Essays on the effects of government spending in the presence of skill accumulation through Learning-by-Doing

Abstract

In this thesis I present a theoretical and empirical investigation of the effects and propagation mechanisms of a government spending shock with Learning- By-Doing (LBD). A positive government spending shock increases hours worked. By the LBD mechanism the increase in hours yields an increase in productivity and hence a decline in inflation and interest rates. The fall in the long-term real interest rate generates an inter-temporal effect leading my model results closer in line with the data. In chapter 1 I first provide a literature review on the effects of a change in government spending. I then present a DSGE new Keynesian model with LBD and comment its main features. I show that, by including LBD, the model generates an increase in productivity and consumption in line with the empirical evidences. Chapter 2 analysis the effect of a government spending shock on the real exchange rate. I show that including LBD makes the model able to reproduce the real exchange rate depreciation observed in the data. This result derives from the increase in consumption and the assumption of international risk sharing condition. Chapter 3 investigates the effect of a government spending shock on housing market. I find Vector autoregression (VAR) evidence that house prices increase after a government spending shock. In a model where housing can be used by credit constrained households as collateral to borrow, the increase in housing wealth introduces an additional propagation mechanism for a government spending shock. I present a model with two sectors, heterogeneity in the households' discount factor, and credit constrained agents. I show that introducing the LBD mechanism, by contrast with a model where LBD is absent, makes the model able to replicate the observed increase of real house prices. The model is estimated by matching DSGE and VAR impulses responses.