In the Wake of the Crisis: Leading Economists Reassess Economic Policy

23. Liquidity and the International Monetary System

International Monetary Fund
Published Date:
March 2012
  • ShareShare
Show Summary Details
Maurice Obstfeld

In this chapter, I touch on points that are inspired by the crisis of 2008 to 2009, which displayed a number of stresses in the international system. Not surprisingly, these stresses are related to the problems that motivated the founding of the International Monetary Fund nearly seven decades ago.

Global liquidity needs, exchange rates, and external imbalances were fundamental problems in the interwar period and before, and the original design of the IMF devised ways to address these coordination problems within a framework that was appropriate to the economic and financial conditions of the time. But conditions have changed dramatically since the Bretton Woods era that ended in 1973. In fact, even over the course of the Bretton Woods era, the international economy evolved dramatically in ways that called into question the relevance of the settlement achieved at Bretton Woods in the 1940s. Here I focus on only one of many important issues—the liquidity issue.

One of the biggest changes to have occurred since World War II is financial globalization, the extent of which is unparalleled in history and far beyond what we had in the 1970s when the Bretton Woods system broke down. There are benefits to financial globalization but also immense risks, as we have seen recently. One of the key indicators of these risks is the very high level of gross external asset and liability positions in the world economy. These have grown explosively in the last couple of decades, and they bring currency mismatches and financial counterparty risks. If these risks are socialized, as they have been broadly across the world, they can become sovereign debt risks. Think of the case of Ireland. The globalization of capital markets has facilitated larger current-account imbalances, and those can also carry risks. The cases of Greece and arguably of the United States, too, illustrate some of those risks.

Consider the graphs of these gross positions in figure 23.1, showing external assets plus liabilities as a percentage of gross domestic product, from the updated Lane and Milesi-Ferretti (2007) data. The numbers for Switzerland, for example, approach fourteen times Swiss GDP toward the end of the data sample in 2007. We know the problems that the Swiss banks experienced. The major financial centers are bigger than other countries in this regard; smaller countries, like Ireland, can also have big ratios. Given these huge numbers, the implications of even smallish percentage losses, when the tab is picked up by governments, are frightening from a fiscal standpoint. Ireland well illustrates this problem.

Figure 23.1Gross external positions as a multiple of gross domestic product, 1970 to 2007.

Source: World Economic Outlook (2010).

In this context, the official institutional framework for providing international lending support has come to the fore. To safeguard financial stability, the crisis showed that we might need lender-of-last-resort support in multiple currencies, primarily the dollar but also other widely used international currencies.

An example of this need originated in the behavior of many European banks in the run-up to the crisis. They piled into the market for dollar-denominated asset-backed securities issued in the United States, funding their acquisitions with short-term wholesale dollar borrowing. You might have thought that their foreign-currency positions were hedged and that there was no mismatch, so no problem. But in fact, the liquidity and maturity transformation involved in these positions turned out to be very worrisome in the crisis, given the credit-market stresses that developed. Many banks in Europe found themselves unable to refinance their short-term dollar liabilities when credit markets and foreign-currency swap markets broke down. To make short-term dollar finance available to these banks, the Federal Reserve stepped in with dollar swap lines to foreign central banks. Other central banks established swap lines in their currencies.

Although the history of central-bank cooperation goes back at least to the early nineteenth century, these swap facilities were very different from most that we have seen before. In the 1960s, for example, the U.S. Treasury pioneered swap lines meant for balance-of-payments support, a fairly elaborate and, for a time, effective setup, but in the crisis we saw swap lines meant to channel true lender-of-last-resort assistance directly to beleaguered financial institutions.

The Federal Reserve became the dollar lender of last resort to the world, but this ad hoc role is unlikely to be preserved indefinitely because there are huge political obstacles to such an outcome. At one time, the perceived creditworthiness of the advanced markets made the lender-of-last-resort problem seem to be one of simply expanding conventional IMF lending and almost exclusively a problem of emerging-market and developing economies. But the crisis showed us that this is not so; the problem today is much broader than this.

For various reasons, depending on their individual experiences, emerging markets accumulated large reserves during the 2000s and found self-insurance to be one way of dealing with the risk of liquidity problems (figure 23.2) . Emerging-market and developing-country reserves grew rapidly and have overtaken industrial-country reserves by quite a big margin, as this figure illustrates. Some reserves were used in the crisis, but the trend of rapid accumulation has hardly been affected. Some of this accumulation represents the side effects of intervention, and some of it is purposefully precautionary. It is hard to divide observed accumulation between the two motives.

