4 Sources of Systemic Stress

Atish R. Ghosh; Jonathan David Ostry; Charalambos G Tsangarides

Jump to Content Jump to Main Navigation

Author:

Mr. Atish R. Ghosh

Mr. Atish R. Ghosh
Search for other papers by Mr. Atish R. Ghosh in
Current site
Google Scholar
PubMed

Mr. Jonathan David Ostry

Mr. Jonathan David Ostry
Search for other papers by Mr. Jonathan David Ostry in
Current site
Google Scholar
PubMed

, and

Mr. Charalambos G Tsangarides

Mr. Charalambos G Tsangarides
Search for other papers by Mr. Charalambos G Tsangarides in
Current site
Google Scholar
PubMed

Language:: English

Keywords:: OP; regime; country; IMF staff estimate; EME; exchange rate; reserve currency status; inflation benefit; regime case; EMES' incentive; country's interaction; Exchange rate arrangements; Floating exchange rates; Emerging and frontier financial markets; Inflation; Real exchange rates; Global

Download PDF (351.5 KB)

Abstract

The member countries of the International Monetary Fund collaborate to try to assure orderly exchange arrangements and promote a stable system of exchange rates, recognizing that the essential purpose of the international monetary system is to facilitate the exchange of goods, services, and capital, and to sustain sound economic growth. The paper reviews the stability of the overall system of exchange rates by examining macroeconomic performance (inflation, growth, crises) under alternative exchange rate regimes; implications of exchange rate regime choice for interaction with the rest of the system (external adjustment, trade integration, capital flows); and potential sources of stress to the international monetary system.

The discussion in Chapter 3 alluded to possible tensions between individual countries’ choice of regime and systemic stability. This chapter broadens the discussion by considering three interrelated but distinct sources of potential stress: a possible scramble for reserves following the current global crisis, persistent global imbalances, and a “tipping point” in the reserve currency status of the U.S. dollar.

Reserves Accumulation

A key feature of the past decade has been the accumulation of foreign exchange reserves by emerging market countries (Figure 4.1). A number of reasons have been put forward to explain this trend, including export-led growth strategies and precautionary demand against the risk of sudden stops and financial crisis. Indeed, following an initial increase in the early 1990s as capital flows to emerging market countries resumed after the 1980s debt crisis, there was a trend increase in reserve holdings during the 1990s, particularly after the 1997/98 Asian crises.

In the aftermath of the current crisis, developing and emerging market countries may well conclude that they need both to replenish and to further stockpile large foreign exchange reserves to safeguard against both “home-grown” financial crises and—what is more novel—global financial crises that originate in advanced economies. This could be a natural conclusion to the extent that larger initial reserves may have reduced the “output costs” associated with the current crisis (see Figure 4.2, which suggests a positive and statistically significant association between countries’ reserves as of end-2007 and their projected dollar-GDP growth during 2008–09, relative to their historical growth performance).¹

Looking forward, baseline projections for a range of metrics (imports, short-term debt, broad money) suggest that merely restoring reserve ratios to levels prevailing at end-2007 would imply that emerging market and developing countries (excluding China and fuel exporters) would need to acquire some $500 billion to $750 billion over the next five years. Extrapolating these trends (under the assumption of constant growth rates of the scale variables) suggests that these countries would need to acquire a further $500 billion to $1.2 trillion over 2015–19 to maintain the respective end-2007 coverage ratios. How much additional coverage emerging market countries will desire as a result of the current global crisis is difficult to tell because it depends on, among other things, how crisis risks are reassessed and whether risk aversion increases.² But it certainly seems possible that many developing and emerging market countries, under a broad range of exchange rate regimes, could demand greater “country insurance” cover in the aftermath of the current crisis.

In terms of efficiency, the best solution would be market-based insurance mechanisms (i.e., instruments whose payoff compensates for the shock). But in practice, the scope for such instruments is limited by moral hazard and counterparty risk.³ A second best alternative is borrowing in the face of shocks, though this is not genuine insurance in the sense of compensation for the shock, and there may be uncertainty about the availability of finance when it is needed. Third best from the efficiency perspective is self-insurance through reserve accumulation. This is not only costly for the individual country (which forgoes the opportunity cost of holding reserves) but—as elaborated upon below—would also be likely to add to systemic strains by preventing a narrowing of global imbalances or, if extended for a sufficiently long period, amplifying risks of a disorderly stock adjustment in the composition of reserve currencies. To help meet the demand for precautionary reserves, a number of mechanisms could be considered, as outlined below.

Size of the IMF

Because the IMF has near-universal membership, it is an efficient reserve pool that allows member countries to borrow multiples of their foreign exchange contributions from the pool in the event of a balance of payments crisis. The larger the IMF, the smaller the need for countries to self-insure through the accumulation of reserves. By almost every relevant metric (GDP, trade, capital flows, external debt), however, the IMF has been shrinking in recent years (Figure 4.3). Reversing this trend would make IMF membership a better substitute for self-insurance. Recent proposals to increase available resources are an important step in this direction. For a given size of the IMF, increasing the relative quota of developing and emerging market countries would provide them with greater incentive to rely on IMF membership for country insurance.

Access to IMF Resources

For a given quota, increasing the amount and predictability of access to IMF resources again reduces the need for self-insurance. Recent reforms to the IMF’s lending toolkit, including the introduction of the Flexible Credit Line—a precautionary/contingent financing instrument that provides uncapped access and up-front disbursements—should go some way toward persuading emerging market countries to rely more on the IMF for contingent support. Moreover, the very existence of this instrument provides an incentive for stronger policies, reducing the need to stockpile reserves to safeguard against crises. For countries that do not qualify for the Flexible Credit Line, the formalization of High Access Precautionary Stand-By Arrangements likewise provides significant access to IMF resources on a contingent basis, reducing the need for self-insurance.

Special Drawing Right Allocation

Allocations of Special Drawing Rights (SDRs) increase countries’ reserves, without requiring any country to run the corresponding deficit. As such, a substantial SDR allocation would increase developing and emerging market countries’ reserves without exacerbating global imbalances or increasing the stock of U.S. dollar assets held by central banks. The currently proposed SDR allocation of some SDR 167 billion is an important step, because it would increase emerging market countries’ reserves by about 6 percent, on average. However, if viewed against the scenarios for accumulation described at the start of this section, it may not fully meet the potential underlying demand.

The effectiveness of the above mechanisms could be substantial if pursued together and also in conjunction with reforms designed to strengthen perceptions among the membership that the IMF is a reliable and efficient source of external funds in periods of financial stress.

Global Imbalances

Global imbalances—large current account surpluses and deficits on a systemically important scale—emerged during the first years of the current century. Emerging market countries, especially in Asia, ran current account surpluses both as a byproduct of export-led growth strategies and in order to accumulate foreign exchange reserves in the aftermath of the capital account crises of the 1990s. Coupled with rising surpluses in Japan and among major oil producers, by 2008 the sum of current account surpluses amounted to some 2.5 percent of GDP for the Group of Twenty countries (G-20)—which largely corresponded to the deficits of the United States (1.5 percent of G-20 GDP) as well as some developing and emerging market countries (especially in Europe) (Figure 4.4).

By 2003, the IMF was warning about the dangers inherent in imbalances on such a global scale, particularly the risk of a sudden unwinding. The “disruptive scenario” described in the staff report for the Multilateral Consultation (IMF, 2007) highlighted the substantial costs for the global economy and global financial markets of an abrupt decline in worldwide demand for U.S. assets. While going forward the large current account imbalances are projected to shrink from their 2007 peak, they are expected to be at least as large over the next few years as they were in 2004, and thus may continue to pose a risk.

Although there is not a necessary connection between the exchange rate regime and imbalances (the current account reflects all economic policies, not just the exchange rate), the empirical finding that inflexible regimes are associated with larger and more persistent current account imbalances is suggestive. Indeed, theoretical models show that one way countries may promote exports and have a (larger) surplus is by maintaining a pegged exchange rate at an “undervalued” rate through sterilized intervention. These models also suggest, however, that under a floating exchange rate, fiscal policy (such as a consumption tax) can equivalently affect the current account balance (Ghosh and Kim, 2009). Together, these arguments imply that greater exchange rate flexibility in key surplus countries, coupled with measures to raise national saving in systemic deficit countries, should shrink the global imbalances, and thereby reduce the systemic risks that they may pose.

Reserve Currency Issues

A third potential risk to systemic stability, related to those discussed above, is a sudden loss of the U.S. dollar’s reserve currency status. It bears emphasizing that this status is not mandated, nor is it likely to be an all-or-nothing characteristic; rather, it refers to the willingness of central banks to hold reserves in the form of (very high-quality and liquidity) U.S. dollar assets. If the dollar were to suddenly lose its reserve currency status on a flow basis (i.e., central banks stop accumulating further reserves in dollars), this would likely imply a “hard landing” for the dollar; if central banks sought to convert their stock of dollar reserves into other currencies, the impact on the dollar—and the international monetary system—could be highly disruptive. But is that likely to happen?

There are two schools of thought on this. According to Chinn and Frankel (2008), while economic size, rate of depreciation, and importance in global financial transactions all matter for the share of a currency in world official reserve assets, there is considerable inertia in the currency composition of world reserves. Therefore, although their model suggests network effects, because most of these determinants (except the rate of depreciation) move slowly, even the long-run share of the dollar—to which central bank portfolios will gradually adjust—is not likely to change very abruptly (although in some of their simulations, the dollar loses its dominant share by 2020).⁴ An alternative view is that rapid switches in reserve currencies are in fact possible (Eichengreen and Flandreau, 2008), and indeed suggested by the switches in sterling’s reserve currency status during the interwar period.

There is historical precedent for abrupt changes in the dominant reserve currency from the interwar gold exchange standard period, but this episode may be of limited relevance to today’s debate (Box 4.1). In particular, while there was little inherent reason to hold multiple reserve currencies under the gold exchange standard, in an era of floating exchange rates among major currencies, central banks will want to hold a diversified portfolio of reserve assets (while accounting for factors such as the availability of high-quality assets, liquidity and size of the market, network effects, trade patterns, and the currency denomination of trade). Assuming that expected returns are pinned down by uncovered interest rate parity (in other words, expected depreciations are compensated by higher interest rates), central banks will still want to diversify risk in order to preserve the “real” value of their reserves portfolios. An obvious choice of deflator for calculating the real value of the portfolio is the country’s import price deflator, because, ultimately, the purpose of international reserves is to be able to undertake net imports.⁵ Using this assumption, it is possible to calculate optimal shares of dollars, euros, and yen for each central bank’s reserve portfolio. Aggregating across central banks then yields each currency’s optimal share in global reserves. Such an exercise yields an optimal share for the dollar of about 60 to 65 percent (Figure 4.5). Although illustrative, the simulation suggests that the dollar is not grossly overweight in central bank portfolios: according to the most recent Currency Composition of Official Foreign Exchange Reserves (COFER) data, the dollar’s share in reporting central banks’ portfolios is 65 percent.⁶

Box 4.1.

Dollar versus Sterling: The Interwar Antecedent of Switching Dominant Reserve Currencies

Do historical antecedents hold any clues about the evolution of the reserve currency status of the U.S. dollar? Chinn and Frankel (2008) model the composition of central bank reserves since 1973 and find that the reserve currency’s country size, inflation, rate of depreciation, and exchange rate volatility are all important determinants—but that the single most important variable is the lagged portfolio share. Therefore, although the dollar may soon be approaching the tipping point at which the euro becomes the dominant reserve currency, the switch will not be a dramatic dumping of dollars for euros in central banks’ reserve holdings, but rather a gradual shift (the rate at which the shift occurs depends on, among other things, the economic and financial market size of the euro, and would be considerably faster, they argue, if the United Kingdom joined the euro area).

Eichengreen and Flandreau (2008) counter that very rapid switches in the dominant reserve currency have been observed in the past—specifically from sterling to the dollar and back to sterling during the interwar period (see Figure 1).

