Capital flow episodes and emerging market economies

Understanding capital flows in emerging market economies has been important in the literature of international economics with the growing integration of these economies with the global financial markets. After the recent financial crisis and the U.S. monetary policy hitting the “zero lower bound,’’ there has been a resurgence of interest in analyzing the capital flow movement in these economies. Historically, the flows in these economies have been episodic. There are times when they experience a huge inflow of capital (“surges’) and there are times when there is huge outflow of capital (“flights”). The existing literature, however, lacks proper modeling of such extreme episodes. They rely on ad-hoc definitions of surges and flights.

This paper attempts to identify such extreme episodes using formal statistical models. In particular, I use a Markov switching model to characterize the different episodes of net capital inflows for a sample of 46 emerging markets economies. I allow the net capital inflows to follow different regimes and characterize the regimes by the values of the mean of the net capital inflows. I refer to the regimes as ‘high inflow’(surges) and ‘low inflow’ (no- surges) regimes.

The preliminary findings suggest, for some countries , a two state model best fits the data while for some countries a three state model is more suited to identify periods of extreme inflows. For the three state model, I characterize the states as ‘high inflow’(surges), ‘medium inflow’ and ‘low inflow’ states. This suggests that there is some heterogeneity in characterization of the regimes of net capital inflows for these countries. Hence, using an ad-hoc algorithm to identify surge and non surge episodes in these economies may not be appropriate.