Journal Issue

Fourth Annual IMF Research Conference: What drives capital flow cycles?

International Monetary Fund. External Relations Dept.
Published Date:
December 2003
  • ShareShare
Show Summary Details

Economists from the IMF and elsewhere gathered at the IMF’s Fourth Annual Research Conference on November 6-7. The conference had an overarching theme of capital flows and macroeconomic cycles but also dealt with other issues. (Papers are available on the IMF’s website (

This year’s conference, noted Managing Director Horst Köhler in opening remarks, had come at an opportune time. Over the past year, financing conditions for developing and emerging market economies had steadily improved, with spreads compressed to near all-time lows. But was the world in for another bout of exuberance? Had lessons been learned? Good research, Köhler said, is crucially important if the IMF is to better understand the mechanisms that drive capital flow cycles and help member countries shape the policies needed to meet new challenges.

Current account imbalances

Sebastian Edwards (University of California, Los Angeles, and National Bureau of Economic Research) delivered the Fourth Mundell-Fleming Lecture, “Current Account Imbalances: History, Trends, and Adjustment Mechanisms.” The issue could not have been more topical given the ongoing debate on the United States’ current account deficit and its eventual correction.

Edwards reviewed the distribution of current account imbalances in the world economy during the past 32 years and analyzed the pattern of adjustment that countries followed in dealing with large payments disequilibria. He focused, in particular, on the connection between adjustment and exchange rate regimes; and between the cost of current account deficit reversals and “sudden stops” of capital inflows. Edwards also considered whether openness, the extent of dollarization, and the exchange rate regime affect the cost of reversals.

Edwards’ findings indicated that throughout the sample period (1970-2001), the vast majority of countries ran current account deficits, but, with the exception of a few countries, large current account deficits did not persist for significant periods. Edwards found that a larger number of countries ran persistently large surpluses, implying that the nature of the adjustment process was asymmetrical.

Also, major reversals in current account deficits were strongly associated with “sudden stops” of capital flows and with a high probability of exchange rate crises. That said, Edwards found no statistically significant relationship between reversals and banking crises or between reversals and IMF-supported adjustment programs within a three-year window. Edwards concluded that current account reversals had a negative effect on real growth that went beyond their direct effect on investment, that the size of the effect of reversals on growth depended on the country’s degree of openness (more open countries suffer less), that there appeared to be no link between a higher degree of dollarization and the negative effects of reversals, and that countries with more flexible exchange rate regimes were better able to adapt to reversals than countries with more rigid exchange rate regimes.

“Trilemma” alive and well?

Are policymakers really forced to choose no more than two policy options from among fixed exchange rates, free capital mobility, and monetary policy independence? After studying historical data from the gold standard era to the present, Maurice Obstfeld (University of California, Berkeley), Jay C. Shambaugh (Dartmouth College), and Alan M. Taylor (University of California, Davis) came to believe that the “trilemma” forces stark choices. Using differentials between national and international short-term market interest rates to measure monetary independence, they found that countries that opted to peg their exchange rate lost considerable monetary independence, absent capital controls, and countries that did not peg had a fair amount of monetary policy autonomy even without capital controls. Whether they chose to exercise it was another matter—quite a few did not.

The authors focused on three periods: the pre-World War I gold standard (1870-1913), the convertible Bretton Woods system (1959-73), and the post-Bretton Woods era (1974-present). They saw the gold standard as “a period of mostly fixed exchange rates, unfettered capital mobility, and, hence, limited monetary independence.”

The Bretton Woods regime, by contrast, was designed to preserve the monetary independence that had been emerging in the interwar period. Relatively stable fixed-but-adjustable exchange rates coupled with the pervasive capital controls of the Bretton Woods period succeeded in giving more room for monetary autonomy. But as capital controls diminished, “the combination of exchange rate pegs and monetary independence became untenable.” Unpegging rates has at least partially restored monetary independence, if central banks choose to use it.

