Chapter

3 Lessons and Elements of Best Practice

Editor(s):
William Alexander, Jeffrey Davis, Liam Ebrill, and Carl-Johan Lindgren
Published Date:
August 1997
Share
  • ShareShare
Show Summary Details
Author(s)
Claudia Dziobek and Ceyla Pazarbaşioḡlu 

Chapter 3 provides background material for Chapters 1 and 2. It analyzes the experiences of selected countries that have undertaken systemic bank restructuring in recent years, with a view to drawing lessons about successful restructuring strategies and best practices. “Best practices” are defined as those that are observed, based on experience over time and across a broad group of countries, to contribute to a successful bank-restructuring outcome. Accordingly, a basic requirement of best practices is that they should be robust to a reasonably wide range of conditions faced by restructuring countries. At the same time, best practices also involve being able to modify elements of the restructuring strategy to conform to the particular circumstances of the problem in each individual country.

The determination of best practices in systemic bank restructuring is an ongoing process that is being shaped to an important extent by the IMF’s experience in dealing with particular countries. This volume reflects the IMF staff’s experience in technical assistance in restructuring countries as well as extensive discussions with experts on the subject. Both the detailed analysis of country cases and the broader empirical analysis of the characteristics of systemic bank restructuring in a selected sample of countries undertaken in this chapter are intended to build on that effort. Nevertheless, systemic bank restructuring is an area of operational work where there are few guiding principles. Accordingly, the IMF will continue to review experiences with systemic bank restructuring to refine aspects of the approach.

The eight case studies analyzed here have been selected to illustrate either particular aspects of best practices, or the consequences of not following best practices. Additional evidence on best practice policies is provided through a statistical analysis of the experience in a sample of 24 countries that have undergone systemic bank restructuring during 1980–92. The intent is to confirm that the conclusions drawn from the case study analysis are valid across a broad group of countries, so as to ensure that the various conclusions are robust and applicable to a wide range of circumstances. Unless tests of robustness are performed, there is clearly a risk that what is inferred as best practice is only valid in the particular circumstances of an individual country. Finally, a summary of best practice policies that can be confirmed from both the case studies and the statistical analysis is provided.

Country Case Studies

The countries selected for the case studies include Chile; Côte d’Ivoire; Mauritania; the Philippines; Poland; Latvia; Spain; Sweden; the United States; and the Baltic States, Russia, and other countries of the former Soviet Union; and Sweden.

Côte d’Ivoire: Importance of a Comprehensive Approach

In 1987, prior to bank-restructuring operations, Côte d’Ivoire’s financial system consisted of 5 state development banks and a fairly diversified system of 14 commercial banks with a developed national branch network. Four commercial banks, of which three are affiliated with major French banks and one is state owned, accounted for 90 percent of deposits in 1991. Côte d’Ivoire is a member of the West African Economic and Monetary Union (WAEMU) whose member countries share a common currency and a common central bank, the BCEAO (Central Bank of West African States). Due to sharply declining world market prices for cocoa and coffee, Côte d’Ivoire suffered a deterioration of more than 40 percent in its terms of trade between 1986 and 1990.

The deteriorating economy had massive repercussions on Côte d’Ivoire’s banking sector, leading to a rapid rise in arrears by public and private borrowers, deposit withdrawals, and decreased lines of credit from foreign parent banks. Rigid interest rate structures, weak banking regulation, and lax supervision added to these problems. A deficient legal frame-work and judicial system prevented the banks from initiating action against delinquent borrowers. Well before the weakening economic situation began to affect the commercial banking sector, Côte d’Ivoire’s five development banks were plagued by high levels of nonperforming loans. Failed attempts to rehabilitate these banks in the mid-1980s led the authorities to initiate liquidation procedures for two housing and two industrial development banks in 1988.

In addition to the problems of development banks, the dramatically worsening performance of the commercial banks began to surface in the late 1980s. Between 1988 and 1990, deposits of the commercial banks declined 22 percent while credits from foreign banks fell 14 percent. By December 1990, roughly half of the commercial bank credit portfolio was nonperforming. This decline resulted in growing illiquidity and insolvency of these banks, partly offset by substantial credit from the BCEAO.

To avert further deterioration and a potential outright crisis in the commercial banking sector, the government developed a comprehensive restructuring plan for the financial sector. Implementation began in 1991 in conjunction with a Financial Sector Adjustment Program from the World Bank. The program provided for the rehabilitation of the commercial banking sector through the settlement of government arrears, recapitalization of banks—mostly by private owners—and operational restructuring. Prior action by the authorities included introducing legislative reforms to facilitate the management and foreclosure on nonperforming loans. Prior actions also included a move to market-based monetary policy instruments by the BCEAO.

The only remaining development bank (Agricultural Development Bank, BNDA) was placed in liquidation in September 1991, and an international expert was hired to oversee the liquidation. The private depositors of the BNDA were only partially refunded, and a portion of the refunds was contingent upon successful loan recovery. The signaling effect of these actions was considered an important part of the bank-restructuring strategy. For instance, the owners of each of the insolvent commercial banks were presented with the option of either closure and liquidation or an obligation to inject fresh capital. Thus, the government did not protect any bank under a “too-big-to-fail” provision.

A centralized, government-owned asset recovery unit was established in late 1992 with a mandate to recover nonperforming loans from the liquidated state banks. As expected, the agency’s recovery activities were not very successful, owing in part to the bankruptcy of major debtors, inadequate credit files, and slack demand in the export and real estate markets. Between May 1993 and June 1995, about 45 percent of targeted recoveries were realized (less than 4 percent of the outstanding portfolio). Net proceeds of the recoveries were used to repay depositors of the institutions affected. In addition, a special government fund was used to settle a large share of the payments due to private depositors of the agricultural development bank. The government is now being partially reimbursed through loan recovery from former BNDA debtors.

The settlement of the government’s arrears to commercial banks (about 7 percent of 1990 GDP) was a major component of the recapitalization program. About 20 percent of the settlement of government arrears was accomplished through a series of cash payments from the government to the four large batiks starting at the end of 1991. The remainder was securitized over 15 years at 3 percent interest with a 2-year grace period. The concessional interest rate was accompanied by a special refinancing facility at the BCEAO, under which up to 90 percent of the new bonds could be discounted at a 3 percent rate, compared with the normal discount rate of 11 percent. The banks immediately made full use of this facility. Thus, central bank refinancing was an important element of the bank-restructuring strategy. The recapitalization of the largest banks was concluded in mid-1992, with both private parties and government contributing their shares to the absorption of past losses, in part through cash payments. Government support was limited to the equivalent of 20 percent of bank capital, thus, providing the incentive for private owners to contribute capital. As a result of the recapitalization, the liquidity position of the banks improved and lending resumed on a limited scale in 1992.

