Essays on empirical macroeconomics [Elektronische Ressource] : convergence and risk sharing / Falko Jüßen

Administrator

Informations

• Publié par: technischen_universitat_dortmund
• Publié le: 01 janvier 2006
• Nombre de lectures: 33
• Poids de l'ouvrage: 2 Mo

Extrait

Essays on Empirical Macroeconomics:
Convergence and Risk Sharing
Inauguraldissertation
zur Erlangung des akademischen Grades
Doctor rerum politicarum
der Universität Dortmund
Falko Jüßen
May 2006Contents
Preface viii
1Introduction 1
1.1PartI:RegionalConvergenceinGermany............ 2
1.2PartI:RiskSharinginGermanyandtheUS.......... 8
I Regional Convergence in Germany 15
2 Convergence in Unemployment Rates 16
2.1 Introduction............................. 16
2.2Theoreticalconcepts........................ 18
2.3Dataandgraphicalanalysis.................... 20
2.3.1 Data............................. 20
2.3.2 Graphicalanalysis..................... 2
2.4Unit-roottestswithoutstructuralbreaks............. 24
2.4.1 Univariateunit-roottests................. 24
2.4.2 Panelunit-roottests.................... 26
2.5Unit-roottestswithstructuralbreaks............... 31
2.5.1 Testprocedure....................... 32
2.5.2 Testresults......................... 34
2.5.3 Spedofconvergence. 37
2.6Interpretationanddiscusion.................... 38
2.7 Conclusion .............................. 41
3 Regional GDP Convergence in Germany 43
3.1 Introduction........ 43
3.2 Literature . 45
3.3Data................................. 47
3.4Empiricalanalysis.......................... 49
iCONTENTS ii
3.4.1 DensityfunctionsofrelativeGDP............. 50
3.4.2 Intradistributionmobility................. 54
3.4.3 Long-termanalysis..................... 58
3.4.4 Robustnestests...................... 61
3.5 Conclusion ......... 64
II Risk Sharing in Germany and the US 67
4 Risk Sharing and Fiscal Redistribution 68
4.1 Introduction............................. 68
4.2Data................................. 71
4.3 The short-term stabilizing e ffects of interregional risk sharing . . 74
4.4 The long-term redistributive e ffects of interregional risk sharing 81
4.4.1 Estimatingdistributiondynamicsandtheimpliedergodic
density............................ 82
4.4.2 Estimationresults...................... 85
4.5Discusion.............................. 8
4.6 Conclusion . 90
5 Home Bias and Neighborhood Bias 92
5.1 Introduction........ 92
5.2 Risk sharing, factor income flows,andlocalbiases 95
5.2.1 Risk sharing through factor income flows. 95
5.2.2 Risk sharing and biases in capital income flows...... 97
5.3Aspatialmodelofcapitalmarketrisksharing..........10
5.3.1 Summaryoftheempiricalstrategy............10
5.3.2 Model specification.....................101
5.4Estimatingrisksharingandtheneighborhoodbias.......105
5.4.1 Data.............................105
5.4.2 Cros-sectionalanalysis.106
5.4.3 Panel-dataestimation.10
5.5 Accounting for commuter flows15
5.6 Extensions for future research .123
5.7Conclusion..............................124
6 Concluding Remarks 126
Bibliography 129CONTENTS iii
A Appendix to Chapter 2 146
B Appendix to Chapter 3 153
B.1 Univariate adaptive kernel estimation . ..............153
B.2Bivariateadaptivekernelestimation................154
B.3Standarddeviationoftheergodicdensity155List of Figures
2.1 Relative unemployment rates in West Germany, 1960-2002 . . . 23
2.2 RelativeunemploymentratesinWestGermany,sub-periods1960-
1979and1980-202......................... 24
3.1 GDP per worker across German labor market regions, relative
to the German-wide average GDP per worker. Left: 1992, right:
2002.................................. 48
3.2 Densities of relative GDP per worker across 271 German labor
marketregions,192and202. 51
3.3 DensitiesofrelativeGDPperworkerforWest-andEastGerman
regionsseparately,1992and2002.................. 53
3.4 Surface and contour plots of g ( z| x) and g ( z| x). Left: yearly1 10
transitions between 1992 and 2002. Right: Ten-year transitions
(obtainedbymultiplyingtheone-yeartransitionprobabilitiesten
times)................................. 56
3.5 Boldline: ErgodicdensityofrelativeGDPperworker,calculated
on the basis of g ( z| x) (yearly transitions between 1992-2002).1
Thin line: Actual density of relative GDP per worker in 2002. . 60
3.6 Summaryofvariousrobustnesstests. ‘Outliersremoved’: regions
with the 10 highest and 10 lowest values of relative GDP per
worker in 1992 are excluded in all years. ‘Di fferent bandwidth’:
Sheather and Jones (1991) plug-in bandwidth criterion. . . . . . 61
3.7 SeparateanalysisoftheWestandEastGermaneconomies. Top:
Implied ergodic densities of West and East German GDP per
worker relative to the German-wide average GDP per worker.
