Abstract
We investigate the effects of financial development on recession while controlling for potential recession factors using data of about 129 countries covering the 1990-2010 period. To the best of our knowledge, this is the first study examining this relationship using a plural and innovative methodology along with a newly primary and hitherto almost unexploited “Rare macroeconomic disasters” data from Barro and Ursua (2012) which allow us to build a more specific proxy of the variable “economic recession”. We executed step by step “Feasible Generalized Least Squares”, “Locally Weighted Scatterplot Smoothing”, “Local Linear” and “Iteratively Reweighted Least Squares” regression methods joint with a Sasabuchi test to verify the inverse U-shape and estimate the extreme point. Strong evidence of a nonlinear and U-shaped relationship between both phenomena with a threshold effect of 1.1528 emerged. Financial development presents an expansionary impact for countries with financial performance less than the threshold, while there is a recessionary impact for countries with financial performance above it. This last result is robust to the control of extreme values and heterogeneity within the sample. The semiparametric regression shows that the results of the parametric part converges with the previous results in general and reveals with illustration the functional form of the nonlinear relation between both indicators. Moreover, controlling the heterogeneity within our sample permitted us to demonstrate that fuels for South Asia (SASIA) and Latin America and Caribbean (LAC) countries and financial openness for sub-Saharan Africa (SSA) countries are negatively related to recessions. Related and relevant policy recommendations are highlighted and discussed.
