the econometrics of macroeconomic modelling Other Advanced Texts in Econometrics ARCH:SelectedReadings EditedbyRobertF.Engle AsymptoticTheoryforIntegratedProcesses ByH.PeterBoswijk BayesianInferenceinDynamicEconometricModels ByLucBauwens,MichelLubrano,andJean-Fran¸coisRichard Co-integration,ErrorCorrection,andtheEconometricAnalysisofNon-StationaryData ByAnindyaBanerjee,JuanJ.Dolado,JohnW.Galbraith,andDavidHendry DynamicEconometrics ByDavidF.Hendry FiniteSampleEconometrics ByAmanUllah GeneralizedMethodofMoments ByAlastairHall Likelihood-BasedInferenceinCointegratedVectorAutoregressiveModels BySørenJohansen Long-RunEconometricRelationships: ReadingsinCointegration EditedbyR.F.EngleandC.W.J.Granger Micro-EconometricsforPolicy,Program,andTreatmentEffect ByMyoung-jaeLee ModellingEconomicSeries: ReadingsinEconometricMethodology EditedbyC.W.J.Granger ModellingNon-LinearEconomicRelationships ByCliveW.J.GrangerandTimoTer¨asvirta ModellingSeasonality EditedbyS.Hylleberg Non-StationaryTimesSeriesAnalysisandCointegration EditedbyColinP.Hargeaves OutlierRobustAnalysisofEconomicTimeSeries ByAndr´eLucas,PhilipHansFranses,andDickvanDijk PanelDataEconometrics ByManuelArellano PeriodicityandStochasticTrendsinEconomicTimeSeries ByPhilipHansFranses ProgressiveModelling: Non-nestedTestingandEncompassing EditedbyMassimilianoMarcellinoandGrayhamE.Mizon ReadingsinUnobservedComponents EditedbyAndrewHarveyandTommasoProietti StochasticLimitTheory: AnIntroductionforEconometricians ByJamesDavidson StochasticVolatility EditedbyNeilShephard TestingExogeneity EditedbyNeilR.EricssonandJohnS.Irons TheEconometricsofMacroeconomicModelling ByGunnarB˚ardsen,ØyvindEitrheim,EilevS.Jansen,andRagnarNymoen TimeSerieswithLongMemory EditedbyPeterM.Robinson Time-Series-BasedEconometrics: UnitRootsandCo-integrations ByMichioHatanaka WorkbookonCointegration ByPeterReinhardHansenandSørenJohansen The Econometrics of Macroeconomic Modelling GUNNAR B˚ARDSEN ØYVIND EITRHEIM EILEV S. JANSEN AND RAGNAR NYMOEN 1 3 GreatClarendonStreet,Oxfordox2 6dp OxfordUniversityPressisadepartmentoftheUniversityofOxford. ItfurtherstheUniversity’sobjectiveofexcellenceinresearch,scholarship, andeducationbypublishingworldwidein Oxford NewYork Auckland CapeTown DaresSalaam HongKong Karachi KualaLumpur Madrid Melbourne MexicoCity Nairobi NewDelhi Shanghai Taipei Toronto Withofficesin Argentina Austria Brazil Chile CzechRepublic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore SouthKorea Switzerland Thailand Turkey Ukraine Vietnam OxfordisaregisteredtrademarkofOxfordUniversityPress intheUKandincertainothercountries PublishedintheUnitedStates byOxfordUniversityPressInc.,NewYork (cid:1)c GunnarB˚ardsen,ØyvindEitrheim,EilevS.Jansen,andRagnarNymoen2005 Themoralrightsoftheauthorshavebeenasserted DatabaserightOxfordUniversityPress(maker) Firstpublished2005 Allrightsreserved. Nopartofthispublicationmaybereproduced, storedinaretrievalsystem,ortransmitted,inanyformorbyanymeans, withoutthepriorpermissioninwritingofOxfordUniversityPress, orasexpresslypermittedbylaw,orundertermsagreedwiththeappropriate reprographicsrightsorganization. Enquiriesconcerningreproduction outsidethescopeoftheaboveshouldbesenttotheRightsDepartment, OxfordUniversityPress,attheaddressabove Youmustnotcirculatethisbookinanyotherbindingorcover andyoumustimposethissameconditiononanyacquirer BritishLibraryCataloguinginPublicationData Dataavailable LibraryofCongressCataloginginPublicationData Dataavailable TypesetbyNewgenImagingSystems(P)Ltd.,Chennai,India PrintedinGreatBritain onacid-freepaperby BiddlesLtd.,King’sLynn ISBN 0-19-924649-1 978-0-19-9246496 ISBN 0-19-924650-5 978-0-19-9246502 3 5 7 9 10 8 6 4 2 E.S.J.: To Kristin G.B.: To Tordis R.N.: To Kjersti-Gro Ø.E.: To Gro This page intentionally left blank Preface At the European Meeting of the Econometric Society in Santiago de Compostela in September 1999, Clive Granger asked if we would like to write a book for the Advanced Texts in Econometrics series about the approach to macroeconometric modelling we had adopted at the Research Department of NorgesBankoverthepast15years. Ithastakenus5yearstocomplywithhis request, and the result is found within these covers. Thisbookisaboutbuildingmodelsbytestinghypothesesofmacroeconomic theories–rather than by imposing theories untested. This is quite a crucial distinction in macroeconometric model building. For an empirical model to be useful, be it as a basis for economic policy decisions or for forecasting, it needs to describe the relevant aspects of reality. Simplification is the main virtue of theoretical model building. In empirical modelling it might easily become a vice. A theoretical model is often reduced to just those equations that are required to make it work for the problem at hand. A good empirical model shouldalsobeabletoexplainproblemsthatmightoccur. Einstein’sadvicethat ‘everything should be as simple as possible...but no simpler’ is as relevant as ever. If a model does not describe the data, it may just be too simple to be used as a tool for macroeconomic decision making. The main target group for the book