International Institutions for Reducing Global Financial Instability

Kenneth Rogoff

I t is hard to open a business newspaper or magazine these days withoutconfronting another sweeping proposal to reform the “international financialarchitecture.” George Soros (1998) has called for the formation of an inter- national deposit insurance corporation, while Jeffrey Sachs (1995) advocates the formation of international bankruptcy court. Paul Krugman (1998a, b) suggests that economists need to rethink their traditional antipathy towards controls on capital controls outflows, whereas Barry Eichengreen (1999) is among many who advocate Chilean-style controls on capital inflows. Henry Kaufman (1998) recom- mends creating a single global super-regulator of financial markets and institutions, and Jeffrey Garten (1998) proposes a world central bank with responsibility for overseeing a new global currency. Stanley Fischer (this issue) makes the case that, with a range of improvements in the system, a multilateral lender can effectively perform the main functions of a lender of last resort, even without being able to issue currency. Many of these ideas are not new, but they are being vented more forcefully, and taken more seriously, than at any time since Harry Dexter White and John Maynard Keynes masterminded the creation of the World Bank and the International Monetary Fund at the Bretton Woods conference at the end of World War II.

Is there a “crisis in global capitalism”? Is the current system actually in desper- ate need of repair? In this paper, I will provide an overview of some of the main problems and critically assess some illustrative alternative plans for dealing with them. The first part of this paper gives an overview of the current system, and a brief discussion of some of the conceptual issues. I then proceed to consider a range of plans that would purportedly improve things. My focus is more on ambitious grand

y Kenneth Rogoff is Professor of Economics, Harvard University, Cambridge, Massachusetts.

Journal of Economic Perspectives—Volume 13, Number 4—Fall 1999—Pages 21–42

schemes than on small marginal changes. Even though such schemes tend to be impractical, especially in the absence of a genuine world government, they throw the problems facing global leaders into sharp relief. I try throughout to highlight important research questions and show how they relate to evaluating the various plans. The third section of the paper reviews reforms that developing countries can implement unilaterally to reduce the costs of capital flow volatility. The final section highlights the importance of correcting the bias towards debt financing and bank intermediation in sovereign lending.

Problems with the Status Quo

Before turning to proposals for radical change of the international financial system, it is important to give a brief critical assessment of the main issues and motivations for change.

Alternative Perspectives on the Global Financial System Whether one views technology-driven innovation in the global financial system

as an engine of growth or as an agent of destruction depends on where you sit. In the United States, where financial markets are the deepest and most sophisticated in the world, their benefits seem obvious. Despite having one of the lowest savings rates in the industrialized world, the U.S. economy has enjoyed a remarkable period of sustained growth over the past eight years. The efficacy with which financial markets have helped lever a small pool of savings into a large effective increase in capital is remarkable, even when one takes into account the help of foreign capital inflows. Hyper-efficient U.S. financial markets can also be credited with helping to fuel the extensive corporate restructuring of the 1980s, thereby laying the foundations for the sustained rapid growth of the 1990s. Europe, with its introduction of the euro and its efforts at stimulating innovation and competition in financial services, clearly recognizes the importance of deep, sophisticated asset markets. True, the stunning volatility of stock and exchange rate markets is of genuine concern to policymakers in industrialized countries. The August 1998 collapse of Long-Term Capital Management underscored how a single relatively small hedge fund could threaten to bring down a much wider circle of financial institutions. But in the United States, those voices seeking to quash capital markets are typically drowned out by those who argue that a better solution is for such markets to become broader and more deeply entrenched.

Matters look very different to citizens of the developing world, many of whom rue the day their governments started taking down barriers to international capital mobility. Starting with Mexico in 1994, and including a score of countries in Asia in 1997, one high-growth achiever after another has been leveled by sudden withdrawals of short-term capital. (This is not to say that low-growth achievers have been spared, but capital withdrawals from countries such as Russia are less difficult

22 Journal of Economic Perspectives

to explain.) Countries which had become accustomed to seeing GNP double every 10 to 15 years suddenly saw their currencies and stock markets collapse and their economies go into deep recession. The 1990s financial crises have brought a sharp contraction of lending to the developing world, and there is serious concern that the fallout will continue to inhibit international capital markets for some time to come. The exact timing and nature of speculative attacks on emerging market economies is a topic of great debate, as we shall see. But in the majority of cases, there is little question that the attacks were exacerbated by the way that many developing country governments chose to open their capital markets radically to the rest of the world during the early 1990s. Critics of “excessive” capital market liberalization, whose numbers include such influential economists as Jagdish Bhag- wati (1998) and Dani Rodrik (1997), can point to countries such as China and India whose capital controls, however repressive, did seem to make them relatively resistant to the Asian flu. Bhagwati, in particular, has argued that the benefits of a high level of international capital market integration are grossly overrated, and that the parallels between the gains to trade in capital, and the gains to trade in goods, are quite thin. He criticizes the U.S. Treasury and the International Monetary Fund (IMF) for rushing too many countries into bringing down their controls on international capital mobility, without sufficient consideration of whether domestic regulation was adequate to deal with the changes that rapidly ensued.

Are the Benefits to International Capital Market Integration Overrated? Perhaps a little, but they are important. From a theoretical perspective, there

are strong analogies between gains from intertemporal trade in goods, and standard intratemporal trade (for example, Svensson, 1988; Obstfeld and Rogoff, 1996, ch. 5). In theory, huge long-run efficiency gains can be reaped by allowing global investment to flow towards countries with low capital-labor ratios and high rates of return to capital although, as Ventura (1997) points out, trade in goods of differing capital intensity can achieve part of this gain. Global equity markets allow a small country that produces a relatively narrow range of goods to diversify its very risky income portfolio. In the case of foreign direct investment, benefits can also arise from an accelerated transfer of technology.

If there is a debate in the theoretical academic literature on the importance of gains from international capital market integration, it has mainly to do with whether, given trade in bonds, there is a substantial further gain to introducing complete equity markets. However, researchers have now come to believe that the marginal gains from trade in equity can be very large, once one takes into account the ability to diversify production risk, which encourages small countries to specialize, and more generally to shift production towards higher-risk, higher-return projects (Obstfeld, 1994; Acemoglu and Zilibotti, 1997; Martin and Rey, 1998). Later, in the final section, I will argue that there are other political economy benefits to redirecting capital flows towards equity that are not captured in these models.

Kenneth Rogoff 23

An Unreconstructed Real Business Cycle Interpretation of the Asian Flu Rather than blame international capital markets for the severe recessions in

Asia and elsewhere, a modern real business cycle economist (or an old-fashioned Schumpeterian) might just say “welcome to free market capitalism.” How surprised should one be that economies racing along at 5-7 percent annual growth rates for more than two decades should occasionally experience a significant downturn, or even a severe one? Might not the sudden reversal of capital flows simply reflect underlying real shocks to, say, patterns of global technology progress? For example, if the U.S. economy experiences an extraordinary period of growth, is it surprising that this leads to a temporary redirection of investment away from middle-income countries?1 Besides, Japan had been mired in recession for several years prior to 1997, placing a major drag on the region.

This “unreconstructed real business cycle interpretation” of the developing country debt crisis clearly fails to capture the whole picture. A great deal of evidence suggests that banking system collapses can play an important role in propagating and amplifying recessions, with Japan’s recession of the 1990s being a prime case in point (see also Mishkin’s article in this issue). Relatedly, many of the plans below aim to address either developing country bank runs or runs on government debt. Imperfections in international capital markets, resulting espe- cially from difficulties in enforcing contracts across borders, can sometimes lead to large misallocations in global savings.

But even if the real business cycle interpretation is incomplete, it probably does provide an important part of the picture, a part that is all too often forgotten in policy discussions which tend to blame emerging economy recessions entirely on speculators. One should also bear in mind that the speculative attacks of the 1990s, even if they did cause or exacerbate recessions, may someday be viewed as mere hiccups, a small price to pay if capital market integration puts countries on a faster trajectory towards integration with the industrialized world.

Multiple Equilibria as a Rationale for an International Lender of Last Resort Many have argued that a strong parallel exists between sudden massive with-

drawals of capital from developing countries and bank runs (for example, Cole and Kehoe, 1998; Chang and Velasco, 1998). Banks are vulnerable to runs because they issue highly liquid short-term liabilities (like checking accounts) which their de- positors can, if they choose, all withdraw simultaneously. At the same time, many of their assets are held in the form of highly illiquid long-term loans (for example, to a local construction company) that can only be liquidated prematurely at great expense. One reason why the secondary market might be illiquid is that evaluating loans to local firms requires specialized expertise that banks build up only over a

1 Bulow and Rogoff (1990) argue that the combination of adverse terms of trade shocks, rises in global real interest rates, and recessions in the industrialized world played a much larger role in the poor growth performance of Latin America during the first half of the 1980s than any debt overhang effects.

24 Journal of Economic Perspectives

long period. Given the illiquidity of its assets, a bank may find itself in trouble if all its depositors suddenly decide to withdraw their money, even if the bank is fully solvent in an actuarial sense. Thus, as illustrated in the classic models of Bryant (1980) and Diamond and Dybvig (1983), bank panics can be self-fulfilling.

The parallel with country debt runs is two-fold. First, many country debt runs are intimately linked to their banking sectors, as Chang and Velasco (1998) emphasize. In many developing economies, banks are implicitly insured by the government. A country-wide run on local banks will thus translate into a huge increase in government liabilities, and this in turn can lead to a flight from government securities. But the analogy runs much deeper. Many high-yield projects in developing countries, like building a factory or a new highway, are highly illiquid and have only long-term payoff potential. At the same time, a considerable portion of lending to developing countries is in the form of relatively short-term debt. If creditors suddenly become unwilling to roll over short-term loans as they fall due, a country may find itself in a financial squeeze even if, absent a run, it would have had no problems servicing its debts. Devotees of the “multiple equilibrium” view believe that this is precisely what happened in the case of, say, Mexico in 1994, or Korea in 1997. For example, creditor panic at a relatively small devaluation of the peso in December 1994 suddenly made it impossible for Mexico to roll over its short-term debt, quickly precipitating a crisis. Instead of humming along in a “good” growth equilibrium as Mexico seemed to be doing prior to the crisis, it suddenly was bounced into a “bad” recessionary equilibrium. There was no adverse technology shock a la modern real business cycle theory—just good old-fashioned creditor panic.

If the multiple equilibrium view is correct (a conclusion the reader should not rush to accept), what is the solution? Bryant (1980) and Diamond and Dybvig (1983) show that in a domestic banking context, the problem can be eliminated, at virtually no cost, by having the government guarantee bank deposits—that is, serve as a lender of last resort. If depositors know they will always be paid even if their bank fails, bank runs will not be a problem and, in fact, the government will never (or at least seldom) have to honor its pledge. Thus, their models provide a rationale for the Federal Deposit Insurance Corporation in the United States, and the broader set of implicit guarantees that institution represents. Many of the proposals for reform of the international monetary system draw heavily on this analogy—if a lender of last resort can stop bank runs in a domestic context, why can’t an analogous institution be created to stop country debt runs? What could be simpler?

Chinks in the Theoretical Case for a Domestic Lender of Last Resort The case for having a domestic lender of last resort is far less coherent than

many writers in the “save the global financial system” literature seem to realize. The Bryant-Diamond-Dybvig rationale for a lender of last resort relies on a number of assumptions that have been challenged in the literature (for a recent review, see Freixas and Rochet, 1997). The most obvious omission from the story we have told

International Institutions for Reducing Global Financial Instability 25

is that it neglects moral hazard: government deposit guarantees allow a bank to hold a risky portfolio while still borrowing at a risk-free interest rate. In principle, moral hazard problems can be mitigated through bank supervision, capital require- ments and other devices, though in many countries these checks and balances are patently inadequate. As Caprio and Honohan (this issue) discuss, in 59 worldwide banking crashes in the 20 years prior to the Asian crises, the average cost of government bailouts was over 9 percent of GDP in developing countries and 4 percent of GDP in industrialized countries—hardly evidence in favor of the view that creating a lender of last resort is a free lunch. But even if the moral hazard problem could be substantially ameliorated, the case for having a lender of last resort is still somewhat shaky.

As Diamond and Dybvig (1983) themselves show, allowing banks to suspend withdrawals of deposits temporarily is a fully efficient mechanism for eliminating the multiple equilibrium problem, provided a bank knows when it is seeing the start of a run and not just an unusual surge in withdrawals. Wallace (1988) argues that the informational assumptions can be relaxed, once one allows for sophisticated partial suspension schemes. These involve having the bank start to place increasingly tight percentage caps on withdrawals during periods of abnormally high demand. Wallace shows that deposit insurance cannot improve on an optimal suspension policy, unless the lender of last resort has superior information. In a more general setting, one can imagine many other private sector responses to dealing with bank runs, such as the development of interbank credit agreements to deal with panics. Indeed, many of these have been seen in practice in earlier periods.

Therefore, one should not conclude that theory shows decisively that a lender of last resort is needed. Admittedly, one has the nagging feeling that the govern- ment is better positioned to make credible guarantees concerning its policy for dealing with bank runs than can any private sector agent or network. But at the same time, it is important to be aware that theory does not provide an airtight case for this assertion, despite many efforts to do so.

Finally, even the notion that country debt and currency runs might represent realizations of multiple equilibria can be challenged. In a closely related context, Morris and Shin (1998) argue that introducing a small amount of private informa- tion can eliminate the problem of multiple equilibria in models of currency attacks. In this class of models, government policies that affect transparency and the dissemination of information can be more useful than introducing insurance.

The G-7 as Incumbent Global Lender of Last Resort It would be an overstatement to say that the world financial system has been living

without any lender of last resort. One exists, just not an explicit one. Over the course of the 1990s, the so-called G-7 group of industrialized countries (United States, Japan, Germany, France, United Kingdom, Italy and Canada), acting in concert with the International Monetary Fund, the World Bank and other OECD countries, have found themselves cast in this role. In early 1995, they awarded Mexico an unprecedented

26 Journal of Economic Perspectives