual physical and moral excellence and its very full formation (Tozlu, 1997:93). Along with these usages, term education has come to mean to different things as well. Depending upon the field in which the term is used, the term has differentiated itself. The term has different relative connotations with related limitations and ambiguities. While the daily usage of education has encompassed a wide array of fields with flexible and broad openings, the term possesses a totality and a core of itself where these broad definitions are derived. Henceforth, one should trace the term to its kernel and root (Yılmaz, 2000:19). Knowledge, knowledge production, and education come to occupy an essential place in modern societies. Education is inextricably linked to knowledge and knowledge production processes using providing terrain for such production, the upbringing of generations for the related schemes, the spread of knowledge via different channels (Cerit, 1997: 64). The term education in Turkey has come to mean different sets of meanings corresponding to terms maarif, tedrisat, and terbiye as corresponding to Ottoman Turkish (Başaran, 1984:14). The term education as under the term

174

Özer and Özer

eğitim is inclusive of the aforementioned terms gaining manners, nurturing, fostering, science, discipline.

2.1.1  Types of Education By the legislation No:1739 having come to affect in the Official Newspaper of Turkey No: 14574 dated June 24th, 1973, education came to be defined in two basic categories formal education and standard education, both being complementary and cooperative. The former one is inclusive of primary, secondary, high school, and higher education as processes while the latter type comes to include the kind of education for the people that could not have access to educational opportunities. Those that dropped out of formal education and schooling opportunities, for the people that would like to take advantage while being involved in the legal, educational process and for those groups that would like contribute to their occupational competencies.

2.1.2  Investment Expenditures Educational services have positive social, economic, and political positive externalities in society. Along with the increase of these positive externalities, the quality of education has come to be associated with the related term. The educational services add up to increases in productivity, political stability, social and cultural development, and efforts of industrialization efforts to yield (Şener, 2001:  356–357). Due to these characteristics, education is classified as semipublic goods, and the marginal benefits of these services are lower than the marginal social benefits. Provided that these services are solely left to the market, there will be an under-production phenomenon that requires a bailout by the government through the general budget (Madanoğlu, 1992: 59). In addition to the direct funding of education, there are other types of financing, such as partial and indirect funding programs. In direct financing, the beneficiaries are asked to pay for the courses that students attend to finance certain items in services. The indirect financing, in the meanwhile, refers to the introduction of private partners with granted certain initiatives in the educational sector while the government carries out the public provision of these services (Devrim, Tosuner, 1987: 86–87). The voucher called system is the transfer of a voucher to the beneficiary, which gives a chance for the students to make use of private educational services (Stiglitz, trans.: Batırel 1988: 463–464). On the other hand, educational expenditures facilitate the redistribution of income, the realization of the economic growth, sustaining economic growth which all sum up to the making of investment expenditures (Mutlu, 1997: 249–250).

Subsidies Granted to the Private Educational Institutions

175

3 Subsidies in the Turkish Education System The economic development of a country is interwoven with the personal and social development of a country. By means of attitudes and behaviors gained in the socialization processes via the educational channels, economic growth may be facilitated by such methods (Kaya, 1984:  10). The provision of educational services may be realized by both the public and the private sector. While the private sector may face different challenges in the provision of these services, the public sector may provide remedies to overcome the related problems utilizing subsidies schemes. According to the decision of the Council of Misters dated June 19th, 2012, Number 350, the private sector has come to be supported with different subsidies and exceptions, which the study shall address in the following sections.

3.1 Revenue Exception for the Private Education and Teaching The exception to special education and training is regulated in Income Taxation Law No:  193 Article No:20 and Corporation Taxation Law No:  5520 Article No: 5/1-ı as amended by Corporation Taxation Law Number 5528. This exception have been defined as follows “revenues gained from preschool, private primary school and private secondary school corporations, pending upon the consent of the related ministries, shall be subject five-term-exemption within the framework of regulation and rules to be determined by the Ministry of Finance and exemptions shall commence in the aftermath of the very date of educational services’ beginning and the exemptions to be valid for the subsequent five intervals of taxation” (GTL Art.  20; VATL Art.5/1-ı; GTGL 254)  (5228 provisional Law Article. 1).

3.2 Exception for Research and Development Activities By means of this exception, all players in the process of research and development processes were to be supported. From the perspective of scientific expenditures approach, the ration of research and development of expenditures to GDP and the number of researchers to the total population rates have been significant indicators for economic development (Kızılot, 2000:398) According to the Corporation Taxation Law No:  5521 Article No:1 Clause No: 1 Paragraph, those corporations under the condition of not being an exclusively defined research and development corporations along with other activities carried out shall have the chance of expenditures of research and development activities realized for the objective of creation of new technologies within the

176

Özer and Özer

R&D departments within the company exclusively by a 40% amount (byte Legislation Number February 2nd, 2008 Article Number 5 the rate was determined by 100%) from the revenue earnings under the item of R&D deductions. Yet this provision was annulled from the legislation Number 6728 Article Number 58 dated July 15th, 2016.

3.3 Value Added Tax in Educational Service In the Value Added Tax (VAT) Law, the subject of VAT has been defined in the very first article of the related legislation as those activities: • Services contracts realized within the framework of commercial, industrial and agricultural services • Import of goods and services of any kind. • Contracts and services arising from other activities The casual factor about the formation of VAT comes to emerge in different according to its types. In terms of interactions subject to be taxation, the causal events leading to the formation of VAT have been juxtaposed in the VAT Law (Article 10). As a rule, the main causal event in the structure of private educational services is the completion of the service. The education in the private educational sector is carried out over a while, namely the school year. Henceforth, the primary causal factor for the formation of VAT is realized at the end of each schooling year (Yılmazcan, 1997:82). Meanwhile, the private education sector subject to the Law No: 625 has also been given the right for tax exception for the free educational and teaching services delivered bona fide with the condition that they do not exceed over 10% of their capacity (KDVK Art.17/2-b).

3.4 VAT Rates in Private Education Sector Several definitions have been forth for educational corporations that are ­regulated by Law No: 5580 Private Education Corporation Laws Article No: 2. By means of different amendments, Amendment No: 26742 dated December 30th 2007, by means of list Number II mentioned in the decision of the Council of Ministers” universities and higher education facilities” and e­ ducation and teaching services given by the private education corporations mentioned in the Law 5580 Private Education Corporations, Law Number 2282 Social and Children Services Law and governmental decree Number 576 for the Private Education Corporations and transportation and transfer s­ervices given for the students as indicates by the “Bylaw for School Transpiration Services and the accommodation and

Subsidies Granted to the Private Educational Institutions

177

dormitory services m ­ entioned in the “Private School Accommodation Bylaw”, the VAT rate has been determined as 8%.

3.5 Investment Subsidies for Investments to be Realized by Private Educational Corporations 2012/3305 Decision of the Council of Ministers with the date of June 6th, 2012 No:  28328 stipulates while regulating the “Government Assistance in Investments” that those investments realized by kindergarten, preschool, primary, secondary, high school schoolings shall be deemed as education investments. (Supplementary Table  2/a). Irrespective of the provinces where these investments would be carried out, the investments would take advantage of the 5th Zone, and they shall be subject to 6th zone investments should they carry out the concerning investments in the 6th Zone”. In those provinces where the investors would be investing educational investments, within the framework of regional subsidies implementations, investors shall also be granted land and/or provide the location themselves the necessary area, whereby components of subsidies were juxtaposed in the fourth article third paragraph of the legal regulation. These are: • • • • • • • •

Tariff exemption VAT Exemption Tax Rebates Share of Employer Support for Insurance Premiums Location  Grants Interest Support Income Tax Support (For the 6th Zone Premiums) Insurance Premium Support (For the 6th Zone Premiums)

In order for these components to be used for the first and second regions, minimum investment levels of 1  million Turkish lira are required. The minimum investment amount for the third, fourth, fifth and sixth regions should be 500.000 Turkish Liras.

3.5.1  Tariff Exception Most of the equipment required in the Investment Subsidy License can be purchased domestically while they can be supplied from abroad. Most imported machinery and equipment lists that can be seen in the license mentioned will be provided without any tariff levied according to the provision stated in the

178

Özer and Özer

decision stipulating all machinery and shall not be subject to any tariff whatsoever clause.

3.5.2  VAT Exemption The Turkish private education sector is specified and defined within the II article of the Legislation Number 5580 Private Education Enterprises. According to the stated legislation, preschool, primary school, secondary private education institutions, private courses distance education enterprises, in-house training units student study centers, private education, and rehabilitation centers are included within the stated law (Ozansoy, 2008: 105). According to the VAT Law in Turkey dated October 10th, 1984, Number 305 states that those investors with subsidies license to purchase machinery and equipment shall also be granted VAT exemption during the purchases of this item. On the other hand, all transfers of machinery and equipment within the content of subsidies license content shall also be subject to the exemption while the exemption is also valid for the subsets of equipment and machinery listed in the license Fixed investment limits over 500 million TL shall also be considered as a strategic investment with exemptions granted for expenditures of construction for the infrastructure (Nr. 28328, 2012/3305 Dcs. Nr., Art. 10).

3.5.3  Tax Rebate Along with the VAT Law in Turkey dated October 10th, 1984, Number 305 Article 32/A stipulates for the content of regional subsidies implementations, revenue, and corporation taxes shall be subject to 70% deduction until the amount reaches to the foreseen subsidies contribution rate. According to the subsidies documents required by the subsidies decision until the very date December 31st, 2014 (inclusive of this very exact date) and provided that investment process has commenced the subsidies support will vary from 80 percent to 40 percent (Nr. 28328, 2012/3305 Dcs. Nr., Art. 15).