Bottom: Implied ergodic densities of West and East German
GDP per worker relative to the West and East German aver-
ageGDPperworker,respectively.................. 64
ivLISTOF FIGURES v
4.1 Contourplotof f( ∆gd p − ∆ in c| ∆gd p)(capitalmarketsmoothing
ofidiosyncraticoutputshocks)................... 7
4.2 Contourplotof f( ∆ in c − ∆di nc| ∆gd p)(federalgovernmentsmooth-
ingofidiosyncraticoutputshocks)................ 79
4.3 Ergodic densities of relative output, income, and disposable in-
come, calculated on the basis of g ( z| x) for gd p , i n c , and dinc1
(yearlytransitionsbetwen195-202).............. 86
5.1 Capital market risk sharing of own idiosyncratic output shocks
among USfederal states, 1963-1998. Top: rawpoint estimates of
β against time. Bottom: smoothed sequences of β againstK, t K, t
time..................................107
5.2 Bold line: neighborhood bias in factor income flows among US
federal states, 1963-1998. Thin line: smoothed sequence of β K,t
against time, see Figure 5.1. Top: raw point estimates of β N,t
against time. Bottom: smoothed sequence of β against time. 110N, t
5.3 NetcommuterratesacrossUSfederalstates,1990(top)and2000
(botom)...............................17List of Tables
1.1 Studies which use Quah’s distribution dynamics approach in the
fieldofgrowthandconvergence.................. 7
2.1 ADF test for relative unemployment rates (without trend) . . . 25
2.2 Levin, Lin, and Chu and Breitung and Meyer tests for a unit
rootinrelativeunemploymentrates................ 27
2.3 Pooled AR(1) estimation with fixed e ffects............ 28
2.4 Im, Peasaran, and Shin and Sarno and Taylor tests for a unit
rootinrelativeunemploymentrates 30
2.5 Levin, Lin, and Chu and Im, Peasaran, and Shin tests for a unit
rootinrelativeunemploymentrateswithheterogeneouslaglengths 31
2.6 Perron and Vogelsang unit-root tests, lag length selected by se-
quentialt-tests........................... 36
2.7 Half-lifes (in years) of shocks to relative unemployment rates,
computed from impulse-response functions based on regression
resultsasreportedinTable2.6.................. 38
4.1OLSestimatesofrisksharingchannels(percent)......... 80
5.1 Spatial risk sharingmodel withspatial error autocorrelation and
fixed e ffects, estimated using the estimator developed by Elhorst
(2003)................................111
5.2Standardizedestimatesoftheneighborhoodbias.........15
5.3 Exclude federal states with intensive commuter flows......18
5.4 Impact of commuter flows on the neighborhood bias, Part I . . . 120
5.5 Impact of commuter flows on the neighborhood bias, Part II . . 123
A.1 Phillips and Perron test for relative unemployment rates (with-
outtrend)..............................147
A.2 KPSS tests for level stationarity of relative unemployment rates 148
viLISTOF TABLES vii
A.3 Madalla and Wu and Choi tests for a unit root in relative unem-
ploymentrates...........................149
A.4 Perron-Vogelsang unit-root tests, lag length selected by AIC . . 151
A.5 Perron-Vogelsang unit-root tests, lag length selected by BIC . . 152Preface
This thesis presents the research I undertook during the four years as a re-
searchandteaching assistant at the Chairof Macroeconomics at the University
ofDortmund. Ithasbeenapleasuretowritethisthesisunderthesupervisionof
Professor Heinz Holländer and Professor Mathias Ho ffmann and I am indebted
to both of them for their support. In particular, I would like to thank Math-
ias Ho ffmann for his suggestion to develop a transfer of techniques from the
growth econometrics literature to the macroeconomic literature on aggregate
risk sharing. Both papers presented in Part II pick up this suggestion.
Ihavehighlybenefited from discussions with professors, colleagues, and
fellow doctoral students of the Graduiertenkolleg ‘Allokationstheorie, Wirt-
schaftspolitik und kollektive Entscheidungen’. Moreover, I have benefited from
many comments and suggestions which I have received at various international
conferences.
Avery special mentioning deserves my colleague Christian Bayer. He was a
soundingboardfornewideasandhelpedmetoimprovemyresearchthroughout
the whole process. The second chapter of this thesis presents a paper which we
have written in joint work. A non-exhaustive list of other people that in some
way had an impact on the final version includes Michael Roos, Eckhardt Bode,
Wolfram Richter, Wolfgang Leininger, Kornelius Kraft, Christoph Alsleben,
Julia Angerhausen, Christiane Schuppert, and Frauke Eckermann. Moreover, I
thank several anonymous referees of academic journals who have peer-reviewed
the papers presented in Sections 2, 3, and 4.
viiiChapter 1
Introduction
This thesis presents an analysis of empirical