263 60 7MB
English Pages 286 Year 2004
PRAISEFOR
InsidetheYieldBook TheClassicThatCreatedtheScienceofBondAnalysis BY S IDNEY H OMER AND M ARTIN L.L EIBOWITZ ,P H .D.
“For active ixed-income investors, this book has always been the holy writ,whereitallbegan.Leibowitz’srevisionsmakecertainthatthebook willcontinueinthatrole.Butforallinvestors,ofanystripe,InsidetheYield Book is the essential work for understanding what the entire investment processisallabout.” —P ETER L.B ERNSTEIN AuthorofAgainsttheGodsandCapitalIdeas President,PeterL.Bernstein,Inc.
“It’s pure pleasure to revisit this innovative and authoritative bedrock ofmodernbondanalysis.Theneweditiondemonstrateshowtheauthors’ pioneeringapproachcanbeappliedbroadly.Youcanignoretheanalysis if you’d like; however, the description of Wall Street’s original ‘rocket scientist’encounteringSalomonBrothers’tradingloorisworththeprice ofadmission.” —J OHN L IPSKY ChiefEconomist,JPMorganSecuritiesInc.
“This book transformed the markets’ understanding of bonds. The new materialinthisexpandededitionextendsthoseinsightstoequitiesand otherinvestments—andtheinvestmentworldisindebtedanewtoSidney HomerandMartinLeibowitz.” —F RANK J.F ABOZZI ,P H .D.,CFA EditoroftheJournalofPortfolioManagement FrederickFrankAdjunctProfessorofFinanceatYaleUniversity’s SchoolofManagement
“InsidetheYieldBookmaynotbewhereitallbegan,butitcertainlyiswhere it all began to be understood. The clarity and elegance of language and thoughtisstartling.Thenewmaterialwillnotdisappoint.Thisbookwill liveforever!” —J ACK R.M EYER PresidentandCEO,HarvardManagementCompany
“Assomeonewho’sbeeninthebusinessforforty-eightyearsandstillmakes gooduseoftheoriginaledition, I’mverypleasedtoseeInsidetheYield Bookbackinprintwiththeveryhelpfuladditiononpresentvalue.” —G EDALE H OROWITZ SeniorManagingDirector,CitigroupGlobalMarketsInc.
“Inthe1970sand1980s,webuiltstrongixed-incomefranchisesbothatthe HarrisBankandLincolnCapitalbasedonMarty’sworkcontainedinInside theYieldBook.Few,ifany,mathematicalstudieshavehadtheimpactonan industrythatthisbookhas.Morethanthirtyyearslater,itstillshouldbe mustreadingforeveryinvestmentprofessional.” —KENNETH R.M EYER Chairman,LincolnCapitalManagementCompany ALehmanCompany
“Inside theYield Book brought bond trading out of the DarkAges and the investment world’s brightest stars into bond trading. Its enduring popularity relects the authors’ rare ability to provide sound solutions to practicalconcernsinclear,economicalprose.” —M ARTIN F RIDSON PublisherofLeverageWorld MemberoftheFixedIncomeAnalystsSocietyHallofFame
“Theneweditiongivesusthebestofbothworlds:thecompleteclassicthat helpedlaunchagoldenageforixed-incomeinvesting,andcompletelynew sectionsthatshowhowrelevantandessentialitswisdomremainstoday.” —L EWIS S.R ANIERI Chairman,HyperionPartners
“When I was a young mortgage security trader in the mid 1970s, I found InsidetheYieldBooktobeanabsolutelyessentialreferencetool.Today,as ixed-income markets grow even more complex and global risk management becomes mandatory, every participant in the ixed-income market needstoreadandthenrereadthisauthoritativework.” —L AURENCE F INK ChairmanandCEO,BlackRock
“There have been many developments in ixed-income analysis, and theyallbeneitfromtheseminalpublicationInsidetheYieldBook.The foundation that was established supports and inspires the work which hasfollowed.Marty’sadditionalmaterialaddstothelegacyofhispathbreakingefforts.” —H .G IFFORD F ONG President,GiffordFongAssociates
“Just as thousands of years ago, the Greeks measured the diameter of the earthquiteaccurately;andjustastheChinesecircumnavigatedtheglobe longbeforeColumbusandMagellan;soSidneyHomerchartedtheconstellation of bond mathematics long before modern computers. Since bonds inancehumanactivity,thisbookisfulloffascinatinghistory,anditisnot only Sidney’s. Marty Leibowitz’s ingerprints are all over it, too. Sidney wasthehistorian,andMartyisthepoet,ofthemostimportantinancial marketonearth:bonds.” —A NDREW M.(A NDY )C ARTER ViceChairman,HyperionCapitalManagement,Inc.
“Thisexpandedversionoftheclassicstudycontinuestomakeasigniicant contribution to our understanding of interest rates and debt pricing. It should be required reading for both practitioners and academics. Martyhastheabilitytoexplaincomplexrelationshipsinthedebtmarketsin alogicalandoftenintuitiveway,andtothenbacktheintuitionwithrigorous mathematics.” —M ARTIN J.G RUBER NomuraProfessorofFinance,SternSchoolofBusiness NewYorkUniversity
“InsidetheYieldBookistheixedincomeclassic,andithasbeenupdated withafascinatingchapterofbondmarkethistory. Somebooksgetbetter withage,andthisisoneofthem.” —B RIAN S.O’N EIL ChiefInvestmentOficer,RobertWoodJohnsonFoundation
“WhenmysonearnedhisMBA,Igavehimanengravedwristwatchanda dog-eared copy of Inside theYield Book.The watch will help him get to workontime,butthebookwastherealgift. You cannot succeed as an investor without understanding the concepts Inside the Yield Book teaches. Marty Leibowitz has done a great service to all investors by bringingbackthisclassic.” —B RIAN F.W RUBLE ,CFA GeneralPartner,OdysseyPartners,L.P. PastChairman,InstituteofCharteredFinancialAnalysts
“Withtheglobalnatureoftoday’sinvestmentmanagementprocessandthe increasingcomplexityofinancialinstruments,weseemfarremovedfrom thebondmarketsthatSidneyHomerandMartyLeibowitzirstwroteabout in1972.However,Sidney’sintuitionandMarty’smathematicalrigorgave allofusthefoundationfromwhichalloftoday’sanalyticalapproachesemanate.Theirworkwasinnovativeatthetimebut,ifanything,maybeeven morerelevanttoday.” —T HOMAS E.K LAFFKY ManagingDirectorandHeadoftheYieldBookGroup,Citigroup
INSIDE THE YIELD BOOK
ALSOAVAILABLEFROMBLOOMBERGPRESS FixedIncomeSecuritiesandDerivativesHandbook: AnalysisandValuation byMooradChoudhry (October2004) TheSecuritizationMarketsHandbook: IssuingandInvestinginMortgage-andAsset-BackedSecurities byCharlesAustinStoneandAnneZissu (November2004) PIPEs: AGuidetoPrivateInvestmentsinPublicEquity editedbyStevenDresnerwithE.KurtKim NewThinkinginTechnicalAnalysis: TradingModelsfromtheMasters editedbyRickBensignor
Acompletelistofourtitlesisavailable atwww.bloomberg.com/books
A TTENTION C ORPORATIONS Thisbookisavailableforbulkpurchaseatspecialdiscount.Specialeditions or chapter reprints can also be customized to specifications. For information, please e-mail Bloomberg Press, [email protected], Attention: Director of SpecialMarkets.
INSIDE THE YIELD BOOK TheClassicThatCreated theScienceofBondAnalysis
SIDNEYHOMER andMARTINL.LEIBOWITZ,Ph.D. WithaForewordbyHenryKaufman andtwonewsectionsbyMartinL.Leibowitz
PRINCETON
©2004byMartinLeibowitz,Ph.D.,forForeword,Preface,Acknowledgements,etc. ©Renewed2004byMartinL.Leibowitz,Ph.D,andSidneyHomer’sheirs,LouiseHomerHannum,MarionHomer PainterandGeorgianaHomerDaskais. ©1972Prentice-Hall,Inc. Allrightsreserved.ProtectedundertheBerneConvention.PrintedintheUnitedStatesofAmerica.Nopartofthisbook maybereproduced,storedinaretrievalsystem,ortransmitted,inanyformorbyanymeans,electronic,mechanical,photocopying,recording,orotherwise,withoutthepriorwrittenpermissionofthepublisherexceptinthecaseofbriefquotationsembodiedincriticalarticlesandreviews.Forinformation,pleasewrite:PermissionsDepartment,BloombergPress, 100BusinessParkDrive,P.O.Box888,Princeton,NJ08542-0888U.S.A. BLOOMBERG, BLOOMBERG NEWS, BLOOMBERG FINANCIAL MARKETS, OPEN BLOOMBERG, THE BLOOMBERG FORUM, COMPANY CONNECTION, COMPANY CONNEX, BLOOMBERG PRESS, BLOOMBERGPROFESSIONALLIBRARY,BLOOMBERGPERSONALBOOKSHELF,andBLOOMBERG SMALLBUSINESSaretrademarksandservicemarksofBloombergL.P.Allrightsreserved. Thispublicationcontainstheauthors’opinionsandisdesignedtoprovideaccurateandauthoritativeinformation. Itissoldwiththeunderstandingthattheauthors,publisher,andBloombergL.P.arenotengagedinrenderinglegal, accounting,investment-planning,businessmanagement,orotherprofessionaladvice.Thereadershouldseekthe servicesofaqualifiedprofessionalforsuchadvice.Theauthors,publisher,andBloombergL.P.cannotbeheldresponsibleforanylossincurredasaresultofspecificinvestmentsorplanningdecisionsmadebythereader.
