Controls on Capital Outflows On September 28, 1998, Paul Krugman posted on the web a thoughtful and

provocative article on the use of controls on capital outflows to combat a specula- tive attack.8 The following day, Malaysia’s prime minister Mahathir imposed such controls. And they say no one listens to economists! True, by February 1999, Malaysia had lifted most of its controls, and it is not obvious that the country has fared any better than other similar Asian countries in emerging from the region’s crisis. But the episode raises the broader question of whether the simplest solution to speculative attacks is for countries to “put some sand in the wheels” of interna- tional capital markets, to borrow Tobin’s (1978) famous analogy.

The crux of Krugman’s (1998) argument is that emergency controls on outflows might be the least bad choice for a country whose currency and debt is under severe attack from domestic and foreign speculators. A nation that attempts to protect its currency through sharp rises in interest rates, a remedy the IMF has often prescribed in the past, puts tremendous pressure on its economy and espe- cially on its banking system. Allowing a sharp depreciation of the exchange rate, as advocated by Sachs (1998), also wreaks havoc with the domestic banking system. Developing country banks often have heavy offshore borrowing in foreign cur- rency, but loans in domestic currency, which means that depreciation renders them insolvent. So, Krugman argues, perhaps capital controls are sometimes the best alternative, however abhorrent they are to economists.

The first reaction of most academic economists is that policies that prevent international investors from repatriating their funds can’t possibly be a good idea for any country that desires future investment from abroad. Countries with a track

8 See also Krugman’s column in Fortune, September 7, 1998.

34 Journal of Economic Perspectives

record of imposing capital exit controls will surely drop to the bottom of most international investment “buy” lists.

This initial reaction may well be the right one, but economists should also recognize that the issues are quite subtle and complex. I have already argued that, in theory, a temporary payments standstill may sometimes be the best response to a run, absent a lender of last resort. Moreover, in multi-period models of interna- tional borrowing, it is by no means the case that an efficient contract always calls for full debt repayment in every state of nature. Several authors have developed models in which the implicit contract between country debtors and international creditors calls for only partial repayment when growth is unexpectedly low (Grossman and Van Huyck, 1988; Bulow and Rogoff, 1989a; Obstfeld and Rogoff, 1996, ch. 6).

But there are also a variety of powerful reasons why the international commu- nity should not be too happy about seeing pervasive use of restrictions on capital outflows. Controls may scare off investors, who find them arbitrary and unpredict- able, far more than a bankruptcy court or a crisis manager. Controls are an open invitation to corruption, as investors with huge sums of money at stake will be tempted to try to bribe local officials. Thus, although it is a false reading of the theory literature to conclude that temporary outflow controls are absolutely never an optimal response to a run, the problems may well outweigh the benefits.

Controls on Capital Inflows Another less radical school of thought holds that the international community,

as embodied in the actions of the G-7 and IMF, should allow and even encourage developing countries to place taxes on short-term capital inflows; Eichengreen (1999) is one recent advocate of this approach. Chile, which is generally held as the most successful economy in Latin America over the past two decades, is the poster country for capital inflow taxes. From May 1992 to May 1998, the Chileans required that all nonequity foreign capital inflows be accompanied by a noninterest bearing one-year deposit equal to 30 percent of the initial value of the investment. Since the restricted account must be held for only a year, the effective tax rate imposed by this restriction is larger for a short-term investment and smaller for a long-term investment. The rationale for the Chilean tax is that it discourages locals from relying too heavily on short-term borrowing, and thereby mitigates the problem of maturity mismatch—that is, heavy short-term borrowing and long-term lending— that seems to underlie many episodes of speculative attack. Because the tax is completely transparent, it does not suffer from the arbitrariness that many investors associate with capital outflow taxes. Admittedly, Chilean-style controls must be very comprehensive to be effective. For example, domestic banks must be prevented from writing offshore derivative swap contracts with foreign holders of long-term Chilean debt. By including suitable margin and call conditions, such contracts can effectively make a Chilean bank the true holder of the long-term income stream, and the foreign bank the holder of a short-term loan.

There are various concerns with trying to apply the Chilean lesson too broadly.

Kenneth Rogoff 35

Chile has been relatively successful in avoiding speculative pressures, but as Edwards (1998, this issue) argues, this probably has had less to do with its system of capital controls than with a variety of other favorable conditions, especially the country’s relatively well-developed system of prudential banking regulation. It may be the case that for Chile, lenders were willing to advance long-term loans at rates only slightly higher than for short-term loans. Many developing countries, however, may find that foreign investors demand a much higher premium. In this case, the borrower will have to choose between accepting short-term loans or not being able to borrow from abroad at all. Indeed, presently even Chile is not employing “Chile-style” controls on capital inflows: by September 1998, the tax had been reduced to zero in response to a persistent current account deficit. (When a country needs to borrow to pay for current consumption, it is less well-positioned to impose taxes on foreign investors.)

In sum, capital inflows taxes may work for a small select number of countries, but most developing countries will find them a quick route to a sharp reduction in lending from international capital markets. Still, if short-term capital inflow taxes can be enforced cleanly and transparently, a big qualification in countries where official corruption is a major problem, it is hard to see why the IMF should take a strident position against them.