Figure 23.2Global reserve holdings.

Source: COFER database, International Monetary Fund (various years).

Note: Years marked with * indicate a change in country average.

This mechanism of self-insurance was advantageous at the individual-country level during the crisis, motivating policymakers in developing countries to continue to accumulate. But what is good for the individual is not necessarily good for the community, and there are many systemic drawbacks to self-insurance. There are also individual-country drawbacks, but the systemic drawbacks—drawbacks for the global financial system as a whole—are worrisome.

What are these systemic drawbacks? When you use reserves, you reallocate liquidity from other locations, so there is no creation of true outside liquidity in the system, as might be necessary in a crisis. Reserve accumulation can have asset-price effects—for example, when countries shift their reserve-currency choices on a large scale. And some reserves do not materially enhance financial stability. If you intervene to hold the exchange rate steady in the face of a short-term capital inflow, you increase short-term liabilities by the same amount as reserves. When the liabilities go, the reserves go. There is no increase in stability in that setting, but the volume of reserves is higher, and some of the perils of instability in world asset prices might be higher as well.

There are also neighborhood effects at work in determining reserve demand. If your neighbor holds more reserves than you, you may be viewed by markets as a more tempting target—the better place to sell assets in a crisis. For that reason, there will be an arms race—a tendency for countries to overaccumulate reserves. Finally, reserve accumulation may be deflationary if it is effected through higher current-account surpluses or policies that give rise to those.

Another systemic threat has been less recognized, and this is a modern-day version of the Triffin paradox. In the 1960s, the United States supported a growing stock of world dollar reserves on a narrow base of gold holdings. Although the United States promised to redeem those official reserves for gold at a fixed gold price, the system was unsustainable because to fulfill the underlying promise became impossibly expensive. A clear statement of Robert Triffin’s updated paradox appears in a 2011 paper by Emmanuel Fahi, Pierre-Olivier Gourinchas, and Hélène Rey.

The reason for associating this new paradox with Triffin’s name is that it likewise results from a similar inexorable dynamic. As a direct result of satisfying growing world demand for a reserve asset, an asset that is supposedly safe, the issuer eventually becomes unable to guarantee the reserve asset ‘s safety. If reserve demanders prefer safe government debt, for example, then governments have to issue more debt. The assets the governments might hold as a counterpart will inherently be more risky. These might, for example, be claims on the private sector or foreign-currency claims. So there are fiscal limits on the ability of governments to satisfy the demand for safe reserves, just as there were limits in the 1960s on the ability of the United States to satisfy the world’s demand for reserves while guaranteeing the reserves’ value in terms of gold. The internal contradiction originates in some global economic asymmetries. First of all, considering economic growth rates, if we view the developing and emerging countries as the major demanders of reserves, they are simply growing much faster than the supposedly more creditworthy advanced economies. (The stronger credit of the advanced countries, making their governments’ debts more eligible as reserve assets, is a second global asymmetry.) Lower-income countries now account for more than 50 percent of the world economy at purchasing power parity (PPP), and they are forecasted by the IMF and others to grow even bigger. If these countries keep accumulating reserves at the rate they have been, and if present growth trends continue as we expect, how will this demand for reserves possibly be satisfied?

One can get a rough idea of where the modern Triffin paradox stands today by looking at data like those in figure 23.3, which tracks reserves compared with the general government debts of the euro zone and the United States. This comparison is very crude, however, for a number of reasons. These are general and not central government debts, and certainly some euro zone central government debts (including those of Italy, Spain, Portugal, Ireland, and Greece) would hardly be viewed as safe reserve assets in the near term. Furthermore, countries may hold reserves in commercial banks. Those banks may be viewed as safe if there is an implicit bailout guarantee, but otherwise, they might be viewed as less than completely safe. (Again, the government’ s fiscal capacity is central to that assessment.) Syria and Brazil might have different views about where it is safe to hold their reserves. There are many interesting nuances here, but the basic point remains that the question of future reserve adequacy is a question of current relevance.

Figure 23.3Global nongold reserves compared with gross general government debt.

