This dissertation explores the dynamics adjustment of output, output components, labor market variables and output growth after currency crises. In the first essay, I examine the dynamic response of output and its components to large devaluations in 159 countries over the 1970-2005 period. Recent empirical literature on measuring the effects of currency crises shows permanent drops in output following these events. However, theoretical models of contractionary devaluations predict only temporary drops in output due to nominal wage rigidities. I examine whether one can reconcile early theoretical predictions of full long-run recovery of output with the data after isolating the twin crisis episodes. Our estimates show that the output fully recovers to its pre-crisis level in about four years when a currency crisis is not associated with a banking crisis. The analyses of the components show that the initial contractionary effects are due to drops in consumption and investment with slow recovery. The net exports do increase though not sufficient to offset the drops in the other components. In the second essay, I use a panel of 66 countries over the 1990-2010 period to examine the short and long-term impact of currency crises on key labor market variables including real wages. Previous studies are either short-term, case study analysis of the issue or focus mainly on employment. The results, suggest that the adjustment in labor market following a currency crisis is mainly through real wages. While employment and labor productivity initially decline, these decreases are far less than the decrease in real wages and the recovery is also swifter. The recovery in real wages is sluggish and usually takes at least three and at most eight years. The results are robust to different definitions of currency crises and estimation methods. The third essay concerns the shape of GDP growth recovery following currency crises. While theoretically the shape of output growth recovery is ambiguous, the empirical literature employs univariate approach and reports mixed results. Therefore, I re-examine the issue further with a multivariate approach. I estimate "U-shape", "V-shape" and "L-shape" models of growth recovery first proposed by Kim et al. (2005) in the context of post-recession recovery. Thereafter, based on the information criteria I allow the data to select the appropriate model. The methodology enables us to test for the significance of higher than average growth rate or 'bounce-back' term in the recovery phase. I consider currency crashes and reserve crises, as the two forms of currency crises, separately. The results suggest that there is a significant and positive bounce-back in growth rates after currency crashes. Moreover, the shape of the recovery is "U-shaped" across the entire panel. Alternatively, after reserve losses, the recovery resembles a long "U-shaped" model and there is no bounce-back in growth rates. Moreover, I found that a large pre-crisis short term debt relative to total foreign debt and pre-crisis share of export sector in an economy are the best predictors of post-crisis recovery.
展开▼