is researchers and practitioners of macroeconomicmodelbuildinginacademia,privateagenciesandgovernmental services. As a textbook it can be used in graduate courses on applied macro- econometrics in general and—more specifically—in courses focusing on wage and price formation in the open economy. In that context it is obvious that a companion text on econometric methods and practice will be useful, and we recommend Dynamic Econometrics by David F. Hendry (Hendry 1995a) and Empirical Modeling of Economic Time Series by Neil R. Ericsson (Ericsson 2005) for this purpose. The work on the book has formed a joint research agenda for the authors since its conception. Hence, we draw extensively on our published papers, manyofwhichwaswrittenwiththedemandsofthisbookinmind: Section1.4 and Chapter 2 are based on Jansen (2002); Sections 5.6 and 6.7.2 on B˚ardsen etal. (1998);Sections6.1–6.3onKolsrudandNymoen(1998)andB˚ardsenand Nymoen (2003); Section 6.8 on Holden and Nymoen (2002) and Nymoen and Rødseth (2003); Chapter 7 on B˚ardsen et al. (2004), Section 8.4 on Eitrheim vii viii Preface (1998); Chapter 9 on B˚ardsen et al. (2003); Section 11.2 on Eitrheim et al. (1999, 2002a) and Section 11.3 on B˚ardsen et al. (2002a). Also,wehaveusedmaterialfromunpublishedjointworkwithotherauthors. In particular we would like to thank Q. Farooq Akram, Neil R. Ericsson and NevaA.Kerbeshianfortheirpermissiontodoso: Akrametal. (2003)underlies Chapter 10 and we draw on Ericsson et al. (1997) in Section 4.4. The views are those of the authors and should not be interpreted to reflect those of their respective institutions. Throughout the book our main econo- metric tools have been the programs developed by Jurgen A. Doornik, David F. Hendry and Hans-Martin Krolzig, i.e., the Oxmetrics package (provided by TimberlakeConsultants),inparticularPcGive,PcFIMLandPcGets. InChap- ter7andSections9.5and10.3wehaveusedEviews(providedbyQuantitative Micro Software) and the simulations in Section 11.2.2 are carried out with TROLL (provided by Intex Solutions). Data documentation, data series, programs and detailed information about the software used are available from a homepage for the book: http://www.svt.ntnu.no/iso/macroectrics. We are indebted to many colleagues and friends for comments, discussions and critisism to the various parts of the book. The editors of the series—Clive W. J. Granger and Grayham E. Mizon—have given us advice and constant encouragement. David F. Hendry and Bjørn E. Naug have read the entire manuscript and given us extensive, constructive and very helpful comments. In addition to those already acknowledged, grateful thanks goes to: Q. Farooq Akram,OlavBjerkholt,NeilR.Ericsson,PaulG.Fisher,RogerHammersland, Steinar Holden, Tore Anders Husebø, K˚are Johansen, Søren Johansen, Adrian Pagan, Asbjørn Rødseth, Timo Tera¨svirta, Anders Vredin, Kenneth F. Wallis, and Fredrik Wulfsberg. Last, but not least, we are indebted to Jurgen A. Doornik for his generosity with both time, patience, and effort throughout the project. While working on the book Gunnar B˚ardsen has visited the School of Economics and Finance, Queensland University of Technology (November 2000–January 2001) and Department of Economics, University of California SanDiego(March2003),andEilevS.JansenhasbeenavisitoratDepartment of Economics, University of Oslo (August 2001–January 2003), DG Research, EuropeanCentralBank,Frankfurt(February2003–June2003)andDepartment of Economics, University of California San Diego (August 2003–July 2004). The hospitality and excellent working conditions offered at those institutions are gratefully acknowledged. Finally, we are grateful to our respective employers—Norges Bank, Norwegian University of Science and Technology, and University of Oslo—for allocatingresourcesandtimeforthisproject. Thatsaid, thetimespentonthe book has often gone beyond normal hours, which is but one reason why this book is dedicated to our wonderful and wise wives. Trondheim/Oslo, November 2004 Gunnar B˚ardsen, Øyvind Eitrheim, Eilev S. Jansen and Ragnar Nymoen Contents List of Figures xv List of Tables xix List of Abbreviations xxi 1 Introduction 1 1.1 The case for macroeconometric models 1 1.2 Methodological issues (Chapter 2) 4 1.3 The supply-side and wage- and price-setting (Chapters 3–8) 7 1.4 The transmission mechanism (Chapters 9 and 10) 11 1.5 Forecast properties (Chapter 11) 15 2 Methodological issues of large-scale macromodels 17 2.1 Introduction: small vs. large models 17 2.2 The roles of statistics and economic theory in macroeconometrics 20 2.2.1 The influx of statistics into economics 20 2.2.2 Role of economic theory in macroeconometrics 22 2.3 Identifying partial structure in submodels 24 2.3.1 The theory of reduction 24 2.3.2 Congruence 26 2.4 An example: modelling the household sector 29 2.4.1 The aggregate consumption function 30 2.4.2 Rival models 31 2.5 Is modelling subsystems and combining them to a global model a viable procedure? 32 3 Inflation in open economies: the main-course model 35 3.1 Introduction 35 3.2 Cointegration 37 3.2.1 Causality 41 ix