3.5.4  Insurance Premium Employers’ Share Support In the regional support programs, upon the completion approval realized subsidies license under the condition that required employment is not exceeded • For the completely new investments realized with investment subsidies licenses • In the other types of investment in the aftermath of completion of investment, or in the period six months before the end of the investment (for the seasonal

Subsidies Granted to the Private Educational Institutions

179

investment the very recent year’s averages are taken under consideration) for the premiums and service documents passed to the Social Security Institution depending on the average number of workers Premiums support shall also be granted in favor of employers based on the minimum wage levels from the budgetary allocations granted to the Ministry. The concerning support program shall be given for the projects that have started since December 31st, 2012, for years and five years for those that have started after January 1st, 2015. The benefitted insurance premium support levels shall not surpass 25% of the investment (Nr. 28328, 2012/3305 Dcs. Nr., Art.12).

3.5.5  Investment Location Support In the related decision document dated June 29th, 2001, Number 4706 supplementary third article corporations may be supported with investment land and location grants for those investment project licenses (Nr. 28328, 2012/3305 Dcs. Nr., Art.16).

3.5.6  Interest Support In the case of demands for investments realized through regional support programs and strategic aids along with the content of R&D and environmental projects, the credits to be utilized from banks with one-year period interests shall be supported with 70 percent of fixed investment rates for the interests incurred. For the Turkish lira based interests within the 5th Zone Investment region up till 5 points shall be bailed by the government while for foreign currency found credit 2 points shall be supported within the same region. Yet the support limits will not exceed over TL700000 for the interest support. For the 6th Zone of Investments, the Turkish lira based support will be around 7 percent while the foreign currency based credit support shall not be over 2 points with limits not to exceed over TL 900.000 (Nr. 28328, 2012/3305 Dcs. Nr., Art.11).

3.5.7  Income Tax Support (For Investments in the 6th Zone) For the 6th Zone of Investments as indicated in the Investment Subsidies Decision, for the extra employment provided that the amount shall not exceed over the recorded employment level income tax for the workers calculated over the minimum wage amounts shall be not to taxation over the declaration to be given in the aftermath of 10 years of completion of the investment Project fully or partially (Nr. 28328, 2012/3305 Dcs. Nr., art. 14).

180

Özer and Özer

3.5.8  Insurance Premium Support (For Investments in the 6th Zone) For the 6th Zone of Investments as indicated in the Investment Subsidies Decision, for the extra employment provided that the amount shall not exceed over the recorded employment level insurance premiums paid to the Social Security Institution shall be funded by the Ministry in the aftermath of 10 years of completion of the investment Project fully or partially. (Nr. 28328, 2012/3305 Dcs. Nr., Art.13).

4 Concluding Remarks Education is a sector that plays a quintessential role in the lives of individuals with its semi-public good characteristics. The private sector plays within the game in the appendix to the public sector’s involvement. The increase in competitiveness and quality increases in education have come to be two championing claims of the private face of the education sector. In the mid of these processes, tax systems are multitasking schemes. One of these tasks is to increase public finance for public expenditures and contribute to the enhancement of societal benefit. Yet a commodification of the education system shall be hazardous to the equality of opportunity. In the name of supports given to the private sector, the government should be regulating the subsidies given the impact of social and economic equilibrium to the fifth zone of investment within the regional support programs under the title of “Government Subsidies in Investments” June 9th 2012, 23138 Number3305, with an elaboration of VAT exemptions, tariff reductions and exemptions, tax rebates ınsurance premium supports, employers support, interest supports along with Zone & implementations. The effort has also dealt with matters including earnings exemption research and development exemptions and other VAT exemptions. It is without any doubt that the education sectors need to be supported for quality increases in content and quality. It will be of utmost importance that related support and subsidies programs be regulated and monitored closely, which constitutes a subject for a new study.

References 193 Sayılı Gelir Vergisi Kanunu (31.12.1960) Resmi Gazete (Sayı: 10700). http:// www.mevzuat.gov.tr/MevzuatMetin/1.4.193.pdf, (05.05.2019). 254 Seri No’lu Gelir Vergisi Genel Tebliği. (25.11.2004). Resmi Gazete (Sayı: 25651). http://www.gib.gov.tr/node/87689, (05.05.2019).

Subsidies Granted to the Private Educational Institutions

181

1739 Sayılı Milli Eğitim Temel Kanunu. (14.06.1983). Resmi Gazete (Sayı: 14574). http://www.mevzuat.gov.tr/MevzuatMetin/1.5.1739.pdf, (05.05.2019). 2012/3305 Karar Sayılı Yatırımlarda Devlet Yardımları Hakkında Karar. (19.06.2012). Resmi Gazete (Sayı: 28328). http://www.resmigazete.gov.tr/ eskiler/2012/06/20120619-1.htm, (07.05.2019). 3065 Sayılı Katma Değer Vergisi Kanunu. (25.10.1984). Resmi Gazete (Sayı: 18563). http://www.mevzuat.gov.tr/MevzuatMetin/1.5.3065.pdf, (07.05.2019). 5228 Sayılı Bazı Kanunlarda ve 178 Sayılı Kanun Hükmünde Kararnamede Değişiklik Yapılması Hakkında Kanun. (16.07.2004). Resmi Gazete (Sayı: 25539). http://www.resmigazete.gov.tr/eskiler/2004/07/20040731. htm#1, (07.05.2019). 5520 Sayılı Kurumlar Vergisi Kanunu (13.06.2006). Resmi Gazete (Sayı: 26205). http://www.mevzuat.gov.tr/MevzuatMetin/1.5.5520.pdf, (07.05.2019). 5580 Sayılı Özel Öğretim Kurumları Kanunu. (08.02.2007). Resmi Gazete (Sayı: 26434). http://www.mevzuat.gov.tr/MevzuatMetin/1.5.5580.pdf, (07.05.2019). Başaran, İ.E. (1984). Eğitime Giriş. Sevinç Matbaası. Ankara. Bıyık, R., Kıratlı, A. (2001). Vergi Teşvikleri ve Korumaları. Maliye Hesap Uzmanları Derneği. Ankara. Cerit,Y. (1997). Bilgi Toplumu ve Bilgi Üretiminde Yükselen Değer: Eğitim, Milli Eğitim Dergisi. Milli Eğitim Bakanlığı Yayınları, Ankara, 135, 64–67. Devrim, F., Tosuner, M. (1987). Türkiye’ de Eğitim Hizmetlerinin Finansmanında Son Gelişmeler, 3. Türkiye Eğitim Maliye Sempozyumu Bildiri Kitabı, İstanbul. Doğan, M. (1996). Büyük Türkçe Sözlük. İz Yayıncılık. İstanbul. Duran, M. (2003). Teşvik Politikaları ve Doğrudan Sermaye Yatırımları, T.C. Başbakanlık Hazine Müsteşarlığı Ekonomik Araştırmalar Genel Müdürlüğü, Araştırma ve İnceleme Dizisi No: 33, Ankara. Ertürk, S. (1982). Eğitimde Program Geliştirme. Meteksan Lmt. Şti. Ankara. Eser, E. (2011). Türkiye’de Uygulanan Yatırım Teşvik Sistemleri ve Mevcut Sistemin Yapısına Yönelik Öneriler, (Uzmanlık Tezi). DPT. Ankara. Giray, F. (2008). Vergi Teşvik Sistemi. Ezgi Kitabevi. Bursa. Giray, F., Koban, E., Gerçek, A. (1998). Avrupa Birliği ve Türkiye’de Yatırımlara Yönelik Vergi Teşvikleri ve Karşılaştırmalı Değerlendirmesi. BOSYÖD. Bursa. İncekara, A. (1995). Türkiye’de Teşvik Sistemi (Genel Değerlendirme). 10. Baskı, İstanbul Ticaret Odası. İstanbul.

182

Özer and Özer

Karabıçak, M. (2013). “Türkiye’de Uygulanan Ekonomik Teşvik Politikalarının Boyutu, Ulusal, Bölgesel ve Yerel Kalkınma Üzerine Olası Etkileri”, Süleyman Demirel Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi, C: 18 (3), ss. 263–280. Kaya, Y. (1984). İnsan Yetiştirme Düzenimiz. H.Ü. Sosyal ve İdari Bilimler Döner Sermaye İşletmesi Yayını. Ankara. Kızılot, Ş. (2000). Kurumlar Vergisi Kanunu ve Uygulaması. Yaklaşım Yayınları. Ankara. Küçüktürkman, U. (2007). “Teşvik Politikaları ve Doğrudan Sermaye Yatırımları, Sektörlere Etkileri ve Etkilerinin Kıyaslanması Üzerine Bir Araştırma Teşvik Politikaları Ve Doğrudan Sermaye Yatırımları, Sektörlere Etkileri ve Etkilerinin Kıyaslanması Üzerine Bir Araştırma”. Dokuz Eylül Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 9 (1), 61–79. Mutlu, A. (1997). Türkiye’de Konsolide Bütçe Harcamalarının Değerlendirilmesi 1981-1985, Türkiye’de Kamu Ekonomisi ve Mali Kriz, XXI. Türkiye Maliye Sempozyumu, Maliye Araştırma Merkezi, Yayın No: 83. Nadaroğlu, H. (1992). Kamu Maliyesi Teorisi. 10. Baskı, Beta Basım Yayım. İstanbul. OECD. (2001). Corporate Tax Incentives for Foreign Direct Investment, OECD Tax Policy Studies, No. 4. Ozansoy, A. (2008). Türk Vergi Hukukunda Eğitimin Vergilendirilmesi, (Yüksek Lisans Tezi). Gazi Üniversitesi SBE. Ankara. Stiglitz, J. E. (1982). Kamu Kesimi Ekonomisi. Çeviren: Ömer Faruk Batırel, Marmara Üniversitesi İktisadi ve İdari Bilimler Fakültesi Yayınları, Yayın No: 396. Şener, O. (2001). Kamu Ekonomisi. Beta Basım Yayım. İstanbul. Thomas, P. (2007). Investment Incentives: Growing Use, Uncertain Benefits, Uneven Controls, For the Global Subsidies Initiative (GSI) of the International Institute for Sustainable Development (IISD), Geneva, Switzerland. Tozlu, N. (1997). Eğitim Felsefesi. MEB Yayınları. İstanbul. Tuncer, S. (2008). Türkiye’de Vergi Teşvikleri ve Uygulaması (Genel Değerlendirme). Yaklaşım Yayıncılık. Ankara. Yılmaz, M. (2000). “Türk Atasözlerinde Eğitim Anlayışı”. Ankara Üniversitesi Eğitim Bilimleri Fakültesi Dergisi, Cilt 33, Sayı 1–2, 137–145. Yılmazcan, D. (1997). “Okul, Dersane ve Kurslarda Eğitim ve Öğretimde K.D.V.”. Yaklaşım Dergisi. S. 51, 79–85.