PrintedintheUnitedStatesofAmerica 13579108642
LibraryofCongressCataloging-in-PublicationData
Homer,Sidney,1902–1983 Insidetheyieldbook:theclassicthatcreatedthescienceofbondanalysis/SidneyHomerandMartinL. Leibowitz;withaforewordbyHenryKaufmanandtwonewsectionsbyMartinL.Leibowitz. p.cm. Includesbibliographicalreferencesandindex. ISBN1-57660-159-5(alk.paper) 1.Bonds--UnitedStates.2.Bondmarket.I.Leibowitz,MartinL.,1936–II.Title. HG4936.H652004 332.63'23--dc22 2004006036
To the memory of my coauthor, Sidney Homer, who irst introduced me to the inancial marketplace with its intriguing challenges, its colorful personalities, and its many fascinating problems that I am still struggling to more fully understand. Martin L. Leibowitz
This page intentionally left blank
Contents ForewordbyHenryKaufman ...................................................... ix Prefacetothe2004Edition:AHistoricalPerspective .................. xiii Acknowledgments .................................................................. xxiii SomeTopicsThatDidn’tMakeItintothe1972Edition ............. xxv TechnicalAppendixto“SomeTopics” ...................................... lxiii References ............................................................................ lxxiii InsidetheYieldBook(OriginalEdition) .............................. 1–205 Prefacetothe1972Edition.............3 Contentsofthe1972Edition ..........7 ListofTables ................................13
AbouttheAuthors ................................................................... 207
This page intentionally left blank
Foreword byHenryKaufman
ISTILLRECALLwhenin1969SidneyHomeraskedmetomeetwithMartin Leibowitzinordertoascertainhowhemightitintoourresearcheffortat SalomonBrothers.SidneyHomerhadjoinedSalomonin1960tosetupa bondmarketresearchdepartment.Hehadhiredmeayearlater.Myrole eventuallyexpandedtoassumingoverallresearchresponsibilities.Bythe timeMartyarrived,Sidneyhadalreadypublishedhismonumentalbook, AHistoryofInterestRates,followedbyanumberofpapersonthebehaviorofinterestratesfromaportfoliomanagementperspective.However, SidneyandIconcludedthatthetimewasnowripeforamorequantitative approachtoevaluatingtheopportunitiesandpitfallsinthebondmarket. Withadoctorateinmathematics,Martywaswellpreparedtojoinwithus inthisnewpursuit. IshouldexplainthetitleInsidetheYieldBook.Marketveteranswill understand the title, but newcomers generally will not. Back in those days,the“yieldbook”wasacompilationofnumericaltablesofprices and yields for a wide range of bond maturities. Traders and investors wouldagreeontheyieldbasisforatradeandthenlaboriouslyplowinto thesetablestodeterminethecorrespondingprice. Thereweretwoproblemsthatbecameincreasinglytroublesomeas thebondmarketbegantoexperienceenormousgrowthinthesizeand breadthofnewissues.Theirstproblemwasthat,withtheincreasing rangeofinterestratesduringthe1960sand1970s,thesevolumesbecame thicker and thicker, resulting in an ever longer time required to indtheinterpolatedpriceandthennegotiateagiventrade.Today,the computerhasbasicallysolvedthisproblem.Thesecondproblemwas that bond market participants—portfolio managers as well as broker/ ix
x
FOREWORD
dealers—had come to rely upon these yield books as gospel, with all too little understanding of the underlying mathematical and inancial concepts. Sidney Homer had long been urging market participants to move toward a more total-return orientation that would be consistent withmoderninancialtheory. When Sidney’s insight and experience were combined with Marty’s abilitytoprobebeyondtherigidtabulationsoftheyieldbook,theywere abletocoauthoraseriesofstudiesthatopenedawiderandmorecompellingvistaintothisnewworldofbondanalysis.Thesestudiessooncame tohaveaprofoundimpactonthebondcommunity,anditwasthatreceptionthatledtothepublicationofthisbook’sirsteditionin1972. TheinsightsandobservationsinInsidetheYieldBookareastruetoday astheywerethen.Virtuallyeveryoneoftheifteenchapters,startingwith theirstpartthatdealswith“BondYields,BondPrices,andBondInvestment”andconcludingwiththechapterson“TheMathematicsofBond Yields,”havestoodthetestoftime.Inparticular,thechaptersonpresent value,interestoninterest,andthedifferentratesofreturnhaveoftenbeen citedasamongtheclearestexpositionofthesekeyconcepts,whichare fundamentaltounderstandingtheanalysisofcashlowsnotonlyinthe bondworldbutinanyareaofinance. Following the publication of Inside the Yield Book, Marty’s career blossomed. He became the head of an important division (in research) knownastheBondPortfolioAnalysisGroup.AfterIleftSalomonBrothersin1988,Martybecamedirectorofglobalresearch,responsibleforall oftheirm’sresearchactivitiesinequitiesaswellasixedincome.Marty attractedanumberofhighlycompetentquantitativeanalysts,manywith Ph.D. degrees, not just in economics, but in mathematics, engineering, andevenastronomy.Theirworkbecamevitaltotheirm’stradingdesk andwasresponsibleformanyclienttransactions.Manyoftheseanalysts wentontoachievewiderecognitionontheirown. Martyledtheresearcheffortbyexample.Overthecourseoftheyears, therewasanoutpouringofwritings.Manywerepublishedinbookform inInvesting:TheCollectedWorksofMartinL.Leibowitz(1991),Franchise Value and the Price/Earnings Ratio (1994), and Return Targets andShortfallRisks(1996).Therangeofsubjectscoveredinthesepages is truly breathtaking. Throughout their combined 2,000 pages, Marty’s greatintellectandbroadrangeofknowledgeiseverpresent.
FOREWORD
xi
In1995,Martyjoinedoneofthelargestretirementfundsintheworld, TIAA-CREF,aschiefinvestmentoficer.Inthiscapacity,hewasableto putmanyofhisinvestmentideasandtheoriesintopractice. Marty’s career has been characterized by the persistent search for a deeperunderstandingofthemostbasicinvestmentconcepts,anapproach thatirstcametolightinhisworkonInsidetheYieldBook.
This page intentionally left blank
Prefacetothe2004Edition: AHistoricalPerspective byMartinL.Leibowitz WHENFIRSTPUBLISHEDIN1972,InsidetheYieldBookmadeabigsplashin averyshallowpond.Fromtoday’svantagepoint,ixed-incomeactivities back in those days can seem somewhat naïve, and perhaps even rather dull.ButwhenIirstfoundmyselfonthebondtradinglooratSalomon BrothersandHutzler(SB&H),itwasanythingbut.Infact,ithadsome asylumcharacteristics,withtradersandpartnersshoutingandscreaming, bangingphonesdowninfrustration,arguingbitterlywithtradersacross theaisle,sometimesholdingtheirheadsindespairor,whenadynamite tradewasconsummated,givingeveryone“highives”andevenoccasionallyjumpingonthedeskforavictorydance. Perhapsthemoststrikingfeaturewasthetraysofhalf-eatenlunches, severaldaysold,thatcouldbestackedthreeandfourhighonthetrading desks.Thetradersalwaysateattheirdesksandrarelyhadtimetoclear awaytheiruneatensandwiches.AsafreshlymintedPh.D.inmathematics,thiswasnotexactlymyexpectationofhighinance.Especiallynot after riding up in the oak-paneled elevator, entering the elegantly oakpaneled foyer, and stealing a yearning glimpse into the elegantly oakpaneledPartner’sDiningRoom,withits—yes,elegantoak-paneledtable. NorwasthiswhatIwasexpectingafterlearningaboutSalomonBrothers andHutzlerfromSidneyHomer. SidneyHomerwasremarkableonmanycounts.Firstofall,hewasmy wife’suncle,whichishowIirstmethim.Infact,whenmywife,Sarah, andIweremarriedin1966,itwasSidneywhoaccompaniedherdownthe aisleinlieuofherfatherwhohadpassedawayyearsearlier.Sidneywasa manwithapatricianpresence.Hewasveryintelligent,inishinganedu-
xiii
xiv
PREFACETOTHE2004EDITION
cationinclassicsatHarvardinthreeyears,andwasanembarrassingly giftedwriter(atleast,heembarrassedmeonrepeatedoccasionswithhis uncannyknackforturninganinitiallyawkwardphraseintoaneloquent statement).Hisparentswerebothmusiciansoftheirstrank.Hisfather, alsonamedSidneyHomer,wasacomposerofsuperbclassicalartsongs, andhismother,MadameLouiseHomer,wasarenownedmezzo-soprano with a long career at the Metropolitan Opera, singing with the greats suchasEnricoCaruso.Sidney’sparentsweresoexaltedintheirartistic sphereandsobroadlyreveredthattheymingledwiththeuppercrustof NewYorksociety.Alas,asisalmostalwaysthecasewiththearts,talent andfundingfollowseparatepaths:Sidneymayhavehadanaristocratic upbringing,buthisfamilywasfarfromwealthy.SowhenSidneyfellin loveandmarriedataveryyoungage,hehadtogotowork.Andtheonly workthathecouldindatthetimewastojoinaWallStreetbondirm— acrassdescentintocommercialismbyhisfamily’sstandards. Butqualityshinesthrough.Sidneybecameahighlyskilledbondmanager,spendingthelargerpartofhiscareeratScudder,StevensandClark. Alongtheway,hisinquiringmindledhimnotjusttoparticipateinthe bondmarket,buttostudyitdeeplyandwriteabouthisindings.Sidney’s literary skills were exceptional by inancial market standards, and his bond market studies gained a wide following. He soon became known asthe“bardofthebondmarket”—anhonorarytitlethatnooneelsehas everheldsince(orperhapseveraspiredto).Hewroteseveralbooksthat becameclassicsintheirday.Oneofhisbooks,TheBondBuyer’sPrimer,1 wasatongue-in-cheekstorydescribinghowabondsalesmanshouldgo aboutsellingbonds,andhowabondbuyershouldgoaboutresistinghim. Itisnowoutofprintbuthighlyprizedbythosewhohaveacopy. One of Sidney’s enduring works is his monumental study, The HistoryofInterestRates.2Mostfinancialwriters would have been content withcoveringthelasttwocenturies,butSidney’sclassicaltrainingand avocational interest led him to extend his history back to pre-Biblical times.Hemanagedtotakeapotentiallytedioussubjectandmakeitinto afascinatingstoryrelatingcyclicalsweepsininterestratestograndsocietalchanges.Thebookhasbeenthroughseveraleditions,withthelatest updatedbyHenryKaufmanandRichardSylla,aneminentinancialhistorianfromNewYorkUniversity.ThepublicationofTheHistoryadded atouchofclasstothebondmarket,andeveryseriousparticipanthada