Besides settling government arrears, the restructuring programs for the four largest banks included a number of components supporting the banks’ operational restructuring. Reductions in staff, branches, and overall operating expenses took place, and management was changed. While public sector loans were settled, nonperforming private sector loans were provisioned, and commercial banks took responsibility for collecting on the nonperforming loans of private borrowers. Past losses were absorbed by each shareholder in proportion to his share of bank capital. Thus, capital was restored to levels consistent with prudential requirements mainly by private owners. Realization of these measures, which forced bank owners to assume part of the bank restructuring costs, was greatly facilitated by the fact that three of the four major banks were subsidiaries of large foreign banks that were solvent and hence able to provide the needed additional capital.

The package of restructuring measures was designed not only to address the immediate problems of the banks but also to improve the regulatory, accounting, fiscal and legal, and monetary framework. Banking regulation and supervision were tightened through the new banking legislation introduced in 1990. Prudential rules were strengthened in mid-1991, including the introduction of new norms on capital adequacy, liquidity matching, and lending limits to single borrowers. A regional Banking Commission was formed to conduct on-site and off-site supervision for the member countries of the WAEMU. An extensive monitoring system was set up in cooperation with external donors. The monitoring included regular bank audits, prudential enforcement, and other elements of the restructuring strategy.

In the area of legal reforms, new legislation strengthened the banks’ ability to collect on collateral and to foreclose on delinquent borrowers. Court fees charged by the state amounting to 5 percent of the collateral value payable prior to the collection procedures were reduced. Banks were exempt from transfer fees for the acquisition of buildings subsequent to foreclosing on a guarantee. These reforms were completed by 1992, thus, laying the groundwork for improved bank performance. A set of arbitration procedures was included in the civil code. In 1993, recovery procedures for commercial and civil debt were also simplified.

During the first three years after the onset of bank restructuring, macroeconomic recovery proceeded slowly, with real GDP growth stagnating. The fiscal deficit declined, though not steadily, from a peak of 17 percent of GDP in 1989 to about 8 percent in 1994. Inflation remained at low levels, only rising to 4 percent in 1991. Strengthened macroeconomic policies and structural reforms, coupled with the devaluation of the CFA franc in January 1994, put Côte d’Ivoire back on a sustainable growth path beginning in that year.

An Enhanced Structural Adjustment Facility (ESAF) arrangement with the IMF was concluded in 1994. The adjustment program supported by the ESAF was designed to reactivate the economy, to improve public finances, and to restore price stability after a sharp rise in inflation to 32 percent in the aftermath of the devaluation. It was specifically designed to support the authorities’ bank-restructuring efforts and financial market development. An explicit goal of the program to strengthen macroeconomic performance was to enhance the financial results of the national recovery agency.

Despite the severity of the adverse terms of trade shocks that Côte d’Ivoire experienced in the late 1980s, the combination of a comprehensive bank-restructuring strategy and a vigorous macroeconomic adjustment effort was relatively successful over the medium term in restoring the conditions for both a sound banking system and a sound macroeconomic performance. From 1992 to 1994, banks’ return on capital rose to almost 10 percent, operating costs fell, and banks’ loan-pricing systems were revised so as to reflect actual risks. By 1994, most banks had returned to full compliance with prudential rules.

The experience of Côte d’Ivoire illustrates that a comprehensive approach is an essential ingredient of a successful bank-restructuring strategy. Clear incentives, including a firm exit policy, loss sharing, particularly on the part of private owners, and an emphasis on operational restructuring, loan workout, and continuous monitoring were central to the achievements of the restructuring package. While the macroeconomic conditions were weak in the initial phase, the economic recovery following the devaluation in early 1994 was an important underlying condition for restoration of banking sector soundness.

Mauritania: Consequences of Neglecting Operational Restructuring

In 1989, the banking system in Mauritania consisted of one development bank and four commercial banks, in each of which the state held a majority interest. The banking system was very fragile, mainly owing to weak portfolio management by the commercial banks and the mismanagement of banks following the nationalization of the early 1980s. In particular, since the economy was undiversified, banking system loans were heavily concentrated in the agricultural and fishing sectors and the distress in the banking system was intensified in 1988–89 due to adverse exogenous shocks experienced in these two major sectors.

With technical and financial assistance from the World Bank, a bank-restructuring program was implemented over 1988–90 with a view to transforming the commercial banks into viable and profitable institutions. To resolve the stock problem of the banks that had built significant overdrafts with the central bank, some 7.7 billion ouguiyas in treasury bills (9.4 percent of 1990 GDP) was issued and the liabilities of the banks to the central bank were cancelled. This operation was part of the measures to improve the balance sheets of the banks before their assets could be offered for acquisition by the private sector.

The program included measures to strengthen regulation and supervision by the central bank and adequate internal procedures for credit assessment. These measures were not effectively implemented, however, owing to a lack of political consensus on the need to address the problems of the banking sector. In effect, the bank-restructuring program mainly took the form of financial restructuring, with little attention given to operational restructuring. In particular, the program did not include (1) enactment of structural reforms regarding the accounting, legal, and regulatory framework, which would also help to strengthen prudential supervision; (2) establishment of proper incentives through conditionality measures for support programs as well as a firm exit policy; or (3) the establishment of loan workout schemes. Thus, important principles of a comprehensive restructuring package were not in place. As a result, banks continued to grant loans to the financially troubled agricultural and fishing sectors, failed to make appropriate provisions for delinquent loans, and generally did not attempt to improve the quality of their loan portfolio.

In the central bank’s capacity as supervisor of the banking system, it was unable to enforce compliance with prudential ratios and experienced other difficulties that prevented it from exercising effective supervision of banks, thereby weakening the motivation of banks to address their problems. Furthermore, the quality of the claims that the central bank accepted as collateral remained weak. The privatization plans and recovery of nonperforming loans continued to lag far behind plans, owing to long and costly legal procedures and difficulties in selling seized properties.

As part of the restructuring program, an interbank money market was created in 1990, allowing participating banks to adjust their liquidity positions. However, the higher level of transactions made possible by a situation of increased liquidity in the system was not sustainable as it was largely based on the accumulation of arrears by banks. The blocking of these arrears by the central bank in late 1991 led to a collapse of the interbank market and to a sharp decline in credit to the private sector.72

By the end of 1991, it was evident that the bank-restructuring efforts of 1988–90 had not succeeded. The two largest banks that already had been recapitalized were again facing a difficult financial situation. Furthermore, the problems in these banks had also spilled over to two other banks, and nonperforming loans had reached 50 percent of total outstanding credit to the private sector. The escalating banking sector problems, the buildup by the treasury of claims on banks in the form of uncleared checks, and the interruption of import financing due to the persistence of external arrears all contributed to the disruption of financial intermediation.