Özgür Biyan and Güneş Yılmaz

Artificial Intelligence: If It’s Taxed, But How? “Artificial intelligence will be either the best or the worst thing ever to happen to humanity.” Stephen Hawking

Abstract: In today’s world of rapid digitalization, the widespread use of artificial intelligence has reached such a level that it will have some consequences in terms of public finance. The change in employment policies due to the external factors resulting from the prevalent use of artificial intelligence, and therefore the possibility that budget revenues might be affected, has led to the discussions about the taxation of artificial intelligence. This study discusses the issues of how artificial intelligence can be taxed in accordance with the discussions going on about the same. The main point derived implies that it does not seem plausible that artificial intelligence could become a taxpayer as per the applicable legal system in force. Keywords: Artificial Intelligence, Robot tax, Digitalization, Taxation JEL Codes: D62, H23, K34

1 Introduction Non-stop growth of technology beginning with the invention of the steam engine continues tremendously owing to Industry 4.0 today, with direct effects on human living. Widespread use of the internet, fast-paced of robotization, and development of artificial intelligence have risen some question marks in the minds as to where the social and economic life would head. Internet of things (continuously connected to the net), smart cities, smart buildings, autonomous (self-driving) cars and, ultimately projects involving the participation of Al robot in the workforce; have been included in the agenda of the governments as the topics to be discussed regarding the future policies. All these developments, which we tried to outline briefly, have also been reflected in the field of public finance. Inevitable radical changes in the ways of work and business manners soon have led to the fact that employment, economic, and fiscal policies should be reconsidered. This study deals with such a massive increase in technology from the perspective of the “externality” theory as part of public finance. Although the term “technological development” sounds excellent, the discussions began whether it has created positive or negative externality, considering the results it created or it might result in in the future.

184

Biyan and Yılmaz

On the other hand, legal science has also had to accommodate itself to technological developments. Likewise, the changes in business models and working methods required the adoption of harmonized arrangements in terms of private and public law. The fact that whether artificial intelligence, related to our study and analyzed in great detail, can be given personhood and then be assigned a responsibility has been the main topic of most branches of law. Also, the tax law is dealt with as part of these discussions and thus has to produce solutions. The conclusions derived at the end of these discussions would lead to the clarification of the technological developments in the face of taxation regimes. Therefore, according to the questions concerning the discussions focusing on the core of this study, it will be necessary to make assessments as to whether the artificial intelligence should be taxed and based on the arguments made and seek solutions. Before we go ahead and address the discussions and our opinions about the taxation aspect of the study, the concept of artificial intelligence will be outlined, and the predicted results will be expressed, after which the research will address the suggestions on how to impose a tax, referring to the opinions in the doctrine.

2 Artificial Intelligence and Characteristics 2.1 Artificial Intelligence Conceptually The emergence of the concept “artificial intelligence”, formed with the combination of the individual terms artificial and intelligence, dates back to the 1940s. It was in the 1940s that McCulloch and Pitts attempted to express the concept of “intelligence” mathematically for the first time. In 1948, William Gray Walter built two small robots called “Elmer” and “Elsie” and enabled them to respond to obstacles they hit (Huang Z, 2018: 1818). In 1950, Alan Turing carried out some studies on computerized machines and intelligence. He indicated that communication could be established between man and machine, thanks to his experiment known as “Turing Test”. In the Turing experiment, he enabled a man to talk with both a man and a machine simultaneously; and he proposed that the computer passed the test when he noticed that the man could not distinguish between the device and another person. So, it was how the first steps were taken towards the evolution of artificial intelligence (AIR, 2018: 2). At the Dartmouth workshop in 1956, the term “artificial intelligence” was proposed, claiming that it was a discipline (DDAIT, 2018: 1818). First used by McCarthy in 1955, Artificial intelligence can be defined as “a machine capable of solving problems that individuals can solve with natural intelligence”. But, the part which underlies the concept of artificial intelligence

Artificial Intelligence: If It’s Taxed, But How?

185

with a broad insight, and is concerned with our study, is that the machines have learning skills. In other words, a machine with artificial intelligence corrects the errors or mistakes made by trial and error; namely, it can learn. Statistical models, created with inspiration from the neural networks in the human brain, comprise the artificial neural networks. These neural networks enable deep learning method can be implemented (Yüksel, 2018: 588–589). Artificial intelligence is a system that performs normal human intelligence functions such as perception, learning, development, creativity, communication, decision-making, conclusion, etc. (Zorluel, 2009: 308). Artificial intelligence is considered one of the leading events in the 4th stage of industrialization (Industry 4.0) (Marwala, 2018: 2). In another saying, the “Fourth Industrial Revolution” is begun with a new technological wave that had profound economic effects within the scope of closely-related features, such as robot dexterity, machine learning, processing power, and sensor capabilities (Ooi & Goh, 2019: 2).

2.2 Characteristics of Artificial Intelligence Artificial intelligence, which is used to express the techniques that render the machines “smart”, make use of automation by developing or reproducing the human intellect to improve the analyzing and decision-making capabilities of machines and enable them to perform research and implementation. It catalyzes structural transformation in various industries, offering the managers unprecedented opportunities and tools to facilitate the complexity of decision-making. Besides, it allows otherwise complicated and time-consuming tasks to be completed more effectively and efficiently (DDAIT, 2018: 1818–1819). The concept of artificial intelligence is a concept that refers to information systems inspired by biological systems and is accepted as an umbrella term involving several technologies such as deep learning, machine vision, natural language processing (“NLP”), and machine reasoning (AIR, 2018: 1). Artificial intelligence is divided into two. The first one is a reliable artificial intelligence (deep learning). This intelligence is implemented by imitating the human brain. A  technology that can think like a human is aimed. The second one is weak artificial intelligence (machine learning). This intelligence is, though, the one that performs predetermined movements based on rules (Yüksel, 2018: 591; Zorluel, 2009: 308). Artificial intelligence is a kind of computer program. It is a program that uses the situations affecting the sensory organs, such as an image, sound, touch, hears, smell, or taste, as an input through the sensors. The inputs are processed and evaluated algorithmically and transformed into a movement or a thought in the long run. Meantime, the samples are matched; the research is conducted,

186

Biyan and Yılmaz

the reasoning is made, and thus the learning activity is performed (Yüksel, 2018: 592). Large amounts of data are processed quickly and accurately using the transactions carried out via algorithms within the scope of artificial intelligence. Humans have been replaced by artificial intelligence programs owing to the natural speed, reliability, and scalability of algorithms (Ooi & Goh, 2019: 3).

2.3 Difference between Artificial Intelligence and Robot Artificial intelligence does not mean a robot. While artificial intelligence is a software, a robot is an object, and some machine made up of mechanical parts. Not all robots have artificial intelligence. There is not even such a rule claiming that artificial intelligence exists in robots only (Yüksel, 2018: 593). The robots are the mechanisms capable of managing themselves, moving independently, and performing specific tasks assigned to them (Zorluel, 2009: 309). Robots can walk, perform their jobs, and have artificial intelligence to make better decisions. Even though the automation systems (Industry 2.0) are thought together with artificial intelligence, they are entirely different from artificial intelligence, too. For example, when traffic lights were introduced, they replaced the traffic police. What’s more, also the term “robot” was used for traffic lights in South Africa. However, that does not necessarily mean that traffic lights can be regarded as artificial intelligence (Marwala, 2018: 2). The Google search engine is a kind of artificial intelligence, but it is not considered a robot. As you search through the Google search engine, Google first detects the subject being explored. Once it detects the topic, it carries on the process of reasoning thanks to its algorithms, by using the information it learned previously and presents the most relevant websites to the user in a hierarchical manner. In this way, Google fulfills the perception, learning, reasoning, and deduction processes of intelligence. However, that does not necessarily mean that it can be considered a robot (Zorluel, 2009: 309). So, based on the reasons mentioned above, the report on “European Civil Law Rules in Robotics”, which was issued by the European Parliament Policy Department for Citizens’ Rights and Constitutional Affairs, stated that a smart autonomous robot was required to possess the following characteristics (ECLRR, 2016: 8): (a) Ability to move to utilize the sensors and/or by analyzing and using the peripheral data; (b) Self-learning; (c) Use of physical support; (d) Able to keep up with the environment owing to its movements and behaviors.

Artificial Intelligence: If It’s Taxed, But How?

