Advanced Mathematical Economics Astheintersectionbetweeneconomicsandmathematicscontinuestogrowinboth theoryandpractice,asolidgroundinginmathematicalconceptsisessentialforall seriousstudentsofeconomictheory. Inthisclearandentertainingvolume,RakeshV.Vohrasetsoutthebasicconcepts of mathematics as they relate to economics. The book divides the mathematical problems that arise in economic theory into three types: feasibility problems, optimality problems and fixed-point problems. Of particular salience to modern economic thought are sections on lattices, supermodularity, matroids and their applications.Inadeparturefromtheprevailingfashion,muchgreaterattentionis devotedtolinearprogramminganditsapplications. Ofinteresttoadvancedstudentsofeconomicsaswellasthoseseekingagreater understandingoftheinfluenceofmathematicson‘thedismalscience’.Advanced MathematicalEconomicsfollowsalongandcelebratedtraditionoftheapplication ofmathematicalconceptstothesocialandphysicalsciences. Rakesh V. Vohra is the John L. and Helen Kellogg Professor of Managerial Economics and Decision Sciences at the Kellogg School of Management at NorthwesternUniversity,Illinois. RAKE:“fm” — 2004/9/17 — 06:12 — page i — #1 Routledgeadvancedtextsineconomicsandfinance FinancialEconometrics PeijieWang MacroeconomicsforDevelopingCountries2ndedition RaghbendraJha AdvancedMathematicalEconomics RakeshV.Vohra AdvancedEconometricTheory JohnS.Chipman RAKE:“fm” — 2004/9/17 — 06:12 — page ii — #2 Advanced Mathematical Economics Rakesh V. Vohra RAKE:“fm” — 2004/9/17 — 06:12 — page iii — #3 Firstpublished2005 byRoutledge 2ParkSquare,MiltonPark,Abingdon,OxonOX144RN SimultaneouslypublishedintheUSAandCanada byRoutledge 270MadisonAve,NewYork,NY10016 RoutledgeisanimprintoftheTaylor&FrancisGroup This edition published in the Taylor & Francis e-Library, 2009. To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk. ©2005RakeshV.Vohra Allrightsreserved.Nopartofthisbookmaybereprintedor reproducedorutilizedinanyformorbyanyelectronic, mechanical,orothermeans,nowknownorhereafter invented,includingphotocopyingandrecording,orinany informationstorageorretrievalsystem,withoutpermissionin writingfromthepublishers. BritishLibraryCataloguinginPublicationData Acataloguerecordforthisbookisavailable fromtheBritishLibrary LibraryofCongressCataloginginPublicationData Acatalogrecordforthisbookhasbeenrequested ISBN 0-203-79995-X Master e-book ISBN ISBN 0-203-68209-2 (Adobe ebook Reader Format) ISBN0–415–70007–8(hbk) ISBN0–415–70008–6(pbk) RAKE:“fm” — 2004/9/17 — 06:12 — page iv — #4 Contents Preface viii 1 Thingstoknow 1 1.1 Sets 1 1.2 Thespaceweworkin 1 1.3 Factsfromrealanalysis 2 1.4 Factsfromlinearalgebra 6 1.5 Factsfromgraphtheory 9 2 Feasibility 13 2.1 Fundamentaltheoremoflinearalgebra 13 2.2 Linearinequalities 15 2.3 Non-negativesolutions 15 2.4 Thegeneralcase 19 2.5 Application:arbitrage 20 2.6 Application:co-operativegames 24 2.7 Application:auctions 25 3 Convexsets 33 3.1 Separatinghyperplanetheorem 34 3.2 Polyhedronsandpolytopes 40 3.3 Dimensionofaset 46 3.4 Propertiesofconvexsets 47 3.5 Application:linearproductionmodel 49 4 Linearprogramming 53 4.1 Basicsolutions 56 4.2 Duality 60 RAKE:“fm” — 2004/9/17 — 06:12 — page v — #5 vi Contents 4.3 Writingdownthedual 64 4.4 Interpretingthedual 64 4.5 Marginalvaluetheorem 68 4.6 Application:zero-sumgames 69 4.7 Application:Afriat’stheorem 72 4.8 Integerprogramming 75 4.9 Application:efficientassignment 78 4.10 Application:Arrow’stheorem 80 5 Non-linearprogramming 87 5.1 Necessaryconditionsforlocaloptimality 89 5.2 Sufficientconditionsforoptimality 95 5.3 Envelopetheorem 100 5.4 Anasideonutilityfunctions 103 5.5 Application:marketgames 105 5.6 Application:principal–agentproblem 109 6 Fixedpoints 117 6.1 Banachfixedpointtheorem 117 6.2 Brouwerfixedpointtheorem 118 6.3 Application:Nashequilibrium 125 6.4 Application:equilibriuminexchangeeconomies 127 6.5 Application:Hex 133 6.6 Kakutani’s fixedpointtheorem 136 7 Latticesandsupermodularity 142 7.1 Abstractlattices 149 7.2 Application:supermodulargames 151 7.3 Application:transportationproblem 152 7.4 Application:efficientassignmentandthecore 154 7.5 Application:stablematchings 159 8 Matroids 163 8.1 Introduction 163 8.2 Matroidoptimization 165 8.3 Rankfunctions 166 8.4 Deletionandcontraction 169 RAKE:“fm” — 2004/9/17 — 06:12 — page vi — #6 Contents vii 8.5 Matroidintersectionandpartitioning 170 8.6 Polymatroids 175 8.7 Application:efficientallocationwithindivisibilities 179 8.8 Application:Shannonswitchinggame 186 Index 191 RAKE:“fm” — 2004/9/17 — 06:12 — page vii — #7 Preface