Within the framework of the 1992–94 ESAF arrangement and IMF technical assistance, the government undertook a second bank-restructuring program. The main elements were the closure and liquidation of the previously recapitalized government-owned development bank; the recapitalization and full privatization of the four commercial banks; the launching of a loan recovery program supported by enactment of new legislation; and a strengthening of the supervisory role of the central bank. The development bank was closed in June 1993, a liquidator was appointed, and rules for compensation were established. The government’s role in the banking sector was further reduced through the completion of privatization of the four commercial banks. These banks were recapitalized in late 1993 on the basis of the external audits undertaken with assistance from the World Bank.

An agreement was reached between the private shareholders of the fourth largest bank and the government, whereby the government relinquished its 46 percent share in the bank, gave up treasury claims in the form of uncashed checks, assumed responsibility for certain claims on public enterprises it had guaranteed, and compensated the central bank for its exposure to this bank. In exchange, the private shareholders committed to augment the bank’s capital as indicated by the audit (and did so by 1994) and to assume full responsibility for any future increases in capital that might be called for. In the meantime, a loan recovery law approved by parliament in February 1993 strengthened the rights of banks and lowered the cost of seeking recourse through the courts. Following enactment of the law, a loan recovery agency was established, and the loan recovery target for 1995 was achieved.

Overall, the budgetary impact of the second bank-restructuring program was estimated as UM 8.9 billion (7.6 percent of 1992 GDP) increasing the overall deficit of the consolidated government to 11 percent in 1993—up from 5.4 percent of GDP in 1992. The government incurred a revenue loss of about UM 700 million from relinquishing its claims on treasury checks; it paid out UM 1.3 billion in cash and UM 6.5 billion for the capital value of treasury bonds. While cash payments were made to individual depositors and international organizations, the public enterprises were issued treasury bonds. The cost of the liquidation of the development bank alone accounts for more than half of the overall costs—estimated as UM 5 billion (4.4 percent of GDP).

The case of Mauritania demonstrates that if financial restructuring is not accompanied by operational restructuring, a recurrence of banking problems is unavoidable. In contrast to the first bank-restructuring program, the second one stressed operational restructuring and structural reforms as central components. The main features of the latter included the establishment of loan workout schemes, enactment of structural reforms regarding accounting, legal, and regulatory framework and strengthening of prudential supervision, and establishment of proper incentives through conditionality measures for support programs as well as a firm exit policy. Since the second round of the bank-restructuring program took place in late 1993 and 1994, it is too early to judge its success. There are signs, however, that the banking sector increased its profitability in 1995.

Sweden: Role of Political Consensus and Need for Rapid Action

Financial sector difficulties surfaced in Sweden in late 1990, following sizable credit losses by a number of finance companies due to lending to real estate companies. The problems in the real estate sector spread to the banking sector in 1991 and intensified in 1992. The surge in loan losses was particularly abrupt, rising from 0.3 percent of total loans in 1989 to 7 percent in 1992. While losses on real estate loans represented a significant share of the problem, other sectors also experienced financial distress as the recession deepened. In these conditions, the government stepped in to provide support in the form of capital injections and loan guarantees. In December 1992, the Swedish Parliament passed a bill affirming that the state would provide guarantees that banks and certain other credit institutions would meet their commitments on a timely basis. To support the objectives of the new law, Parliament set up a separate authority, the Bank Support Authority (BSA), as the lead restructuring agency.

Unlike a number of countries, where special government bonds were exchanged for nonperforming bank loans, Sweden’s main form of assistance consisted of the abovementioned guarantees of banks’ payments and capital injections to be financed from the budget as required. A number of institutions (including foreign subsidiaries in Sweden) were covered by the parliamentary guarantee. The guarantee was rescinded on July 1, 1996. No applications for support were received and no support was provided in 1994, 1995, and 1996. Several of the major banks availed themselves of these guarantees. This required approval from the Bank Support Authority and was given only after thorough investigations as to the need for receiving support. Only a few banks actually used the government assistance offered. As a result, nearly 98 percent of all the state financial assistance was to two large banks (Gota Bank and state-owned Nordbanken) and their associated asset-management companies, Retriva and Securum.73

Total commitments (including all guarantees) amounted to 85 billion kronor (5.9 percent of GDP). However, since most guarantees were not paid out, budgetary support was mainly in the form of capital injections (86 percent), with lesser amounts provided by share subscriptions or share purchases (10 percent) and interest subsidies (2 percent). The direct cost to the budget was Skr 61 billion (4.2 percent of GDP). Moreover, the net fiscal cost is diminishing over time, since loans are being recovered by the asset-management companies and share prices of Nordbanken have risen considerably, reflecting the bank’s high profitability. The state is withdrawing from bank ownership by selling its share in the remaining Nordbanken at a substantial premium.

The decision to adopt a comprehensive strategy enabled Sweden to weather a severe crisis, and maintain the payment system’s stability and the country’s credit rating, while minimizing the costs of the restructuring program. The main principles underlying the success of the bank-restructuring strategy of Sweden can be summarized as follows:74

  • The key element was the formation of a broad political consensus. The government provided extensive information to Parliament to foster bipartisan support for its restructuring policies. The opposition was involved in all of the main meetings and was represented on the board of the BSA.
  • Transparency and disclosure of information were crucial for regaining confidence domestically and abroad. The implications of support measures for depositors and investors were extensively reported.
  • The formation of an explicit institutional framework clarified the respective roles for the Ministry of Finance, the Riksbank, the Financial Supervisory Authority, and the BSA. Exchange of information among these institutions was continuous. A separate institution was created to implement the bank-restructuring strategy.
  • Diagnosis was the first step in the restructuring process. A common yardstick based on capital adequacy rules and other financial ratios was designed to measure the extent of the problem faced by each individual bank. Initially, the banks were divided into two categories according to whether they could be restored to viability or not. The first category was eligible for financial support whereas banks in the second category could be closed or merged with other institutions. Support payments would be used to pay off the banks’ creditors.
  • Proper incentives were established through the conditionality embedded in support agreements. The foremost condition was the need for a change in management and an upgrading of internal control and risk-management systems (which had been judged inadequate in most cases). Owners’ equity was not covered by the parliamentary guarantee; in the case of Gota Bank, owners lost their equity.
  • Structural reforms were enacted to strengthen the accounting, legal, and regulatory frameworks, and prudential supervision. Clear guidelines on valuation of assets and banks’ holdings of collateral (real estate and other assets) were set, and compliance with these procedures was strictly enforced.
  • Establishment of institutions for loan workout was given high priority. Nonperforming assets from each of the two problem banks were transferred to its associated management company at an estimated market price. This had advantages as the banks could continue their normal business without having to handle a large volume of workout cases. As with other types of support provided to banks, strict conditionality (as discussed previously) was attached to the cleaning up of balance sheets. The asset-management company could recruit the specific expertise needed for the work of transforming the bad assets into salable assets. The government funded and capitalized the asset-management companies and was the sole owner.

