16. The Case for Regulating Cross-Border Capital Flows
- International Monetary Fund
- Published Date:
- March 2012
José Antonio Ocampo
Let me start this chapter with two remarks. The first one is that cross-border capital flows are one of the expressions of finance. It is thus peculiar that the Group of Twenty excluded this topic in its discussions about reregulating finance, as if cross-border finance were not finance. In fact, when we talk about regulating domestic finance, we always use the word regulation, but when we talk about cross-border finance, we use the word controls. I prefer and have long used the concept of capital-account regulations because they are regulations. Some are quantitative, such as the prohibitions on certain activities or limits on them, which is also true of some domestic prudential regulations. Some are price-based, such as unremunerated reserve requirements, again like domestic regulations, such as capital or liquidity requirements. They all belong to the same family of what we now call macroprudential regulations.
My second introductory point is that I come from a tradition of macroeconomic thinking in which the balance of payments is always placed at the center, as the source of both positive and negative shocks and therefore as the essential source of business cycles. This tradition is the best starting point to analyze the macroeconomic dynamics of developing countries and particularly of emerging-market economies. Furthermore, a broadly accepted principle is that the best policies are those designed to intervene directly at the source of the shocks. A clear case is a terms-of-trade shock. The best intervention in this case deals directly with the shock by saving a significant amount of a commodity price boom and using those savings later when prices fall. This is what Chile has been doing with its copper revenues and what Colombia did with coffee for a long time. This principle is widely accepted.
The same thing should be true of capital accounts. They are the main source of shocks that emerging markets have experienced in recent decades, both positive and negative. Therefore, policies should also focus in this case on the source of the shock. Furthermore, by referring to them as positive and negative, I am also expressing the fact that capital flows have a temporary component, which authorities may want to smooth out. Indeed, such smoothing out is the essence of what we mean by a countercyclical macroeconomic policy.
The starting point here is the recognition that finance is procyclical, particularly for those agents that are considered to be riskier by markets. This reflects the fact that there is a significant segmentation in financial markets and that riskier agents are subject to stronger procyclical shocks than the average market pattern. This is true of small and medium-sized enterprises and consumer credit in all economies. And it is true of developing countries in global markets. Developing countries are considered to be riskier borrowers, so they are subject to the phenomena of flight to quality and sudden stops during crises, but they are also subject to risk appetite during booms such as the one several emerging economies are experiencing now. That strong procyclicality is the particular issue that policies should address.
On top of that problem, in developing countries financial markets are more incomplete, and this is reflected in the variable mixes of currency and maturity mismatches in portfolios. These problems are more important for developing countries and emerging markets than they are for industrial economies. As countries develop deeper financial systems, markets will partly solve these problems. In any case, they should be the subject of specific attention by regulatory authorities.
Procyclicality is not only a question of short-term volatility. In a sense, short-term volatility is the easiest to administer through active foreign-exchange reserve management. The most difficult problem is managing the medium-term cycle of capital flows. We have experienced three such medium-term cycles in recent decades and may be starting a fourth. There was one that started in the mid-1970s and went through the 1980s. Then we had the 1990 to 1997 boom (briefly interrupted by the December 1994 Mexican crisis) and the long crisis that started in Asia in mid-1997. The most recent cycle was the boom from 2003 to mid-2008, followed by the sudden stop as a result of the Lehman Brothers collapse. This was a shorter cycle, and a fourth one may have started in mid-2009.
The basic problem of these medium-term cycles is that they drive the macroeconomics of emerging and developing economies while simultaneously constraining the capacity to undertake countercyclical macro-economic policies. They drive the exchange rate, spending, and domestic demand. They also drive fiscal policy in a procyclical way as revenues increase and access to capital markets is made easier during booms, and both are cut sharply during crises. They also drive interest rates in a procyclical way because markets tend to reduce risk spreads and thus interest rates during booms and increase them during crises. Therefore, countercyclical macroeconomic policy has to lean against those strong procyclical market trends. If authorities want to adopt countercyclical macroeconomic policies, they have to ask themselves why they would not want to intervene at all at the major source of the business cycle—the cyclicality of capital flows.
In fact, if authorities do not manage the capital account, they will really be making a choice about the particular way they want the pro-cyclical effects of financial markets to be reflected in their economy. If they control the exchange rate, they will have to give up managing the interest rate, and then the policy package would be clearly procyclical. If they decide to control the interest rate, they have to give up managing the exchange rate, but appreciation during booms and depreciation during crises also have procyclical effects. The countercyclical effects that operate through the current account of the balance of payments are well known. But if we allowed this effect to run fully, exchange-rate appreciation during booms will tend to generate overvaluation and current-account deficits that increase the risk of a crisis later on. So it is a double-edged sword. Furthermore, procyclical effects operate through balance sheets in countries where the private sector has net external liabilities in foreign currencies and through income distribution (increases in real wages as a result of appreciation and reductions due to depreciation). These procyclical effects tend to predominate in practice.
In both cases, authorities are not controlling procyclicality: they are really choosing which effect of market procyclicality they are allowing into the economy. This looks very much more like Robert Mundell’s analysis of a fixed exchange-rate system: authorities control the composition of the quantity of money, but they have no control over the quantity of money. In the case we are discussing, authorities are choosing whether they want the procyclicality of global financial markets to be reflected in interest rates or exchange rates, but they are not controlling procyclicality. That is why, if authorities are thinking of doing countercyclical management, they should start by looking at the source of procyclicality—the capital account. And they will have to think of combining this with aggressive foreign-exchange reserve management. One way of understanding this is that, since the instruments they normally have are insufficient to run a countercyclical policy, they have to look for additional ones.
The developing-country and emerging-market authorities have actually been quite wise in this regard. They have discovered that they can obtain more degrees of freedom to adopt countercyclical macroeconomic policies by adding policy instruments, either intervening massively in the foreign-exchange market (buying and selling depending on the phase of the cycle) or regulating capital flows—or both. In this regard, I agree with a point that Rakesh Mohan has raised about the impossible trinity being a somewhat confusing way of understanding the policy choices (chapter 15 in this book). In fact, all the interesting things happen inside the impossible trinity.
What does the empirical work on the effectiveness of regulations of the capital account indicate? The most comprehensive analysis I have read on this issue is a 2000 International Monetary Fund study by Akira Ariyoshi and others. The one that was done in 2010 by Jonathan Ostry and collaborators was provocative and an excellent way to stimulate the appetite for the debate, but I think the IMF should take a broader look at what countries experience and what those experiences reveal about the effectiveness of different types of regulations.
Considering those studies and many others, capital-account regulations can be said to have two different effects. The first effect is that they operate as a liability policy. They are designed to improve the term structure of the country’s liabilities, and in this regard, the evidence is strong that regulations are effective, whether they are price- or quantity-based.
The second effect is that regulations are a complement or support to countercyclical macroeconomic policies. In this regard, the evidence has been subject to much more debate, particularly in the case of price-based controls, such as the unremunerated reserve requirements used by Chile and Colombia in the 1990s and Colombia again in the mid-2000s. Studies have shown, however, that they affect the quantity of flows or domestic interest rates or both, thus allowing authorities to increase rates during booms without attracting additional capital flows. Either way, they are effective, and monetary policy will then determine whether the regulations affect the quantity of flows or the domestic interest rate or a mix of both. Regulations are effective as long as the margin between the domestic interest rate and the foreign interest rate can be raised without attracting additional capital flows or as long as flows can be reduced at given interest-rate differentials.
The evidence also indicates that some of those effects are temporary. In my writings, I have referred to capital-account regulations as speed bumps rather than permanent restrictions because market agents learn how to avoid them. This implies that authorities have to be equally dynamic, strengthen regulations over time, and close loopholes to make them effective. This is true for any type of prudential regulation. Authorities always have to see how the market is evolving and adjust regulations to make them more effective.