Iwantedtotitlethisbook‘LeisureoftheTheoryClass’.Thepublishersdemurred. Mysecondchoicewas‘Feasibility,OptimalityandFixedPoints’.Whileaccurate, itdidnotidentify,asthepublishernoted,theintendedaudience.Wesettledatlast ontheanodynetitlethatnowgracesthisbook.Asitsuggests, thebookisabout mathematics.Thequalifier‘advanced’signifiesthatthereadershouldhavesome mathematical sophistication. This means linear algebra and basic real analysis.1 Chapter 1 provides a list of cheerful facts from these subjects that the reader is expectedtoknow.Thelastwordinthetitleindicatesthatitisdirectedtostudents ofthedismalscience.2 Threekindsofmathematicalquestionsarediscussed. Givenafunctionf and asetS, • Findanx suchthatf(x)isinS.Thisisthefeasibilityquestion. • Findanx inS thatoptimizesf(x).Thisistheproblemof optimality. • Findanx inS suchthatf(x)=x;thisisthefixedpointproblem. These questions arise frequently in Economic Theory, and the applications describedinthebookillustratethis. Thetopicscoveredarestandard.Exceptionsarematroidsandlattices.Unusual forabooksuchasthisistheattentionpaidtoLinearProgramming.Itiscommon amongstcyclopeanEconomiststodismissthisasaspecialcaseofKuhn–Tucker.A mistakeinmyview.Ihopetopersuadethereader,byexample,ofthesame.Another unusualfeature,Ithink,aretheapplications.Theyarenotmerelycomputational, i.e.,thisishowoneusesTheoremXtocomputesuchandsuch.Theyaresubstantive piecesofeconomictheory. Thehomeworkproblems, asmystudentswillattest, arenotforthefainthearted. Ofthemakingofbooksthereisnoend.3Theremarkisparticularlytrueforbooks devotedtotopicsthatthepresentonecovers.So,someexplanationisrequiredof howthisbookdiffersfromothersofitsilk. The voluminous Simon and Blume (1994), ends (with exceptions) where this bookbegins.InfactaknowledgeofSimonandBlumeisagoodprerequisitefor thisone. RAKE:“fm” — 2004/9/17 — 06:12 — page viii — #8 Preface ix Thethick,square,Mas-Colleletal.(1995)containsanappendixthatcoversa subsetofwhatiscoveredhere.Thetreatmentisnecessarilybrief,omittingmany interestingdetailsandconnectionsbetweentopics. Sundaram’sexcellent‘FirstCourseinOptimization’(1996)isperhapsclosestof themorerecentbooks.Buttherearecleardifferences.Sundaramcoversdynamic optimizationwhilethisbookdoesnot.Ontheotherhand,thisbookdiscussesfixed pointsandmatroids,whileSundaramdoesnot. This book is closest in spirit to books of an earlier time when, giants, I am reliably informed, walked the Earth. Two in particular have inspired me. The first is Joel Franklin’s ‘Methods of Mathematical Economics’ (1980), a title that pays homage to Courant and Hilbert’s celebrated ‘Mathematical Methods of Physics’(1924). Franklin is a little short on the economic applications of the mathematicsdescribed.However,theinformalanddirectstyleconveyhisdelight inthesubject.ItisadelightIshareandIhopethisbookwillinfectthereaderwith thesame. The second is Nicholas Rau’s ‘Matrices and Mathematical Programming:An IntroductionforEconomists’.Rauisformal,preciseandstartlinglyclear.Ihave triedtomatchthisstandard,goingsofarastofollow,insomecases,hisexposition ofproofs. ThebookbeforeyouistheoutgrowthofaPhDclassthatallgraduatestudents inmydepartmentmusttakeintheirfirstyear.4 Itsrootshowevergobacktomy saladdays.Ihavelearntanenormousamountfromteachersandcolleaguesmuch ofwhichinfusesthisbook.Innoparticularorder,theyareAilsaLand,SaulGass, Bruce Golden, Jack Edmonds, H. PeytonYoung, Dean Foster, Teo Chung Piaw andJamesSchummer. Four cohorts of graduate students at Kellogg and other parts of Northwestern have patiently endured early versions of this book. Their questions, both puzzled and pointed have been a constant goad. I hope this book will do them justice. Denizens of MEDS, my department, have patiently explained to me the finer points of auctions, general equilibrium, mechanism design and integrability. In returnIhavesubjectedthemtoendlessspeechesabouttheutilityofLinearPro- gramming.Thebookisasmuchareflectionofmyownhobbyhorsesasthespirit ofthedepartment. This book could not have been written without help from both my parents (FaqirandSudeshVohra)andin-laws(HarishandKrishnaMahajan).Theytook it in turns to see the children off to school, made sure that the fridge was full, dinner on the table and the laundry done. If it continues I’m in clover! My wife, Sangeeta took care of many things I should have taken care of, including myself. Piotr Kuszewski, a graduate student in Economics, played an important role in preparing the figures, formatting and producing a text that was a pleasure to lookat. Finally, this book is dedicated to my childrenAkhil and Sonya, who I hope will find the same joy in Mathematics as I have. Perhaps, like the pilgrims in RAKE:“fm” — 2004/9/17 — 06:12 — page ix — #9