PREFACETOTHE2004EDITION
xv
copy. Its most thorough reader was undoubtedly the young Dr. Henry Kaufmanwho,afterbeinghiredbySidneyatSB&H,actually“volunteered”toproofreadall472pages. Intheearly1960s,SB&Hwasthepremierbondtradingirm,butit wasbasicallyjustthat—abondtradingirm.Oneoftheseniorpartners, CharlesSimon,realizedthatfortheirmtolourish,itneededtoprovide servicesbeyondthebestpricetoitscustomers.Hehitupontheideaof creatingabondmarketresearchdepartment—theirstonWallStreet— andheenticedSidneytoheaditup. Back in those days, bond trading was an arcane backwater. I had literallyneverheardofSB&H,norhadalotofothergenerallywell-informedpeople.MywifeandIhadvisitedSidney’sGramercyParkhome on numerous occasions, but we never discussed Wall Street until one eveningwhenSidneylearnedthatIwouldsooncompletemydoctoratein mathematics.Hepulledmeasideanddugoutailethatcontainedabout iftyhandwrittenpagesofanever-inishedbookentitledTheMathematicsofBonds.Heexplainedthatthisprojectwasbegunmanyyearsbefore, butithadfounderedintoiledom.Hehadstartedthebookbysettingforth a number of principles that he (and virtually every other bond market participant)wassureweretrue.Oneoftheseprincipleswasthatlongermaturitybondshavegreaterpricevolatilitythanshorterbonds.However, ashedevelopednumericalexamples,hefoundthattheycontradictedhis “rock certain” principles. After years of letting his manuscript collect dust,henowwonderedifIwouldtakealookatittoseeifmymathematicalbackgroundcouldhelpuntangletheparadox. Imustconfessthat,atthatpoint,myknowledgeofbondswasnonexistent.ButIgamelytookthepageshomeandworkedupthe(relatively straightforward)algebrathatdeinedyield-to-maturityandrelatedittoa bondprice.WhenIpresentedmyindings,hewasdulyappreciativeand gracious,ashealwayswas.However,atthispointneitherSidneynorI wasparticularlyexcitedbymyexplanations.Theymightbeilluminating, butneitherofussawhowtheycouldprovereallyuseful.Outofcuriosity, IaskedSidneywhyhehadn’ttakentheproblemtothe“housemathematician”atSB&H.Hefoundgreathumorinmyquestion,becausenooneat SB&Hcameclosetoittingthatdescription.Thatsurprisedme.Iwould havethoughtthatthepremierbondirm,tradinginstrumentsthathadso many mathematical facets, would surely have some in-house expertise
xvi
PREFACETOTHE2004EDITION
ofthatsort.AsIponderedhisresponse,itdawneduponmethatmaybeI couldbecomethat“housemathematician.” MyrouteintoSB&Hwasmorecircuitousthanonemightimagine. Atirst,Sidneywasnotencouraging.Hisresearchdepartmentwasvery compactatthatpoint,andhecertainlydidnothaveroomforsomeonenot steeped in the inancial markets. But persistence pays off (sometimes), and one day Sidney heard that two of his associates, Morris Ofit and Harry Peterson, were looking for someone to develop computer-based analysesthatcouldfacilitatevarioustradingactivities.Onethingledto anotherandIeventuallyfoundmyselfmanningthesingle,time-shared computerterminalonthetradingloorat60WallStreet(theonewiththe stacksofoldlunchtrays). My irst two weeks at SB & H were spent going through their rudimentarytrainingprogram.Thisconsistedofsittingnexttovarioustraders, pluggingintotheirphonelines,andlisteningtotheirdialogues.Because trade talk is almost always highly compressed, clipped, super-fast, and repletewithmarketjargon(andotherspecialtywords),thiswasanarduous learning experience. Coincidentally, on my very irst day, I was assignedtositnexttoayoungbutclearlyup-and-comingequitytraderby thenameofMikeBloomberg. Workonthetradingloorwashectic,butitgavemeagreateducation abouttheinancialmarketsandthetransactionprocess.Thetradersand salesmen were generally kind to me. They became even kinder when Iwasabletodevelopapackageofcomputerprogramsthatfacilitated anumberoftrades.Also,withmylittletime-sharingterminal,Icould determinetheyieldforanygivenpricewithgreatspeedandaccuracy. However, the traders were themselves very adept at using the look-up tables—theirso-called“yieldbooks”—toindtheyieldvaluesrequired tocompletetheirtrades.So,atirst,my“high-tech”yieldcalculatorwas justacuriosity.Butin1970,wheninterestratesmovedhigherthanthe levels available in any of the traders’ yield book tables, I became the only game in town. Senior partners lined up in front of my terminal, desperateforthenumberthatcouldconirmtheirlatesttrade.Needless to say, this boosted my standing on the loor, although it put me in a harrowingpositioninwhichanymistakecouldprovefatal.Inacurious sense, one might say that I beneited from interest rates moving “outside”theyieldbook.
PREFACETOTHE2004EDITION
xvii
Itissaidthatneedisthemotherofinvention.Intheinancialworld, the gestation period can be very compressed. It was not many weeks before computer terminals and clunky special-purpose hand-held yield calculators sprouted up all over the trading loor. My reign as the sole “yieldkeeper”cametoanabruptend,allowingmetoreturntodeveloping models for analyzing portfolio improvements involving corporate bondsandconvertibles. Thebroad-baseduseofcomputerstoreplacetheyieldbooktables didnothingtofurtherthegeneralunderstandingofwhatabondyield wasallabout.Infact,thegreatfacilityofthecomputermayhavebeen astepbackward.Traderscouldpunchinafewnumbersandthedesired yield value would pop up. There was no need to ponder what it all meant,orwhatwouldhappenifyouchangedthecouponorthematurity.Atleasttheoldyieldbookprocedurerequiredatablelook-upand aninterpolationthatforcedthetradertomoveaingerupanddownthe yieldrowsandacrossthematuritycolumns.So,inasense,theadvent ofthecomputercapabilityactuallyreducedtheneedforanappreciationofwhatayieldreallymeant.Thismaybeageneralproblemofour computerage.Thecomputercanbeaneffectivefacilitatorinallsorts ofareas,butitsuseonarotebasisalsodullsthedesiretoseekadeeper understanding. Atthattime,mostbondportfolioswerelong-termoriented,andour analyticalmodelstypicallyfocusedonthelong-termbeneitsofholding onetypeofbondversusanother.Thealternativebondswereusuallyofthe samecreditqualitybutdifferedincoupon,maturity,sinkingfund,and/or callfeatures,therebycreatingdifferentpatternsofcashlowovertime.In ordertomakefaircomparisons,wehadtoassumethatcashreceiptswere reinvestedatacommonsetofhypotheticalinterestrates.Wesoonbegan to notice that the return from a given bond investment depended criticallyontheassumptionofacommonreinvestmentopportunityandon thechoiceofthatreinvestmentrate.Thisindinggreatlysurprisedmany ofthebondveterans. Basically,theirconfusionstemmedfromthewidespreadbeliefthata bond’syielddescribedtheaccumulationofwealththatwouldbegeneratedoveritslife.FrommyearlierdiscussionswithSidney,Iknewthat thiswasnotthecase.Nowthecomputermodelsvalidatedtheideathat reinvestedratesplayedacriticalroleinthewealthbuildup.Moreover,the
xviii
PREFACETOTHE2004EDITION
computer output made this inding visible through numerical examples thatwouldbehardtoarguewith,althoughthatdidn’tstopmanyofmy tradingloorneighborsfromtrying.Anumberofpathsconvergedwhen Irealizedthat(1)thereinvestmenteffectwastheoreticallyimportant,(2) thiseffectwasalsopracticallyimportant(i.e.,itcouldaffectinvestment decisions), (3) virtually every bond trader, salesperson, and portfolio manager was woefully unaware of this fact, and (4) perhaps most important,mycomputercouldprovideacompellingdemonstrationofthis result.Thisrealizationsignaledopportunity,andIwenttovisitSidneyin hisoff-the-loorofice(oak-paneled,naturally). Sidney was very interested. When I expressed my surprise at how littleknownthisreinvestmenteffectseemedtobewithinthebondmarket (afterall,thiswasnot“rocketscience”),Sidneyobservedthatthereare many myths and half-truths (some of them useful) embedded in daily practice.Sidney’svastexperiencegavehimauniquevantagepoint:He knewwhatwas“known”andwhatwas“notknown.”Itwasthismeeting thatreallygavebirthtoInsidetheYieldBook,althoughIdidn’tknowthat atthetime. A week later, while I was on vacation in Florida, Sidney tracked me downtotellmethathehadbeenruminatingonmyreinvestmentresults. Hehadalreadyputtogetheradraft“MemorandumtoPortfolioManagers,” ashetermedhisresearchreports.Hesaidthatwewouldbecoauthorsof theinishedproduct. Theirstmemorandumintheseries,entitled“InterestonInterest,”was publishedonOctober5,1970.Itwasviewedbymanyreadersasanattack onthesanctityofthestandardyieldmeasure.Therewasconsiderableoutrageamongmanyofthecrustiermembersofthebondcommunity(and there were a lot of crusty members!). Sidney received many indignant calls and letters from valued friends and even more valued customers. Allofthesecommunicationswereturnedovertome,andSidneycharged me with the job of responding to—and convincing—each and every complainant. The bad news was that undercutting the long-held views ofourbestcustomersdidnotexactlyendearmetotheirm’sseniorpartners.Thishardlyseemedliketheidealwaytolaunchaledginginancial career.ThegoodnewswasthatasImethodicallychewedthroughthecorrespondencethatSidneypiledonmydesk,Ifoundmyselfcominginto contactwiththepillarsofthebondcommunity.
PREFACETOTHE2004EDITION
xix
Someweresoothedbymywrittenexplanations,butmanywerenot, whichresultedinsomeprettynervousgrumblingwithinSB&Habout potentialdamagetotheirm’sstanding.Sidneywassteadfastinhissupport.Moreover,heputhisprestigeonthelinebypersonallymakingappointmentsforustovisittheoficesoftheremainingrecalcitrantcustomers.Withmymathematicalarguments,thankfullyreinforcedbySidney’s credibility,wewereinallyabletoattainalevelofacceptancesuficient tokeepSB&H’sreputationintact. The next two memoranda, on price volatility, received a more graciousreception,eventhoughtheysurprisedmanyreadersbypointingout thatlow-couponbondscouldbemorevolatilethanparbondswithmuch longermaturities. ThereisoneparagraphinChapter2thatwasbasedontwoquestions Sidney posed to me. First, if a Roman centurion at the time of Christ hadinvestedonedrachmaandallowedittocompoundatjust4percent through the centuries, what would be the accumulated amount? I was abletoperformthiscalculationanditturnedouttobeahugenumberof drachmas, which, at virtually any exchange rate, would exceed all the capitalwealthnowvisibleintheworld’sinancialmarkets.Hissecond questionIfoundnotsoeasytoanswer:Whathappenedtoallthatpotentialwealth?SidneyhadawaywiththeBigIdeasandansweredthissuccinctly,ifnotwhollysatisfactorily,onpage32ofInsidetheYieldBook: “Aside from the destructive effects of wars, revolutions and inlations, andtheincidenceoftaxes,thereisaveryhumanpropensitytoconsume.” Thisquestionisstillwellworththepondering,notwithstandingitsrather depressingimplications. Subsequentmemorandadealtwithavarietyofsubjects—zero-coupon bonds(longbeforezero-couponbondsactuallyexisted),callablebonds, andthetotal-returnconceptforbondsofdifferentmaturitiesandcoupons. Theseneweffortswerereadilydigestedbyagrowingreadership. Itwasthelastmemorandumthathadthegreatestimpactontheactual practiceofbondportfoliomanagement.Atthetime,virtuallyeverytrade thatinvolvedsellingonebondandbuyinganotherwascalledaswap.The failuretodifferentiateamongdifferenttypesofswapsoftenledtoserious confusionamongmarketparticipants.Theinalmemorandumproposed aclassiicationsystemthatsegregatedthesetradesintofourdistinctcategories:(1)yieldpick-upswaps,(2)substitutionswaps,(3)sectorswaps,
xx
PREFACETOTHE2004EDITION
and(4)rate-anticipationswaps.Thisterminologyprovedusefulasaway ofdistinguishingoneactivityfromanotherandrapidlyworkeditsway intothestandardvocabularyofthebondmarket. Astheygainedbroaderacceptance,theivememorandawerewidely redistributed both in the United States and internationally (they were quicklytranslatedintoJapaneseandGerman).Theyalsofoundtheirway intoinvestmenttrainingprograms,notonlyatSB&Hbutatmanyother WallStreetirms,sometimesinphotocopiedformwiththeSB&Hbannerremoved.ItwasnotlongbeforetheNewYork Institute of Finance andPrentice-Hallurgedustoexpandthememorandaintoabook. Weaddedafewmorechaptersandatechnicalappendixthatdescribed thebasicmathematicsinvolvedincalculatingpresentvalues,yields,and rates of return. Such appendices are usually backwaters rarely dipped into,soweweresurprisedtohearmanyreaderscommentthatoursimple, step-by-stepmathematicaldevelopmentintheappendixhelpedthemto understandthepresentvalueconceptfortheirsttime. THERESULTINGVOLUMEwaspublishedin1972underthetitleInsidetheYield Book.Itsubsequentlywentthroughmorethantwenty-ivereprintings. Theoccasionofthisneweditionisanopportunityforsomeadditional acknowledgments.Asweproceededfromonememorandumtothenext, wecameintocontactwithmarketparticipants as well as scholars who had given deep thought to the fundamental nature of the bond market. We learned from all these individuals things that enlightened us and enrichedourendeavors.Amongmanywhohelpedtomoveusforward, therewereseveralwho,foronereasonoranother,playedaparticularly specialrole. First of all, it turned out that the London gilt irms were far ahead oftheU.S.marketintermsoftheirsophisticationandevenintheiruse ofcomputertools.UnlikeU.S.irms,thoseintheUnitedKingdomhad many senior staff members who were broadly trained actuaries with powerfulmathematicalbackgrounds.ThroughhisnetworkintheUnited Kingdom,SidneyHomerwasabletosendmetoLondonwithintroductionstokeybondpeopleatirmssuchasGreenwellandCo.,Phillipsand Drew, and Grieveson Grant. Our British friends were not used to such visits,buttheyreceivedmewithgreatwarmth.(Thethree-hourluncheon meetings were unlike any I ever experienced—before or since. I think
PREFACETOTHE2004EDITION
xxi
they may now be ancient history in London.) Everyone in these irms was extraordinarily forthcoming about their analytic approaches to the market.Althoughmostoftheirtechniqueswouldnothaveworkedinthe UnitedStateswithoutconsiderablereworking,wewenttogreatlengths toincorporatewhatwecouldoftheirthinkingintoouranalyticaltools, andsomeoftheirideassurelyimprovedourlatermemoranda. ImentionedearlierthatSidneyHomerhadspentthebulkofhislong career as a bond manager at the irm of Scudder, Stevens, and Clark. OneofhisintellectualsoulmatestherewasHermanLiss,abrilliantand creativestudentofthebondandconvertiblemarkets.Atanearlypoint, SidneyHomersentmetomeetwithHermanforaseriesoflunches.Hermanspokesoquicklyandsparkledwithsomanyintriguingideasthathe washardtofollow.IsoonlearnedthatIhadtoexcusemyselffromthe lunch table for a few minutes and sneak off to some corner to quickly scribbledownnotesonhismanyoutpourings.Ihesitatetoconjectureas tohowHermaninterpretedthoseinterruptions;needlesstosay,manyof histhoughtsfoundtheirwayintoourwork. Intheacademicsphere,therewasatthattimerelativelylittleinterest inbonds,althoughthereweresomenotablestudiesbyPeterWilliamson atDartmouth3andbyLarryFisherandRomanWeilattheUniversityof Chicago.4The classic work of Frederick Macaulay5 was, of course, invaluable.AndmanyofuscameamicablytotermswiththeTreasuryyield curvethroughawonderfulbookbyPrinceton’sBurtonMalkiel,TheTerm StructureofInterestRates.6 Finally,itwouldbeunfairnottorecognizethatmanySB&Hcustomers werewellaheadofthepack.Wewerefortunatetobeabletobeneitfrom ourdialoguewiththesethoughtleaderswhohelpedtoshapeourwork. Tosumitupinthelanguageofthebondworld,weallowemanymore debtsofintellectualgratitudethanwecaneverredeem.
This page intentionally left blank
Acknowledgments
I WOULD LIKE TO EXPRESS my deep gratitude to my associates Dr. Brett HammondandDr.StanleyKogelmanfortheirmanyvaluablecomments and suggestions that have found their way into the new material presentedinthisvolume. AtBloombergL.P.,IamgratefultoThomasKeene,BloombergNews editor at large, for championing this new edition, and to my editor at BloombergPress,JaredKieling,andassociateeditorTracyTaitforguidingussmoothlyaroundtheinevitableproblemsthatariseinanyeditorial andpublicationprocess. I must also express my deepest appreciation to my assistant, Celia Blancalor,withoutwhosetirelessandmeticulouseffortinthepreparation ofthismanuscriptthisnewworkwouldneverhaveseenthelightofday. Finally,IwouldliketosalutemyformerpartneratSalomonBrothers, Michael R. Bloomberg, for having the vision to create an organization thatcanproduceanddisseminatesuchworthypublicationsfortheinancialcommunity. MARTIN L. LEIBOWITZ
This page intentionally left blank
SomeTopicsThatDidn’tMakeIt intothe1972Edition byMartinL.Leibowitz THE CONCEPT OFpresentvalue(PV)isbasicallyasimpleonethatplaysa keyroleinvirtuallyeveryareaofinance.Yetsomesurprisingpointsof confusion and gaps in understanding remain, even among experienced inancialanalysts.InInsidetheYieldBook,weaddressedsomeofthese issueswithinthebonddomain.Theoriginalappendixrepresentedadeliberateattempttomovebeyondthesimplemathematicsofpresentvalue andprovideamorecomprehensivefeelingfortheunderlyingmotivations andassumptions.Inparticular,wetriedto relate PV to the more intuitiveconceptofabasiccompoundingprocessthatgeneratesfuturevalue (FV) over time. The following discussion explores a number of topics regardingthePVconceptandhowitinteractswithacashlow’sfuture valueoveraprescribedtimehorizon.Forcompleteness,theequationsare developedintheTechnicalAppendix. In the inal section of this chapter, we go beyond the ixed-income realm and suggest some generalizations of the PV concept that can be usefulinthinkingaboutthevalueofequitiesorvirtuallyanyothertypes ofinvestment.(Somereadersmaywishtoirstmovetothesegeneralizations and then later back into the more mathematical treatment of the ixed-incometopics.)
