Few industrial and emerging market countries over the past two decades have been immune to banking sector turbulence. Full-blown crises, in particular, have exacted a tremendous toll. The frequency and high cost of these episodes prompted the World Bank and the IMF to set up a Bank-Fund Financial Sector Liaison Committee to strengthen the exchange of information on financial sector issues and better coordinate their advice to member countries.
On January 11, the committee sponsored a seminar on managing banking crises to share their accumulated experience. The seminar, held at the World Bank, focused on the dynamics of banking crises and the lessons learned from efforts to resolve major episodes in Indonesia, Korea, Mexico, and Russia, and others. Among the recommendations of the seminar were the creation of a single agency to deal with a crisis, having authorities assume a private sector approach to the management of banking crises, and ensuring adequate resources are assigned to the resolution process.
The seminar, chaired by Stefan Ingves, Director of the IMF’s Monetary and Exchange Affairs Department, included presentations by Stijn Claessens, Lead Economist of the World Bank’s Financial Sector Strategy and Policy Department; Charles Enoch, Assistant Director, and David Hoelscher, Advisor, of the IMF’s Monetary and Exchange Affairs Department; Fernando Montes-Negret, Sector Leader for the World Bank’s Latin American Finance Group; and David Scott, Advisor in the Bank’s Financial Sector Department. Following is a summary of the seminar presentations. The slide projections that accompanied these presentations are available on the World Bank’s website (http://www.worldbank.org/finance).
Dynamics of banking crises
In an opening presentation, Claessens outlined the scope and structure of banking crises, and detailed their global reach and the significant costs involved in recapitalizing and restructuring banking systems and, often, in coping with major output losses. After surveying the macroeconomic and microeconomic missteps that can lead to a banking crisis, he discussed the "phases of distress" of a crisis. These he segmented into a containment phase (a critical period when appropriate steps can halt the spread of a banking crisis); a restructuring phase (when various institutional, rehabilitation, and recapitalization measures are called for); and a reform phase (when the deeper causes should be addressed through fundamental and necessarily longer-term reforms).
Claessens further divided the containment and restructuring phases into discrete stages. The containment phase, he noted, has an onset stage when bad financial conditions are accelerating; a period of open crisis, which sometimes features bank runs and often includes large liquidity support, currency crises, and interest rate spikes; an effort to limit losses, which frequently entails suspending weak banks, imposing con-servatorships, and issuing government guarantees of depositor holdings; and a stabilization program.
The restructuring phase, he said, typically has five stages: diagnosis, in which both the scope and the source of the problems are determined; development of appropriate institutional tools and adaptation of the legal framework as needed; creation of a strategy that outlines the authorities’ vision of a reformed banking sector; loss allocation and government support (including such steps as write-offs, closures, and the use of public funds); and reprivatization and normalization of the financial system.
Indonesia’s banking crisis, Enoch noted, was characterized by particularly deep insolvency, a large number of banks in the system, limited management capacity, and little scope for resolution through outside acquisitions. In November 1997, in the face of widespread runs on the banking system, the authorities closed 16 banks (3 percent of the banking system). The closures themselves were handled efficiently, but there was a lack of clarity as to why the banks had been selected. The authorities failed to demonstrate commitment to the concomitant economic program, and the limited deposit insurance scheme prompted concerns about losses among large depositors. Over the following weeks runs on the banks resumed and became pervasive, leading to a generalized run on the currency.
At the end of January 1998, and facing imminent financial meltdown, the authorities announced a blanket guarantee for all depositors and creditors and established the Indonesian Bank Restructuring Agency (IBRA). In mid-February, IBRA intervened in 55 banks that had borrowed more than twice their capital from the central bank. But this very "soft" open bank resolution, which did not replace owners and managers or announce these steps to the public, did little to stem the crisis. The next month, IBRA took over seven large banks that had borrowed heavily (over $240 million), changed management and suspended owners’ rights. IBRA also closed seven small banks that had borrowed over 500 percent of their capital, subsequently closing three and identifying the largest remaining bank as a “platform bank” into which the remaining banks of this group would be folded. Finally, in March 1999, the authorities were able to take a comprehensive approach to resolving the remaining private banks. They categorized 74 small banks as sufficiently strong to survive unaided; jointly recapitalized, with the owners, 9 middle-performing banks, and closed 38 very weak banks.
Enoch observed that banking system resolution typically involves both bank closures (where the doors are closed and assets and liabilities transferred) and open bank resolution (where ownership and management generally change but the bank stays open). In either case, he said, intervention is only the first stage, and the final cost of intervention will be determined as much by the efficiency of follow-up activities. The authorities’ approach will evolve as information becomes available, but initial interventions are likely to be on banks with protracted runs or deep insolvency, or where there is evidence of fraud.
For the succeeding period, Enoch suggested, the authorities may have to rely on indicators of liquidity, which is a good proxy for insolvency in conditions of limited information. They may also need to contain central bank liquidity support to avoid losing monetary control. Outside auditors can help provide a fuller picture of banks’ underlying condition and provide the basis for identifying banks strong enough to survive without official support, banks who should be given official support under stringent conditions, and banks with no viable future that should be closed. Overall a banking crisis will lead to fewer banks and—at least temporarily—an increase in the share of state ownership, he noted.
According to Enoch, Indonesia’s experience underscores the importance of uniform and well-explained criteria for invention. It also suggests that when there is fear of systemic problems, the introduction of blanket guarantees, despite their potential cost and moral hazard implications, is likely to be a part of any resolution strategy. The choice between closure and open bank resolution is likely to have to be determined case-by-case and will vary across countries, but where there is pervasive insolvency, many banks, and evidence of fraud, Enoch concluded, bank closures are likely to be part of the program.
According to Montes-Negret, the roots of Mexico’s 1994 banking crisis lay in the government’s emphasis on maximizing revenue from the privatization of banks that already had portfolio problems. Banks that were nationalized in 1982 had long ago lost private sector initiative and their best staff. When Mexico reprivatized these banks in 1991-92, excluding foreign banks, it awarded licenses without proper tests of new and unproven bankers and permitted a number of banks to be set up with little capital and highly leveraged assets. A 1991-94 credit boom, a large number of foreign-currency-denominated loans, inadequate accounting standards, weak supervision, and distorted incentives set the stage for a banking crisis even before political and economic shocks led to a run on the currency in late 1994.
Prolonged denial and very poor accounting standards characterized the first phase of Mexico’s banking crisis. Montes-Negret cited serious conflicts of interest in which supervisors also directly involved themselves in bank restructuring deals. Efforts to attract foreign investment led to great deals for foreign banks. Also, he said, some insolvent banks continued to operate far too long and at great fiscal cost. Bank managers were replaced too slowly, even when there had been looting of bank assets. Lax legal arrangements induced nonpayment, and little information about the depth of the crisis was available to other government agencies. One positive feature, however, was that Mexico experienced no runs on its banks.
A second phase in 1997 saw the introduction of a new accounting framework, but the resolution process became overpoliticized in 1998, with political discussion continuing while costs continued to rise. A more decisive phase began in 1999 with the formation of a new resolution and deposit insurance agency.
Among the chief lessons are, Montes-Negret said, that hiding a problem does not solve it, and as Mexico has become more democratic, there has been more disclosure. Further, supervisors entered into complex ad hoc deals without a global framework, resulting in far too much discretion and excessive secrecy. Clearly also, he emphasized, a least-cash solution is not always the least-cost solution. Mexico’s accounting solution merely postponed the financial solution, and its overly complex "engineering," which provided government paper to banks in lieu of cash, actually worsened the banks’ cash-flow situation in some cases. Finally, he said, beware of early declarations of victory and of tailoring a solution to fit the cash on hand.
The Russian banking system grew with astonishing rapidity, Hoelscher noted, burgeoning from 10 banks to 1,600 between 1992 and 1997. With most banks meeting Russia’s prudential regulations, the authorities assumed that all was well in the banking system, but existing statistics did not reflect actual conditions, he said. The banking system was highly concentrated (10 banks held 80 percent of the system’s assets), and the portfolios of these large banks were heavily concentrated (almost 60 percent were concentrated in GKOs, the Russian government bonds) and had significant exposure to foreign exchange risk.
While a forced restructuring of the GKOs and exchange rate depreciation provided the immediate impetus for the 1998 banking crisis, its roots, Hoelscher suggested, lay in the banks’ significant level of nonperforming loan portfolios, as well as their vulnerability to exchange rate volatility and the heavy concentration in GKOs. The crisis immediately left large banks illiquid. As a result, deposits were frozen, the payment system ground to a halt, and banks defaulted on forward exchange contracts and foreign debt-service obligations.
But the Russian authorities, who remained unconvinced that this was an insolvency problem, called for a moratorium on foreign debt payments and permitted households to shift deposits to Sherbank, the state-owned savings bank. In the absence of reliable data and in the face of severe budget and institutional constraints, a joint IMF-World Bank team of advisors identified the need to collect more accurate data on the major banks, strengthen the legal framework for bank restructuring and liquidation, and establish an institutional framework for restructuring. A World Bank-financed due diligence review subsequently found 15 of the 18 major banks deeply insolvent, with a net negative worth of 173 percent of assets (in individual cases, net negative worth was the equivalent of 400 percent of assets). The review also found that the default on the GKOs was not the principal culprit—provisions for nonperforming loans and foreign exchange losses figured much more prominently.
As a result of this review and a series of technical assistance missions and activities related to the use of IMF and World Bank resources, the Central Bank of Russia removed the licenses of 6 of the 18 banks and put 3 others under management of a newly established bank restructuring agency, Hoelscher explained. Both bank bankruptcy and bank restructuring legislation were signed into law in 1999. In addition, the central bank began modernizing supervisory regulations and consolidating and improving the management of these supervisory functions.
A joint World Bank-IMF mission has developed an action plan for the coming year that calls for continued consolidation of the banking system, development of a core banking system, and improvement of the operating environment, including continued modification of prudential regulations and a revision of bank disclosure requirements.
The Russian crisis, Hoelscher indicated, taught the IMF and the Bank that authorities’ perceptions can critically influence the sequencing of reform measures and that close Bank-IMF cooperation can produce a more coherent and more successful program.
Minimizing the cost of crises
Scott described the approach taken in advising Korea on how to reduce the costs of crises. These costs, he explained, grow with time and are to a large extent determined by the behavior of debtors, creditors, and the authorities. Often unnecessary bailouts and theft also contribute to the costs. While governments can do much to control costs, he argued, they are normally not well equipped to do so and frequently get it wrong. He recommended the use of a specialized team focusing exclusively on resolving the crisis. The team should view its job as akin to investment management, with the goal of minimizing the amount of government outlays needed to get the job done and maximizing the value of that investment over time. Ultimately, he said, the team can get the best results by acting like private investors.
Scott asserted that acting like private investors would entail, among other things, moving fast to minimize the negative carry associated with nonperforming assets, avoiding bailing out those who can service their debt, avoiding moral hazard (which, he said, arises when the authorities allow support to be squandered and then provide additional rounds of support), and stopping the looting of banks by getting inside the banks, both before and after support is given. Governments can maximize their investment (that is, their support) by promoting a corporate debt restructuring that is sensitive to growth and does not leave banks hobbled, ensuring that banks carry out operational restructuring, and leveraging the crisis to achieve structural change that will make the market more attractive for new investment by the private sector.
Scott reiterated the well-understood policy that government support should be provided to banks only after existing shareholders have absorbed existing losses. These recapitalized and perhaps nationalized banks should then be sold as rapidly as possible. Pursuing rapid sales, he emphasized, gives focus to all restructuring actions, makes transparent the true financial situation of the bank, and forcefully demonstrates the intentions and resolve of the authorities.
Finally, Scott stressed that having access to adequate resources for investing in banks and assets was essential to doing the job right the first time. Failure to provide adequate support will force the government to intervene again, he said, triggering moral hazard and raising long-run costs.