187

2.4 Possible Effects of Artificial Intelligence on the Future In parallel with the rapid developments in artificial intelligence, it began to be used in more and more fields. It is highly likely that artificial intelligence will be used more widely in the future. Moreover, discussions began on where certain professional groups today would end up. While the technological developments in the 1980s and 1990s enabled the employees to speed up and carry out their works comfortably, recent developments began to replace the employees by machines or robots. Technological progress has increased so much so that there was nearly no need to employ qualified personnel, apart from the limited number of trained staff in charge of monitoring the automation system. Less and fewer qualified staff was used, and also a decrease was observed in the name of the entire team employed. Furthermore, the level of wages further reduced as more unqualified employees were recruited (Korinek, 2019: 2–3). The market size of artificial intelligence globally is approximately $6 billion, according to the gross value added estimates as of 2016. This figure is expected to rise to $60 billion, with a 10-fold increase by 2025 (AIR, 2018: 3). Today, artificial intelligence has begun to be used widely in transport, education, employment, defense and security, health, virtual reality and virtual assistant, internet of wearables and objects, commercial intelligence, and robotization. Also, artificial intelligence lends substantial support to painting, story writing, and scripting, composing, computer programming, filmmaking, recipe making (AIR, 2018: 5–12). According to the 2017 Statistical Report on Internet Development in China, 2.542 artificial intelligence companies worldwide have come into operation. 1.078 of these companies are based in U.S.S, and 592 are in China (Huang Z, 2018: 1819). Therefore, it will be fair to say that artificial intelligence will be seen in many more areas in the near future. Therefore, with all these developments, it is clear that the legal status of artificial intelligence is used instead of humans. Because it is important to determine the status of artificial intelligence (or not) in life and especially in-laws. The assessment of the state of artificial intelligence in legal, financial, social, and even political life should be discussed internationally.

3 Taxation Size: To Be or Not to Be… In order to put an answer to the question “can artificial intelligence be taxed?”, it is necessary to set forth its essence in terms of fiscal and legal sciences dealing with tax as a profession. In this context, it will be appropriate to determine the status of artificial intelligence against the externality theory in terms of public

188

Biyan and Yılmaz

finance and to determine its essence and status in terms of the law, so that its taxation aspect can be adequately addressed.

3.1 Artificial Intelligence in Terms of Externality Theory The reflection of technological developments manifested itself under two headings. Firstly one is the increase in the amount of production. As a result of technological progress, the amount of output in the economy has increased, and especially the businesses that had innovation policies have benefited significantly from this deal. Secondly, technological development has led to the fact that the share of income received from the economy was distributed again. These new developments also referred to as the sharing economy, have influenced the way the funds are exchanged between the persons (Korinek & Stiglitz, 2017: 6). Technological improvements have significantly contributed to the increase of efficiency, reduction in the cost of operational transactions, and facilitation of data transfer to/from machines (Kavoya, 2018: 52). Nevertheless, technological revolutions are also regarded as the cause of the mass replacement of human labor, which has been restored by technological advances (Ooi & Goh, 2019: 4). Public finance has faced two fundamental problems in the face of these technological developments. First, the market mechanism has lost its production efficiency upon the inclusion of information technologies. The firms producing financial information have started to steer the economy by creating a monopoly effect on the one hand and started to change the useful point of demand with high prices by guiding the consumers in this direction, on the other. At this point, it is recommended that the public sector establish a fund for the creation of information technologies and allow individuals to access them at affordable prices. The second main problem affecting public finance is the intensive use of these public funds by the private sector. As an example, when Steve Jobs designed iPhone, U.S. Defense, Advanced Research Projects Agency, had already been established. This agency could not have outsourced (even if it wanted) the goods with a design, which could not yet be imagined by visionaries like Steve Jobs. And that began to restrict the areas where effective results could be achieved with the public investments made in the sources of information (Korinek, 2019: 6). According to some approaches, the development of artificial intelligence and Al robots made it possible to use artificial intelligence in areas where humans cannot obtain efficiency or involve the robots or mechanisms with artificial intelligence in dangerous works. This is described as a positive externality. Whereas those with other approaches suggest that artificial intelligence, which is capable of developing their skills and mimicking intelligent behavior, can increase

Artificial Intelligence: If It’s Taxed, But How?

189

unemployment, as it will replace humans. Those that have the second approach argue that Al robots must be taxed or a tax must be imposed on the robot (AIR, 2018: 26). The idea that artificial intelligence will affect employment negatively forms the basis of the need for taxation. Although it was known that technological revolutions in the previous years also contributed to unemployment, a current wave of automation caused in parallel with Industry 4.0 is likely to be more destructive than the previous ones for several reasons. Previous technological innovations did not eliminate the need for human labor in operation and control technology. The autonomous nature of the existing technologies removes the need for human intervention, thus threatening the place of social work to a great extent. In addition, unlike previous technological innovations that are limited in terms of applicability, the autonomous technology is a general-purpose technology that has a broader set of various capabilities, including physical action, information processing, etc. So, inevitably, it has the potential to have a devastating impact on a broader range of sectors. Due to the speed of development in automation technology, it is stated that there is hardly time for the governments to respond to automation. Otherwise, the consequences can be severe unless quick actions are taken (Ooi & Goh, 2019: 5). While the Bank of America’s Merrill Lynch argues that artificial intelligence will save $9 trillion in employment costs by 2025, a report by the World Economic Forum estimates that 5.1  million people will lose their jobs due to artificial intelligence automation by 2020. Deloitte, a consulting firm, claimed that thirty-five percent of people employed in the UK were at risk of layoff due to improvements to be made in automation systems over the next ten to twenty years (Abbott & Bogenschneider, 2018: 153). Therefore, it is stated that as automation and artificial intelligence prevail, productivity will increase, and new businesses will emerge. Still, on the other hand, unemployment and inequality will inevitably be experienced (Abbott & Bogenschneider, 2018: 154). Hence, job losses, increasing inequality, and a decrease in tax revenues are seen to be inevitable (Mazur, 2018: 6–17). On the other hand, it may not be possible for an artificial intelligence automation system to create an equal effect in all sectors. It will be difficult for most of the businesses within the same industry or occupational class to switch to new businesses if they are highly sensitive to automation. Put it differently; a radical restructuring of the company or business policy will be needed, even though it is not required at the moment. It is necessary to take policy measures to reduce the impact of automation on specific sectors or occupational classes. In some industries, technological developments can make it plausible yet and financially demanding for companies to automate their entire line of business entirely.

190

Biyan and Yılmaz

A significant problem will arise if the workers doing these jobs do not need to have the skills to allow them to perform alternative performance typically. For instance, the effect of self-driving trucks on truck drivers makes a good example. If self-driving vehicles replace truck drivers who do not have a different job alternative, then there will be a long-term probability of structural unemployment. Therefore, it is stated that the need for intervention, particularly in such sectors, is necessary (Ooi & Goh, 2019: 5). As a result of all these discussions, it is understood that the fact of artificial intelligence brings about either positive or negative externalities, as the case may be. In such a case, it is necessary to clarify the taxation or promotion of machines, robots, or articles using artificial intelligence. It is also clear that a policy should be developed based on the findings of the analysis of sectors and their impact on employment. The general opinion is that artificial intelligence hurts work (Ooi & Goh, 2018: 5, Bottone, 2017: 12, Englisch, 2018: 7–8). Therefore, with the increase in artificial intelligence and the gadgets having artificial intelligence, which will affect employment negatively and replace humans, it seems to be inevitable to impose a tax or similar financial obligations. Of course, it is controversial on what or whom it will be imposed and how it should be applied. As elaborated in the following sections of the study, it has to be clarified whether it will be imposed on income or via a Pigovian tax application. On the other hand, the decrease in the employment of real people is of particular concern to the public budgets in macro terms. Namely, if artificial intelligence starts to work instead of real people, it will lead to a decline in public revenues, as it will not be possible to realize the tax obligations collected based on social security payments and wages. That is because the employment of “real person” is what makes them be paid. However, artificial intelligence needs neither social security nor income. This result reveals that public finance will be affected, more or less, as the artificial intelligence is used widely. Therefore, even just for this reason, it can be said that a taxation study should be done regarding artificial intelligence.

3.2 Artificial Intelligence in Terms of Law 3.2.1  Should Artificial Intelligence Be Personified? The hottest debate in legal terms is whether artificial intelligence needs to be given legal personhood is given. In the legal system, personhood is divided into two: real and legitimate. For individuals not having legal personhood, solutions are produced by making special regulations in the laws. Artificial intelligence is

Artificial Intelligence: If It’s Taxed, But How?

