Mercredi | 2014-11-05
salle B103, 12h-14h
This paper empirically analyses the relationship between public spending and GDP in a panel of Emerging (EMEs) and Advanced Economies (AEs) with quarterly data from 1990 to 2013. Using a Panel VAR model, we compute spending multipliers in AEs and EMEs and we confirm previous results in the literature: EMEs have smaller spending multipliers than AEs, and the persistence is weak in both cases. We then use a Panel Conditionally Homogenous VAR model (Georgiadis, 2012) in order to test if the development level modifies the role of the main determinants of fiscal multipliers. This method allows the relationship between GDP and public spending to change across countries according to one or two conditioning variables. First of all, we find that traditional determinants (imports, public debt, unemployment, savings, interest rates, financial development) act the same way in EMEs and AEs. Secondly, in relative terms and considering the weakness of fiscal multipliers in EMEs, public spending efficacy is more sensitive to the considered determinant in EMEs than in AEs. Thirdly, each tested determinant is individually insufficient to reach the same efficiency of public spending in EMEs than in AEs. This last result makes us conclude that we cannot directly extend conclusions about AEs to the case of EMEs, there are others factors on which government have to act in order to improve their efficiency.