Philippines: State Banks and the Problem of Incentives

Unfavorable external conditions, coupled with growing political uncertainty, led to a sharp deterioration of the Philippines’ balance of payments at the end of the 1970s. As real interest rates in international markets rose, the servicing of the Philippines’ large external debt became increasingly difficult. The government tried to stem a larger crisis by stepping up its assistance to troubled enterprises. In 1983, amidst growing political uncertainty and a continued deterioration in the balance of payments, the announcement of a moratorium on external debt service triggered a financial panic and led to bank runs, a flight to currency, and capital outflows.

In addition to the macroeconomic causes of the banking crisis, a number of weaknesses in the supervision of banks had contributed over time to the deterioration of the banks’ conditions. Existing rules were not enforced even though the Monetary Board was informed promptly of problem banks.75 Some of the accounting rules were not codified in a way that ensured transparency and consistency. Furthermore, instead of obliging banks to make provisions for bad debt so as to restore confidence, the Monetary Board chose to waive enforcement of regulations in order to give troubled banks time to overcome their financial difficulties. In the event, this regulatory forbearance served to accelerate the deterioration of bank balance sheets.

Over 1982–85, the government provided assistance to other sectors mainly to nonfinancial public enterprises. It accelerated subsidies to unprofitable state enterprises and increased its emergency lending and equity contributions to public corporations. These policies proved insufficient, however, to prevent the escalation of difficulties in the banking system. In addition, the government directed banks to continue lending to enterprises in financial distress, thereby accelerating the deterioration of bank portfolios.

By the end of 1985, the two largest state-owned banks, Philippines National Bank (PNB) and the Development Bank of the Philippines (DBP), which together accounted for about half of the banking system’s assets, were recognized as insolvent; nonperforming loans amounted to about 70 percent of their combined portfolios or about 21 percent of the assets of the banking system.76 A comprehensive rehabilitation program for both banks was initiated for the downsized banks by assuming an equivalent amount of deposit liabilities, which became part of government debt and was paid off over three years. At the core of the rehabilitation program was a transfer of nonperforming assets from banks’ balance sheets.

Following an external audit to evaluate the assets and liabilities of the two banks, in November 1986, 108 billion pesos of nonperforming assets (17 percent of GDP) were transferred to the Assets Privatization Trust (APT) that had been established. Selected staff from the PNB and DBP participated in loan-workout task forces and were charged with managing the nonperforming assets transferred to the APT. These tasks included preparing legal documentation preliminary to the sale or seizure of collateral as well as preparing these assets for sale or recovery. For the nonperforming accounts that were to remain on the books of the DBP and PNB, additional task forces reinforced loan recovery. Overall, loan recovery rates were reported to have been satisfactory.

The government recapitalized the two banks by injecting P 15.9 billion in cash (3 percent of GDP) and wrote off its own deposits (1 percent of GDP). As a result of these measures, and the transfer of assets to APT, the balance sheets of the two banks were scaled down sharply: total assets of PNB were reduced by 54 percent and those of DBP by about 87 percent. New management was introduced, the branch networks were cut down, and major cost-reduction programs—including significant staff cuts—were implemented (PNB staff was reduced by one-fourth and that of DBP by two-fifths). New charters were established for both banks. It was announced that all tax privileges would be withdrawn, public deposits at these banks would be restricted to working balances, the banks would be subject to private external audits, and the government would extend no further guarantees. By 1987, both banks had returned to profitability and were able to improve their capital-asset ratios. In 1989, 30 percent of PNB’s outstanding shares were privatized. By 1996, this figure had increased to 57 percent.

Although political and macroeconomic developments of the late 1970s and early 1980s increased financial fragility, the root of the Philippines banking crisis lay within the financial sector. Weaknesses in the regulatory framework and lax banking practices triggered and magnified the crisis. As with other successful bank-restructuring programs, that of the Philippines focused on resolving the underlying causes of the banking crisis by incorporating the following fundamental principles: proper diagnosis of the problem; financial restructuring complemented by operational restructuring; and a strengthened regulatory and prudential framework. In particular, measures to rehabilitate the state-owned banks, which accounted for most of the problems, concentrated on substantial operational restructuring, including replacement of the management, and a substantial downsizing of staff. Such comprehensive rehabilitation paved the way for a successful privatization.

United States: Role of the Regulator in Dealing with Moral Hazard

The United States, in the wake of catastrophic financial crises that it suffered during the Great Depression, augmented the safety net for its commercial banking industry and created a separate safety net for thrift institutions.77 During the widespread financial sector distress in the 1980s, the safety net for the commercial banks succeeded in handling weakness in the banking sector, while that of the thrift institutions failed to do so. The following study seeks to explain the contrasting performance.

The safety net for banks now consists of the Federal Reserve as lender of last resort (LOLR), deposit insurance funded by banks and administered by the Federal Deposit Insurance Corporation (FDIC), and a system of regulation and supervision conducted by the Office of the Comptroller of the Currency (OCC) for federally chartered banks. This overall safety net also relies on cooperation between state and federal regulators for banks chartered by the states. The existing framework proved sufficient to overcome the severe problems that commercial banks experienced during the 1980s and early 1990s as a result of economic recessions and weaknesses in regulation and in the configuration of the industry.78 By professionally executed regulation and supervision, it held moral hazard in check, unlike the situation in the thrift industry. Consequently, there was no need to create special institutions or use public funds to handle the large number of bank failures.