TheBasicConceptofPresentValue Everyinvestmentisanexchangeofcurrentresourcesforsomefuture lowofpayments.Inthebroadestsense,theconceptofPVisagaugeof thevalueofthosefuturepaymentsincurrentterms.Onecouldarguethat thePVconceptisatwork(atleastimplicitly)ineveryinvestmentdecixxv
xxvi
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
sion,bothintheprimaryandthesecondarymarket.ThePVideaisavery oldconcept,andeveryinvestorhassomeintuitivesenseofhowitworks. Infact,itissuchabasictool,andsowidelyusedandtaught,thatitsapplicationhasbecomealltoosecondnature.Theproblemwiththisisthat useofcommonyardsticks(aswesawinInsidetheYieldBook)caneasily becomerote,thatis,usedroutinelywithoutanythoughtfulapplicationof theirroots—orlimitations. ThefundamentalelementinthePVcalculationisthediscountrate— therateofinterestthatrelateswhataninvestoriswillingtopaycurrently toreceiveafuturepaymentatsomespeciiedpointintime.Thissubject ofdiscountratescanquicklybecomevery complicated. Discount rates canvaryduetoavarietyoffactors:thetimetoeachcashpayment,the riskassociatedwiththepayments,thevolatilityofthediscountrateitself, andsoforth.Forclarityofexposition,weshallkeepitsimpleandassume throughoutthatthemarketappliesthesingle,latdiscountrateof8%to allinvestments. Withthisheroicassumption,anylowofpaymentscanbediscounted atthis8%ratetodetermineaPVthatshouldexactlycorrespondtoits fairmarketprice.Inaveryrealsense,amarketdiscountrateisameasureofsociety’stimevalueofmoney,andmoregenerally,thetimevalue ofscarceresourcesingeneral.(Andonecouldobviouslygoonatgreat lengthabouttherelationshipofinlation,growthprospects,resourcescarcity,consumptionpatterns,etc.) Seenanotherway,thediscountrateisequivalenttothebasicmarket returnonafairvalueinvestment.Thisreturnsviewofthediscountrate (justtheotherfaceofthesamecoin)leadstoaslightlydifferentinterpretation of the PV: The PV is the dollar amount that, if invested and compoundedatthediscountrate,couldproducetheexactsamepattern offuturelowsastheoriginalinvestment.Itshouldbenotedthatbothof the above interpretations of the PV—as a time exchange of current for futuredollars,andasaninvestedamountthatwouldmimictheoriginal investment’s low—make no reference to what happens to those future lows once they are received. The future payment may be spent, reinvested,orjustgivenaway.Whateverthefateofthefuturepayments,the PVwouldbethesame. Asanexample,throughoutthisdiscussion,wewillconsiderthesimplestpossiblecashlow:a10-yearannuityconsistingof10annualpay-
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xxvii
TABLE 1
Cumulative Present Value 10 Annual Payments of $10 8% Discount Rate
Horizon, H
Payments
Present Value of Each Payment, PV (H,H )
Cumulative Present Value, PV (1,H )
0 1 2 3 4 5
$0 10 10 10 10 10
$0.00 9.26 8.57 7.94 7.35 6.81
$0.00 9.26 17.83 25.77 33.12 39.93
6 7 8 9 10
10 10 10 10 10
6.30 5.83 5.40 5.00 4.63
46.23 52.06 57.47 62.47 67.10
11 12 13 14 15
0 0 0 0 0
0.00 0.00 0.00 0.00 0.00
67.10 67.10 67.10 67.10 67.10
mentsof$10each,subjecttoourdiscountrateof8%.Wewillusethis samecashlowexampletoillustrateanumberofanalyticpoints,mostof whichapplyquitegenerallytoanycashlow.InTable1,thethirdcolumn labeledPV(H,H)showsthePVofthepaymentreceivedinyearH.(More precisedeinitionsandmorecompletedevelopmentofthemathematical concepts are presented in the TechnicalAppendix.) Thus, the irst $10 payment at the end of year 1 has a PV(1,1) of $9.26 in current dollar terms.Wecouldalsoturnthisaroundandnotethata$9.26depositwould havegrownto$10wheninvestedforoneyearatan8%rateofinterest: $9.26×1.08=$10.
xxviii
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
Thesecond$10paymenthasaPV(2,2)of$8.57,andsoonupuntilthe lastpaymentinthe10thyearwhichhasaPV(10,10)ofonly$4.63.Thenext column,labeledPV(1,H),istheaccumulationofallpaymentsfromtheirst yearuptoandincludingtheoneinyearH.Thus,thePV(1,2)fortheirst two$10paymentsis$9.26+$8.57=$17.83.ThePV(1,5)ofthelow’sirst 5yearsis$39.93,andthePV(1,10)fortheentirecashlowis$67.10.
TheReinvestedFutureValue However,itisveryusefultomovebeyondthisstrictly“discounting” frameworktothinkintermsofageneralizedfuturevalue(FV)ofaninvestment.Aswithmostthings,it’seasytofallintoanoverlycomplicated discussion.So,keepingitsimple,wecandeineareinvestedfuturevalue RFV(1,H)atsomehorizonHasthetotalfundsthatwouldhaveaccumulatedifallofaninvestment’spaymentsfromyear1toHwerereinvested (andcompounded)atagivenreinvestmentrate.Forthemoment,assume thattimehorizonHcoincideswiththelastpaymentfromtheoriginalinvestment.Returningtoourbasicannuityexample,Table2illustratesthe reinvestmentprocessforthe10-year$10annuity.Theirst$10payment isreceivedattheendofyear1andisimmediatelyreinvestedattheassumed8%rate.Thisreinvestmentgeneratesanadditional$0.80interest, sothattogetherwiththesecond$10payment,theaccumulatedreinvested valueRFV(1,2)=$20.80.Thereinvestmentprocessthencontinuesyear byyearuntiltheendofthe10thyear,atwhichpointtheRFV(1,10)= $144.87.Notethatthissumimpliestheinvestorwillreceive$44.87ininterestinadditiontotheunderlying$100fromtheoriginal10payments. TheRFVconceptpavesthewaytoaparticularlysimple(anduseful) interpretation of an investment’s PV. Consider the total RFV(1,H) that wouldbebuiltupasoftheHthyear.WhenthisRFV(1,H)amountisdiscountedbacktothepresent,itwillalwaysjustequalthePV(1,H)ofthe originalinvestment.Turningthisideaaround,anyinvestment’sPV(1,H) relectsthemagnitudeoftherequiredinvestmentthatwouldgrowona fully compounded basis to the given RFV(1,H), again assuming that a singlemarketinterestrateisusedforbothdiscountingandreinvestments. (SeetheTechnicalAppendixfollowingthissection.) Forthesimplestpossibleexample,considertheRFV(1,1)=$10for theirstpayment.Whendiscountedbacktothepresent8%,thePV(1,1)
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xxix
TABLE 2
Reinvested Future Value 10 Annual Payments of $10 8% Discount Rate
New Interest
Reinvested Future Value, RFV(1,H )
0 1 2 3 4 5
$0 10 10 10 10 10
$0.00 0.00 10.00 20.80 32.46 45.06
$0.00 0.00 0.80 1.66 2.60 3.60
$0.00 10.00 20.80 32.46 45.06 58.67
6 7 8 9 10
10 10 10 10 10
58.67 73.36 89.23 106.37 124.88
4.69 5.87 7.14 8.51 9.99
73.36 89.23 106.37 124.88 144.87
11 12 13 14 15
0 0 0 0 0
144.87 156.45 168.97 182.49 197.09
11.59 12.52 13.52 14.60 15.77
156.45 168.97 182.49 197.09 212.86
Horizon, H
Payments
Carryforward Amount
=$9.26isobtained.Andasintheprecedingsection,thisPV(1,1)=$9.26 grows to exactly the RFV(1,1) = $10 when invested at 8%. Similarly, forH=7,Table2showsthattheRFV(1,7)=$89.23.FromTable1,this 7-year low has a cumulative PV(1,7) = $52.06.A simple computation showsthat $52.06×(1.08)7=$89.23. Moregenerally,asdemonstratedintheTechnicalAppendix, PV(1,H)×(1.08)H=RFV(1,H).
xxx
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
Like the PV, the FV has the appeal of great simplicity. Rather than thinkthroughacomplexpatternofpayments,wecanjustobservethat all investments with the same last payment date will have the same RFV(1,H)foreachdollarofPV(1,H):
RFV(1,H) =(1.08)H PV(1,H)
In other words, any investment’s RFV(1,H) can be reproduced by simplydeployingitsPVamountintoasavingsaccountthatisthencompoundedforwardatthegivenrate. Inthepreceding,theRFV’shorizonwasdeinedtocoincidewiththe lastpaymentdate.Supposethatisnotthecase,thatis,supposewewantto considera15-yearhorizonbuttheinvestment’spaymentsonlycover10 years?Thereisaneasyix.Afterthelastpaymentinthe10thyear,theaccumulatedvalueRFV(1,H)wouldsimplybereinvestedandcompounded forwardatthemarketrateuntilthe15th-yearhorizon,therebygrowingto theRFV(1,15)=$212.86asshowninTable2.
TheHorizonPresentValue(HPV) Intheprecedingdiscussion,whenthehorizonHmatchesorexceeds the investment’s life, there is no further cash low beyond the horizon. However, a somewhat more complex situation arises when the horizon date falls before the cash low’s last payment. For example, consider a 7-yearhorizonthatfallsinthemidstofour10-yearcashlow.Inthe7th year, there will remain a 3-year “tail” consisting of the three $10 paymentsinyears8,9,and10.Now,themostnaturalwaytoputanumberon thistailistoagainusethePVapproach.Thus,atthetimeofthe7th-year horizon,thetailistreatedasanew3-yearinvestmentandtheremaining lows are discounted back to a PV, which we may call the horizon PV, withthesymbolHPV(8,10). Ingeneralterms,foraninvestmenthavingitslast(maturity)payment inyearM,wecanexpressthehorizonPVofthenext(M-H)payments asHPV(H+1,M),representingthePVasoftimeH(i.e.,justafterthe Hthpayment). Forthebasiccaseofalevel-payannuity,theHPVisaprettysimple
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xxxi
TABLE 3
Horizon Present Value 10 Annual Payments of $10 8% Discount Rate
Horizon, H
Payments
Present Value of Each Payment, (H,H )
Cumulative Present Value PV (1,H )
Horizon Present Value, HPV (H+1,10)
0 1 2 3 4 5
$0 10 10 10 10 10
$0.00 9.26 8.57 7.94 7.35 6.81
$0.00 9.26 17.83 25.77 33.12 39.93
$67.10 62.47 57.47 52.06 46.23 39.93
6 7 8 9 10
10 10 10 10 10
6.30 5.83 5.40 5.00 4.63
46.23 52.06 57.47 62.47 67.10
33.12 25.77 17.83 9.26 0.00
11 12 13 14 15
0 0 0 0 0
0.00 0.00 0.00 0.00 0.00
67.10 67.10 67.10 67.10 67.10
0.00 0.00 0.00 0.00 0.00
calculation.Attheoutset,H=0andtheentire10-yearannuityremains ahead of us, so that the HPV(1,10) is just the same as the PV(1,10) fortheentireannuity.InTable1,PV(1,10)=$67.10,andinTable3, HPV(1,10)isalso$67.10.However,attheendoftheirstyear,forH =1, there are only 9 remaining payments. In other words, Table 3’s HPV(2,10)=$62.47isthesameasTable1’sPV(1,9),thatis,thePV foralevellowwith9annualpayments.WecancontinueinthisfashionuntilwereachahorizonH=9,atwhichpointthereisonlytheone remaining$10paymenttobereceived.Thus,Table3’sHPV(10,10)= $9.26isjustthediscountedvalueofa$10paymentoneyearforward, PV(1,1),whichwecouldhavereadfromthethirdcolumnofTable1.
xxxii
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
TheexceptionallysimplerelationshipbetweenTables1and3holds onlyforlevelcashlows.Formorecomplexcashlows,theHPV(H+1,M) mustbeadjustedtorelectthePVofthecashlow’sremainingpayments fromyear(H+1)tothematurityyearM.