191

neither an actual nor an authorized person. So, first of all, the answer to the question, “Should personhood be given” has to be/has been sought. The basic approach concerning the debates regarding the concept of personhood and legal status in terms of artificial intelligence, which survived today, is that artificial intelligence is an “item/tool” and should be accepted in the ownership of its producer. However, in view of the fact that artificial intelligence is increasingly a more significant part of human life and acquiring more and more humanlike features, the idea to accept them as tools or items only is increasingly being abandoned. The most striking proposal regarding the legal status of artificial intelligence is the one suggesting that “Electronic Personhood” should be granted to artificial intelligence (Leroux & Labruto & Boscarato, 2012: 61). And, that stems from the concept of legal personhood; and development of electronic personhood is discussed within the frame of the fact that artificial intelligence, which is capable of making autonomous decisions and communicating with people, should be registered in a special register. Thus it is intended that it could have some acquired and individual rights and obligations; the responsibilities of related parties (users, vendors, manufacturers, etc.) can be determined. In this way, electronic personhood, similar to the legal personhood, should be developed (Zorluel, 2009: 344). Legal personhood is always related to individual autonomy. The question “can an artificially intelligent asset be given legal personhood?” is a matter of whether such an asset set will possess any legal rights and obligations. The essence of the legal personhood is based on the fact that whether such an asset has the right to ownership and the capacity to file an action or engage in a lawsuit (AIR, 2018:  13). What is taxed within the scope of the proposal of electronic personhood was not the robot itself but rather the companies that use it (Mazur, 2018: 18). According to an argument put forward as to whether or not the legal entity (personhood) given to the companies in the doctrine can also be given to the artificial intelligence; it is possible to establish the legal construct, which was created for the companies regarded as an essential example of fake person, for artificial intelligence, as well. On the other hand, it is also accepted that there is not an absolute similarity between companies and artificial intelligence. That is because, while the companies are considered autonomous institutions constructionally and their stakeholders decide on their activities, artificial intelligence does not have any stakeholders and makes decisions directly by themselves (AIR, 2018: 13). From the viewpoint of the capacity to have the right and ability to act in Turkish law, many discussions come to surface. In another saying, if personhood

192

Biyan and Yılmaz

is recognized, then some questions have to be answered. For example, when making agreements, it will be difficult, at least for today, to find the answers to the questions such as, whether they will be a party to the agreement, get the approval of the other party and whether the intentions of Al robots are measurable (Akbilek, 2017: 227). Because, to be a “person” requires some consequences (Gözler, 2014: 179). The most moderate view, among others, put forward in the doctrine today, is the one suggesting that the operator of Al robot should be held responsible (Chopra & White, 2004: 3). Considering some of the examples relating to the subject in the world, the studies conducted by South Korea and Estonia emphasize the idea of giving personhood to the robots. In 2012, South Korea introduced a legal regulation in 2012, restricting that humans must always control the robots and ruled out the possibility of giving separate legal personhood. On the other hand, Estonia discussed whether special regulations should be made for the robots within the scope of the Civil Law, and some proposals were made on responsibility. The most remarkable suggestion was one offering that the robots should be allowed to acquire legal personhood in a way to include also the authority to represent their owner. Aside from the ongoing discussions, according to an agreement we advocate, it is not possible to agree with the idea that robots can act since they are not autonomous enough to perform their actions and operations by their own will, i.e., without any external intervention (Akbilek, 2017: 231–232). However, the subject will be put under discussion again in the future, if we have such Al robots that can make decisions and act on their own and do not need human intervention whatsoever. Since the assets in question are not “real persons”, it causes a dilemma as to whether it would be possible to assess them with the attributes of a “legal person”. Real persons ultimately guide legal persons, and the responsibilities can be shared with the real persons. However, there is no human intervention in artificial intelligence. On the other hand, artificial intelligence does not have a body, soul, nationality, feelings, consciousness, interests, and curiosity, nor a free will as real people do. It should be noted, however, that Saudi Arabian granted citizenship to the artificial intelligence, named Sophia, which was produced by a Hong Kong firm Hanson Robotic (AIR, 2018: 13–14). And, that is a clear indication that it does not mean artificial intelligence will not have what it doesn’t today, over time. It seems plausible in the forthcoming years that it will be possible to impose tax by giving a statue of personhood subject to the law, rather than to an artificial intelligence programmer or user (AIR, 2018: 25).

Artificial Intelligence: If It’s Taxed, But How?

193

3.2.2  Artificial Intelligence in Terms of Responsibility and Punishment The concept of “responsibility” is the main reason for the discussions about whether artificial intelligence should be given personhood. In particular, this subject of responsibility was further highlighted as a result of the accident that claimed the life of a 49-year-old person in Arizona, U.S.A., during the test drive of autonomous vehicles, which was carried out by UBER. So, the discussions began as to whom should be held responsible since there was no legal regulation as to whether Uber Technologies Inc. should be held accountable or whether the artificial intelligence operating the autonomous car would be held responsible alone (AIR, 2018:14). Therefore, artificial intelligence is not naturally responsible at the moment. However, it might be confusing whether it will be considered reliable as the potential operations increase over time. Moreover, given the fact that the decisions could also be appealed, in other words, the fact that it can implement its own decisions raises the issue of whether or not a legal responsibility can be assigned. The fact that each Al robot has different autonomy levels creates problems in producing solutions regarding the law of responsibility and makes matters even worse. Do the acts, actions, or functions performed by robots result from the design and programming, or do they improve and evolve depending on the features they own? If a person guides the actions of anAl robot or has a part in its activities, then we can talk about some aspects such as a fault or intention in terms of liability law. However, it is argued that some regulations should be made under a different approach, for Al robots, which are capable of moving autonomously (Akbilek, 2017: 219–220). Yet another point reached in these arguments is the question of whether artificial intelligence can be punished or not. It must have a bank account and pay from its account in case it is served with an administrative fine. But, artificial intelligence does not have a bank account (for the moment at least). Nor will it be eligible to open a bank account. Therefore, it will not be possible to punish it either legally or effectively. According to some recommendations on this topic (AIR, 2018: 24); (a) If an act of artificial intelligence requires death penalty, deleting the artificial intelligence can be an option; (b) If the law of artificial intelligence requires a prison sentence, it may be temporarily put out of service; (c) If the action of artificial intelligence requires a public service, it can be allocated to that particular the public service;

194

Biyan and Yılmaz

(d) If the act of artificial intelligence requires an administrative fine, the user or creator of that artificial intelligence may be served with a fine. However, this is controversial in terms of the personality principle of penalties. Criminal and financial responsibility is closely related to giving personhood to artificial intelligence. In general, criminal liability belongs to natural persons, while financial responsibility belongs to natural or legal persons (Gözler, 2014: 223). If artificial intelligence is to be given personhood and responsibility is to be assigned accordingly. Its results must be taken into consideration because artificial intelligence can be served neither with imprisonment nor a monetary fine. Hence, if a criminal or financial action is to be considered for an act of artificial intelligence, it should be the person who is the creator/operator of the artificial intelligence that must be held responsible. Breaking off the relationship between artificial intelligence and its creator/operator may lead to an uncontrollable point where it will be impossible to find the wrongdoer (offender) and impose a sanction. Although it can be argued that artificial intelligence can take an autonomous decision, and thus it must be punished, the fact that no action can be taken against it, either criminally or effectively, can lead the argument to a meaningless point. What is more, in our opinion, suggestions such as taking actions to delete and/or remove artificial intelligence (AIR, 2018: 24) does not make any sense.

3.2.3  The Fate of Copyrighted Work Produced by Artificial Intelligence It is also argued in legal terms how it will be dealt in case a copyrighted work is produced by artificial intelligence produce. If artificial intelligence provides any task which requires copyright, such as painting, story, and scriptwriting, composing, computer programming, and filmmaking, artificial intelligence, which is not a real person, does face the same problem of granting personhood. Even though the real person takes the first step in creating the copyrighted work, it is still unclear how to determine the actual owner of such copyrighted work as artificial intelligence is involved in and affects the process of creation. The same applies to all transactions that create intangible rights. Does the copyright belong to artificial intelligence or the real person who creates it? (AIR, 2018: 17). In one of its decisions, the U.S. Copyright Office decided that works produced without the contribution of any creative people could not be regarded as works. English law recognizes the author as the person who makes the necessary adjustments for the creation of the work. In this case, computergenerated literary, dramatic, musical, or artistic works are not considered works. As in the U.S.  legal system, the fact that the products of artificial

Artificial Intelligence: If It’s Taxed, But How?

195

intelligence are not accepted within the scope of copyright, subjected to direct public use, is seen as an obstacle preventing the progress of the works from being carried out in this field (Zorluel, 2009: 325–326). On the other hand, the products that can be considered an action, according to Law on Intellectual and Artistic Works in Turkish Law, can only be produced by humans. In other words, the owner of the works created by artificial intelligence may not be artificial intelligence. However, it is stated in the doctrine that this situation is not sustainable.

3.3 Discussions on Taxation Regime As mentioned in the previous chapters of the study, how to tax the artificial intelligence in terms of finance theory and legal science, and what methods to use in taxation per the existing rules, if artificial intelligence is to be taxed was discussed. The subject will be considered under this heading by also including our opinions based on these discussions.

3.3.1  In Terms of Income Tax 3.3.1.1 Principle of Financial Power and Identification of Taxpayer: Artificial Intelligence? Or Its Creator/Operator? The majority of the discussions on the taxation of artificial intelligence are by and large on whether or not artificial intelligence can be identified as a taxpayer. This issue has been addressed both in terms of giving personhood to artificial intelligence and whether it should be vested with financial power. According to positive law, a person must be a real person to be an income taxpayer. Besides, considering the principle of financial power as per the constitutional provisions, it must also have a taxable income. Real persons, who obtain the elements making up the income, submit tax returns according to their particular circumstances and qualifications, -or even if they do not-, they are (often) taxed through withholding. The taxes to be paid are assessed by taking the specific situations of real people, such as disability, marital status, number of children, education and health expenditures, donations, and charities. It is unclear yet whether artificial intelligence will make such expenditures as real people do. In other words, artificial intelligence does not make donations and give charities, nor does it spend on health and education. And, that raises the question of whether a certain fee can be set and accepted as a levy for the expenses that are not made. What is more important above all is how to impose a tax following the principle of fiscal ability. In other saying, do robots have a fiscal ability?