The United States did, however, make some adjustments to its existing arrangements regarding supervision, last-resort lending, and deposit insurance in the FDIC Improvement Act (FDICIA) of 1991. For example, prior to FDICIA the Federal Reserve had discretion in its last-resort lending, but a study by the House Banking Committee revealed that the Federal Reserve had consistently given liquidity assistance to the weakest, even insolvent, banks; sometimes for extended periods—as in the case of Continental Illinois Bank, which failed in 1984.79 Such lending was shown to have added to the costs that the FDIC incurred in dealing with failed banks. Consequently, FDICIA limited the Federal Reserve’s discretion to lend (even with collateral) to troubled banks for periods over 60 days. In addition, FDICIA instituted a system to reinforce regulators’ incentives to take prompt remedial actions against undercapitalized banks and to close those whose equity capital had declined to 2 percent of total assets. Under FDICIA, the intention was that regulators should close the bank before it became insolvent and imposed losses on the FDIC and uninsured depositors. In addition, the Act required the FDIC to use the “least-cost” form of bank resolution.80

The same success cannot be claimed for the handling of the large number of thrift institutions that failed in the United States between 1980 and 1992.81 The safety net for savings and loan associations (S&Ls), established during the depression, consisted of the Federal Home Loan Bank Board (FHLBB) as the regulator of federally chartered thrifts, which collaborated with state regulatory agencies in supervising state-chartered S&Ls; a subsidiary called the Federal Savings and Loan Insurance Corporation (FSLIC) that insured S&L deposits; and 12 Federal Home Loan Banks that acted as the thrifts’ lenders of last resort.

The thrift safety net failed in its responsibilities, partly because it did not properly supervise weak S&Ls, or close them when they became insolvent, so that it allowed moral hazard to prevail. Consequently, major changes were made in the Financial Institutions Reform, Recovery, and Enforcement Act (FIRREA) of 1989. These changes included the abolition of the thrift regulator and its deposit insurance subsidiary, and some limitation of the Home Loan Banks’ LOLR powers. A new agency (the Office of Thrift Supervision) was created to regulate and supervise S&Ls; the deposit insurance fund had to be recapitalized at an estimated $130 billion cost to the taxpayer; and the insurance function was transferred to the FDIC. The role of LOLR was (to some extent) transferred to the Federal Reserve. Furthermore, a separate temporary agency, the Resolution Trust Corporation (RTC) was created to take control of failed thrifts and dispose of their assets.82

Weaknesses in the configuration of both the banking and thrift industries in the United States made them vulnerable to shocks. Both were locally based and, until recently, were prevented by law from branching across state lines.83 Given that these two industries were simultaneously exposed to similar climatic and economic shocks, a question naturally arises why the banks’ safety net proved adequate to its task while the thrifts’ did not.84

Several factors explain the difference in outcome for the two industries. In particular, the thrift industry faced greater structural impediments. At the start of the period, when the difficulties of savings and loans emerged, the industry’s loans were by law confined almost exclusively to residential mortgages (usually made with a 30-year maturity) at a rate that was fixed at the time the mortgage was initiated. As deposit rates were progressively freed from government control during the late 1970s and early 1980s, thrifts’ portfolios suffered heavy losses in market value when the Federal Reserve moved vigorously to quell inflation starting in 1979 and interest rates rose sharply. By contrast, the banking industry was much less exposed to interest rate risk because its principal assets were commercial loans made at variable rates; that is, it was not legally obliged to engage in such a high degree of maturity transformation.

Regulators and politicians decided at the beginning of the 1980s not to close, restrain, or closely supervise thrifts that were insolvent at market value85 but to let them continue operating as their book value declined. They were permitted to remain in operation even when they became book-value insolvent, by several officially approved opportunities for accounting legerdemain.86 Insolvent thrifts also received extensive liquidity support from the Federal Home Loan Banks.87 Banking industry regulators, on the other hand, did not permit insolvency to be disguised by creative accounting and closed failed banks much more promptly.88 Moreover, while the Federal Reserve was criticized for lending to insolvent banks, forebearance was contained by the FDIC’s stricter closure policy.89

Institutions without capital have incentives to grow and gamble, and this is what weak U.S. thrifts did.90 The legislation of 1980 and 1982 gave thrifts additional opportunities to engage in more risky business.91 Consequently, losses due to the industry’s exposure to interest rate risk in the early 1980s were replaced by massive losses from bad loans later in the decade. And many owners of nonviable thrifts looted their institutions.

The environment that permitted these abuses was the configuration of thrift regulatory agencies. From their creation in the 1930s, the federal thrift agencies by law were assigned the role of promoting home ownership in the United States. Consequently, thrift regulators saw themselves as both thrift supervisors and promoters of the housing industry. They were, therefore, prone to much more “regulatory capture” than were the banking regulators.92 In addition, thrift regulators were subject to political interference from both the executive and legislative branches of government. While federal regulators at both bank and thrift agencies were short of resources in the early 1980s, the thrift regulators seemed more affected than the banking agencies. For example, the salaries of thrift supervisors in the early 1980s were lower than those at the banking agencies despite the need for strong supervision during this period of financial liberalization.93

When regulators fail in their responsibilities, the markets can provide necessary discipline by demanding a higher interest rate for placing funds in weak thrifts or denying funds entirely to nonviable institutions. But the markets did not do so in the 1980s. While weak institutions, particularly insolvent thrifts, did offer somewhat higher rates than strong banks, the 1980 increase in the limit on federal deposit insurance to $100,000 (over eight times per capita GDP at that time), the FDIC’s preference for resolving failed banks by mergers (where the acquiring institution assumed all deposits), and the brokering of deposits enabled insolvent thrifts to pay less than risk-adjusted rates for funds.94

The following lessons relevant to countries that are restructuring their banking industries can be drawn from the restructuring experience of U.S. banking and thrift institutions:

  • Regulatory restrictions on the configuration of an industry—particularly on S&Ls’ portfolio composition and both banks’ and thrifts’ geographical location—can reduce the ability to withstand macroeconomic shocks.
  • Because bank regulators closed insolvent banks more promptly than thrift regulators, internal governance by bank management was generally stronger than at thrifts.
  • Regulatory independence and accountability for supervisory actions are important to protect creditors and, ultimately, taxpayers. The thrift regulator practiced forebearance partly as a result of political interference.
  • Whereas the Federal Reserve was criticized for lending too readily to nonviable banks, such lending by the Federal Home Loan Banks was vastly more problematic.
  • The thrift regulators also practiced forebearance because the thrift deposit insurance fund was undercapitalized—and in fact became insolvent in the mid-1980s—so that it did not have the resources to meet its obligations. It can be concluded that an insurance scheme that is so underfunded that it becomes heavily insolvent does more harm than good.
  • Market discipline can be valuable in controlling undue risk taking. It failed at both banks and thrifts because the deposit insurance limit was too high, preferred forms of failure resolution protected all depositors, and deposit brokering extended coverage to virtually all deposits.
  • Market discipline was weaker at thrifts than at banks because the thrift regulators released far less data (in fact they concealed information) on the condition of individual thrifts and the industry in general than did the bank regulators.
  • Regulators in both industries lacked resources (money and staff) at a critical time, but the lacuna was worse at the Federal Home Loan Bank Board.
  • The latest estimate of the cost of the S&L debacle is $130 billion, which is considerably less than first anticipated.

Chile and Spain: Role of Central Banks in Systemic Bank Restructuring

Both Chile and Spain experienced systemic bank problems in the early 1980s. Banking crises in the two countries occurred as a result of a combination of external shocks and endogenous factors. The Spanish economy suffered from the consequences of the oil shocks of the 1970s, while the Chilean economy experienced a severe recession with rapidly declining terms of trade. In both countries, systemic banking problems arose following several years of liberalization, expansion, and increased risk taking in the banking sectors without adequate legal, regulatory, and accounting frameworks necessary for proper risk management. In both countries the central bank assumed a lead role in devising, implementing, and financing the bank-restructuring strategy. In Spain, the central bank conducted some of its operations through the Deposit Guarantee Fund, while the Central Bank of Chile’s role was supplemented by the Bank Superintendence a separate entity responsible for banking supervision.

Chile

Following several years of liberalization and rapid expansion of the financial sector, and amidst a severe recession, a financial crisis broke out in late 1981. Liberalization had taken place amid lax supervisory standards that were conducive to excessive lending to related parties and other imprudent banking behavior, such as foreign exchange lending to domestic borrowers who had no access to foreign exchange earnings. But problems were exacerbated when the macroeconomic environment began to deteriorate dramatically in 1981 after U.S. interest rates were pushed to historically high levels to combat domestic inflation and the industrial countries began to fall into recession. Chile experienced a sharp deterioration in its terms of trade (with copper prices falling by more than 30 percent between 1980 and 1982). Capital began to flow out of Chile, forcing the authorities to abandon the fixed exchange rate as Chile lost about $500 million in reserves between October 1981 and June 1982. The Chilean peso depreciated against the U.S. dollar by almost 90 percent in 1982. The exchange rate depreciation in turn led to a rapid buildup of arrears on the foreign exchange indexed loans outstanding to domestic borrowers. This and rising real interest rates increased commercial bankruptcies, leading to increased levels of nonperforming loans and further weakening the banks’ condition. A Stand-By Arrangement with the IMF began in early 1983 accompanying the bank-restructuring strategy.

Between 1981 and 1984 the Superintendency of Banks took over 14 (out of a total of 26) private domestic banks and 8 (out of a total of 17) private domestic finance companies. While there was a substantial time lag between the occurrence of banking problems and intervention by the authorities, once the strategy was determined, the authorities moved quickly in taking control of the banks. In 1981, they began to develop what amounted to a comprehensive bank-restructuring program. Measures were implemented to address the immediate problems of the banks as well as weaknesses in accounting, disclosure, and banking regulation and supervision. The liquidation of eight of the banks and all of the finance companies placed under the Superintendency’s control, as well as a number of mergers arranged by the authorities, reduced the number of banks by one-third and reduced the number of finance companies by two-thirds.

During the crisis and throughout the bank-restructuring process, the Central Bank of Chile (CBC) provided short-term and long-term financial resources to banks. These loans were granted at subsidized rates and many of the advances were never repaid. As a result, the CBC incurred substantial operational losses. These were equivalent to 2.9 percent of GDP in 1983 and 4.6 percent in 1984. At the start of the crisis, the CBC granted emergency loans to all financial institutions placed under the control of the Superintendency of Banks. These included emergency loans to banks in liquidation that were granted to ensure that commitments to domestic and foreign creditors were met, thereby preventing a potential crisis in the payments system. Some of the emergency loans were later repaid using funds from the government through the Development Corporation (Corfo). Promissory notes were issued by the CBC to depositors of liquidated banks. In this way, all deposits were reimbursed in accordance with a 1982 law that overturned limited deposit insurance.

Between 1981 and 1984, the CBC acquired all assets and liabilities of banks and financing companies in liquidation. The CBC became actively involved in loan management in the form of debt-relief and debt-rescheduling schemes. These schemes entailed the conversion of loans to domestic borrowers denominated in foreign exchange into loans denominated in domestic currency, with the CBC assuming the losses attributable to the devaluation. The CBC’s objective was to alleviate the debt burden of borrowers, thereby avoiding further bankruptcies, which would have led to further deterioration of the banks’ asset quality.

For the same reasons, debt-relief schemes were instituted for other borrowers in the form of long-term credits involving large subsidies to firms in the manufacturing and transportation sectors, mortgage debtors, and preferential exchange rates for foreign currency debtors. These loans were made at subsidized rates and amounted to about 21 percent of domestic credit in 1984. Technically, these were “pass-through” loans from the CBC to the insolvent banks. Nonetheless, the CBC set the terms and conditions for the reschedulings and thus became directly involved in the lending operations. These loans remained on the CBC’s books for over ten years, and their management appears to have entailed high costs for the CBC. The CBC also provided loans to banks for debt-relief schemes involving direct credit lines to consumers, and trading companies, intermediated by the intervened banks. These were mainly short term in nature and were settled in subsequent years.

To alleviate the burden of nonperforming loans on the books of insolvent but potentially viable banks, the CBC began to purchase past-due loans from them. This program was started in 1982 and expanded in subsequent years as the severity of the crisis became manifest. The CBC purchased these loans at par value (up to a maximum of 1½ times the banks’ capital and reserves, later raised to 2½ times and then to 3½ times capital and reserves). The CBC purchased nonperforming loans from 24 financial institutions (19 banks and 5 financial companies) for an amount of over $3.4 billion, or about 23 percent of GDP (30 percent of outstanding loans). The banks receiving CBC support initially agreed to repurchase their nonperforming assets over a 10- to 40-year period. The CBC could request the liquidation of the financial institutions that did not repurchase their nonperforming portfolio on time. Financial institutions were not allowed to pay dividends until they had repurchased their nonperforming portfolio in full; however, this restriction did not apply to newly issued stock so as not to hamper investment in the sector.95

Between the end of 1984 and the end of 1989, the banks repurchased about 18 percent of the indexed value of their nonperforming loan portfolio, leaving a balance of $3.1 billion (12 percent of GDP), owed by 13 financial institutions.96 However, by the end of 1989 the cost-sharing arrangement was modified: annual payments to the CBC were limited to an agreed share of annual profits and the repayment period was extended indefinitely until such time as banks made enough profits to pay for the debt. This reduced substantially the present value of the debt (now called the “subordinated debt”) and thus the costs to be borne by the banks. Between the end of 1989 and 1995, seven institutions repaid their subordinated debt from the CBC in full. During the period, however, the total outstanding balance declined at a snail’s pace due to the impact of the 1989 modifications. It even increased in years when banks’ profits were lower than the interest due on their outstanding debt. The arrangements were again modified in 1996 with a view to increasing the share of costs to be borne by the banks, and a maximum 40-year time frame was established for repayment. Under this arrangement, the CBC expects to recoup more than 55 percent of the book value of the subordinated debt outstanding as of the end of 1995 ($4.5 billion).

More successful components of the bank-restructuring process were the strengthening of bank regulation, the modernization of bank supervision, and the intensification of on-site inspections. Personnel were increased and specially trained to conduct comprehensive examinations. These measures were effective in establishing an enhanced and forward-looking environment for resumed bank activity.

The return to a sound banking system in Chile was slow and exceedingly costly. Four years after the crisis, bank capital ratios had not recovered to their precrisis level, bank profitability continued to be low, and operating expenses did not decline. Moreover, four years after the onset of bank restructuring, bank assets had expanded in real terms by almost 60 percent while the number of bank employees had grown by 20 percent. This result is consistent with the approach taken by the CBC whereby operations of insolvent banks were in effect continued. The active engagement of the central bank in debt rescheduling and direct involvement in commercial lending appears to have slowed down the bank-restructuring operations.

Spain

Spain’s banking problems resulted from a combination of external and internal factors. In the 1970s, liberalized bank legislation without adequate adjustments in the regulatory, supervisory, and accounting systems led to highly speculative, growth-oriented banking activity. Extensive ownership of banks by industrial conglomerates leading to excessive lending to related parties was also a contributing factor. Structural adjustment problems resulting from the oil shocks of 1973–74 and 1978 were important factors triggering the banking crisis.

Problems in the Spanish banking sector began to surface in 1977 when the Bank of Spain was faced with growing liquidity problems that reflected, in some cases, much larger solvency problems. Rapidly growing levels of nonperforming loans led to massive bank insolvencies in the early 1980s. The Spanish banking crisis affected 51 of the country’s 110 banks, accounting for more than 20 percent of deposits. This led the authorities to engage in a comprehensive bank-restructuring process. An important component of the bank restructuring was tightened banking supervision, including enhanced reporting requirements and the implementation of an early warning system for banks.

The Bank of Spain assumed the lead role in the bank-restructuring process, mainly acting through the Deposit Guarantee Fund (FGD), jointly owned by the Bank of Spain (50 percent) and private banks. It was financed by private banks with matching funds from the Bank of Spain. Thus, a loss-sharing arrangement was part of the restructuring strategy. In 1980, the FGD was restructured and its powers substantially increased to ensure the solvency and operation of the banks in the interests of depositors and of the fund itself. With these changes, the FGD became the lead agency in the bank-restructuring process.

During the crisis, the Bank of Spain provided some direct support to the banks by granting exceptions to reserve requirements or mandatory investments for banks. The mandatory investments yielded 3 percentage points below market rates.

Bank rehabilitation was done in the following sequence, referred to by the authorities as “accordion” recapitalization. First, existing bad debts were written off against remaining capital. Then, the FGD acquired a controlling interest in the bank, and later subscribed to new capital issues (i.e., injected cash for equity stakes), and finally sold the bank to new shareholders. The system implied strong incentives toward improved corporate governance as the old owners lost their ownership stakes.

A different approach was taken to restructure the Rumasa group (a holding company that included 20 banks as well as other industrial and commercial enterprises). In view of the high concentration of loans to related borrowers who were overindebted, and the potential employment effect (the Rumasa group had 50,000 employees, 11,000 of them in banks), a more gradual approach to restructuring was taken. The authorities took effective control of all related companies in 1983, conducted audits of the banks and firms, and then began to sell the companies. Management and sale of the banks was assumed by the FGD. The largest bank of the Rumasa group was sold to a foreign consortium of banks, 2 smaller ones were purchased by domestic investors, and the remaining 17 were absorbed by major domestic banks who, in return, assumed part of the bank-restructuring costs.

In the case of the Rumasa group, the central bank provided direct support in a number of ways. As part of the Rumasa banks’ rescue, the Bank of Spain made a long-term loan to the Rumasa banks carrying the guarantee of the group of banks. The loan amounted to 131 billion pesetas (0.6 percent of GDP), maturing in 12 years at an interest rate of 8 percent. Another loan amounting to Ptas 269 billion was renewed. The Rumasa banks placed the sum of both items (Ptas 400 billion, about 2 percent of GDP) with the group of banks in a 12-year deposit yielding 13½ percent interest. Furthermore, the 12 banks purchased government bonds issued at below market rates (9½ percent). For the involved group of banks, the cost was the difference between the interest received from the bonds (9½ percent) and that paid for the deposit (13½ percent), estimated at about Ptas 192 billion at the time. Thus, the cost of this operation was shared between the Bank of Spain and the banking community.

For each bank undergoing rehabilitation, an operational recovery plan was developed and approved covering a period of several years (frequently five years). The plan specified goals and deadlines for achieving these. Banks periodically reported on progress to the supervisor and the FGD. At the end of the plan, the supervisor carried out a detailed inspection. Interim inspections were also conducted at the convenience of the bank supervisor. Thus, the financial restructuring measures were accompanied by operational restructuring. By the mid-1980s, banking sector soundness was largely restored. Nonperforming loans had declined to low levels, and capital ratios improved although bank lending resumed cautiously. The cost of bank restructuring was estimated at about 15 percent of GDP.

Conclusions

Bank restructuring in Spain was generally considered successful, while less so in Chile. Initially, the scope of the problems in Chile and Spain appears to have been similar, affecting about 20 percent of total loans, but the cost of bank restructuring in Chile is estimated at 33 percent of GDP compared with an estimated cost of 15 percent of GDP in Spain (calculated as the ratio of average government outlay to GDP in each year).97 The differing roles of the central bank in the two cases may explain some of this outcome. In Spain, the central bank placed considerable emphasis on cost sharing with the banking community. Other appropriate incentives for improved corporate governance were in place. Moreover, the Bank of Spain’s activities in bank restructuring were somewhat removed from its monetary policy and supervisory activities by working through the Deposit Guarantee Fund.

In contrast, the Central Bank of Chile engaged in ongoing quasi-fiscal expenditures by assuming the financial costs of bank rescue operations. The cost-sharing system employed in Chile deferred the cost to bank owners to the future, and in the end much of the cost was absorbed by the Central Bank of Chile. The bank-restructuring approach was complex and not sufficiently transparent. Nonetheless, perhaps in large measure owing to a comprehensive approach, including strengthening of banking supervision, Chile succeeded in restoring bank soundness, although at a slow pace and at a high cost.

