The Mystery of Capital Flows

Economists, or at least this one, spend quite a lot of time in trying to understand what drives capital flows and what impact they have on growth and welfare. I started writing this column last Friday but then had to hunker down,” as Texans say, for hurricane Ike. We lost power along with lots of capital in the city of Houston. We had to relocate to Austin for a couple of days. On the way back to Houston I saw a lot of capital coming into the city from 30 dierent states and Canada!So, why does capital flow where it does? Productivity shocks (positive or negative), broadly defined, are clearly a driver. Legal, and more importantly, national barriers are also important. Economic models predict that capital will move from regions where the marginal product of capital is relatively low to regions where it is relatively high. The marginal product of capital is negatively related to the level of capital stock, but positively related to the level of total factor productivity, a rough gauge of overall productivity in the economy. Hence, we cannot simply assume that capital flows to capital scarce countries (i.e. developing nations), Recent research shows a positive relation between capital flows and various proxies for productivity, such as good institutions (Alfaro, Kalemli-Ozcan, Volosovych, 2008), and low risk of sovereign default (Reinhart and Rogo, 2004). As a result capital might flow uphill” from poor, capital-scarce countries to rich, capital-abundant, but more productive countries. If we assume fully integrated nancial markets, and perfectly diversied ownership of capital, then the economic models imply that the capital stock will be highest in regions with the highest level of productivity.

Can we test this? Is capital really flowing uphill?” If we would like to give a chance to the above outlined theory we really have to test it in a fully integrated nancial market, such as the U.S., or the EU. In Kalemli-Ozcan, Reshef, Srensen, and Yosha (2005), and in Ekinci, Kalemli-Ozcan, and Srensen (2007), we showed that capital flows between regions within countries such as the U.S. and Germany are consistent with the predictions of the models, while flows between European countriesare not. Specically, within a single country capital does flow to productive places, but this is not true between countries, even when these are member states of Europe and not a disconnected group of emerging markets. Should we expect EU member states to behave like the U.S states at some point? In Kalemli-Ozcan, Sorensen, and Turan (2008), we compare the integration of the U.S. capital markets since 1950 with that of EU countries since 1970. When long-standing barriers to capital flows are removed capital flows to relatively poor regions during a catch-up growth” phase and when this phase is over capital flows to regions with high productivity. Barriers to capital flows within the U.S. disappeared in the mid-20th century and for a period of time capital flowed from the wealthy northern states to poorer states in the old Confederacy. Today, U.S. states are past the catch-up” phase and capital flows to states with positive productivity shocks|which tend to already be the rich” states. During past week, with the destruction of capital in Houston due to the hurricane, we just have witnessed a mini catch-up” phase. EU countries still appears to be in this type of catch-up” phase: Within the EU capital flows to poorer, but fast growing countries, such as Greece and Portugal.*

What happens outside of the EU and U.S., especially among the emerging markets? In a sample of 100+ countries all bets are o. There are stark dierences in fundamentals among these countries in addition to the fact that some governments target the current accounts, which all might be hard to account for in a standard analysis. A rst look at the data (gure below) shows that there is no robust relation between capital flows (measured as the negative of current account to GDP) and GDP growth for the largest sample of 106 developing countries (black line), in the 1990s. But if we concentrate on a sample of 67 developing, Asian countries (red line), poor and low growth countries, such as Madagascar, seem to be net borrowers and not so rich but high growth countries, such as China, seem to be net lenders to the rest of the world. And yet when we look at the sample of Eastern European countries (blue line), it seems that not so rich but high growth places, such as Latvia, seem to be net borrowers as opposed to low growth Russia who is a net lender. Exact opposites. If we add developed countries to the gure, the picture gets even more complicated. The fact that China runs a current account surplus and the U.S. a decit does not mean the rest of the emerging markets are all net lenders (as shown in the gure) and the rest of the rich world are all net borrowers. Is China done with the catch-up growth phase? They sure did not seem tobe capital-scarce at the Olympics opening ceremony, whereas the U.S. is in urgent need of capital right now!

*Also shown by Blanchard and Giavazzi (2002)




Related RGE Content: