Essays on Fiscal Policy, Private Consumption and Non-Stationary Panel Data

Abstract: In the first essay of this thesis, we examine whether the effects of fiscal expansions and contractions exert a symmetric influence on the relationship between fiscal policy and private consumption. Our results indicate a considerable asymmetry across fiscal expansions and contractions. More specifically, during fiscal contractions, transfers seem to influence private consumption in a neoclassical way, which indicates that contractions alter the expected permanence of transfer changes. The results imply that focus has to be directed towards situations where the effects of fiscal policy are not altered symmetrically across expansions and contractions. In the second essay, we look at one possible explanation for the asymmetry found in the first essay. To specify an empirical model, we utilize a theoretical model that suggests that the effects of fiscal policy on private consumption depend on the level of government debt. If the theoretical model can be used to explain the asymmetries in fiscal policy effects, the effect of changes in transfers should vary with the level of government debt. Using a time-varying parameter model, to study if the Swedish experience is consistent with the prediction of the theoretical model, we find evidence indicating that government debt can be used to explain asymmetries in fiscal policy effects. In the third essay, we investigate if government consumption, by acting either as a substitute for, or a complement to, private consumption, can help explain the non-Keynesian effects of fiscal policy that have occurred in Denmark, Germany, Ireland and Sweden. By using econometric methodologies, that both exogenously and endogenously recognize that government consumption can act as either a substitute for or a complement to private consumption, we find that the empirical evidence indicates that the complement/substitution dichotomy can help explain the non-Keynesian effects of fiscal policy that have been previously documented in Denmark, Ireland and Sweden. In the fourth essay, we investigate the effects of cross-sectional correlation in a homogeneous panel data unit root test. We find that cross-sectional correlation causes size distortions, and investigate how well one existing remedy works. We find that the remedy is inappropriate in small samples, and suggest a small-sample homogeneous panel data unit root test that has good size properties in the presence of cross-sectional correlation. To illustrate the adverse effects of cross-sectional correlation in empirical applications of panel data unit root tests, we test the random walk hypothesis for private consumption. In the fifth essay, we extend the scope of the analysis in the fourth chapter to include stationarity testing in panel data. We show that a panel data stationarity test is size-distorted when errors are cross-sectionally correlated, and suggest an orthogonalization procedure to eliminate the size distortions. By applying both the test that does not account for cross-sectional correlation, and the test suggested in the fifth essay, to test for output convergence, we show that inference from panel data stationarity tests can be adversely affected if cross-sectional correlation is not accounted for.