Sitikantha Pattanaik*

In the context of the growing international perception that the International Monetary Fund (IMF) cannot and should not function as an International Lender of Last Resort (ILOLR), this note presents a perspective on why the emerging market economies need the IMF as an ILOLR. It argues that the risk of financial instability and associated loss of output and employment is a more serious concern than the risk of moral hazard, and that an effective and credible ILOLR can make the current process of globalisation less painful and disorderly. It highlights that a transparent and financially empowered IMF can enhance the effectiveness of national Lender of Last Resort (LOLRs) in dealing with twin crises. It also emphasises the point that international initiatives on crisis prevention/resolution undertaken as part of the work programme on new international financial architecture should not be viewed as a substitute for ILOLR. The ILOLR would continue to be relevant, irrespective of the degree of progress that can be achieved on these initiatives.

JEL Classification: F330; F020

Key words: ILOLR, Liquidity, Solvency, Moral hazard.

Introduction

The conclusion of the Twelfth General Review of Quotas in January 2003 with a decision against any increase in the general resources of the IMF clearly weakened the capacity of the IMF to function as an International Lender of Last Resort (ILOLR). The decision, however, helped in formally validating the argument that has been gaining ground since the Mexican crisis of December 1994 that the IMF cannot and should not function as an ILOLR. According to this view, the Bagehot (1873) principle would require that like a domestic LOLR, the ILOLR should be in a position to lend freely, at a penal rate, and on good collaterals. Other than possibly using a penal rate on its lending, IMF cannot meet the other requirements. Hence, IMF cannot be an ILOLR (Schwartz, 1999). Also, the IMF should not perform the role of an ILOLR because that would accentuate the moral hazard problem and thereby encourage more imprudent lending and borrowing and ultimately give rise to financial crises (Calomiris, 1998). Some even argue that any increase in the general resources of the Fund may be made, if necessary, only after the present work on constructive Private Sector Involvement (PSI)/Sovereign Debt Restructuring Mechanism (SDRM) achieves the envisioned objective, since an early increase in general resources could in fact raise the moral-hazard problem and weaken the response of both creditors and debtors to such initiatives.

The argument that the IMF cannot and should not be the ILOLR essentially reflects the position of advanced countries who in fact do not require an ILOLR, because their domestic LOLRs can fill the vacuum due to their ability to create international liquidity. Central Banks of emerging market economies, however, cannot create international liquidity and they generally find it the most difficult to arrange international liquidity from any source amidst a crisis. In the absence of a credible ILOLR, they often try to strengthen the capacity of their domestic LOLRs by pursuing a policy of maintaining high foreign exchange reserves. Such a self-insurance centric domestic financial architecture not only comes at the expense of less flexible exchange rates, but also entails several other costs which could have been avoided to some extent with an ILOLR. The moral hazard argument has been conveniently used to stall appropriate financial empowerment of the IMF, despite the fact that IMF liabilities have rarely been defaulted by emerging markets and that countries actually approach the IMF only as a last – not first – resort due to the fear of intrusive conditionality package that accompanies Fund assistance. The growing impact of contagion on the prospects of emerging markets and the general unpredictability of the timing and intensity of any crisis suggest that liquidity needs of a country at times could change dramatically, and could be large as well. In this context, a role for an ILOLR is justified on the ground that it may be more appropriate to support the country where crisis strikes first with quick provision of adequate international liquidity, instead of allowing the contagion to show its full strength and then assist a number of countries with larger liquidity.

The objective of this paper is to present an emerging market perspective on the need for an ILOLR. For developing the rationale behind an ILOLR, one needs to examine first the relevance of all the arguments that support the view that the IMF cannot and should not take up the role of an ILOLR. Accordingly, Section I sets out the basic functions of any domestic LOLR and evaluates the extent to which IMF can perform those functions at the international level. Section II encapsulates all the arguments on moral hazard and presents a comparative assessment of the costs, with or without ILOLR. Section III presents the arguments as to how the IMF can, and why it should function as the ILOLR. Section IV offers a few concluding observations.

Section I

The Difference Between LOLR and ILOLR

Since the origination of the idea in the 19^th Century by Thornton (1802) and Bagehot (1873), the architecture of an ideal LOLR has been extensively debated, even though the essential elements of any LOLR continue to be more or less the same even today. According to Saxton (1999):

(i) The LOLR has a macroeconomic rather than microeconomic responsibility. Only when the failure of an insolvent individual institution threatens the stability of both the financial system and the value of money that the LOLR should make available adequate liquidity. Unsound institutions should be allowed to fail, unless they have important spillover effects.

(ii) The fractional reserve system of banking and the government monopoly over issuance of legal tender suggest that while the former creates a need for LOLR the latter provides the means for satisfying the need.

(iii) The LOLR function, which is a short-run stabilisation role, should not conflict with the medium- to long-term goal of a Central Bank, particularly price stability.

(iv) An important function of any LOLR being to avert panic, it should be transparent. To minimise the moral hazard problem, however, it must make the point loud and clear that liquidity will not be provided to insolvent banks directly, but only to the market or at best to the solvent and sound banks directly who can offer appropriate collateral while receiving the liquidity support at a penal rate. Transparent assurance to "lend freely to the market at a penal rate on good collateral" ensures that lending remains short-term and that borrowers exhaust all other private sources of fund and approach the LOLR genuinely as a last resort.

In practice, most of the above essential elements are difficult to realise. The most important challenge that any LOLR encounters is to distinguish between solvent and insolvent banks. As underscored by Freixas et al (1999), "a central bank may not always be able to make this distinction, particularly in the short time-scale in which a lending decision may have to be made". The argument that the LOLR should lend only to the market through open market operations also presumes that the markets must be efficient, so that all solvent banks can get enough liquidity from the inter-bank market. Insolvent banks, in turn, would automatically be eliminated from the system. In reality, the inter-bank market suffers from information problem and it cannot differentiate the banks on the basis of their solvency in any way better than a national regulator/ supervisor, even when information disclosure and transparency norms are in place. Coordination problem could also prevent access to liquidity for some solvent banks amidst a run on the bank. In view of the moral hazard problem associated with a policy of transparency, LOLRs often maintain "constructive ambiguity" under which the liquidity support is determined on an ad-hoc and ex-post basis. As emphasised by Niskanen (2002), "central banks are, in the final analysis, faced with a time inconsistency problem. Even if an LOLR facility were beneficial ex-post, i.e. after a liquidity shortage has occurred, expectations of support from such a facility may weaken the bank’s incentives ex-ante in a way that makes the occurrence of a liquidity shock even more probable". This justifies a policy of constructive ambiguity. Rochet and Vives (2002), however, suggest that transparency can be better than constructive ambiguity, provided the LOLR policies could be based on the principles that guide the "private lines of credit" of commercial banks. In such a case, LOLR would assume a contractual character which need no longer provide liquidity support "free of charge, without limits, and at the complete discretion of the central bank". The LOLR function then should be guided by: (i) payment of commitment fees ex ante by the banks, (ii) explicit limits on the access to liquidity support, and (iii) compliance with preconditions (say, proper risk management and prudential regulation). In this context, it is interesting to note the findings of a study by Herrala (2001) which compares the social desirability of three alternative governance structures of LOLR: a public lender of last resort, a mutual clearing house that formulates policy by voting, and a profit maximising private LOLR scheme. Voluntary/private schemes succeed as long as all recipients of liquidity repay loans with interest, so that the LOLR does not suffer losses. When the LOLR suffers losses, good quality banks would tend to stay away and only weaker banks would join the scheme. Hence, LOLR in the public sector may remain as the only viable option.

In the context of the debate on transparency versus constructive ambiguity, Guttentag and Herring (1983) differentiate between three different forms of commitment that the LOLR can do in advance: (i) it explicitly states what it is able and prepared to do and what it is not prepared to do, (ii) it promises less than what it is able and prepared to do, and (iii) it promises more than what it is able and prepared to do. The last one is a clear prescription for disaster, and should always be avoided. Between the first two, even though the Bagehot prescription would favour the first option, the more appropriate policy of any LOLR could be based on the second option. Constructive ambiguity that represents the second option, however, could provide some unavoidable advantage to large banks who may be too big to fail.

Irrespective of these above operational challenges, every national economy operates with a LOLR, and as highlighted by Guttentag and Herring (1983), real life experience suggests that " central to the LOLR function is a willingness to accept a risk unacceptable to other lenders…Because of capital market imperfections, the LOLR function requires direct credit extensions to individual banks. …The social costs associated with bank failures during a financial crisis are larger than private costs." To avoid the social costs, a LOLR could willingly assume risks that the private market would refuse to bear. Hence, "banking crises are preventable by an LOLR at relatively small social cost". Economic historians, such as Kindleberger (1989) and Bordo (1990) also found that the development of LOLR has helped in reducing the frequency and severity of national banking crises.

Unlike any national LOLR, the IMF cannot lend freely, constrained by the lack of ability to issue international liquidity. Decisions like the one taken under the Twelfth General Review of Quotas can further constrain its capacity to lend freely. It also cannot lend quickly, due to the requirement of Board approval and acceptance of conditionality by the national authorities. It is more difficult for the Fund than any national LOLR to differentiate solvency crisis from liquidity crisis. Most importantly, it can at best lend against conditionality, but not against good collaterals, since a country in the midst of a crisis may not possess internationally acceptable collaterals at its disposal. It can lend at penal rates, as it has done against purchases under the Supplemental Reserve Facility (SRF), but SRF is neither backed by any additional Fund resources (hence it competes with conventional Fund facilities for limited resources at the disposal of the Fund) nor is it subject to the usual access limits (enlarging thereby the scope for discretion, which could potentially be used to favour a few preferred Fund members). The lack of any progress on the Contingent Credit Lines (CCL) also clearly shows that penal rates and intrusive pre-qualifying norms may not be acceptable to a vast majority of Fund members, even when CCL could have the typical features of an ILOLR. All these elements which constrain the IMF’s ability to function as an ILOLR are examined in greater detail in Section III in the context of the analysis justifying an ILOLR role for the IMF.

Section II

Greater Evil - Moral Hazard or Global Financial Instability?

Balancing the moral hazard risk against the meltdown risk is the key challenge for an ILOLR. The international perception, which is essentially a pro-market view advanced and propagated by the first world, seems to have overemphasised the risk of moral hazard. Litschig (2001) went to the extent of arguing that moral hazard was the key determinant of financial crises in emerging markets in the 1990s. Expectation of an official bailout induced overtly risky behaviour by creditors. Since the Mexican bailout "creditors have learned that there is an implicit IMF guarantee on capital flows to large emerging markets and that a risk premium can be collected without incurring the risk". As a result of the Mexican bailout, international investors could collect a risk premium of about 6 per cent without assuming any risk, because the bailout essentially equated the 12 per cent Mexican sovereign debt with the 6 per cent US Treasury securities. According to Calomiris (1998), an ardent supporter of the moral hazard argument, "if the risk taking bankers know that future gains from taking on risk will be private, but losses will be borne by taxpayers, that amounts to a government subsidy for risk, which thereby encourages excessive risk taking (the so called moral hazard problem)". Based on these arguments, IMF bailouts, as the experience of Mexico and Indonesia would suggest, lead to emergence of the wealthy and the politically influential risk-takers as winners and the taxpayers as biggest losers. Furthermore, recent IMF bailouts, instead of preventing panic, actually facilitated transfer of wealth from taxpayers to insolvent financial institutions and international creditors. He recommends, therefore, that policy makers should realise how counterproductive IMF bailouts could be and why IMF’s resources should not be expanded for financing bailouts.

More than an ILOLR, therefore, what the global financial architecture requires is stronger and more effective crisis prevention measures. As stressed by Kaufman (1999), "the need is not for a lender of last resort, the need is for improved supervision and regulation over the major risk-takers and the major markets….no matter what happens in the emerging countries, the result will never be a systemic risk (to the global system)". Tietmeyer (1999) also echoed a similar perception by viewing that neither every financial crisis can have systemic implications nor is systemic risk a given quantity. Rather it is an endogenous variable, which is influenced by the structure of financial markets, supervisory framework at the national and international levels, and the decisions of the national political and monetary authorities. The emphasis, therefore, should be on avoidance of systemic risk rather than an ILOLR.

Calomiris (1998), like Kaufman (1999) also viewed that "irrational financial contagion" is not the real concern because history and theory of banking panics suggest that panics have always been rational phenomena. "Thus, concerns of irrational contagion spreading from one country to another without any fundamental explanatory link connecting the countries are unwarranted. Such concerns should not be used to justify financial bailouts." Contagion, however, has been a major source of financial crises in countries having sound fundamentals in the past decade. As underscored by Mishkin (2000), " a successful speculative attack on one emerging market country does lead to speculative attacks on other emerging market countries, which can lead to collapses of additional currencies….An ILOLR has the ability to stop contagion by providing international reserves to emerging market countries threatened by speculative attacks so that they can keep their currencies from plummeting." The formation of the G-20 after the Asian crises is a clear demonstration of the international recognition of the importance of contagion, particularly if it originates in one of the systemically important countries. Besides the concern of contagion and the large output and employment costs associated with banking and currency crises that justify the need for an ILOLR, the following arguments also provide considerable support to the relevance of an ILOLR for the emerging market economies.

i) Sovereign Debt Restructuring Mechanism (SDRM) and constructive Private Sector Involvement (PSI) can prevent unsustainable sovereign debt induced crises; but crises arising from over-lending and over-investment can not be prevented by such initiatives.

ii) Those who target IMF bailouts as the source of moral hazard generally assume that all instances of past IMF financial packages were dealing with sovereign insolvency, rather than illiquidity.

iii) IMF always gets back the crisis time finance it provides to its members. Outstanding arrears at any point of time have been negligible, and with a view to safeguard Fund resources, even the impact of such arrears is completely offset by the burden sharing arrangement that shifts the burden to the Fund members.

iv) As rightly emphasised by Litschig (2001), "the governments of Mexico, Thailand, Korea, Indonesia or Brazil were merely illiquid, rather than insolvent, as witnessed by their success in avoiding default on private claims and their early repayment of the IMF financing that was made available".

v) Most importantly, the ILOLR has to lend to sovereign governments, not to the banks/private creditors directly. Unlike the creditor moral hazard, there is hardly any sovereign moral hazard (unless it is made out to be so through biased arguments). Moreover, to determine as to whether a sovereign is insolvent or merely illiquid is an art rather than a science. Hence, in the absence of sovereign moral hazard, the ILOLR should provide international liquidity without being motivated by the concern of creditor moral hazard. Absence of ILOLR would not necessarily make domestic LOLRs to stop doing what they have been doing.

vi) International initiatives on crisis prevention are no substitute for an ILOLR. Progress on PSI/SDRM, effective Fund surveillance, promotion of transparency and compliance with international standards and codes, etc. could reduce the need for an elaborate ILOLR, but cannot deliver what an ILOLR can. No emerging market needs an ILOLR as a substitute for reform. They, however, feel that despite pursuing sound macroeconomic policies and having strong financial systems, they can still encounter situations of liquidity crises.

vii) Ramifications of a default on Fund liability are too costly; that itself is the best collateral against which the Fund should lend freely.

viii) To sustain and deepen the process of globalisation (which almost all the supporters of the moral hazard argument want), national financial and corporate entities have to necessarily increase their exposures to different international currencies. In the international context, national financial institutions have to not only ensure appropriate maturity transformation but also prudent transformation of currency exposures. Liquidity problem arising from asset-liability mismatch in terms of both maturity and currency exposure can be addressed only to a limited degree by national central banks, depending on the exchange rate regime and the size of the foreign exchange reserves in relation to the aggregate exposure. An ILOLR, thus, is a global safety-net that is essential to deal with the challenges of growing globalisation.

The above assessment clearly suggests that the need for avoiding global financial crises is a much more serious concern than the risk of moral hazard. The objective of the debate on the new global financial architecture, therefore, should be to strengthen the crisis prevention framework and to undertake measures that could contain the scope for moral hazard. Such initiatives, however, should not undermine the importance of an ILOLR. As noted by Frankel (1999), "moral hazard is not to refrain from ameliorating the effects of a given crash …moral hazard cannot be the fundamental (source of) market failure". Little and Olivei (1999) echoed a similar concern that "the absence of a reliable ILOLR may have increased the volatility of investor sentiment, aggravating herd behaviour". A well functioning international capital market cannot do the job of an ILOLR. As the pattern of private capital flows to emerging markets in the last decade reveals, there has been high degree of concentration of such flows in few emerging markets and when any of these countries needs such capital the most it may actually experience a rush for exit by the private creditors. A large number of countries do not get adequate foreign private capital when they need, but face the pangs of contagion once it starts. As noted by Goodhart and Huang (2000), "when there is an international inter-bank market, the total amount of necessary liquidity to meet pure liquidity shocks to be provided by an ILOLR is smaller than that when there is no international inter-bank market". The risk of contagion, however, increases when the international inter-bank market exists and grows in importance over time. An ILOLR, therefore, is required to deal with the uncertainty arising from the failure of international inter-bank market.

Section III

Empowering the IMF as an ILOLR.

Lack of adequate resources has been a major factor that severely constrains the IMF to assume the role of an ILOLR. Further financial empowerment of the Fund, however, was deliberately stalled in the Twelfth General Review of Quotas. Sufficient financial empowerment of the Fund is crucial for enhancing its capacity to undertake credible crisis prevention and resolution measures. It needs to be noted that crises in the capital account have completely changed the dynamics of demand for Fund resources. A small number of large arrangements can take the Fund to exhaustion limit. Past recent arrangements in Mexico, Thailand, Korea, Russia, Brazil, Argentina and Turkey suggest that individual arrangements were much in excess of what any quota linked financing would have warranted. For Korea and Turkey the arrangements were 1939 per cent and 1560 per cent of their respective quotas. Further, more than 70 per cent of the agreed resources under these high access arrangements were made available within one year. As several countries like Mexico, Brazil, Korea and Russia effected repurchases ahead of schedule, that relieved the pressure on Fund’s resource position to a great extent. But there are cases like Argentina that requested for an extension of repurchase and Turkey that financed its SRF obligations by using stand-by resources. It may be noted that if the above seven countries alone (with an aggregated actual quota of about 17.4 billion SDRs) get into a crisis simultaneously and require ten times their respective quotas under the high access arrangements, the Fund would not be able to meet the demand, even after tapping the entire amount available under General Agreement to Borrow and New Arrangement to Borrow (GAB and NAB). If these countries together request for extension of repurchase, the Fund’s revolving financing mechanism itself may completely breakdown. There is a high degree of discretion and flexibility in dealing with arrangements under exceptional access facilities like the SRF and there has been evidence of high degree of credit concentration in few countries in the 1990s, which may also threaten the financial position of the Fund given the mandated need to safeguard its resources. It was just a coincidence that 88 per cent of the arrangements entered into after 1995 were within the normal access limits of stand-by and EFF. If the number of exceptional arrangements increases in future and if the CCL is also activated after suitably streamlining the scheme, Fund’s present level of resources may appear grossly inadequate.

The costs associated with any shortfall in Fund resources to deal with future crises are also asymmetric as far as the creditors and debtors are concerned. The creditors, who may like to delay any early augmentation in Fund resources, have little to loose even when they agree to an augmentation because they earn a remuneration at market related rate and make available their resources only when needed by the Fund, depending on the Financial Transaction Plan (FTP) for any quarter. The potential costs for the debtors stemming from inadequate Fund resources, however, could be substantial, particularly in a globalised system where channels of contagion are stronger and more disruptive.

The argument that an early augmentation of fund resources may raise the moral hazard problem and disrupt the ongoing work on PSI/SDRM is not entirely correct. Only a financially stronger and more credible IMF can broker deals between debtors and private creditors; the incentive to enter into any IMF brokered deal will be higher for both private creditors and debtors when the size of Fund financing under an arrangement would be expected to be higher. The debtors can be better disciplined and the private sector can be more effectively bailed-in to share the burden of adjustment during the stabilisation programmes only when the Fund will have the resources and the willingness to extend financial assistance to countries depending on their need during a crisis. Again, on the ground of containing moral hazard, stiffer conditionality should not be attached to higher access facilities like the SRF. Rather the catalytic role of the Fund be strengthened so that Fund’s resources can be better safeguarded by creating the fear among the debtors that unless the repurchase schedules of the IMF are met, the access to private capital could be significantly constrained. PSI and SDRM can help in strengthening the international architecture for crisis prevention and resolution. But they cannot be viewed as substitutes of what an ILOLR is expected to do.

It may be correct that Fund’s resources proved sufficient to deal with the series of emerging market crises in the 1990s. But in the absence of the 45 per cent augmentation through the Eleventh Review, one is not sure whether the condition would have been the same. Fund resources aim at dealing with uncertainty and a key fundamental fact of the globalisation process in the recent years has been greater uncertainty.

Other than the argument based on possible disruption of the SDRM/PSI initiative, several other arguments have also been offered to stall any augmentation of the Fund’s quota-based general resources. One view would suggest that the future demand for Fund resources is difficult to quantify in view of the uneven behaviour of the key determinants of the quotas (i.e. the variables used in alternative quota formulas) in the recent period, as also the heightened uncertainty about the timing and the nature of future crises. Another view may emphasise that financial globalisation has led to increasing reliance on private capital. Hence, mechanisms that could ensure greater participation of the private creditors in crisis prevention/resolution may be more credible and effective than a mere augmentation of Fund resources. The Fund could contribute to strengthen the effectiveness of such mechanisms by adapting its surveillance policies more appropriately, ensuring improved vulnerability assessment, greater transparency, compliance with international standards and codes, and rigorous external assessment of the strength and resilience of national financial systems. It may be correct that the Fund’s current liquidity position has generally been viewed as adequate by historical standards. But, as noted by Fisher (1999), " if the IMF were today the same size relative to the output of its member States as it was in 1945, it would be more than three times larger …If the quota formula applied in 1945 were used to calculate actual quota today, the Fund would be five times its size; and if the size of the Fund had been maintained relative to the volume of world trade, it would be more than nine times larger – that is, the size of the Fund would be over 2.5 trillion dollars ( as against current total quota size of SDR 212.7 billion, or about US $ 290 billion)." The Fund size may have to be even larger if one takes into consideration the demand for Fund resources arising from capital account crises, since private capital flows were not that very important in 1945 as they are today. One also has to keep in perspective that even though newer facilities like SRF and CCL have been introduced/contemplated, they compete with the conventional facilities (like stand-by and EFF) for the general resources of the Fund. Any new facility, however, must be additionally funded –whether through augmentation of the general resources, or through creation of separate Funds as has been done in the case of Poverty Reduction and Growth Facility (PRGF).

If the IMF has to function as an ILOLR, the polar alternatives would indicate that it should be in a position to either "issue an indefinite amount of its own currency that could be widely acceptable at the global level" or "it must have a pool of reserve currencies which is large enough to deal with international liquidity crises". In case of the latter, the key issue is how large should be the pool and what factors should decide the size of the pool. An important consideration in this respect is the distinction that is often made between "lending in last resort as an input to monetary policy" and "lending in last resort as an input to banking policy" (Jeanne and Wyplosz, 2001). In case of the former, monetary and financial stability of the system becomes the overriding objective. Hence, ensuring availability of liquidity to the system as a whole, not just to any specific institution, is what a national LOLR intends to achieve. Through open market operations, liquidity is injected to the system. Institutions that are solvent, get liquidity from the market and not from the LOLR. Insolvent institutions, however, are allowed to perish. This helps in addressing the moral hazard problem ( i.e. the problem of insolvency being wrongly interpreted as illiquidity by the LOLR). In the global context, for an ILOLR to inject liquidity into financial systems of the affected countries, whether directly or through national central banks who would be conducting the Open Market Operations (OMOs), it must be in a position to issue its own liquidity. This suggests that only the Federal Reserve (Fed) in the USA can operate as the ILOLR.

This option would not meet the requirements of all emerging markets for a number of reasons. If the Fed injects enough liquidity through the global financial market, many countries may not get the desired amount of international liquidity due to the reasons stated earlier, particularly the observed lopsided pattern of capital flows and the manner in which the inter-bank market for global liquidity functions. Most importantly, if the Fed functions as the ILOLR, it is discriminatory treatment that will become the rule. Unlike Mexico, not many countries could get US $ 12 billion from the Exchange Stabilisation Fund. Nor can they hope to get the type of support that was extended to Korea - an OECD country (Korea received a total package of US $ 57 billion – of which US $ 21 billion that came from the IMF represented 1939 per cent of Korea’s quota, another US $ 14 billion was provided by the World Bank and the ADB, and the G-7 agreed to arrange a second line of defence of about US $ 22 billion). Thus, it is the bilateral relationship with the US rather than the genuine liquidity needs of solvent countries that would guide the functioning of the Fed-run ILOLR.

Due to the inability of the inter-bank market to discriminate solvent from the insolvent, the Fed has to operate through the national central banks, who in turn would inject the dollars made available by the Fed into the system (not to specific institutions) against high quality collaterals (like domestic currency denominated bonds and securities). As the objective would be to deal with a liquidity crisis, over time, the swap could be reversed (i.e. bonds and securities could be sold and dollars acquired when normalcy restores in the system) and the ILOLR be paid back. Under this arrangement, effectively the ability of a central bank to operate as a LOLR would be enhanced through liquidity support from another central bank, most importantly in foreign exchange. The key dilemma as per this variant of ILOLR is that why should the Fed take up this responsibility: (i) unless its own economic linkages with a country are strong enough to cause systemic concerns for its own financial system, and (ii) when IMF arrangements in any case bailout the creditors from developed countries at the expense of tax payers of the countries who use conditional Fund resources. The official position of the US has been that as against its 17.7 per cent share in IMF quota, its share in the usable resources of the Fund is as high as 26 per cent, and its share in GAB credit line is also about 25 per cent. The US has also made available these resources to the Fund at more favourable rates than the cost of money to the Government (i.e. US Treasury rates), involving a subsidy of hundreds of millions of US dollar per year. IMF remunerations also do not adequately reflect the increased riskiness of IMF lending (Frenze and Keleher, 1999). The essence of this official position is that the US already contributes substantially to the IMF. It would assume additional responsibility, it appears, only if that can help in serving its national interest. There are ample evidences of that as well. When Mexico was extended US $ 12 billion from the Exchange Stabilisation Fund at a penal rate, the Treasury got back the entire amount over a short time span fetching an interest income of about US $ 500 million. When the IMF bails out a country in crisis, it is the private creditors of advanced countries whose imprudent lending behaviour is rewarded and their profit prospects are protected. Crisis also opens up the window to acquire banks and corporates of the crisis affected countries at depressed prices. An ILOLR role for the IMF has been discarded by the advanced countries on the ground that the ILOLR, with capacity to create international liquidity, can give rise to a situation of excess liquidity. It has been a fact, however, that the reserve currency countries often pursue easy monetary policy stance keeping in view the national growth concerns; but such easy national policies have also given rise to excess international liquidity. The Asian crisis was a manifestation of such excess liquidity created by the advanced countries, which were redirected by private creditors to the East Asian countries in search of higher return. Excess liquidity argument has, thus, been conveniently used to stall an ILOLR role for the IMF, though the global financial system has already experienced situations of excess liquidity even without an ILOLR.

In case of lending in last resort as an input to banking policy, on the other hand, failing banks/institutions are extended liquidity support directly. Every country maintains some variant of domestic banking safety nets and an ILOLR has to only supplement the national efforts. Hence, the resource requirement of the ILOLR need not be large, possibly should be just sufficient to meet the liquidity gaps in the domestic banking sector (liquidity gap defined as the difference between the banking sector’s short-term foreign exchange liabilities and their liquid foreign exchange assets). Under this arrangement, the ILOLR could finance the discount window of the domestic central bank or directly offer limited guarantees on foreign exchange liabilities of domestic banks. Solvent banks could get dollars from the central bank under the discount window against good collaterals. To enhance the effectiveness of this arrangement one has to address the agency cost problem (given that the domestic LOLR has to operate as an agent on behalf of the ILOLR -the principal- under this arrangement). As per the fear of agency problem, national LOLRs can bail-out insolvent domestic banks when the ILOLR finances the operation. Over a period of time, the ILOLR has to be paid back by the domestic LOLR. But to safeguard the interest of few insolvent domestic bankers who otherwise should have been allowed to perish, the authorities may transfer the burden to domestic taxpayers while repaying to the ILOLR. To deal with such possible imprudent behaviour of national authorities, the ILOLR may have to possess the maximum possible information about each and every bank/financial institution (implying associated intrusion into regulatory and supervisory frameworks of national governments).

One needs to recognise in this regard the fact that the ILOLR has to necessarily deal with the sovereign government, and the IMF has already strengthened its tools of surveillance to identify the country specific vulnerabilities. Along with regular Article-IV bilateral surveillance, Financial Sector Assessment Programmes (FSAPs) and preparation of Reports on Observance of Standards and Codes (ROSCs) should provide enough information about the member countries, and it should not aim at obtaining information on each bank/financial institution while providing liquidity support. Macro level information should be sufficient for the ILOLR to lend to the domestic LOLR, who in turn could monitor individual banks and decide whether to bailout the solvent/insolvent banks. Decision to bailout an insolvent bank at the expense of tax payers is a decision that is internal to every national economy, and even in the absence of an ILOLR, such decisions would continue to be taken, though ideally should be avoided. While functioning as an ILOLR, thus, only the relationship between the ILOLR and the LOLR should be important and as long as a country maintains clean track record of honouring all Fund obligations on time, the ILOLR should not be guided by the consideration agency problem.

In practice, the world has seen ILOLR interventions of both variants in the past. In the aftermath of the December 1994 Mexican crisis, US $ 12 billion was made available in January 1995 to Mexico by the US from its Exchange Stabilisation Fund (which was initially created with the objective of stabilising the external value of the US dollar, not Mexican peso). Effectively, the Fed assumed the role of an ILOLR. In other cases, IMF has provided conditional resources to national authorities to deal with twin crises. When its resources proved insufficient, additional resources were mobilised under the General Agreement to Borrow (GAB) and the New Arrangement to Borrow (NAB). Under the GAB, IMF arranged SDR 6.3 billion for Russia in July 1998 and again in December 1998 it arranged SDR 9.1 billion under the NAB for Brazil.

For the IMF to take over the second variant of ILOLR function, sufficient augmentation of its general resources appear critical. Even though the aggregate quota of all members amount to about SDR 212 billion, usable resources at the disposal of the Fund at any point of time are about 30 to 40 per cent less since a large part of the quota based resources are not usable. NAB and GAB can supplement the Fund resources to a maximum of SDR 34 billion. The IMF has the option of assuming the ILOLR function of the first variant also, as under Article XVIII it can allocate SDRs "to meet the long-term global need, as and when it arises, to supplement existing reserve assets". India has also raised this issue on several occasions that IMF could issue SDRs to itself to augment its resources at the time of need and relinquish the additional liquidity so created as and when the member countries effect the repurchases (Jalan, 1999). The complex issues involved in implementing this suggestion are: (i) 85 per cent majority support, (ii) de-linking allocation of SDRs from quotas and linking the disbursement to a country’s liquidity needs, and (iii) the overall international acceptability of SDR when larger volumes of SDRs are created (so far the cumulative allocation amounts to only SDR 21.4 billion, even though the initial objective was to enhance the role of SDR as the principal reserve currency). If this arrangement can be implemented, IMF can effectively create unlimited liquidity and support the national initiatives in bridging any liquidity shortfall.

There is also a perception that lending by the IMF against good collateral may not be feasible (excluding exceptional cases, like the option to use future oil revenue of oil exporting countries as good collateral). IMF has to, therefore, lend against policy, implying thereby the fact that an ILOLR can only function effectively if greater policy intrusion becomes acceptable to the members. As mentioned earlier, IMF has already expanded the scope and coverage of its regular surveillance of members and the spectre of conditionality already haunts the members which have accepted IMF programmes in the last decade or so (Pattanaik and Misra, 2002). ILOLR lending against policy, therefore, is already in vogue. As mentioned earlier, the ILOLR need not seek institution/bank-wise information to ensure that only solvent institutions get the liquidity support from the ILOLR, since that will be the role and purview of national LOLRs.

Section IV

Concluding Observations

The role played by the IMF in dealing with the series of emerging market crises in the last decade has led to a near polarisation of international perceptions on the need for an ILOLR. While the advanced countries feel that the IMF is too generous and that it creates moral hazard through its generous lending which is detrimental to the smooth functioning of the international financial markets, the emerging market economies generally feel that the IMF provides too little too late, and that it is too severe on crisis countries that often gives rise to economic recession. Reflecting the concern of the advanced countries, financial empowerment of the IMF has been deliberately stalled. At least, such a decision has been postponed till the progress on SDRM/PSI becomes clearly visible. The argument suggests that a stronger crisis prevention architecture would reduce the need for an ILOLR. The experience of the emerging markets, in turn, suggests that inter-bank markets do not allocate international liquidity as per the liquidity needs of countries, and that contagion and uncertain behaviour of private markets that often give rise to rush for exit entail large costs for the countries that participate in the process of globalisation. An ILOLR can make the globalisation process smoother and less painful, even if that gives rise to the problem of moral hazard. Despite the risk of moral hazard, every country has its own national LOLR. In the international context, the ILOLR has to necessarily operate through the national LOLRs while providing the liquidity support, and it should be up to the national LOLRs to address the liquidity problems of national financial institutions. Unlike the constructive ambiguity that characterises national LOLRs, the policies of the ILOLR, however, should be transparent so as to avoid the scope for misuse arising from unfair discrimination of countries.

Initiatives that can strengthen the global crisis prevention and resolution architecture, such as SDRM/PSI, stronger Fund surveillance, private Contingent Credit Lines, declaration of Fund approved standstills, market friendly measures such as transparency and compliance with international standards and codes, and even appropriate exchange rate regimes and domestic safety nets should not be viewed as a substitute of an ILOLR; instead, the ILOLR should continue to have a role irrespective of the degree of progress on these initiatives. The IMF has already assumed the role of an ILOLR, and at times it has lent even to countries which are clearly identified as insolvent due to their unsustainable debt, such as the HIPC/PRGF countries. Lending under the SRF is not constrained by any access limit and such financing is also provided more quickly and at a penal rate. IMF lends against policy and conditionality serves the role of collateral. Irrespective of a country’s performance under conditionality, IMF obligations are always paid back and, in that sense, the implications of a default on Fund liability represent the best form of collateral against which the IMF can lend freely. Its capacity to lend freely, however, is constrained by its limited resources. A few systemically important emerging markets can take the IMF to the point of exhaustion if they together approach the Fund for assistance. The present policy of highlighting the current comfortable liquidity ratio as the indicator of adequate availability of resources with the Fund clearly disregards the potential for sudden spurt in demand that may arise in the eventuality of sustained and simultaneous crises in several systemically important emerging market economies.

References:

Bagehot, W. (1873), " Lombard Street – A Description of the Money Market", John Wiley and Sons, Inc. (New Edition in 1999).

Bordo, M.D. (1990), " The Lender of Last Resort: Alternative Views and Historical Experience", Federal Reserve Bank of Richmond Economic Review, Jan./Feb.

Calomiris, Charles W. (1998), " The IMF’s Imprudent Role as Lender of Last Resort", The Cato Journal, Vol.17, No.3.

Fischer, Stanley (1999), "On the Need for an International Lender of Last Resort" Speech delivered at the Joint Luncheon of the American Economic Association and the American Finance Association, New York, January 03.

Frankel, Jeffrey A. (1999), " International Lender of Last Resort", paper presented at the conference on Rethinking the International Monetary System, Federal Reserve Bank of Boston, June 7-9.

Frenze, Christopher and Robert Keleher (1999), Statement before the International Financial Advisory Commission, September 09.

Frexias, Xavier, Curzio Gianini, Glenn Hoggarth and Farouka Soussa (1999), " Lender of Last Resort: A Review of the Literature", Bank of England Financial Stability Review, November.

Giannini, Curzio (1999), " Enemy of None but a Common Friend of All", Princeton Essays in International Finance, No. 214, June.

Goodfriend, M. and J.M. Lacker (1999), " Limited Commitment and Central Bank Lending", Federal Reserve Bank of Richmond Working Paper.

Goodhart, Charles A.E. and Haizhou Huang (2000), " A simple Model of an International Lender of Last Resort", IMF Working Paper WP/00/75, April.

Guttentag, J. and R. Herring (1983), " The Lender of Last Resort Function in an International Context", Princeton Essays in International Finance, No. 151, May.

Herrala, Risto (2001), " An Assessment of Alternative Lender of Last Resort Schemes", Bank of Finland Discussion Papers, No.1

Humphrey, Thomas (1975), " The Classical Concept of the Lender of Last Resort", Federal Reserve Bank of Richmond Economic Review, Vol. 61.

Jalan, Bimal (1999), " Interim Committee Statement", IMF, April 27.

Kindleberger, C. (1989), " Manias, Panics, and Crashes: A History of Financial Crises", Basic Books, New York.

Litschig, Stephen (2001), " Blueprint for an International Lender of Last Resort: An Assessment", Graduate Institute of International Studies, November 27.

Mann, Catherine L.(1998), " An International Lender of Last Resort and the International Financial Markets", Institute for International Economics, April.

Mishkin, Frederic S. (2000), " The International Lender of Last Resort: What are the Issues?", paper presented at the Keil conference on the "World’s New Financial Landscape: Challenges for Economic Policy", June 19-20.

Niskanen, Mikko (2002), " Lender of Last Resort and the Moral Hazard Problem", Bank of Finland Discussion Papers, No. 17.

Pattanaik, Sitikantha and Sangita Misra (2002), " The Spectre of IMF Conditionality: Some Streamlining Options", The Indian Economic Journal, Vol.50, No.1.

Rochet, Jean-Charles and Xavier Vives (2002), " Coordination Failures and the Lender of Last Resort: Was Bagehot Right After All?", FMG Discussion Paper DP0408, Financial Markets Group.

Rogoff, Kenneth (1999), " International Institutions for Reducing Global Financial Instability", Journal of Economic Perspectives, Vol. 13.

Sachs, Jeffrey D. (1999), " The International Lender of Last Resort: What are the Alternatives?", paper presented at the Conference on Rethinking the International Monetary System, Federal Reserve Bank of Boston, June 7-9.

Saxton, Jim (1999), " An International Lender of Last Resort, the IMF, and the Federal Reserve", Joint Economic Committee Study.

Schwartz, A.J. (1997), " From Obscurity to Notoriety: A Biography of the Exchange Stabilization Fund", Journal of Money, Credit and Banking, No.29, May.

Schwartz, A.J. (1999), " Is There a Need for an International Lender of Last Resort?", NBER, Shadow Open Market Committee, March.

Thornton, H (1802), " An Enquiry into the Nature and Effects of the Paper Credit of Great Britain".

Tietmeyer, Hans (1999), " Systemic Risk and Lender of Last Resort Facilities", Speech delivered at the Center for Financial Studies, June 11.

Wyplosz, C. and O. Jeanne (2001), " The International Lender of Last Resort: How Large is Large Enough?", NBER Working Paper No. 8381, July.

---

* Sitikantha Pattanaik is Assistant Adviser in the Department of Economic Analysis and Policy of the Bank. The views expressed in the paper are those of the author and not of the institution to which he belongs.