Capital Inflow Surges in Emerging Economies: How Worried Should Latin America and the Caribbean Be?

Abstract

Listen

Translate article

In the years leading up to the global economic crisis, Latin America and the Caribbean received large capital inflows. Nonetheless, the region survived the (albeit short-lived) dramatic events of 2007-09 reasonably well, with no financial crisis in any of the larger economies. After the crisis, capital inflows rose again: in 2010 gross inflows to seven Latin American countries (Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Venezuela) amounted to more than 6 percent of gross domestic product (GDP), and net inflows exceeded 3 percent of GDP. More recently following the May 2013 speech of then-Chairman of the U.S. Federal Reserve, Ben Bernanke, there has been concern regarding the potential effects of tapering asset purchases by the U.S. Federal Reserve System and rising interest rates, possibly spelling the end of large capital inflows to the Latin American and Caribbean region.The literature on capital inflows suggests that while these flows represent an opportunity to spur investment and growth, they may also increase vulnerability to financial crises, macroeconomic instability, and ultimately recession. A relevant question, then, is whether the region can declare victory in being able to manage large capital inflows successfully or whether risks remain that justify strong measures.Our starting point is that the region has enjoyed strong surges in capital inflows. We therefore decided to condition the analysis on surges occurring. We do not seek to explain the reasons for such a surge. One justification for this decision is that capital flows appear to be driven as much, if not more, by push rather than pull factors. More specifically, we seek to understand what determines whether such surges end in either a banking crisis or a recession. While there is a substantial literature on the determinants of capital inflows, their potential impacts have received much less attention. The contribution of this paper is to analyze when inflow surges may end in severe economic problems, including financial or macroeconomic instability, rather than to identify the determinants of those surges.We must first tackle the question of how to define a capital inflow surge. We are helped in this regard by a small but growing literature. However, the current definitions of inflow surges in the literature may be injudicious. We propose a new definition which considers the total size of the inflow episode across countries. We also employ a more traditional definition based on yearby-year inflows relative to a country-specific threshold. We contrast relevant statistics on the identified episodes using these two definitions and compare our episodes with others found in the literature.We also consider whether the analysis should be conducted using gross or net inflows. We take the view that employing net flows may discard useful information and that gross flows are relevant where financial instability is a potential concern. Even if net flows are zero, a domestic financial system may intermediate large gross inflows when domestic residents purchase assets abroad. We choose to focus on gross inflows, but we include gross outflows as a potential explanatory variable, so as not to lose information.1The empirical analysis in this paper develops a set of parsimonious models that discriminate reasonably well between the different economic outcomes. We then analyze the case of a typical country from the region with a recent capital inflow surge, providing a set of simulations. To illustrate the economic magnitude of our results, we simulate the changes in the estimated probabilities of negative outcomes given changes in certain critical variables. We suggest that, depending on the country/inflow characteristics, different policy measures may then be justified. However, a caveat is that emerging economies in general and Latin America and the Caribbean in particular have been experimenting with new policy measures, and we are not able to evaluate their potential effectiveness in this paper. …