TheTotalFutureValueTFV(H) ToindthetotalfuturevalueatagivenhorizonTFV(H),wemustadd the going-forward HPV(H+1,M) of the tail lows to the accumulated RFV(1,H) from the reinvestment process. Note that two concepts are combinedhere:(1)reinvesting(andcompounding)thepaymentsfrom theirstHyears,thatis,RFV(1,H),and(2)discountingthetailpayments forthenext(M-H)years,thatis,theHPV(H+1,M).However,withboth reinvestinganddiscountingtakingplaceatthesamemarketrate,wewill obtainaconsistentTFV(H)foranyinvestmentstretchingoveranyspan ofyears: TFV(H)=RFV(1,H)+HPV(H+1,M)
ThenumericalillustrationoftheTFVconcept(Table4)requirescombiningthereinvestedvalueRFV(H+1,M)fromTable2withthegoingforwardHPV(H+1,M)valuesfromTable3.Attheoutset,whenH=0, therewillhavebeennopaymentsasyetandhence,noRFV(0,0)=0,and theTFV(0)issimplythePV(1,M).Similarly,atthe10th-yearhorizon, therearenofurtherpayments,sothatHPV(11,10)=0,andtheTFV(10) consists solely of the accumulated reinvestment RFV(1,10) = $144.87. At the intermediate horizon H = 3, there is a reinvested accumulated RFV(1,3)=$32.46andanHPV(4,10)=$52.06,sothatTFV(3)=$32.46 +$52.06=$84.53. The Technical Appendix following this section demonstrates that, whenbothdiscountingandreinvestmenttakeplaceatthesamerate,the TFV(H)canbedirectlydeterminedfromthePV(1,M)fortheentirecash low: TFV(H)=(1.08)HPV(1,M) or
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xxxiii
TABLE 4
Total Future Value 10 Annual Payments of $10 8% Discount Rate
Horizon, H
Payments
Reinvested Future Value RFV(1,H )
Horizon Present Value HPV(H+1,10)
Total Future Value TFV(H )
0 1 2 3 4 5
$0 10 10 10 10 10
$0.00 10.00 20.80 32.46 45.06 58.67
$67.10 62.47 57.47 52.06 46.23 39.93
$67.10 72.47 78.27 84.53 91.29 98.59
6 7 8 9 10
10 10 10 10 10
73.36 89.23 106.37 124.88 144.87
33.12 25.77 17.83 9.26 0.00
106.48 115.00 124.20 134.13 144.87
11 12 13 14 15
0 0 0 0 0
156.45 168.97 182.49 197.09 212.86
0.00 0.00 0.00 0.00 0.00
156.45 168.97 182.49 197.09 212.86
TFV(H) =(1.08)H PV(1,M)
With the preceding assumptions and the deinition of TFV(H), we have achieved the ultimate in oversimpliication: Per dollar invested, everyinvestmentwillproducetheexactsameTFV(H) atanyspeciied horizondateH!AndjustasanyinvestmentwillhavethesameFVper dollarinvestedtoday,soatagivenhorizon,anyinvestmentwillhavethe samePVperdollarofFV.Thishypotheticalworld—withacommonixed discountratethatsuitseverymarketparticipantanditsallinvestments— isaverydulloneindeed.Atanyfuturepointintime(includingtoday),a
xxxiv
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
dollarinvestedinanyvehiclewillalwaysgeneratethesameTFV.Thus, everyinvestmentoutcomecouldbereplicatedbythemosttrivialaction ofsimplydeployingthecomparablePVintoasavingsaccountthatthen compoundsatthespeciieddiscountrate. Insuchamarket,therewouldbebasicallynopointinchoosingone investmentoveranother,andhencenorealincentivetotrade.Itwouldbe theinancialequivalentofanultimatestateofentropy,whereinformed judgmentcountedfornaught—likeableak,latdesertwithoutanydistinguishingfeatures.(Somemightarguethatitwouldbetheultimatein eficientmarkets,butmostofuswouldwanttobesparedanysuchdeadeningformofeficiency!) Giventhisinding,onemightwonderabouttheutilityoftheRFVand HPV measures. However, the RFV and HPV are very general concepts that can apply to general cash lows subject to different (and possibly morecomplex)discountingandreinvestmentrates.Inthesemoregeneral cases, the appropriately deined RFV(1,H) and the HPV(H+1,M) will stillsumtotheTFV(H),evenwhenthecompoundingrelationship, TFV(H)=(1.08)HPV(1,M), mayfailtohold.
TheFVversusthePV Insomeways,theFVismoreintuitivelyappealingthanthePV.However,theFVdoeshavethedisadvantagethatitmustbepinnedtoaspeciicpointintime,whereasthePVisalwaysuniquelydeinedincurrent dollarterms.Atthesametime,itcouldbearguedthattheFVisthemore generalconcept,withthePVjustbeingaspecialcaseoftheFVwiththe currenttimespeciiedasthereferencedate. OnequestionthatarisesiswhathappenstotheFVconceptwhenfuture payments are consumed rather than reinvested? The theoretician’s answerwouldbetolookbacktothefundamentalideaofadiscountrate as the exchange between a current dollar (which could be invested or consumedtoday)andafuturepayment(whereconsumptionwouldhave beendeferred).Totheextentthatapaymentisconsumed,theinvestoris makingatrade-offimplyingthattheconsumptionisworththe“sacriice”
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xxxv
ofthatmanyFVdollarsatafuturedate.Withthisgeneralizednotionof valuecombiningliteraldollarswithpsychicreward,thePV(orFV)of anyinvestmentwillalwaysbethesame,whetherthepaymentsarereinvestedorjustconsumed.
BondPricesandYields Thereaderwillnoticethatwe’vegottentothispointwithoutspeciicallyaddressingthesubjectofabond’syield—thatwhichInsidetheYield Bookispresumablyallabout.Abond’syield-to-maturity(YTM)isthat discountratewhichgeneratesaPVequaltothebond’sprice.Inourlat worldofasinglemarketdiscountrate,thebond’spricewouldalwaysbe setbythatdiscountrate,sotheYTMwouldalwaysjustbethisdiscount rateitself. Oneofabond’sbasicattributesisitsparvalue,which(roughly)correspondstotheinitialfundsreceivedbytheissuer.Inthemostidealized bond with neither call features nor sinking funds, the inal “principal payment”atmaturitywillalsobegenerallyequaltothisparvalue.The parvalueistypicallysetat$1,000,withthebond’spriceandcouponpaymentthenexpressedasapercentageofthis$1,000standard.Thus,fora couponratethatcoincideswithourixedmarketdiscountrateof8%,the PVof$1,000wouldjustmatchtheparvalue,thepriceratiowouldjust be100%andthebondwouldbecalled—notsurprisingly—a“parbond.” Forcouponrateshigherthanthediscountrate(generally,forbondsthat hadbeenissuedearlierduringahigherrateenvironment),thePVwould exceedtheparvalue,sothepriceratiowouldbegreaterthan100%,and suchabondwouldbecalleda“premiumbond.”Similarly,lowercoupon ratesandlowerPVswouldgiverisetopriceratiosbelow100%—hence, “discountbonds.” Now,allthisisprettyoldhat.Whygothroughthisentirediscussion of a lat discount world only to come to these standard descriptions of thethreebondtypes?Thepointisthatthesecharacterizationsarereally somewhatmisleading.Allthesebondsarefairlypricedinthesensethat theirPVcorrespondstotheirmarketprice.A“discountbond”isnotabargain,norisa“premiumbond”reallyworthmorethanaparbond.Only whensomedifferentiatingfeaturesareincorporatedintotheanalysisdoes anyinvestmentlookcheapordeartoaspeciicinvestor.
xxxvi
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
To really understand the YTM, our lat-world assumption must be replacedbythemorerealisticsituationwherebondpricesarebasedon ahostofdifferentiatingfactorssuchascreditquality,maturity,coupon level,sinkingfunds,callfeatures,marketliquidity,andsoforth.Inthis environment,onecanargueaboutwhetherthebond’spriceortheYTM istheprimarydeterminantofvalue.Thebasicfactisthatabond’sprice, anditsYTM,aredeinedinacircularfashion.7Thus,foranygivenbond, the YTM is the speciic discount rate that generates a PV equal to the bond’smarketprice.Differentpricingeffectscanthenalsobeexpressed asdifferentYTMs.Overtheyears,thisYTMapproachhasproventobea veryconvenientcomparativeyardstick.Forexample,ithasnowbecome commonplacetocharacterizetheincrementalreturnofcorporatebonds intermsoftheirYTM“spread”overtheU.S.Treasurybondcurve.