196

Biyan and Yılmaz

Since artificial intelligence is not a real person, it cannot be a taxpayer for income at present. However, if artificial intelligence were to be given personhood and this personhood can be included in the taxpayer group in terms of income tax, and it can be taxed naturally as soon as it acquires one of the income elements. In this case, it rises to different questions. For example, if artificial intelligence is to be taxed, it cannot be determined how much of the income will belong to artificial intelligence and how much to the creator/operator. Also, whether the user/creator creates the added value or it will be attributed entirely to artificial intelligence. For today, the general approach is to tax the income earned by creators/operators (AIR, 2018: 25). Since the real persons declare their salary as per the income tax based on assessment upon declaration, it is also important how to subject the robots to this regime in this scope. One of the arguments outlined in this context is to grant Al robot a “special status”. It is proposed to grant new legal personhood as per the tax law and shape up the system accordingly. In this context, the robot can be regarded as a separate person. A system called “electronic payment power” can be created. The primary basis of those who claim that robots should be taxed in this way is that robots replace humans. Robots’ income is considered as a fee and can be withheld. Besides, if an Al robot is regarded as a separate taxpayer, it will be highly likely to cause double economic taxation in case the income of the robot and its owner is taxed separately. In this case, a result, as in the profit shares, will eventuate (Englisch, 2018: 4–6). In our opinion, it might be undesirable to establish liability in terms of tax law unless it has a criminal and financial responsibility, whether or not artificial intelligence is given personhood. Primarily, it is probable that artificial intelligence can commit tax misdemeanor and revenue offense as it might mimic human behaviors or display similar acts. In this case, it may not make any sense to imprison artificial intelligence or impose a fine for loss of tax/fraud (irregularity). Although it is considered for a moment that a liability similar to the criminal liability of a legal entity may be given in France or Belgium, it is more appropriate to designate the creator/operator as its legal representative. Otherwise, tax authorities may have to bear the consequences of crimes or offenses committed by artificial intelligence. If its creator/operator is specified as the legal representative, there will be a real person to address in case of misdemeanor and crime. As a result, instead of establishing liability for artificial intelligence, it will be a good practice to accept its creator/operator as the taxpayer and attribute the income earned to the real person. In this way, the criticism of double taxation will be eliminated.

Artificial Intelligence: If It’s Taxed, But How?

197

3.3.1.2 Assessment of Income and Base: What Type of Earnings? One of the essential aspects under discussion in this context is how to determine the income earned/to be received by artificial intelligence and what the type of income will be. How will the income earned by an Al robot be determined? The studies on taxation of artificial intelligence suggested the idea that the income earned through artificial intelligence should be considered as the continuation of the business of its creator/operator. Put it differently; also, the income generated by artificial intelligence, which is created/used by a business earning commercial income, should be considered commercial earning. In such a case, taxation shall be made based on the business under the applicable provisions (AIR, 2018: 25). Within the scope of this possibility, all incomes earned by artificial intelligence will be regarded as the commercial receiving of the business. According to another view, if the robots of artificial intelligence are classified, the income should also be determined accordingly. When artificial intelligence or Al robot is classified according to the intended use as “industrial artificial intelligence/robot” or “serving artificial intelligence/robot”, the income earned in such a scenario can also be accepted as “commercial earnings” or “wage/self-employment earnings”. For example, a service robot is a robot that does useful work for people or tools/equipment, except for industrial automation. On the other hand, a personal service robot or a service robot for personal use is a service robot used for non-commercial purposes or purposes other than commercial use. Domestic servant robots, automatic wheelchairs, own mobility assistance robots, and pet exercise robots are the examples, to name but a few. A professional service robot, or a service robot for professional use, is a service robot used for a commercial task, which is usually operated by a duly trained operator. Examples include cleaning robots for public places, the delivery robot in offices or hospitals, fire fighting robots, rehabilitation robots, and surgical robots in hospitals. In this context, an operator is a person assigned to start, oversee, and stop the intended operation of a robot or a robot system (Bottone, 2017: 4). According to another proposal, if a robot is to be taxed, then the rate of income tax should be based on the possible wage it could have earned had it been a real worker. According to this proposal called “robot income tax, the “economic advantage” obtained by the employer through the use of robots instead of workers can be considered as the criterion (Guerreiro & Rebelo & Teles, 2017: 4). In that case, the robot’s ability to pay can be regulated by law as the technology evolves. As can be seen in the arguments above, it is unclear how to assess the type of income if artificial intelligence is to be held liable for tax purposes. As we

198

Biyan and Yılmaz

mentioned in the previous section, that confirms the fact that rather than establishing liability for artificial intelligence, the creator/operator has to be determined as the taxpayer, considering the current technological and legal circumstances. If the creator/operator is accepted as a taxpayer, we think that it will be appropriate to consider it a commercial earning. As the use/operation of artificial intelligence requires capital, organization, and less labor, this type of gain will be appropriate. Of course, different alternatives can be considered in case an income is earned by leaving it at the disposition of others. For example, with the regulations to be made under Article #70 of Income Tax Law, it will be possible to accept it as real property income, considering that Al robot is rented to another person. On the other hand, if a liability is established for artificial intelligence, the issue will become more complicated. It is uncertain to determine what kind of income the artificial intelligence earns as the actual taxpayer. But still, if a profit is to be attributed under current circumstances, it can be possible to accept it as “commercial earning” or “wage” and to make special regulations in the law.

3.3.1.3 Should Artificial Intelligence Be Considered a Workplace? One of the points under discussion about taxing artificial intelligence and Al robots is seen as a problem under international tax law. Because it is evident that it needs to be clarified where it will be based or considered to be, in terms of “location”. For example, it should be made clear who will be using the taxation authority in cases where the creator/operator resides in country X, but artificial intelligence or Al robot operates in territory Y (Englisch, 2018: 12–13). The workplace, which forms the basis of taxation as a place (location), is defined as the point of affiliation, which allows commercial earnings to be affiliated to the taxation authority of the source country where they are earned (Yaltı Soydan, 1995:  131). The subject of “workplace”, which is one of the affiliation rules in the exercise of taxation authority, is the most important basis through which the tax administrations can impose a tax on corporate income taxpayers and commercial income taxpayers. Those who have a business within the political boundaries of a country shall be subject to the taxation regime of that particular country. The first aspect considered in the implementation of the principle of residence, which is the basic principle in taxation, is whether the person/ corporation has a workplace (Biyan & Yılmaz, 2018: 17). Artificial intelligence can deliver services either online or at a fixed location, as applicable. Therefore, it would be more appropriate to assess a case by case basis. For instance, since an artificial intelligence offering online services will be

Artificial Intelligence: If It’s Taxed, But How?

199

indifferent to a website, it should be evaluated as the businesses operating via a website, in which case the taxation problems will manifest themselves against the artificial intelligence (Biyan & Yılmaz, 2018: 34–36). On the other hand, artificial intelligence affiliated to a fixed location, e.g., an Al robot can be taxed as a fully obligated or limited taxpayer in the country of residence according to the principle of residence or source. Consequently, assigning a status to artificial intelligence in terms of a workplace is one of the main problems caused by the digital economy. Rather than producing national-wide solutions, it would be a good practice to include, address, or refer the matter in tax treaties, to get more proper and practical results.

3.3.1.4 Discussions on Copyright The point of whether artificial intelligence can be a copyright owner also appears in its taxation, as well. Even though artificial intelligence is considered to have the ability to mimic intelligent behavior and process it on its own, it is ultimately made up of computer algorithms and software. And, if that is accepted as the “right to use” for artificial intelligence programmers, then the income earned by artificial intelligence needs considering as copyright. On the other hand, when a profit is obtained by assigning the copyright, it is argued that it can also be accepted as a technical service fee (AIR, 2018: 25–26). In Turkish law, FSEK does not allow the persons, other than real ones, to be a copyright owner; therefore artificial intelligence can’t be copyright owner in Turkey today, as per the positive law. On the other hand, although it may be considered that the artificial intelligence should be given copyright, it seems more appropriate to accept the owner/operator as the copyright owner if a copyrighted work is produced by artificial intelligence. Since it is not clear how all will use the rights that copyright will bring to its owner and how it will get the earnings it will acquire owing to such reasons. Because of the creator/operator of artificial intelligence benefits from earnings. If it is the artificial intelligence that produces such work, then it is the creator/operator who is a real person that creates artificial intelligence. Then again, if the criminal and financial responsibility lies with the creator/operator, in that case, also the creator/operator should be considered a copyright owner. And, the income earned will be taxed as self-employment income. 3.3.1.5 In Terms of VAT It will be inevitable to use the robots as serving gadgets as they were employed instead of real persons. Robots provide several services, such as legal or financial advice, medical assistance or cleaning services, etc. In this case, we come up

200

Biyan and Yılmaz

with the question of whether or not calculate the VAT, which is based on goods delivery and execution of service (Englisch, 2018: 17–8). According to an opinion, value-added tax for activities carried out by robots may be subject to VAT by considering them “service”, as in the case of self-employed traders. However, it is stated that it may not be easy to figure out whether the fee charged by an AI robot is accurate (Oberson, 2017: 256–257). If a robot is used to produce goods or services, they are probably taxed both as intermediate goods and final goods, so it is stated that there is no need to recalculate VAT to avoid the risk of double taxation. However, if the robots are legally considered a “person” with legal and financial capacity, it will be required to levy VAT on the service related to their activities. All proposed solutions are highly controversial in terms of globalization and, consequently, easier circulation/mobilization of capital, the emergence of tax competition between jurisdictions, etc. Therefore, since physical capital tax implies higher costs for national companies and impairs their global competitiveness, it can be stated that the design of a robot tax requires a comprehensive analysis by taking the arguments, especially regarding international taxation at OECD and UN (Bottone, 2017: 17). Since it is evident that the delivery of goods and the execution of the services are included in the subject of VAT, it is clear that the sale of a service or products carried out by artificial intelligence should be subject to VAT in the case of commercial, agricultural or professional activity as well as imports. The VAT will be calculated during the sale of goods or services, which is carried out by an artificial intelligence operated by its creator/operator, and the taxpayer will be the creator/operator. However, if liability is assigned to artificial intelligence, there might be some problems in the performance of the formal obligations of artificial intelligence, even if the VAT calculation of artificial intelligence is programmable. For example, lodging tax returns, in which case its creator/operator fulfills the formal liabilities. In our opinion, it will not pose a problem whether or not artificial intelligence collects VAT or whether the payment is received in full amount since the fees will be calculated both by computer programs and processed through the banks or financial systems. But, some problems can be experienced in case cash transactions are accepted in such situations.