Bayoumi and Vitek (2009), following a methodology similar to that of Chinn and Frankel, conclude that the tipping point may not come until 2063, though there is considerable uncertainty about when it could occur. However, their work also casts some doubt on the specification of Chinn and Frankel’s model, in particular whether it omits variables that could be important for market reasons and that could be subject to more rapid change, reducing the estimated inertia. Based on this, and a reading of the Eichengreen and Flandreau interwar evidence, Bayoumi and Vitek argue that a switch in dominant reserve currencies could happen much sooner—and more abruptly.

At the same time, the interwar historical example may be of limited relevance to today’s debate. First, under the interwar gold exchange standard, gold was the most important reserve asset (accounting for at least 80 percent of reserves), so the choice between dollars and sterling was just a marginal decision about which currency provided the greatest convenience (given trade patterns and political/colonial links) while safeguarding the value against gold. In this setting, there was no inherent reason to hold multiple reserve currencies, and a small change in the relative probability of devaluation could induce a shift in the dominant currency. In today’s world of floating exchange rates among major reserve currencies, by contrast, central banks will in general want to hold a diversified portfolio of reserves.

Second, the switch from sterling to dollars (and back again) did not involve central banks dumping their stock of sterling assets to convert into dollars—rather, as the stock of reserves held in foreign exchange rose, the amount held in dollars increased faster than the amount held in sterling (for example, holdings of both dollars and sterling rose over 1926–28, even as the share of dollars increased). Likewise, the return to sterling occurred against the backdrop of declining reserves, with the amount held in dollars falling faster than the amount held in sterling.

Third, while these marginal (flow) shifts in the composition of reserves took place within a couple of years, the more fundamental shift from sterling to dollars following World War II took much longer (indeed, sterling maintained the dominant share for almost 10 years after the war), as discussed in the text and Schenk (2009a and 2010).

There are two important caveats to both exercises, however. First, the high concentration of reserve holdings (with the top five countries accounting for more than 50 percent of global reserve holdings) means that aggregate optimal shares may be very sensitive to idiosyncratic portfolio decisions on the part of just one or two central banks—though, by the same token, this concentration means that these central banks would likely internalize the impact on their own portfolios of any abrupt portfolio reallocation decision. Second, these simulations assume that there is no abrupt change in the expected return to holding dollar assets. A counterfactual in which there is a loss confidence in the dollar could lead to a “rush for the exit” and a self-fulfilling sharp depreciation that justifies the initial loss of confidence.

The postwar decline of sterling’s reserve status may provide some useful perspective in this regard, because it was precisely this sort of rush for the exit (by holders of sterling reserves) that was of concern to the official community, and that prompted official multilateral action to reduce the risks of an abrupt adjustment. Recognizing their common interest in safeguarding the international monetary system, central banks of the Group of Ten countries (G–10) provided a series of swap arrangements with the Bank of England, which in turn offered a guarantee of the U.S. dollar value of 90 percent of their sterling reserves as long as the countries holding those reserves agreed to maintain certain minimum proportions of their reserves in sterling—thus removing the “first mover advantage” of a scramble out of sterling reserves. These efforts were successful in that there was never an abrupt dumping of sterling as a reserve currency in favor of the U.S. dollar (again, the absence of a historical precedent to an abrupt stock adjustment in reserve holdings may be relevant), but whether they were necessary (i.e., in their absence, a sudden tipping point out of sterling reserves would have occurred) remains an open question.

In sum, neither historical experience nor simulation analysis suggests that an abrupt change in the stock of U.S. dollar assets held as reserves is likely, but the possibility of a sudden and disorderly tipping point cannot be ruled out definitively—perhaps it is best to view such an event as a tail risk, the realization of which would be very disruptive, and as such, needs to be managed. What measures would help reduce the risk of an abrupt tipping point out of dollar reserves? First, well-articulated, credible exit strategies from the exceptional policies (macro and financial) that have been put in place to cope with the crisis would clearly help to sustain confidence of official holders of dollar assets. Second, measures to narrow global imbalances would help to reduce systemic risks, though disruptive scenarios may well differ from those envisaged under the multilateral consultation. Third, providing developing and emerging market countries with a viable alternative to stockpiling reserves (e.g., greater provision of synthetic reserves assets such as the SDR) would mean less accumulation of dollar-denominated liabilities, and thus lower risk of a sudden tipping point. Other measures have been proposed, but from the perspective of today, they seem unlikely to command broad support of the international community.⁷