Sources: International Financial Statistics database (various years); World Economic Outlook database (various years).

The international response to the recent crisis involved an expansion of IMF resources and potentially swifter responses through the flexible and precautionary credit lines. Those are useful innovations that need to be expanded and made more attractive relative to holding reserves, which provide unconditional and immediate liquidity. A major problem, unsurprising in view of past experience, is that very few countries have chosen to pre-qualify for these new IMF facilities.

The central-bank swap lines that sprang up during the crisis could and should be codified, as I and others have suggested (for example, Obstfeld 2009); there will certainly be need for them in the future. They could be run through the IMF directly to central banks that meet accepted standards of independence and supervisory efficacy. Or they could be run through the Bank for International Settlements (BIS). The conditionality and prequalification questions are critical for avoiding moral hazard.

Lending to central banks is different from sovereign lending, which has been the IMF’s main role so far in its history. There is room for both types of crisis lending. If direct central bank loans were routed through the IMF rather than the BIS, the IMF’s articles would need amendment. But there is precedent. When special drawing rights (SDRs) were introduced, they were quite a departure from what the IMF had generally done before.

SDRs amount to a reserve pooling arrangement. An SDR is essentially a claim on other IMF members’ holdings of reserve currencies. It is hard for me to see why more explicit and expanded reserve pooling is not a more attractive option. In other words, why do we need the SDR to accomplish more effective reserve pooling? The SDR has evolved by default into a unit valued by a basket of currencies. That is not how it was originally defined, but when the par-value system disappeared in the early 1970s, there was a need to do something about the SDR’s valuation, and the basket definition emerged as a convenient way to stabilize the unit’s value against a particular reserve portfolio. The defining SDR basket started with sixteen currencies, then went to five, and then to four, which is where it stands now.

An advantage of having some sort of credit-line arrangements (on the model of the central-bank swap lines) is that they would diminish the dollar’s singular role as a reserve currency. If you have credit lines in multiple currencies, you have access to all of those currencies and need not favor holdings of one over another. The dollar still has a dominant role as a vehicle currency in the private markets. It would be harder to dislodge from its vehicle currency role. I share others’ reservations about the role of the SDR as a potential currency. It is not a currency currently, and to make it a currency would be difficult and would not be a natural development.

Discussion of the IMF’s different roles and of the possibility of its acting on a larger scale raises the distinction between liquidity and solvency. The threat of institutional or sovereign insolvency has been an increasingly important factor in financial markets in recent years, and we still need to make progress in resolving sovereign debt and other crises as they come along. Having predictable resolution mechanisms in place is vital for containing moral hazard. Threats to close down big banks or to allow governments to restructure debts, however, cannot be credible without enhanced global supervision of the possibly systemically sensitive entities that lend to them. The crisis in the euro zone illustrates this.

I conclude with what I view as a deep moral of these discussions.

Much of what we mean when we talk about international monetary reform is institution building at a global level. This often seems impossibly hard, although we have made remarkable advances in that realm in the past. An institution like the International Monetary Fund would have been unthinkable without the experiences of the Depression and World War II. We should not need to experience calamities like those again to adapt our earlier achievements to the new and hazardous world that financial globalization has created.

A final question is manifest in the euro zone but also at the global level: whether national sovereignty and self-interest as expressed through democratic processes are inherently friendly to globalization. I believe that they are not inherently friendly to globalization, which is why we have supranational coordinating organizations such as the European Commission, the World Trade Organization, and the IMF. I therefore also believe that if we wish to support expanded globalization in our goods markets and our asset markets and perhaps eventually in labor markets, the globalization of governance institutions must expand as well, not only in liquidity provision but in other areas such as financial supervision.

This is the moral: economic globalization is limited by the globalization of governance.


    FahiEmmanuelPierre-OlivierGourinchas and HélèneRey. 2011. “Reforming the International Monetary System.”Typescript (March).

    LanePhilip R. and GianMaria Milesi-Ferretti.2007. “The External Wealth of Nations Mark II: Revised and Extended Estimates of Foreign Assets and Liabilities, 1970–2004.Journal of International Economics 73 (November): 223250.

    ObstfeldMaurice. Lenders of Last Resort in a Globalized World.2009. Monetary and Economic Studies (Bank of Japan) 27 (November):3552.

    Other Resources Citing This Publication