3.3.1.6 In Terms of Environmental Tax (Pigovian) The widespread use of artificial intelligence and its possibility to affect employment negatively highlights two issues, in particular. First, there is the possibility of a decrease in income tax, which has a significant share in tax revenues in almost several countries. As robots replace human beings, the taxes imposed on them will likely lead to a reduction in tax revenues. The second is to face the fact

Artificial Intelligence: If It’s Taxed, But How?

201

that robots will be employed and thus to train those persons, mainly working in such businesses or jobs that do not require any skills and talents. As a result, the need for public resources will increase (Bottone, 2017:  2). The European Parliament stated that it would be more appropriate to use the income that is generated, to re-train the unemployed workers if it would be necessary to impose a tax on a robot’s work or define a wage for it. Therefore, it was pointed out that sectors, which are at the most risk due to the employment of robots, should be identified (Bottone, 2017: 12). In case an Al robot or automation system is used, it is proposed to apply a high tax rate or impose automation tax or similar tax for businesses with automation. On the other hand, it is also necessary to use tax relief or incentives to companies employing real persons (Abbott & Bogenschneider, 2018: 168–173). The social costs caused by automation create a negative externality as workers or communities cover them. The causal relationship between the transition of the firms to automation and the resulting negativity creates a prima facie situation, in order for the government to intervene by deterring or punishing such automation actions that generate an externality. Such an intervention may return to the party to alleviate such externality as a Pigovian financial obligation. In this case, such tax would impose an automation tax, which will apply to all technologies forming the current wave of technological innovations, on the companies that automate their systems or equipment involved in the production processes (Ooi & Goh, 2019: 6). This tax, which is proposed in the doctrine as a robot tax, is an automation tax and aims to impose a tax on businesses that prefer to operate automation systems by switching to mechanization, instead of employing people. In this context, it serves as a sort of balance (Mazur, 2018: 18). The base of the automation tax can be measured by the total monetary coverage of any decrease in the number of layoffs or employee wages. In addition, this can be easily measured according to financial data, payroll, or other company records. The size of the tax base measured by the layoffs attributed to automation will be proportional to the externality of dismissal based on automation. At this point, however, it is not clear how to assess other possible situations, such as reduced operating conditions or productivity improvements not related to automation, when determining the tax base. It also seems necessary to take into account the analyses on whether the time between automation and layoffs is short enough and whether the automated tasks are similar enough to those performed by those who are discharged (Ooi & Goh, 2019: 11). South Korea is the first country in the world to levy a robot tax. The country in question started to impose robot tax (AIR, 2018: 26) as of August 2017. However, this is not sufficiently considered regarded as a robot tax. This tax in question,

202

Biyan and Yılmaz

which is widely considered as an application that restricts the incentives for the transition to the automation system, is an obligation which is imposed on the corporate taxpayers investing in automated machines and paid in an amount set between 3% and 7% of the investment made. Then, South Korea decreased these rates by two points (Abbott & Bogenschneider, 2018: 149). One of the concrete proposals made in the doctrine is the application of reverse depreciation. It is argued that if the investment made by the taxpayers who invest in automation affects to eliminate employment, the depreciation rate of the investment made will be kept low, not allowing a great majority of the capital expenditures to be deducted from the taxable income. On the contrary, if such investment made is supporting the employment, then in such a case, support can be given, and incentives can be provided by keeping the depreciation rate high (Ooi & Goh, 2019: 18–19).

4 Conclusion The focal point of discussions on taxation of artificial intelligence emphasis on the negative impact it may have on employment due to the negative externality that it creates and on the issue whether or not it should be taxed due to possible losses of revenue in the budget. Although there are opinions that artificial intelligence should be given legal personhood and accepted as a taxpayer -as we have tried to elaborate in the study- it is not a proper approach to provide artificial personhood and assign liability to artificial intelligence, at least for now. In our opinion, whether or not an entity with artificial intelligence is given personhood under the existing technological and legal conditions, it may be inconvenient to establish liability in terms of tax law unless it has criminal and financial responsibility. It is, of course, probable that artificial intelligence can commit tax misdemeanor as it might mimic human behaviors or display similar acts. In this case, it may not make any sense to imprison artificial intelligence or impose a fine for loss of tax/fraud (irregularity). Although it may be considered, for a moment, that financial responsibility may be assigned, it would still be more appropriate to appoint its creator/operator as its legal representative under the current circumstances. Moreover, it is unclear, in the current system, how to determine the type of income if liability is to be assigned to artificial intelligence. It would be appropriate to establish the creator/operator as the taxpayer and accept the income earned as commercial earning, rather than creating a liability for artificial intelligence. Likewise, in case a copyrighted work is produced by artificial intelligence, yet again it will be more appropriate to attribute such income to its creator/operator.

Artificial Intelligence: If It’s Taxed, But How?

203

References Abbott, R., Bogenschneider, B., (2018). “Should Robots Pay Taxes? Tax Policy in the Age of Automation”, Harvard Law & Policy Review, Vol: 12, pp. 145–175, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2932483, (09.05.2019). AIR, (2018). “The Future Is Here: Artificial Intelligence and Robotics”, Nishith Desai Associates, May 2018. https://studylib.net/doc/25209542/artificialintelligence-and-robotics, (16.09.2019). Akbilek, M., (2017). “Teknolojinin Pandora Kutusu: Robotların Hukuki Kişilikleri (!) ve Hukuki Sorumlulukları”, Beykent Üniversitesi Hukuk Fakültesi Dergisi, Cilt: 3, Sayı: 6, Aralık, pp. 215–236. Biyan, Ö., Yılmaz, G., (2018). “A Taxation Problem Caused by Digital Economy: Definition of Virtual Establishment”, Current Perspectives in Public Finance, (Editörler: Selçuk İpek, Adnan Gerçek), Peter Lang, Berlin, 11–42. Bottone, G., (2017). “A Tax on Robots? Some Food for Thought”, MEF Ministero dell’ Economia e delle Finanze, DF Working Papers, pp. 1–21, https://www. finanze.it/export/sites/finanze/it/.content/Documenti/Varie/dfwp3_2018. pdf, (08.05.2019). Chopra, S., White, L., (2004). “Artificial Agents - Personhood in Law and Philosophy”, Proceedings of the European Conference on Artificial Intelligence, http://www.sci.brooklyn.cuny.edu/ ~schopra/agentlawsub.pdf, (09.05.2019). DDAIT, (2018). “Discussion on the Development of Artificial Intelligence in Taxation”, American Journal of Industrial and Business Management, Vol: 8, 1817–1824, http://www.scirp.org/journal/ajibm ISSN Online: 2164-5175 ISSN Print: 2164-5167, https://www.researchgate.net/ publication/327242385_Discussion_on_the_Development_of_Artificial_ Intelligence_in_Taxation, (10.09.2019). ECLRR, (2016). European Civil Law Rules in Robotics, DirectorateGeneral for Internal Policies, Policy Department C: Citizens’ Rights and Constitutional Affairs, http://www.europarl.europa.eu/RegData/etudes/ STUD/2016/571379/IPOL_STU(2016)571379_EN.pdf, (06.06.2019). Englisch, J., (2018). “Digitalisation and the Future of National Tax Systems: Taxing Robots?”, https://ssrn.com/abstract=3244670, (09.05.2019). Guerreiro, J., Rebelo, S., Teles, P., (2017). “Should Robots Be Taxed?”, NBER Working Paper No. 23806, Revised in January 2019, https://www.nber.org/ papers/w23806.pdf, (10.06.2019). Gözler, K., (2014). Hukukun Temel Kavramları, Ekin Kitabevi, Bursa. Huang, Z., (2018). “Discussion on the Development of Artificial Intelligence in Taxation”, American Journal of Industrial and Business Management, Vol: 8,