It is also possible, however, that in the aftermath of this crisis, some EMEs will move toward more flexible regimes to better insulate themselves against future crises, which could lower their demand for reserves. In the past, however, most emerging market countries have maintained significant levels of reserves regardless of their exchange rate regime (for instance, over 2002–07, EMEs with de facto pegged regimes maintained reserves of 15 percent of GDP, whereas floaters kept about 14 percent of GDP).

Models of optimal reserve holding (Jeanne and Rancière, 2009; Kim, 2009) suggest a couple of ways in which the current crisis might affect the demand for country insurance. First, an increase in risk aversion on the part of emerging market countries would raise their demand for precautionary reserves. But if there is a generalized increase in risk aversion, including on the part of creditors, then the interest rates faced by EMEs may increase as well, implying a corresponding increase in the opportunity cost of holding reserves. Unless the reserves are provided at zero cost (e.g., a Special Drawing Right allocation), the net effect on the demand for precautionary reserves could be small. Second, the current crisis may have resulted in a reevaluation of the likelihood or cost of crises. Under plausible calibrations, these models suggest that the additional demand for reserves could be on the order of 10 percent of the reserves required to restore precrisis coverage ratios.

See Becker and others (2007) and IMF (2009).

This is the case when the simulation assumes that the United Kingdom joins the euro area (significantly boosting euro financial markets) and the dollar has a persistent depreciation (over the next 30 years) at the same rate as experienced over 2001–04. This produces a tipping point around 2020, when the dollar’s share falls from about 55 percent to about 25 percent within acouple of years. In their less extreme scenario, the United King domdoes not join the euro are a and the dollar depreciates at the same average rate as experienced over the past 20 years—in this case, the dollar’s share falls from about 70 to 60 percent by 2020 and remains constant thereafter.

This is not identical to matching import shares, because trade with third countries is often denominated in U.S. dollars, euros, or yen. The impact of exchange rate changes on the import price deflator is calculated by matching changes in the import price deflator to changes of the euro-dollar and yen-dollar exchange rates. The use of the import deflator has two potential shortcomings. First, the objective is to protect the country’s ability to make net imports, so the correct deflator should also take account of the impact of exchange rates on the country’s exportrevenues. In the face of financing difficulties (e.g., a suddenstop of capital in flows), most of the adjustment tends to fall on imports, suggesting that preserving the value of the portfolio in terms of the import deflator is more important. In practice, using a weighted average of the import and export deflators leads to very similar results because, for most countries, import and export trade patterns tend to be similar. Second, although the ultimate purpose of reserves is for net imports, central banks may want to hold reserves for short-term debt in the currency in which that debt is denominated. But an exercise in which central banks hold reserves up to the amount of their short-term debt in the currency of that debt (generally dollars, but euros in the case of European countries outside of the euro area), and undertake the portfolio diversification described in the text only for reserves above their short-term debt, yields very similar results for the aggregate demand for dollar, euro, and yen reserves, essentially because global reserves are very concentrated (the top five holders held more than half of the global reserves in 2008), so total reserves in excess of short-term debt account for about two-thirds of total global reserves.

One caveat is that the COFER data do not have information on the currency composition of 37 percent of global reserves (up from 26 percent in 1995), so the true dollar share in global reserves may be greater.

An example is the proposal in the late 1970s of establishing a U.S. dollar-SDR Substitution Account into which central banks could deposit dollar reserves (typically, short-term U.S. treasury bills) in exchange for SDR-denominated claims. The account would convert its assets into longer-term dollar-denominated claims while paying the short-term official SDR interest rate, with the interest differential intended to cover the account’s exchange rate risk. The proposal failed to gain traction, not least because of lack of consensus over who would bear the residual exchange rate risk (Boughton, 2001; and Schenk, 2009b).

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/display/book/9781589069312/ch004.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully