Capital Controls: Capital Idea or Capital Folly?

Barry Eichengreen

November 1998

Since the Asian crisis erupted, undermining the legitimacy of the Suharto Government, Indonesia has been wracked by unrest. In some cities, the authorities responded with an emergency curfew, prohibiting people from going out at night. Some observers have gone so far as to suggest that Indonesia would be a safer place if those measures were maintained indefinitely.

In neighboring Malaysia, meanwhile, Prime Minister Mahathir Mohamed has imposed a curfew on capital. Once the home of one of the world’s most highly capitalized stock markets and open financial markets, Malaysia now controls all purchases and sales of its currency, the ringget, for purposes related to international financial transactions. Not just banks and stock brokers are affected: citizens are prohibited from taking as little as $100 out of the country, and the law is enforced by random searches at the airport.

Mr. Mahathir argues that this infringement of the civil liberties of Malaysia’s citizens, like a dusk-to-dawn curfew, is needed to protect its economy and society against marauding hedge funds mugging innocent bystanders. It would be madness, in his view, to leave currency speculators unrestrained, unnecessarily jeopardizing the health and well-being of the Malaysianeconomy.

This once-radical view has found support in some surprising quarters. Capital controls, long regarded as anathema by respectable economists, are back in fashion. Late this summer, MIT’s Paul Krugman authored a widely-cited article arguing that emergency conditions warranted emergency measures. Warning that the crisis countries of Asia were experiencing a full-scale meltdown, he urged them to consider using controls as shelter behind which to reflate their collapsing economies.

His statement was viewed, rightly or wrongly, as giving Mr. Mahathir intellectual cover for his radical initiative. Harvard’s Dani Rodrik, meanwhile, issued a blanket indictment of capital market liberalization. Rodrik argued that there is in fact no evidence that countries that free capital flows grow faster, while it is self-evident that international financial liberalization exposes them to the danger of debilitating crises.

This apostasy flies in the face of all that is sacred to economists and has predictably provoked harsh criticism of those voicing such unconventional views. The normal presumption in economics is that markets know better than governments and that, left to their own devices, they allocate resources reasonably well. Yet the suspicion remains that there is something different about international financial markets.

The founding fathers of the Bretton Woods System, Harry Dexter White and John Maynard Keynes, certainly thought so: the Bretton Woods Agreement negotiated under their guidance, while encouraging economic liberalization generally, authorized --indeed, encouraged -- countries to retain restrictions on international financial transactions. Only in recent years, responding to pressure from the IMF and the United States, have governments, first in Europe and Japan and now in various emerging markets, finally abandoned capital controls. The Asian crisis now suggests, or so it would appear, that this was a serious mistake.

Not that it is clear why international financial transactions should be treated differently from other transactions, or why the normal presumption that markets know better than bureaucrats is invalid in this case. Is it that the transactions in question are financial, or that they are international? Are capital controls justified as emergency measures in a period of unprecedented crisis, or should they be retained as permanent protection against unreliable financial markets? Should they be considered by all financially-exposed economies or only emerging markets? Unfortunately, the discussion to date has confused these issues more than it has illuminated them.

The Fundamental Case for Financial Liberalization

The most basic insight of modern welfare economics is that self-interested economic actions maximize the collective interest -- or, to put the point more simply, that markets allocate resources in socially-desirable ways. While they may not work perfectly, all the evidence, be it from import substitution in Latin America in the 1950s and 1960s or from central planning in the Soviet Union, is that they produce better outcomes than heavy-handed bureaucratic control.

There is no obvious reason why this presumption should apply less to financial than other markets. Indeed, studies by the World Bank have consistently shown that countries with more developed financial markets grow faster. Experience with policies of financial repression in developing countries and with state monopolies over financial transactions in the Eastern Bloc clearly shows that stifling financial markets can jeopardize growth.

If domestic financial markets have clear benefits, it is not clear a priori why the benefits of international financial markets should be less. International financial transactions transfer resources from high-saving to low-saving countries. They allow economies experiencing business-cycle disturbances to smooth the time profiles of consumption and investment. They allow firms and households to diversify away country-specific risks.

The presumption in favor of markets being so strong, any counter-argument had better be based on incontrovertible evidence. Rodrik’s evidence is widely cited: using data for a cross-section of countries, he finds no association between capital flows and economic growth, which, he argues, seals the case against capital account liberalization.