PVVolatility A central issue in virtually all PV analysis is the level of the PV’s sensitivitytochangesinthebasicdiscountrate.Inthebondworld,such PVchangesinresponsetochanginginterestratesaretheprimarysource ofpricevolatilityinhigh-gradebonds.Naturally,thisquestionhasmajor signiicanceforbondmarketinvestorsandtraders. Let us irst examine PV volatility in the simple case of the 10-year level-payannuity.Table5departsfromthestandardassumptionofaixed discountratetoshow,inthethirdandfourthcolumn,thePVofeachpaymentirstunderourstandard8%,andthena9%discountrate.Thenext columncontainsthepercentagechangeinthePVderivedfrommoving fromthe8%tothe9%rate. Table 5 illustrates two key points about the PV volatility of single lump-sumpayments.Firstofall,atthehigher9%discountrate,ittakes fewerPVdollarstogrowtoamagnitudethatmatchesanygivenfuture payments. Thus, the PV always declines with higher rates. Moreover, moving out towards later payments, these percentage declines become evengreater. The next two columns show the accumulated PV(1,H) for the paymentsfromtheirsttotheHthyear,irstcalculatedatthestandard8% andthenatthenewrateof9%.Asmightbeexpected,withthemoveto 9%, the PV(1,H) value declines for every horizon H, with greater per-
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xxxvii
centagedeclinesassociatedwiththelongerhorizons.Thus,PVvolatility increaseswiththelengthofthetimetotheannuity’slastpayment.This resultistrueforlevelannuities,butitdoesnotholdwhencomparingany twocashlows.Asanillustrationthatlongerlowsarenotalwaysmore volatile,thesinglepaymentinthe7thyearshowsafargreaterpercentage PVchangeof–6.25%(shownintheifthcolumnofTable5)thanthefull 10-yearannuity’spercentagePVchangeof–4.36%. Table6hasexactlythesameformatasTable5,exceptthatthelower rate levels of 4% and 5% have been substituted for the beginning and endingdiscountrates.Bycomparingtheentriesforthepercentageprice declineinthetwotables,onecanimmediatelyseethatalower-rateenvironmentengenderssomewhatgreaterPVvolatility. Theseannuityexamplesprovideaclearbasisforunderstandingthe basicprinciplesthatdeterminePVvolatilityinmorecomplexcashlows suchasbonds.Thereareanumberofsurprisesinthisarea.Forexample, asanextensionofthepointmadepreviously,bondswithlongermaturitiesarenotalwaysmorevolatilethanbondswithshortermaturities.A typicalbondhasastreamofcouponpaymentsstretchingouttoitsultimatematuritypayment.Iftheseinterimcouponpaymentsloomlarge relative to the longer maturity payment (as in a premium bond), then the net effect will be some reduction in the bond’s volatility. On the otherhand,ifthecouponlowsaresmall(asindiscountbonds),thenthe largermaturitypaymentwithitsgreaterPVvolatilitywillbedominant. Atthelimitwhereallcouponpaymentsvanish,onehasahighlyvolatile,purezero-coupondiscountbondthatconsistssolelyofalump-sum paymentatmaturity.ThePVpricevolatilityofthesezero-couponbonds willincreasewitheachextensioninthematuritydate.Moreover,thePV volatilityofzero-couponbondswillalwaysexceedthatofcomparablematurity “normal” bonds with positive coupon lows. It was shown in InsidetheYieldBookthatlongenoughzero-couponordiscountbonds canhaveratesensitivitiesthatexceedthatofthelongestparbonds.Becausetheverylongestbondsaretheso-called“perpetuals”thatprovide thesamecouponlowforever(intheory),thisindingcameasquitea shockinsomepartsofthebondworld(althoughnotintheUnitedKingdom,whereperpetualgovernmentbondshavebeenamarketstaplefor manyyears).
xxxviii
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
TABLE 5
Percentage PV Change under +1% Rate Move 10 Annual Payments of $10 8% Starting Discount Rate
Horizon, H
Payments
Present Value of Each Payment @ 8%
Present Value of Each Payment @ 9%
Percentage PV Change
0 1 2 3 4 5
$0 10 10 10 10 10
$0.00 9.26 8.57 7.94 7.35 6.81
$0.00 9.17 8.42 7.72 7.08 6.50
0.00% –0.92 –1.83 –2.73 –3.62 –4.50
6 7 8 9 10
10 10 10 10 10
6.30 5.83 5.40 5.00 4.63
5.96 5.47 5.02 4.60 4.22
–5.38 –6.25 –7.11 –7.96 –8.80
11 12 13 14 15
0 0 0 0 0
0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00
0 0 0 0 0
TheMacaulayDuration Oneoftheconceptsjustdiscussedisthatabond’smaturitydate—or moregenerally,thedateofanycashlow’slastpayment—isapoorgauge ofthelow’s“life.”Asonemightsuspect,theproblemofindingagood measureofalow’slifeiscloselyrelatedtotheproblemofdetermining itsPVvolatility.Onenaturalwayistosimplycomputean“averagelife” byjustdeterminingthetimetoeachpayment,weightedbythesizeofthe payment.However,alittleexperimentationquicklyrevealsanumberof problemswiththisapproach.Forexample,aperpetualbondhasanini-
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
Cumulative Present Value, PV(1,H ) @ 8%
Cumulative Present Value, PV(1,H ) @ 9%
xxxix
Percentage Change in Cumulative Present Value
$0.00 9.26 17.83 25.77 33.12 39.93
$0.00 9.17 17.59 25.13 32.40 38.90
0.00% –0.92 –1.35 –1.78 –2.19 –2.58
46.23 52.06 57.47 62.47 67.10
44.86 50.33 55.35 59.95 64.18
–2.96 –3.33 –3.69 –4.03 –4.36
67.10 67.10 67.10 67.10 67.10
64.18 64.18 64.18 64.18 64.18
–4.36 –4.36 –4.36 –4.36 –4.36
nitematurityandanininiteaveragelifeaswellonthispayment-weighted basis.But,asnotedabove,theperpetual’sPVvolatilityisgenerallylower thanthatofa15-yearzero-couponbond. A somewhat more sophisticated approach entails again inding the average time to the low’s payments, but now weighted by the PV of each payment. This approach was irst suggested by Frederick Macaulayin1938.8Inhistreatise(whichcoveredabroadrangeoftopics), Macaulayaddressedtheproblemofindingauseful“half-life”measurefor railroadbonds,wherethecashlowswerecomplicatedbythepresenceof strongmandatorysinkingfunds.HeinallydecidedonPV-weightedaver-
xl
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
TABLE 6
Percentage PV Change under +1% Rate Move 10 Annual Payments of $10 4% Starting Discount Rate Present Value of Each Payment @ 4%
Present Value of Each Payment @ 5%
Percentage PV Change
Horizon, H
Payments
0 1 2 3 4 5
$0 10 10 10 10 10
$0.00 9.62 9.25 8.89 8.55 8.22
$0.00 9.52 9.07 8.64 8.23 7.84
6 7 8 9 10
10 10 10 10 10
7.90 7.60 7.31 7.03 6.76
7.46 7.11 6.77 6.45 6.14
–5.58 –6.48 –7.37 –8.25 –9.13
11 12 13 14 15
0 0 0 0 0
0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00
0.00% –0.95 –1.90 –2.83 –3.76 –4.67
agelifeasthebestyardstick.Ithassincebecomeknownasthe“Macaulay Duration,”andithasproventobeanextremelyusefulconcept. TheMacaulayDurationD(1,H)isdeterminedbyweightingthetime toeachpaymentbythatpayment’spercentageofthelow’soverallPV. InTable7,thiscalculationiscarriedoutbyirstmultiplyingthetimeto eachpayment(theirstcolumn)bythepayment’sPV(thethirdcolumn), to obtain a product (the ifth column), which is then accumulated over thehorizonperiod(thesixthcolumn).TheDurationD(1,H)(theseventh column)isthenfoundbydividingthisaccumulatedvaluebythelow’s PV(1,H)tothathorizon(thefourthcolumn).
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
Cumulative Present Value, PV(1,H ) @ 4%
Cumulative Present Value, PV(1,H ) @ 5%
xli
Percentage Change in Cumulative Present Value
$0.00 9.62 18.86 27.75 36.30 44.52
$0.00 9.52 18.59 27.23 35.46 43.29
0.00% –0.95 –1.41 –1.87 –2.31 –2.75
52.42 60.02 67.33 74.35 81.11
50.76 57.86 64.63 71.08 77.22
–3.18 –3.59 –4.00 –4.40 –4.80
81.11 81.11 81.11 81.11 81.11
77.22 77.22 77.22 77.22 77.22
–4.80 –4.80 –4.80 –4.80 –4.80
Thus,theDurationD(1,1)oftheirstyear’spaymentisjustD(1,1)= (1×9.26)÷9.26=1.TheDurationD(1,2)ofthe2-yearlowisfoundby adding1×9.26totheproductof2timesthe$8.57PVofthesecondyear payment,toobtain(1×9.26)+(2×8.57)=26.41,andthendividingby PV(1,2)=$17.83.Inthiscase,theresultD(1,2)=26.41÷17.83=1.48 isclosetothe2-yearlow’ssimpleunweightedaveragelifeof
1×10+2×10 =1.50. (10+10)
xlii
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
TABLE 7
The Macaulay Duration 10 Annual Payments of $10 8% Discount Rate
Cumulative Present Value, PV(1,H )
Time to Payment Multiplied by PV of Each Payment H × PV(H,H )
Horizon, H
Payments
Present Value of Each Payment, PV(H,H )
0 1 2 3 4 5
$0 10 10 10 10 10
$0.00 9.26 8.57 7.94 7.35 6.81
$0.00 9.26 17.83 25.77 33.12 39.93
$0.00 9.26 17.15 23.81 29.40 34.03
6 7 8 9 10
10 10 10 10 10
6.30 5.83 5.40 5.00 4.63
46.23 52.06 57.47 62.47 67.10
37.81 40.84 43.22 45.02 46.32
11 12 13 14 15
0 0 0 0 0
0.00 0.00 0.00 0.00 0.00
67.10 67.10 67.10 67.10 67.10
0.00 0.00 0.00 0.00 0.00
However, as the time is extended, the gap between the weighted Macaulay Duration and the simple average life becomes more pronounced.Forthefull10-yearlow,theDurationD(1,10)=4.87isconsiderably less than the low’s 5.5-year payment-weighted average life. Generally, the longer the annuity, the larger this gap will grow as the moredistantpaymentsarediscountedevermoreseverely.Intheextreme limit,theininiteannuityhasanininiteaveragelife,butitcanbeshown tohaveadurationof13.5years. Theannuityisastreamofixedannualpaymentsoversomespanof time.Withoutalarge,bond-likeprincipalpaymentatmaturity,theannu-
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
Cumulative PV-Weighted Life
Macaulay Duration D(1,H) Cum. PV-Weighted Life as % of Cum. PV(1,H )
xliii
PV Volatility PV-VOL (1,H ) = –D (1,H )/(1+y )
$0.00 9.26 26.41 50.22 79.62 113.65
0.00% 1.00 1.48 1.95 2.40 2.85
0.00% –0.93 –1.37 –1.80 –2.23 –2.64
151.46 192.31 235.53 280.55 326.87
3.28 3.69 4.10 4.49 4.87
–3.03 –3.42 –3.79 –4.16 –4.51
326.87 326.87 326.87 326.87 326.87
4.87 4.87 4.87 4.87 4.87
–4.51 –4.51 –4.51 –4.51 –4.51
itywouldalwayshaveaMacaulayDurationthatisshorterthanitspayment-basedhalf-life.Thisannuityexampleprovidesaclearillustration oftheoriginalMacaulayinsight.Allthepaymentsofanannuityhavethe samedollarvalue(bydeinition).However,whenthePV-weightedaverageofeachpaymentisconsidered,theearlierpaymentsnaturallyloom larger. Consequently, the annuity’s Macaulay Duration will always be shorterthanthemidpointofitsequal-dollarcashlow.(Forgeneralcash lows,allthatcanbesaidisthatthedurationwillneverbelongerthanthe timetothelastpayment.) Theannuityexamplealsoprovidesagoodintuitiveillustrationofhow
xliv
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
durationitselfdependsonthelevelofdiscountrates.Atverylowinterest ratesapproachingzero,thePVofasinglepaymentconvergesonitsraw dollar value. Thus, at ever-lower discount rates, the Macaulay Duration ultimatelydoescoincidewiththeliteralhalf-life.Incontrast,athigherinterestrates,thelaterpaymentsarediscountedmoreseverelyandthestream ofPVsismore“front-loaded,”resultingingenerallyshorterdurations. As noted earlier, this result holds quite generally: The Macaulay Durationofanycashlowbecomeslargerasinterestratesfall.Onemight betemptedtoconcludefromthisobservationthatverylowinterestrate environmentscanbeverytreacherous.Whenratescanonlygoup,and whenthepricesensitivityofanygivencashlowisnearitsmaximum, it’saprettytoxiccombination.
