comments: The Balance Between Adjustment and Financing
- Alexander Swoboda, and Peter Kenen
- Published Date:
- December 2000
My comments are presented in two parts. Part I focuses specifically on the paper by Professors Calvo and Reinhart that addresses the important issue of what policy regimes are most likely to avoid sudden-stop crises of the type experienced by emerging market countries in recent years. Part II discusses some aspects of the balance between adjustment and financing, which is the announced theme of this session.
Avoiding Sudden-Stop Crises
Professors Calvo and Reinhart highlight certain special features of the nature of the crises that have affected emerging market countries and consider alternative policies that might prevent such crises in future. As the authors point out, these crises arise because of a sudden denial of access to capital, the sudden stop, combined with large withdrawals of capital, reflecting either a refusal to roll over a large stock of short-term debt or domestic capital flight. Such crises have a highly disruptive effect on output, with the disruption being greater and the recovery more prolonged when the crisis also involves a banking crisis, as it did in East Asia.1
While the authors document the impact of sudden stops, they do not go into the underlying reasons for this phenomenon, and yet these are surely relevant for preventing crises in future. They rightly point out that currency crises in industrialized countries, such as the ERM crises, differed from crises in emerging market countries because in the former there was no loss of access to capital markets and no disruption of output. Crises in industrialized countries were clearly triggered primarily by the perception that the particular level of the exchange rate was unsustainable, leading to a speculative attack. Once the exchange rate was allowed to drop, however, speculative pressure ceased as soon as the rate reached a level judged to be sustainable. There was no significant overshooting, no denial of access to capital, and certainly no banking collapse.2 Crises in emerging markets on the other hand have been triggered by capital outflows arising from a much broader loss of confidence in economic management, especially in the functioning of the financial system. Weaknesses in the financial sector led to imprudent bank lending, excessive exposure to foreign exchange risk, and an accumulation of short-term debt, all of which increases vulnerability. The extent of investor reaction may also have been exaggerated by lack of information about conditions in these countries, a situation that is very conducive to creating euphoria in good times and panic in bad. Because of these factors, capital outflows did not cease even after the currency had depreciated significantly, and since continuing large withdrawals could not be financed or controlled, the result was a massive overshooting of the exchange rate. This increased the fragility of the already weak financial system and generated severely negative balance sheet effects on corporations and banks, all of which had highly negative effects on the real economy.
If crises are caused by institutional weaknesses in the financial sector, inadequate information, and lack of transparency, then any strategy for crisis prevention must address these deficiencies directly as a first priority. This point is not picked up in the paper, which focuses on other policy options such as capital controls or a different exchange rate regime. These are certainly relevant in reducing vulnerability, but they are not substitutes for correcting the fundamental weaknesses that generate the crises in the first place. Unless these weaknesses are addressed, the problems they give rise to will surface in one form or another, no matter what policy options are adopted.
The authors consider capital controls as a possible means of reducing the probability of crises and come out firmly against them. The results reported in Table 8 are interpreted as showing that capital controls do not influence the level of capital inflows, although they do appear to affect their composition by reducing the proportion of short-term flows. The apparent lack of impact on the total volume of flows in the regressions may reflect no more than the fact that the authorities in the countries concerned may not have wanted to control the level and were concerned only with the composition.
It is difficult to believe that controls have no effect on the level of flows. Reference is usually made in this context to the possibility of leakages from the control system, but this only means that controls are imperfect, not that they have no effect. Whatever the leakages, controls are generally effective in preventing sudden inflows or outflows, and these sudden movements are often the most destabilizing. In any case, the evidence presented shows that controls clearly affect the volume of short-term capital flows, which are usually regarded as the villain of the piece. These considerations suggest that capital controls can be used to reduce a country’s vulnerability to sudden withdrawals of capital.
The authors’ rejection of capital controls probably emanates primarily from the nonmarket nature of the instrument, and this is reflected in their fear that controls may lead to a “gradual reversion to central planning.” This concern is perhaps overdone. Controls probably do generate some efficiency losses and also introduce discretion and nontransparency, both of which are disadvantages, but these costs have to be weighed against the equally worrisome risk of sudden and disruptive outflows. As long as the international financial architecture is unable to provide enough protection to countries against such destabilizing behavior, a case can be made for retaining some controls. Regulatory restrictions on the functioning of individual financial institutions are accepted as part of any market economy. Extrapolating from this, one can argue that well-designed capital controls, especially of the Chilean variety that minimize discretion, serve as a prudential regulation applied at a macro level that can be justified as long as the financial system lacks the strength and sophistication to discourage excessively risky flows on its own on purely prudential grounds. In the aftermath of East Asia, it is difficult to advise developing countries to dismiss capital controls as summarily as the authors have done.
The authors note that soft peg exchange rate regimes are generally viewed as having contributed to many of the recent crises, and, in this context, they discuss the relative merits of flexible exchange rates versus credibly fixed exchange rates. They come out strongly in favor of the latter in its extreme version of dollarization. Surprisingly, this conclusion does not seem to be conditioned by country circumstances. One can easily understand that a relatively small open economy, which wants to maintain full capital convertibility, trades dominantly with one currency area, and has a history of hyperinflation, may find it attractive to opt for dollarization. The abandonment of monetary sovereignty may appear a worthwhile price to pay in purely economic terms to gain relief from currency fluctuations.3 But this may not hold for larger countries with a lower proportion of trade to GDP and also greater diversification across different currency areas.
The authors recognize the advantages of exchange rate flexibility in enabling countries to adjust the relative price of tradables and nontradables without having to rely on downward flexibility in nominal wages and prices, which otherwise becomes essential under fixed exchange rates if full employment is to be achieved. They appear to set too much store, however, by the conviction that dollarization will get rid of currency crises once and for all. As Professor Reinhart put it in her oral presentation: “You cannot attack the peso if it doesn’t exist.” True enough, but the problems that cause currency crises do not necessarily disappear with the peso. Dollarization eliminates speculation about exchange rate changes, but it does not eliminate changes in external demand condition or weakness in the banking system. The key question is whether fixed exchange rates make it easier to handle such problems if and when they arise. This is not at all evident. Maintaining fixed exchange rates in these circumstances only means that the shocks that otherwise cause crises will be transmitted directly either to real wages or to output. The loss of exchange rate flexibility in such circumstances may well be a disadvantage.
The authors claim that exchange rate changes are a poor method of adjusting relative prices of tradables because different sectors may need different degrees of support to become sufficiently competitive to ensure full employment for resources specific to those sectors (the gauchos versus physicists example). They go on to suggest that the necessary adjustment may be better achieved by differential rates of labor subsidies. Apart from fiscal constraints, which are bound to limit the ability to finance such subsidies, differentiated subsidies also take us down a slippery interventionist slope that should be every bit as worrying as the dreaded reversion to central planning! If certain resources are specific to a particular sector in the short or even medium term, it will be necessary for the relevant real wages to adjust to sectoral demand conditions, and this may involve different degrees of adjustment in different cases. Even so, exchange rate variations can take the burden of a great deal of the adjustment needed, leaving less to be done at the level of individual sectors.
For all these reasons, I am unable to agree with the authors’ conclusion that “dollarization would appear to have the edge as a more market-oriented option to ameliorate, if not eliminate, the sudden-stop problem.” They make a much more limited claim earlier in the paper that “dollarization is not the silly idea that conventional thinkers would have us believe,” and this is a position one can readily concede. Dollarization may make sense for some countries, but most developing countries would be better advised to look for salvation in exchange rate flexibility. This is not to say that huge variations in exchange rates are not destabilizing. They are, but the expectation is that a system that allows more frequent exchange rate fluctuations will generate sufficient awareness of foreign exchange risk to avoid the kind of imprudent exposure that creates fragility and leads to crises.
The Balance Between Adjustment and Financing
The Calvo-Reinhart paper, focusing as it does on crisis prevention, does not deal directly with the balance between adjustment and financing, which is essentially a crisis management issue. The role of financing in contemporary crises is particularly important because these crises erupt suddenly and create a large financing gap in a very short time. Since resources are not available to finance such large gaps, countries have often been forced to generate a large surplus in the current account to accommodate outflows in the capital account. Such turnarounds can only be achieved through severe import compression, which in turn implies large output contraction. It is not surprising that there is a widespread perception that the balance has shifted too much toward adjustment and not enough toward financing.
The balance between adjustment and financing traditionally has been decided on the basis of two considerations: the reversibility of the balance of payments shock and the optimum time period over which adjustment should be stretched. Where the shock is temporary and self-reversing, it is generally agreed that the solution lies very largely in financing. Where the shock is likely to be sustained for some time (whether it is due to external factors or internal policy imbalances), it is necessary to adjust by reducing the current account deficit. Even in this situation, there is a need for financing, since the full extent of the adjustment needed cannot be implemented immediately, and the volume of financing needed therefore depends on the period over which the adjustment should take place. It has long been accepted that developing countries need more time to adjust than industrialized countries because structural rigidities in their economies make it difficult for them to increase the output of tradables without addressing longer-term supply constraints, which often require increased investment in critical sectors. Forcing too rapid a pace of adjustment in these circumstances would only lead to excessive reliance on demand management, which would depress investment, output, and growth. These considerations led the IMF to establish the Extended Fund Facility, which made available more resources over a longer period with a more extended period for repurchase.
To this received wisdom to determine the balance between adjustment and financing in contemporary crises, the first step is to determine whether the shock is self-reversing. An extreme view is that these crises are essentially liquidity crises caused by a completely irrational loss of investor confidence leading to a self-fulfilling panic. Using the analogy of central banks coming to the rescue of solvent domestic banks facing a run on deposits, it has been argued that what is needed is not so much adjustment as large amounts of finance, with minimal conditionality, to stem the panic and bring the situation quickly back to normal. This characterization of emerging market crises as instances of pure irrational panics is understandably attractive in countries hit by crisis, and they may also have an element of truth, but it is clearly not the whole story. Panics rarely arise in situations where there is nothing wrong with underlying fundamentals. More likely, they arise when there are some underlying problems, which were either building up over time and were ignored until they reached a critical level, or which arose suddenly to cause alarm among investors. Because of the subjective nature of investor expectations, and the phenomenon of herd instincts, the reaction can be out of proportion to the extent of the weakness, but there is no doubt that in most cases some weaknesses are also involved. The solution in such situations obviously does not lie in providing unconditional financing. Since there are underlying weaknesses, corrective steps are definitely needed to address these. A substantial volume of financing is also needed, however, since the entire burden cannot be borne by adjustment.
The volume of financing needed to handle contemporary crises is much larger than current institutional mechanisms can deliver because of large capital outflows caused by a collapse of confidence. Even if all necessary corrective steps are taken, they may not succeed in restoring confidence and bringing capital flows back to normal levels. This is especially true if the loss of confidence arises because of perceived institutional weaknesses. Corrective steps aimed at strengthening the banking system take considerably more time to show results than corrective steps in traditional areas such as restoring fiscal balance. Investor confidence also depends on many intangible factors that are not easily addressed. Furthermore, even when investor confidence is fully restored, capital flows may not recover to precrisis levels if they were bloated because of euphoria.
The perception that financing is inadequate has to be seen against this background. One could reasonably argue that if a country has taken appropriate corrective steps that will enable restoration of confidence and a return to normal flow levels in a reasonable period—say one to two years—then sufficient financing should be provided in the interim to enable the country to make this transition smoothly. Failure to provide adequate financing will lead to overshooting of the exchange rate with negative effects on the real economy, delaying the return of confidence and reducing the effectiveness of the adjustment program.
The contrast between the experiences of Mexico and East Asia is instructive in this context. The Mexican crisis is widely regarded as having been well handled. The crisis did impose a severe strain on Mexico in the first year, which was perhaps unavoidable (those concerned with moral hazard will even say it was desirable). But confidence was quickly restored, and Mexico made a quick recovery. Much of the credit must go to the large amount of financing made available by the IMF, with substantial financial support from the United States, though the restoration of confidence was undoubtedly aided by Mexico’s proximity to, and special relationship with, a booming U.S. economy.
The East Asian experience was very different. The IMF reacted very quickly to the crisis—more quickly than it did in Mexico thanks to the emergency financing procedures put in place after the Mexico crisis—but the IMF programs failed to stabilize the situation. As is well known, the currency collapse intensified after the IMF programs were put in place. Inadequacy of financing was an important part of the explanation for this failure. The size of the rescue packages for East Asia was much lower (relative to GDP) than for Mexico, especially if account is taken of the lower credibility of the bilateral portion of the East Asian package.4 An IMF staff study attempting to draw lessons from the experience (see Lane and others, 1999) has pointed out that a crucial assumption in the IMF’s East Asian programs was that they would succeed in restoring confidence and halting the capital outflows: the amount of financing provided was adequate only if this assumption proved correct. It did not, and the currencies went into a free fall.
The IMF’s programs have been criticized in many quarters for faulty design. I do not propose to add to this extensive debate on which there is no doubt more to come. With the recovery in East Asia looking more robust than earlier expected, except perhaps in Indonesia where political circumstances have been unusual, some of the criticism of the IMF may appear excessive in retrospect. The point to emphasize is that the IMF programs were calibrated on the assumption that confidence would be restored. But when this did not happen, there was no mechanism by which additional financing could be made available to fill the gap. Whatever the shortcomings in the initial design of the programs (and some of these have been admitted by the IMF), it is entirely possible that had more financing been available, confidence would have returned sooner, reducing the extent of overshooting of the exchange rate and the collateral damage it caused. The recent removal of access limits on IMF financing in certain instances is a welcome development in this context, since it recognizes the need for large volumes of financing in certain types of crises. However this flexibility regarding access limits will need to be matched by a much larger volume of assured resources available to the IMF to meet the financing needs that could arise. Not enough has been done in this area.
IMF financing is not, however, the only way of handling the financing gap that arises in a crisis, and financing from the private sector should also be mobilized. It is difficult to mobilize new private capital in the initial stages of crisis resolution, but it is relevant to focus on the role of standstill arrangements and debt workouts to help stem panic outflows and thus contain the financing problems of crisis-hit countries. At present, we do not have institutionally agreed ways of triggering such workouts at an early stage in crisis resolution, and this is a major factor explaining the severity of recent crises. It is precisely in such situations that it is necessary to have agreed mechanisms for triggering debt-restructuring packages that would halt an otherwise destructive downward spiral. Such mechanisms also have the advantage of forcing a degree of risk sharing by lenders, which is completely absent when public international institutions are seen to be financing the exit of private lenders.
Votaries of the market tend to view debt workouts as an unwarranted intrusion into the functioning of the market, giving defaulting borrowers an easy escape route and thus creating moral hazard. An institutional mechanism for handling these problems in a transparent manner, however, has many advantages. It would ensure an appropriate degree of burden sharing between the borrower and the lender and also create greater transparency in the process. The risk of such debt workouts being triggered would, of course, affect the expectation of lenders at the time they make loans; it could be argued that this would reduce the supply of funds to emerging markets. This prudential moderation of inflows, however, may not be undesirable. The existence of the mechanism would only make explicit a risk that does exist but is typically ignored. If it reduces the flow of credit to developing countries to a more prudent level, this may be systemically desirable. It would certainly reduce the disruptive consequences of the crises that otherwise arise.
To summarize, adjustment has to form an important part of the response to contemporary crises, but financing is at least equally important. Over the years, a consensus did evolve on how to deal with traditional balance of payments crises originating in the current account, and this consensus gave a substantial role to financing, primarily because it was recognized that there are limits to the speed with which developing countries can achieve adjustment in the current account in a short time without disruption in the real economy. The same consideration applies for capital-account-driven crises that force adjustment in the current account, except that the volume of adjustment being forced in such crises is much larger. The five East Asian countries, for example, experienced capital outflows amounting to 10 percent of GDP in 1997. Some of these problems can be forestalled by effective crisis prevention. But we also need much more of both direct financing as well as mechanisms for refinancing through debt restructuring if we want to add stability to the international financial system. The IMF needs to be strengthened to enable it to play a larger role in both areas.
The estimated periods of recovery presented in Table 5 are based on data from crises prior to the East Asian crises. It will be interesting to see if the East Asian economies show a quicker recovery—as now seems likely—reflecting stronger underlying fundamentals.
Banking fragility was evident in Sweden but not to the extent evident in most emerging markets.
This assumes that domestic political sensitivities can be overcome (no easy assumption for most countries).
See World Bank (1998).