This is what Jeffrey Frankel refers to as fail-safe econometrics: “The secret of empirical work is to define your hypothesis so that failure to find significant results can be interpreted as support.” Statisticians can fail to find a relationship between capital account liberalization and growth not because none exists but because they have inadvertently omitted from their analysis other variables that are negatively associated with growth but positively associated with the decision to open the capital account. It is plausible that countries that decide to keep their capital accounts open and closed differ from one another in other ways, including ways for which the statistician finds difficult to control.

In a sense, those who argue that today’s developing countries should resist capital account liberalization are adopting something of a double standard. All of today’s advanced industrial countries have opened their capital accounts. All of them have rendered their currencies convertible for capital account transactions. Doing so is the logical culmination of the process of developing a deep, mature and efficient domestic financial system.

In a fundamental sense, domestic and international financial liberalization go together, since it is hard to liberalize domestic financial transactions and at the same time keep a lid on cross-border transactions.

Some would argue that capital account liberalization is also a concomitant of political liberalization: capital controls necessarily infringe on the economic freedom of residents and are not a policy of a country in which most readers of this paper would themselves prefer to live.

There is a valid argument, as we shall see below, that developing countries should control capital flows while they build deep and diversified financial systems, upgrade prudential supervision, and strengthen their monetary and fiscal institutions — in other words, for using them as transitional measures — but not that such countries should permanently pursue different policies toward the capital account than those which became high-income countries before them.

The transition may be a long one, but the comment made by Marx in another context, that the more advanced economy shows the less advanced economy a picture of the latter’s future, is apposite here.

Capital Controls as Prudential Measures

Markets may be the best mechanism we have for allocating financial resources, but history has also shown that they can be dangerously unstable. Like a trapeze artist, the financial system can perform miraculous tricks but experience a bone-shattering fall if allowed to perform without a net.

Banks in particular share the trapeze artist’s vulnerability. Their investments are less liquid than their deposits; this is what we mean when we say that banks provide “liquidity-transformation services.” They operate in an imperfect information environment; one of their basic functions is to develop long-term relationships with their clients as a way of acquiring proprietary information about their borrowers’ credit worthiness. But the fact that other financial-market participants will not have equally good information about those customers means that banks can raise funds in a crisis only by disposing of assets at fire-sale prices and doing further damage to their balance sheets. Banks do extensive business with one another; hence, problems in one create problems in others. For all these reasons, a sudden loss of depositor confidence can produce a system-wide panic that brings the entire banking system to its knees.

Securities markets share many of the same vulnerabilities. Investors are prone to quick collective reactions. Being imperfectly informed about market conditions, they tend to infer information about the fundamental value of their investments from one another’s actions. Economists have coined the elegant term “information cascades” to denote this phenomenon, which in practice simply means that investors move in a herd, stampeding in and out of markets.

Moreover, investors take positions on credit, so that when the market moves against them, they are required to put up additional collateral. They can thus be forced to sell into a falling market, amplifying asset price volatility. And a large price fall can bring bank and nonbank intermediaries down with it, disrupting the supply of credit to the economy as a whole. The case of Long Term Capital Management reminds us that this scenario is no mere hypothetical.

These are all reasons why governments limit the difficulty of the tricks that banks and other financial-market participants are allowed to attempt. To limit banks’ exposure to market and credit risk, they impose ceilings on concentrated investments and positions in foreign exchange. They limit the amount of margin money, or leverage, that equity markets participants are allowed to use. And they do not allow banks to perform intermediation services without a net. The financial safety net, which takes the form of deposit insurance and the existence of lender of last resort, is designed to catch financial-market participants when they fall.

This is where an open and unregulated capital account poses special risks. Banks enjoying government guarantees and seeking to lever up their bets can do so more readily when the capital account is opened. If they borrow in foreign currency, they strip the authorities of their ability to act as lenders of last resort: a central bank can’t print foreign currency, and its capacity to provide commercial banks the foreign exchange they need to make good on their foreign obligations is limited to its stock of international reserves. Even if the liabilities of the banks are denominated in domestic currency, a central bank trying to peg the exchange rate will find itself between a rock and a hard place. It will have to choose between draining liquidity from the markets to defend the exchange rate, or injecting liquidity to defend the banks. It will not be possible to do both.

These are arguments for using capital controls to backstop other forms of prudential regulation — as reinforcement for other, more conventional measures to limit systemic risk and to prevent banks and other intermediaries from taking on additional risk in response to the provision of the financial safety net.

In a market economy, prudent risk management is first and foremost the responsibility of bank owners and managers themselves; since they are the ones making the investment decisions, they should bear the consequences. This is why banks are required to hold capital and issue subordinated debt -- so that their owners and important creditors have something to lose. The second line of defense against excessive risk taking is the regulators, who monitor and supervise the banks and need to take prompt corrective action when they see evidence of fraud, incompetence, or gambling for redemption. But where risk-management practices are underdeveloped and the regulators lack administrative capacity and insulation from political pressures, it may be necessary to build a third line of defense -- to limit excessive risk taking which threatens systemic stability by limiting the ability of the banks to borrow abroad. And where limits on bank borrowing can be circumvented by having corporations do the borrowing and lend the proceeds to the banks, this will mean controlling or taxing, a la Chile, all capital inflows, whether the borrower of record is a bank or someone else.

These arguments do not justify any and all regulations that governments might be tempted to impose to prevent their citizens from borrowing abroad. Capital controls can be justified on prudential grounds only if they do not arbitrarily discriminate in favor of some banks and residents over others. They are justifiable only where financial markets are thin, the private sector’s risk-management practices are underdeveloped, and the regulators’ capacity to supervise the financial sector is limited — in other words, where the conventional defenses against systemic risk are not enough. In practice, these last three preconditions, and therefore the argument for capital-inflow taxes or controls, apply to the vast majority of developing countries. For emerging markets, an open capital account should be the exception, not the rule.

Eventually, financial markets will deepen, bankers will acquire more sophisticated risk-management skills, and regulators will gain experience, competence and independence. At that point, restrictions on foreign borrowing should be removed, and the economy can graduate to the club of high-income countries with financial systems fully open to international transactions. But here, as in other forms of financial regulation, it is smart to err in the direction of caution -- to be absolutely sure that the necessary preconditions are in place before opening the capital account. After Mexico in 1994 and Asia in 1997, do we really need a third reminder of the dangers of premature and precipitous financial liberalization?

Capital Controls as Emergency Measures

One country after another, from Thailand to Indonesia, to Korea and now Brazil, has been forced to respond to the crisis in emerging markets and the resulting recessionary pressures by cutting its budget deficit, not increasing it as the textbook Keynesian advice would suggest. The single greatest discovery of the Keynesian revolution, namely the importance of fiscal stabilizers, has thus been thrown out the window.

Some would say this simply reflects bad advice by the IMF, which required budget cuts of the Asian crisis countries as a condition for the disbursal of official funds, and which is now demanding the same of Brazil despite forecasts of recession there. In fact, the Fund is merely mirroring market sentiment. Were a country like Brazil to respond to slower economic growth by cutting taxes and increasing public spending, investors would flee, the currency would crash, and the resulting investment collapse and financial distress would only make the recession worse.

Thus, market discipline is perverse. As Krugman puts it, “Brazil, we are informed, must suffer a recession because of its unresolved budget deficit. Huh? Since when does a budget deficit require a recession (which itself will, of course, make the deficit that much harder to bring down).” This is at least part of the rationale for the capital controls imposed by Mr. Mahathir — to provide the leeway to implement a more expansionary fiscal policy and offset an impending recession. It is the realization that has led “otherwise respectable economists” to suggest the use of capital controls to stem capital flight and thereby preserve governments’ freedom to pursue counter cyclical fiscal policies.

Controls have costs — they require a burdensome administrative bureaucracy, reduce the pressure for policy reform, and interrupt access to foreign sources of investment finance — but their benefits may still dominate if they allow the stabilizing use of macroeconomic policy instruments to be regained.

Whether this is a sensible argument hinges on which of two models of market discipline one believes. If investors are irrational and inclined to panic when the government activates its macroeconomic stabilizers, then it can be sensible for countries to use controls to protect themselves from such irrationality. If, on the other hand, investors respond negatively because they correctly anticipate that it is governments themselves that are prone to respond perversely to the crisis, then the solution is not to use controls to relax market discipline but for the government to clean up its act.

The argument goes like this. Some governments lack fiscal discipline, and they are perennially battling the consequences. Like an overweight man, they are continually trying to teach themselves to stay away from the refrigerator. If the fat man says “I’ve had a lousy day; I’m going make myself feel better by having a piece of cake,” his friends are likely to revise downward their estimates of the likelihood that he will stick to his diet.

Governments with a history of fiscal laxity that have a lousy macroeconomic day and respond by increasing their budget deficits similarly run the risk of being re-evaluated in this way -- of being seen as having reverted to their bad old habits of running budget deficits and living beyond their means. And if investors rationally expect budget deficits to be monetized, then deficits today imply inflation tomorrow, encouraging the rational investor to take the first opportunity to get his money out of the country.

This explains the supposedly paradoxical fact that deficit spending in the United States strengthens the currency while deficit spending in Brazil weakens it. In the U.S. case, no one expects the Fed to monetize the deficit; hence, additional government spending pushes up demand, pushes up the real interest rate, and pushes up the real exchange rate. In the Brazilian case, however, monetization is a real possibility (pun intended), implying more inflation and ultimately the need to devalue the currency.

It is also why the other textbook advice for responding to a recession--devaluing the currency in order to switch spending toward domestic goods -- can have such catastrophic effects in emerging markets. Countries weaning themselves from inflation often do so by pegging the exchange rate, which ties the hands of the central bank and signals the government that the inflation tax will no longer be available.

The currency peg is thus the lock on the refrigerator. Countries that devalue are thus seen as having removed the lock from the refrigerator and relapsing to the bad old days of inflationary excess, which leads investors to flee.

The first-best solution in this case is not to impose capital controls but to eliminate the problems leading to the excesses in monetary and fiscal policies in the first place. The most convincing way of signaling that not just current policies but also future policies will be sound and stable is to reform the economic and political arrangements by which they are made.

A large literature now shows quite convincingly that better policy-making institutions produce better outcomes. For monetary policy the point is well known: more independent central banks are better able to resist political pressures to monetize budget deficits and generally run lower inflation rates. For fiscal policy, there are parallel arguments for creating an independent national fiscal council constitutionally empowered to set a ceiling for each year’s budget deficit, along with automatic, legally-mandated procedures for what will be done if deficit spending threatens to broach that limit.

Less ambitiously, fiscal reforms which vest more agenda-setting power in the hands of the prime minister or finance minister, thereby reining in the common-pool problem that arises in the presence of autonomous spending ministries (none of which has an incentive to fully take into account the impact of its additional spending on the deficit as a whole), have been shown to be associated with smaller deficits and debts. Similarly, measures that enhance the transparency of budgeting make it easier for voters to detect politicians who place self-serving goals above the national interest and hence are likely to produce better fiscal outcomes.

With these fundamental institutional reforms in place, markets will not conclude that monetary expansion tomorrow. The freedom to use fiscal and monetary policies counter cyclically will be regained, and capital mobility will no longer be a threat.

Redesigning monetary and fiscal institutions to deliver better outcomes needn’t take forever. A simple law — or better, a constitutional amendment — establishing the independence of the central bank, appointing its governors to long terms in office, and reining in the fiscal autonomy of the spending ministries and lower levels of government can be adopted in short order. Time may be required, however, for the new institutions to establish a track record of delivering sound and stable macroeconomic policies. Still, Argentina’s example suggests that this can be accomplished in a matter of years. By adopting such reforms, Argentina moved from being an inflationary basket case and a capital-market pariah to one of the few emerging market economies able to float bonds on international capital markets and avoid excessive macroeconomic tightening in the global financial turbulence of the second half of 1998.

There may still be a role for capital controls to loosen the constraints on the use of counter cyclical monetary and fiscal policies while a country attempts to get from here to there. But it is reaching the destination that is key.

Conclusion

Developing countries have special financial problems. Their monetary and fiscal institutions lack credibility. Their regulators lack administrative capacity. Their financial markets are shallow. They cannot borrow abroad in domestic currency. However much one tries to assume away these problems, the fact of the matter is that these are the defining features of an underdeveloped economy. And so long as they are present, there are arguments that special deficits today necessarily mean deficits tomorrow, or that monetary expansion today means measures, including capital controls, to limit risks to the financial system and to free up the use monetary and fiscal policies in a slump.

With time, developing countries will develop. Their financial markets will deepen; their macroeconomic and regulatory institutions will grow more robust. With these and other institutional preconditions in place, they will graduate to the club of high-income countries. This will have a important social and economic benefits, not the least of which is that remaining restrictions on international financial transactions can come off. The most critical point of all is therefore that any recourse to capital controls in the meantime should not be taken as an excuse to slow down the fundamental processes of institutional development and policy reform.