Discussant Hélène Rey (Princeton University) suggested that, for periods with high capital mobility, the research might be measuring market integration rather than monetary autonomy or, at least, be unable to distinguish between the two. She also pointed out that the functions of monetary policy have changed greatly over time, leading central banks to view short-term interest rates and exchange rates quite differently in different periods, which could explain why some central banks might choose not to use autonomy when they have it. She added that common shocks could account for many of the observations, a possibility that the authors had also noted.

Does credit follow trade?

Do international trade patterns determine lending patterns? Economists Andrew K. Rose (University of California, Berkeley) and Mark M. Spiegel (Federal Reserve Bank of San Francisco) hypothesized that creditor country banks lend more to sovereign borrowers in trading partner countries because the banks believe borrowers will avoid default for fear of the high cost of the trade reduction likely to ensue. Using a gravity model of sovereign lending, they confirmed that creditors do lend more to the countries with which they trade most.

Discussant Mark Wright (Stanford University) noted that the apparent pattern of credit following trade might be explained nearly as well by models based on the borrowers’ credit market reputation. The authors agreed that their results did not necessarily refute reputation-based models of sovereign debt. And panel chair Gerd Häusler (IMF) pointed to a potential weakness in the bank credit data, which arises from the fact that the big banks that are the lenders of record usually sell the debt immediately rather than holding it as they might have in the past. Wright observed that the model did not control for the possibly large role of bank trade credits.

Emerging market risk without reward?

A study by Christoph Klingen (IMF), Beatrice Weder (University of Mainz), and Jeromin Zettelmeyer (IMF) of “How Private Creditors Fared in Emerging Debt Markets, 1970-2000” yielded some puzzling results. Theory predicts that lenders will demand that risk be rewarded with higher returns. Therefore, loans to sovereigns considered more likely to default ought to earn a premium over the long term. But the authors found little difference between estimated returns on emerging market debt and U.S. Treasury bonds over the 30-year period.

Their finding becomes less surprising when returns for subperiods are considered. The 1970-89 cycle yielded negative or very low returns (ex post spreads over treasuries). Very high returns were realized from 1989 until 1993, and lower but positive returns from 1994 to 2000. A plausible explanation offered by Wright is that the defaults of the 1980s were truly unexpected—a very rare event from which long-term returns are still recovering. The authors agreed that only time will tell. Another surprise is that ex post returns on Brady bonds were quite high—the authors observed that the success of the Brady deals (negotiated debt write-downs with official sponsorship) drove the high returns for the early 1990s. Wright felt this might be evidence that there is a collective action problem in debt rescheduling that needs to be solved.

Shedding light on the quantity puzzle

As countries become more financially integrated, the expectation is that movements in consumption will become more synchronized internationally than movements in output. With capital flows tending to follow high returns, they will tend to boost growth in already fast-growing economies and depress growth where it is already slow, resulting in negative output correlations across countries. By contrast, with consumption decisions being based on fully diversified portfolios of wealth, consumption will be strongly correlated, positively, across countries. However, the data overwhelmingly show that movements in output are more synchronized internationally than movements in consumption, even among financially integrated economies.

It is this anomaly—referred to in the economics literature as the quantity puzzle—that Jean Imbs (London Business School) explored, examining two prominent explanations for the quantity puzzle. The first hypothesis maintains that capital flows are more restricted than popularly believed; as a result, diversification is limited and consumption plans remain largely idiosyncratic and less correlated internationally than GDP fluctuations. The second hypothesis is that capital flows, though unrestricted, are governed by imperfect information and tend to herd rather than respond to differentials in returns. Thus, fluctuations in output can become more synchronized between financially integrated regions.

Imbs argued that the consumption smoothing effect of financial integration is present but overwhelmed by an independent effect of integration on output, the consequence of a “herding” behavior that causes high output correlations across countries.

Other Resources Citing This Publication