204

Biyan and Yılmaz

ss. 1817–1824, https://www.scirp.org/pdf/AJIBM_2018082715192843.pdf, (09.05.2019). Kavoya, J., (2018). “Digital Technologies in the Tax Industry: The Case of VAT”, Tax Administration Review Cıat, AEAT, IEF, No: 43, pp. 51–64, https://www. ciat.org/Biblioteca/Revista/ Revista_43/ingles/2018_TR_43_kavoya.pdf, (10.05.2019). Korinek, A., (2019). “Labor in the Age of Automation and Artificial Intelligence”, Economists For Inclusive Prosperity (ECONFIP), January, pp. 1–9, https://econfip.org/wp-content/uploads/2019/02/6.Labor-in-theAge-of-Automation-and-Artificial-Intelligence.pdf, (09.05.2019). Korinek, A., Stiglitz, J.E., (2017). “Artificial Intelligence and Its Implications for Income Dıstribution and Unemployment”, NBER Working Paper 24174, http://www.nber.org/papers/w24174, (20.05.2019). Leroux, C., Labruto, R., Boscarato, C., (2012). Suggestion for a Green Paper on Legal Issues in Robotics, (Ed. Christophe Leroux, Roberto Labruto), Grant Agreement Number: 248552, 01.01.2010–31.12.2012, https://www.unipvlawtech.eu/files/euRobotics-legal-issues-in-robotics-DRAFT_6j6ryjyp.pdf, (13.05.2019). Marwala, T., (2018). On Robot Revolution and Taxation, Cornell University, 05.08.2018, https://arxiv.org/abs/1808.01666, (09.05.2019). Mazur, O., (2018). “Taxing the Robots”, Pepperdine Law Review (Forthcoming), Vol: 46, pp. 2–48, https://papers.ssrn.com/sol3/papers.cfm?abstract_ id=3231660, (13.05.2019). Oberson, X., (2017). “Taxing Robots? From the Emergence of an Electronic Ability to Pay to a Tax on Robots or the Use of Robots”, World Tax Journal, May, https://www.ibfd.org/sites/ibfd.org/files/content/pdf/wtj_2017_02_ int_3_SeptNewsletter.pdf, (20.08.2019). Ooi, V., Goh, G., (2019). “Taxation of Automation and Artificial Intelligence as a Tool of Labour Policy”, https://ssrn.com/abstract=3322306, (13.05.2019). Yaltı,S.B., (1995). Uluslararası Vergi Anlaşmaları, Beta Publishing, İstanbul. Yüksel, E.B., (2018). “Yapay Zekânın Buluşlarının Patentlenmesi”, Uyuşmazlık Mahkemesi Dergisi, Y. 6, Issue 11, June, pp. 585–622. Zorluel, M., (2009). “Yapay Zekâ ve Telif Hakkı”, TBB Dergisi, Sayı:142, pp. 305–356, http://tbbdergisi.barobirlik.org.tr/m2019-142-1851, (20.05.2019).

List of Figures Fig. 1: Turkey Tax Revenues, 2006M01 – 2018M12. Source: General Directorate of Budget and Fiscal Control and General Directorate of Budget and Fiscal Control. * the t – Statistic Value of the Augmented Dickey-Fuller Test, Indicating Nonstationarity of the Series at Level 0.05 ������������������������������������������������������������������������  17 Fig. 1: Tax Literacy Framework. Source: (Bomman & Wasserman 2018: 7) �������������������������������������������������������������������������������������������������������  60 Fig. 2: Financial Information, Behaviors and Attitudes. Source: (OECD, 2017:8) ��������������������������������������������������������������������������������������������������������  62 Fig. 3: Financial Literacy Level by Education Level. Source: (Lusardi & Mitchell, 2014: 19) ������������������������������������������������������������������������������������  63 Fig. 1: The Dimensions of Digital Transactions. Source: Fortanier and Matei (2017): 10 ����������������������������������������������������������������������������������������  98 Graph 1: Expenditure on Retail Pharmaceuticals by Type of Financing (2017). Source: OECD (2017) ������������������������������������������������������������  114 Graph 2: OECD Pharmaceutical Expenditures – 2017 (%of Total Health Spending). Source: OECD Health Statistics (2017) ������������������������  115 Fig. 1: Original Armey Curve. Source: Armey, 1995: 92 .................................  140 Fig. 2: Armey Curve. Source: Vedder ve Gallaway, 1998: 2 .............................  141 Fig. 3: The Size of Government and Economic Growth Rates in Turkey (1981–2018). Data Source: Ministry of Treasury and Finance, Republic of Turkey (2019) �����������������������������������������������������������������������  146 Fig. 4: The Armey Curve in Turkey ....................................................................  151 Fig. 1: CUSUM Test. Author’s elaboration .........................................................  162 Fig. 2: CUSUMQ Test  ...........................................................................................  163

List of Tables Tab 1: Percentage Distribution of Tax Revenues in Turkey .............................  14 Tab. 2: The results of the ARIMA(0,1,2)(0,1,1)12 mode ....................................  20 Tab. 3: The results of the BATS (0.377, {0,0}, 1, {12}) Mode ............................  21 Tab. 4: Point Forecasts of the Methods for the Testing Data (2016M01–2018M12 ................................................................................  22 Tab. 5: Measures Accuracy of the Methods for Testing Set (2016M01–2018M12 ................................................................................  23 Tab. 1: Comparison of the use of proceeds for 2016 and 2017. Reference: Green Bonds: Review of 2017, s. 11. https://www. environmentalfinance.com/assets/files/Green%20Bonds%20 Review%20of%202017.pdf (16.02.2019) ...............................................  41 Tab. 2: Comparison of Issuer Types (2016–2017). Reference: Green Bonds: Review of 2017, s. 11. https://www.environmentalfinance. com/assets/files/Green%20Bonds%20Review%20of%202017.pdf (16.02.2019) ...............................................................................................  41 Tab. 3: Examples of Tax Incentives Related to Green Bonds. Reference: Climate Bonds Initiative & IISD-International Institute for Sustainable Development 2016: 14; the Author Added France, China, and Turkey ..........................................................  51 Tab. 1: Actors and Determinants in the Formation of Tax Climate. Source: (Alm et al., 2012: 136) ................................................................  58 Tab. 2: Cognitive-Affective-Psychomotor Scope of Tax Literacy. Source: (Yılar & Akdağ, 2017000: 366) ..................................................  59 Tab. 3: Taxjazz Project Processes. Source: http://search.ebscohost.com/ login.aspx?direct=true&db=a9h&AN=48076039&lang=tr&site= ehost-live, The Tax Literacy Project, www.taxjazz.com (01.04.2109) ...............................................................................................  64 Tab. 4: Some Studies Related to Tax Perceptions and Their Target Population in Turkey. Source: Own Elaboration Based on the Studies Above ............................................................................................  67 Tab. 5: Tax Topics and Activities in Social Studies Education in Turkey. Source: (Yılar & Akdağ, 2017: 375) ........................................................  68 Tab. 1: Top 10 Most Complex Jurisdictions for Accounting and Tax Compliance (2017–2018). Source: The Financial ComplexityIndex, 2017, 2018 ......................................................................................  106 Tab. 2: Complexity Issues as Globally and Regional (2017, %). Source: The Financial Complexity-Index 2017: 26 ..............................  108

208

List of Tables

Tab. 1: Review of Literature. Source: Composed by the Authors ...................  116 Tab. 2: The Description of Dat .............................................................................  117 Tab. 3: GMM Result ..............................................................................................  118 Tab. 1: Tax Expenditures and Gross Domestic Product (2006–2019) (Million TL) ...............................................................................................  128 Tab. 2: Tax Expenditures and Public Expenditures (2006–2019) (Million T) .................................................................................................  129 Tab. 3: Tax Expenditures and Tax Revenues (2006–2019) (Million TL). Source: The Author Prepared It According to the Statistics of the Turkish Statistical Institute, the Ministry of Treasury and Finance, and Annual Central Government Budget Figures ...............  130 Tab. 4: Numerical Development of Legislation on Tax Expenditures (2006–2019). Source: The Author Prepares the Annual Central Government Budget .................................................................................  131 Tab. 5: Number Distribution of Tax Expenditures Regarding Tax Laws (2006–2019). Source: The Author Prepares the Annual Central Government Budget .................................................................................  132 Tab. 1: ADF Unit Root Test Result ......................................................................  147 Tab. 2: (2) Numbered Equation Bound Test Result ..........................................  148 Tab. 3: ARDL (3,4,2) Model Long-Term Coefficient Estimation ....................  149 Tab. 4: ARDL Model Error Correction Coefficient Estimation ......................  150 Tab. 1: Economic Growth and Unemployment Rates in Turkey (1980–2016). Source: IMF (International Monetary Fund) (http://www.imf.org, 22.01.2017) ...........................................................  157 Tab. 2: Literature Review ......................................................................................  158 Tab. 3: Descriptive Statistics of Economic Growth and Unemployment Data ............................................................................................................  161 Tab. 4: ADF and PP Unit Root Test Result ........................................................  162 Tab. 5: Diagnostic Tests ARDL (1, 1) Mode ......................................................  163 Tab. 6: Bounds Test ...............................................................................................  164 Tab. 7: Error Correction Mode. Author’s elaboration ......................................  164 Tab. 8: Diagnostic Test Result. Author’s elaboration ........................................  165

Adnan Gerçek/Metin Taş (eds.)

Critical Debates in Public Finance This book examines the main issues discussed in the field of public finance today. These issues are perhaps identified among policy areas that will come to the agenda of many governments over the next decade. Topics covered in the book are as follows; revenue forecasting models, the taxation of sharing economy, tax incentives provided to green bonds in financing of energy efficiency, the importance of tax literacy in tax compliance, the concept of collective investment institutions, digitalization of tax administration and complexity of tax system, macro determinants of pharmaceutical spending, tax expenditures as internal tax bleedıng, the size of the public sector and the Armey Curve, Okun’s Law, subsidies granted to the private educational institutions, and taxation of artificial intelligence. The book consists of twelve chapters on “controversial issues in the public finance” mentioned above. The authors of the chapters also offer some policy recommendations regarding their work.

The Editors Adnan Gerçek is Professor of Fiscal Law at Bursa Uludağ University, Faculty of Economics and Administrative Sciences, Department of Public Finance, Bursa, Turkey. He has a PhD.  from Uludağ University Social Science Institute. He is member of the Turkish Tax Council. His research focuses on tax administration, tax collection procedure, taxpayers’ rights, tax responsibility, discretionary power of tax administration, tax literacy, and the e-taxation system. Metin Taş is Professor of Fiscal Law at İstanbul Gedik University, Faculty of Economics Administrative and Social Sciences, Department of Political Science and Public Administration, İstanbul, Turkey. He has a PhD.  from Uludağ University Social Science Institute. He is the chair of the Department of Political Science and Public Administration. He is also a certified public accountant. His research centers on tax criminal law, tax jurisdiction, Turkish tax system, and tax practices.