PVVolatilityandtheModiiedDuration Macaulaywasquitehappywithhishalf-lifeinterpretationofthedurationmeasure.(Indeed,heonlydevotedafewpagestoitsderivationbefore movingtoothermattersina591-pagebook.)However,itwassubsequentlydiscoveredthatwithaslightadjustmentfactor,theMacaulayDuration couldactasagaugeofthePVvolatilityforageneralcashlow. Table7illustrateshowthedurationD(1,H)canbeusedtoapproximate thepercentagepricesensitivity.DividingD(1,H)by(1+y)whereyisthe discountrateresultsinavaluethatisoftenreferredtoasthe“modiied Duration.”Thusforour8%discountrate,the10-yearMacaulayDuration D(1,10)=4.87isreducedtoamodiiedDuration,
4.87 =4.51. 1.08
Itisworthnotingthatthisadjustmentresultsina4.51valuethatis anotherstepsmallerthantheannuity’ssimpleaveragelifeof5.5years. It can be shown (see the TechnicalAppendix following this section) that this value of 4.51 corresponds to the derivative of the percentagechangeinthePV.InTable5,thePV(1,10) dropsfrom$67.10at the original 8% discount rate to $64.18 at 9%. This percentage loss of –4.36% is closely approximated by the negative of the modiied Durationvalueof4.51.Thus,themodiiedDurationcanbeseentobe
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xlv
areasonableapproximationforthemagnitudeofthepercentageprice lossresultingfroma+1%moveinthediscountrate. Withsmallerandsmallermovesinthediscountrate,thepercentage PVchangeperunitratemoveconverges(inabsolutevalue)tothenegativeofthemodiiedDurationvalue(seetheTechnicalAppendix).Forthis reason,wecanrefertothenegativevalueofthemodiiedDurationasthe PVvolatility,PV-VOL(1,H). In many ways, it was remarkable that Macaulay failed to see the broaderuseofhisdiscoveryasavolatilitymeasure.Itisevenmoreremarkable that more than forty years elapsed before this measure came intocommonusageintheU.S.bondmarket.Infact,thisdoublelevelof discoveryformsafascinatingcasestudyinhowtheoreticalindingscan takeacircuitouspathbeforeultimatelyindingtheirwayintopractice.9,10
ReinvestmentVolatility When we move from the PV, which declines with higher discount rates, to the RFV, which increases with higher reinvestment rates, another“volatility”measurebecomesimportant.Thereinvestmentvolatility (RFV-VOL) is rarely characterized in the same quantitative way as theDurationconcept,butdoingsoleadstosomeinterestingresultsthat shouldbemorewidelyappreciatedandthatmaybeparticularlyusefulfor long-termholdersofixed-incomeexposuressuchasinsurancecompaniesandpensionfunds. The irst step in such a discussion is to focus on some prescribed future date as the RFV horizon. For bonds, as discussed in Inside the YieldBook,thereferenceRFVhorizonHisalmostalwaysthebond’s maturitydate.Wenotedearlierhowpremiumbondswiththeirhigher couponlowsaremorereinvestment-sensitivethandiscountbonds.In contrast,thezero-couponbondistheultimateintermsofreinvestment insensitivity:Whenitsmaturityistakenasthereferencepoint,itwill alwayshaveitsmaturitypaymentastheRFVregardlessofthelevelof interveninginterestrates.Inotherwords,ithasabsolutelynosensitivity to changing reinvestment rates. (Incidentally, for all bonds having thesamematuritydate,thezero-couponhasthehighestPVsensitivity todiscountratechanges,butthelowestRFVsensitivitytoreinvestment rateschanges.)
xlvi
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
ItturnsoutthatthepercentagevolatilityRFV-VOL(1,H)bearsavery simplerelationtotheDurationvalue.Foracashlowstretchingoutto a given horizon H, the RFV volatility can be shown to be just the gap betweenthehorizonandthelow’sMacaulayDuration,adjustedbyone plustheinterestrate:
[H–D(1,H)] RFV-VOL(1,H)= (1+y) H = +PV-VOL(1,H) 1+y
ThisresultisdevelopedintheTechnicalAppendix.Asonemightexpectfromtheearlierdiscussionofthedurationmeasure,thisRFV-VOL indingisessentiallyderivative-based,thatis,itactsasabetterapproximationforever-smallerratemoves. Attheoutset,itcanbeseenthatforH=0,beforeanypaymentwhatsoever,thereisnocashlow,andsotheDurationD(1,0)hasthetrivial valueofzero.Similarly,thereisnoreinvestmentvolatility:
[H–D(1,H)] RFV-VOL(1,10)= (1+y) (0–0) = (1.08) =0.
Moreover,forthesinglelump-sumpaymentatthehorizon,theDurationD(1,H)justequalsthehorizonH:
D(1,H)=D(H,H) =H,
andagainthereisnoreinvestmentsensitivitybecause
[H–D(1,H)] (1+y) [H–H] = (1+y) =0.
RFV-VOL(H,H)=
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xlvii
However,movingfromsinglelump-sumpaymentstowardmoregeneral cashlowsstretchingoutovertime,thereinvestmentvolatilitygrowswith longerhorizons.ThusinTable8,theannuitydevelopsasigniicantexposure tochangingreinvestmentratesasthehorizonlengthens.Takingthe7-year horizonasanexample,Table8providesavalueofD(1,7)=3.69,sothat [H–D(1,H)] RFV-VOL(1,7)= (1+y) 7–3.69 = 1.08 3.31 = 1.08 =3.06.
Toseehowwellthismeasureapproximatesanactualshiftinreinvestmentrates,refertoTable2,whichshowsthatRFV(1,7)=$89.23atthe 8%rate.Ifthereinvestmentrateisraisedto9%,theRFV(1,7)becomes $92.10,a3.22percentageincrease,whichisreasonablyapproximatedby thederivative-basedvalueof3.06. Turningtheaboveindingaround,notethatingeneral,whenalow’s maturity M is taken as the FV horizon, the Duration and the RFVVOL(1,M)volatilityadduptothelow’slife, M=D(1,M)+(1+y)×[RFV-VOL(1,M)]. Thus,fortheannuitywithM=10,Table8providesvaluesofD(1,10) =4.87andRFV-VOL(1,10)=4.75,sothat
M=D(1,10)+(1+y)×[RFV-VOL(1,10)] =4.87+1.08×4.75 =10.
It turns out that this relationship holds for any cash low. For the 10-yearannuity,theDurationandreinvestmentvolatilityturnouttobe nearlyequal.However,thiswillnotbetrueformoregeneralcashlows, eventhoughtheirsumwillalwaysequalthetimetothelastpayment.
xlviii
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
TABLE 8
Reinvestment Volatility 10 Annual Payments of $10 8% Discount Rate
Horizon, H
Reinvested Future Value RFV(1,H )
Macaulay Duration D(1,H )
Horizon-toDuration Gap H – D(1,H )
Reinvestment Volatility RFV-VOL(1,H ) = [H – D(1,H ) / (1+y )]
0 1 2 3 4 5
$0 10.00 20.80 32.46 45.06 58.67
0.00 1.00 1.48 1.95 2.40 2.85
0.00 0.00 0.52 1.05 1.60 2.15
0.00% 0.00 0.48 0.97 1.48 1.99
6 7 8 9 10
73.36 89.23 106.37 124.88 144.87
3.28 3.69 4.10 4.49 4.87
2.72 3.31 3.90 4.51 5.13
2.52 3.06 3.61 4.17 4.75
11 12 13 14 15
156.45 168.97 182.49 197.09 212.86
4.87 4.87 4.87 4.87 4.87
6.13 7.13 8.13 9.13 10.13
5.67 6.60 7.53 8.45 9.38
TheTotalFutureValueVolatilityatLongerHorizons TheprecedingdevelopmentofanRFV-VOL(1,H)volatilityalsoprovidesananswertothequestionofthevolatilityTFV-VOL(H)ofacash low’sTFV(H)withahorizonHthatcoincideswiththelow’slastpaymentM,thatis,whereH=M.Withhorizonsthatmatchthelow’slast payment,therearebydeinitionnotaillowsandsothetotalfuturevalue, TFV,consistsofjustthereinvestment-drivenRFV(1,H).Becausethereinvestmenteffectisalwayspositive,
SOMETOPICSTHATDIDN’TMAKEITINTOTHE1972EDITION
xlix
TFV-VOL(M)=RFV-VOL(1,M) [M–D(1,M)], = (1+y)
higher rates will always lead to higher TFVs (except for the case of a singlelump-sumpaymentsintheMthyear). EvenwhentheFVhorizonisextendedbeyondthelastpaymentdate, this relationship continues to hold. The Duration value remains stable, but the horizon gap increases by the exact length of the extension. For example, if we look at a 12-year horizon with the 10-year annuity, the Duration D(1,12) = D(1,10) remains unchanged at 4.87, but the RFVVOL(1,12)volatilitynowincreasesto6.60: RFV(1,12)=
[H–D(1,M)] (1+y)
12–4.87 = 1.08 =6.60
asshowninTable8forH=12. Moreover,thisTFVvolatilityresultwillholdforanyhorizonlonger thanthelastpaymentdate:
[H–D(1,M)] TFV-VOL(H)=1 1+y
foranyH≥M.
HorizonDurationandtheGeneralizedTFVVolatility Asnotedearlier,onecanalsohavehorizonHdatesthatprecedethe last payment date, that is, H