Université libre de Bruxelles
Faculté des Sciences Sociales et Politiques Solvay Brussels School of Economics and Management
Essays on the Econometrics of Macroeconomic Survey Data
Dissertation présentée en vue de l’obtention du grade de Docteur en Sciences Economiques
par
Cristina Conflitti
Promoteur: Professeur David Veredas Membres du Jury:
Professeur Ingmar Nolte Professeur Catherine Dehon
Professeur Christine De Mol Professeur Domenico Giannone
Année académique 2011-2012
Contents
Introduction 1
1 Measuring Uncertainty and Disagreement in the European Survey of Professional Forecasters 4
1.1 Introduction . . . 5
1.2 The Survey of Professional Forecasters data . . . 8
1.3 Disagreement and Uncertainty . . . 12
1.3.1 The Theoretical Framework . . . 13
1.4 Probability Forecast Estimation . . . 14
1.5 Empirical results . . . 16
1.5.1 GDP . . . 20
1.5.2 Inflation . . . 21
1.5.3 Unemployment . . . 22
1.6 Conclusion . . . 22
1.7 Appendix . . . 32
2 Optimal Combination of Survey Forecasts 42 2.1 Introduction . . . 43
2.2 Optimal combination of point forecasts . . . 45
2.3 Optimal combination of density forecasts . . . 48
2.4 Empirical analysis . . . 50
2.5 Conclusion . . . 53
i 3 Opinion surveys on the euro: a multilevel multinomial logistic analysis 60 3.1 Introduction . . . 61
3.2 Description of the data . . . 64
3.3 The multilevel multinomial logisticmodel . . . 65
3.4 Time panel extension and multivariate extension of the multilevel multinomial model . . . 68
3.5 Results . . . 69
3.6 Concluding remarks . . . 73
3.7 Appendix . . . 86
3.7.1 A.1 . . . 86
3.7.2 A.2 . . . 87
List of Figures
1.1 Survey time line forecast . . . 24
1.2 Example of questionnaire . . . 25
1.3 Example of respondents’ answer . . . 25
1.4 Normal VS Piecewise approximation . . . 26
1.5 Comparison Average point forecast and individual point forecasts . . . 27
1.6 Number of expectation higher than the mean expectation . . . 27
1.7 Disagreement on point estimate: GDP, inflation and unemployment . . 28
1.8 Disagreement and uncertainty: GDP . . . 29
1.9 Disagreement and uncertainty: inflation . . . 30
1.10 Disagreement and uncertainty: unemployment . . . 31
1.11 Inflation aggregate histograms 1999q1- 2003q4. . . . 33
1.12 Inflation aggregate histograms 2004q1- 2008q4. . . . 34
1.13 GDP aggregate histograms 1999q1- 2003q4. . . . 35
1.14 GDP aggregate histograms 2004q1- 2008q4. . . . 36
1.15 Unemployment aggregate histograms 1999q1- 2003q4. . . . 37
1.16 Unemployment aggregate histograms 2004q1- 2008q4. . . . 38
2.1 Incidence of missing data . . . 55
2.2 Density combination . . . 56
3.1 Cook’s distance Model 1. Flash Eurobarometer 193. . . 74
3.2 Cook’s distance Model 2. Flash Eurobarometer 193. . . 75
List of Tables
1.1 Main features of the Survey of Professional Forecasters . . . 9
1.2 Timing European SPF . . . 9
1.3 Summary Statistics 1999-2008 and Forecasts Performance . . . 11
1.4 MZ Regression - Beta Coefficients . . . 12
1.5 Disagreement* on point estimates forecasts. . . 17
1.6 Correlation mean expectation and disagreement . . . 18
1.7 Kolmogorov-Smirnov Test on probability density . . . 19
2.1 MSFE for point forecasts . . . 53
2.2 Average minus-log score for density forecasts . . . 54
3.1 Flash Eurobarometer: Euro perception (Euro Area 12 countries) - percentage value . . . 76
3.2 Covariates summary statistics . . . 77
3.3 Flash Eurobarometer. Dependent variable summary statistics - percentage value . . . 78
3.4 Models’ equations . . . 80
3.5 Public opinion on the euro: advantageous versus disadvantageous. Randomintercept models . . . 81
3.6 Public opinion on the euro: advantageous versus disadvantageous. Randomslopemodels . . . 82
3.7 Interactions analysis. Advantageous vs disadvantageous. . . 83
3.8 Public opinion on the euro: time extension . . . 84
3.9 Public opinion on the euro: multivariate extension . . . 85