Baltic States, Russia, and Other Countries of the Former Soviet Union: Special Issues of Bank Restructuring in Transition Economies

As the Baltic states, Russia, and other countries of the former Soviet Union embarked on market reforms during 1991–92, two-tier banking systems emerged with the creation of central banks and the transformation of specialized banks into notionally autonomous commercial banks.98 Concurrently, in large part owing to lax licensing policies and practices, most of these countries experienced a large increase in the number of commercial banks, and expansion of branch networks. These institutions typically lacked expertise in credit evaluation, and many engaged in aggressive lending to enterprises with which they were associated by ownership or other ties. Early in the transition the consequences of poor credit evaluation were effectively masked by high inflation. The sharp reduction in inflation resulting from the successful stabilization programs, however, revealed the underlying weakness of the banking systems.

For most of the transition countries, the overall financial situation remains fragile. Based on data for 1994–95, the ratio of nonperforming to total loans varies from 14 percent to 63 percent,99 ratios that are high compared with other countries that have experienced banking crises.100 By contrast, partly as a result of demonetization during the initial period of sharp inflation, the household deposit base is small in most of these countries, and the banking sector’s total assets are smaller relative to GDP (on the order of 10–20 percent) than in the transition economies of Central and Eastern Europe (on the order of 40–50 percent). Moreover, the fragility of the banking system remains a major constraint on economic recovery.

The situation in most countries, with the exception of the Baltics, has not yet reached the stage of an acute crisis, nor have there yet been closures or interventions in the large banks.101 However, the problems in the banking sectors in the countries within the Commonwealth of Independent States (CIS) can be considered systemic in nature. In Russia, for example, a large part of the banking system faces serious problems, and many banks are insolvent, Russian commercial bankers estimate that in many banks 50 percent of the loans are nonperforming.102 In most of the other countries, more than 50 percent of the loan portfolio of the five largest banks are nonperforming; it may be inferred, therefore, that most, if not all, of these banks are insolvent. Moreover, as these banks are closely linked through interbank loans and deposits, failure of an individual bank may lead to contagion effects and hence trigger a systemic banking crisis.103 Such possibilities were clearly presaged by the Russian interbank market crisis in 1995.

Poor banking practices underlie the deterioration of the banking system. Shortcomings in the legislation, prudential norms, and accounting standards, together with a lack of skilled bank supervisors have added to, and delayed detection of the problems, which could be substantially worse than the officially available figures indicate. On the other hand, the strengthening of banking supervision itself has been a triggering factor for banking crisis in some countries; when the authorities insisted on compliance with prudential regulations, the weaknesses became apparent to the public and a crisis ensued (such as in Latvia and Lithuania).

In Russia, as well as in most other countries, bank-restructuring measures to date have been mainly directed at the small banks involving increases in minimum capital and its enforcement through license withdrawals, liquidation procedures, and mergers. Between 1991 and May 1996, the licenses of about 450 Russian banks were withdrawn. Of these, 126 were revoked in the period January to May 1996. In addition, capital requirements are being raised and licensing policies are being tightened. However, only 2 of the 12 CIS countries, namely, Kazakstan and the Kyrgyz Republic, have embarked on a systemic restructuring strategy that includes particular measures to deal with the formerly specialized banks that dominate the financial system. In addition, these two countries have both made some progress in enterprise sector reform, which has helped to ease loan problems in the banking sector.

Authorities in most other CIS countries appear reluctant to close the large banks. In Moldova, measures are being designed to rehabilitate the large banks by means of financial and operational restructuring. In Azerbaijan, there has been substantial liquidity support to the Savings Bank (totaling about 1 percent of 1995 GDP). In Belarus, in late 1995 the State Savings Bank and Belarus Bank were merged. Both were experiencing serious difficulties, and the merger resulted in a weaker institution and increased concentration. The merger has had serious negative effects on the whole banking sector since the government, in a move to support the weak new bank, ordered the transfer of all government deposits to the new bank, creating significant liquidity problems for the other banks. Since the deposits of the merged bank are guaranteed by the government, the financial responsibility of the government has increased. In Georgia, the three formerly specialized banks (State Savings Bank, EximBank, and the Industrial Bank) were merged during the second half of 1995. The merged bank became the largest bank in Georgia in terms of branches and assets (accounting for 70 percent of total assets of the banking sector).

In both Georgia and Belarus, the decisions to merge banks appear to have been politically motivated and occurred without consultation and formal concurrence of the supervisory authority. In Turkmenistan, as well, the President authorized the licensing of a special agricultural bank, which is to have a network of 53 branches, in an attempt to restructure the Agroprombank. This created a major increase in the supervisory burden on the Central Bank of Turkmenistan, and because of the severe staffing constraints in banking supervision may pose dangers to effective supervision over the banking sector.

Thus, in most of the CIS countries, banking sector problems have been dealt with as and when they surfaced, and a systemic approach to bank-restructuring has not yet emerged, in part because the magnitude and nature of the problem is not fully transparent. As a first step, these countries are beginning with a diagnostic study to establish the magnitude and the nature of the problems. Two sources of information generally used to establish an approximate indication of the extent of financial distress in the system are monthly off-site supervision reports collected by the central banks and on-site portfolio audits. The gap between the value of existing provisions and general loan-loss reserves of the banks and the size of their reported nonperforming loans (including principal plus interest) provide a first approximation of the potential reserve deficiency.

However, the diagnostic exercise for the CIS countries is limited by weak accounting practices associated with the Gosbank accounting standards. The two main accounting problems are the treatment of overdue interest and the consolidation of branch accounts. Under the Gosbank accounting system, overdue interest is not included in the profit and loss account, and neither are provisions made against the capital value of the nonperforming loan. Unpaid interest is recorded as overdue interest on the asset side, and on the liability side, it is recorded as a component of the future earnings account. This implies a downsizing of a bank’s balance sheet without an impact on its capital, in the event that overdue interest is written off. Removing the interest recorded as future earnings requires a reversal of entries. This does not affect the capital account, since the interest is not recorded as income in the first place. Similarly, for some banks, the nonconsolidated interbranch balances account for a large share of their total assets and liabilities, implying that the consolidation of branch accounts should lead to a significant decline in the banks’ balance sheets.

Most countries have made some progress in improving the legal framework for bank resolution and related actions. The Baltic countries have created the legal framework for and have begun to move toward international accounting standards in banking. However, much work still needs to be done in this area in order to establish the framework for the effective functioning of the banking system. The importance of a smoothly functioning legal and court system is clearly illustrated by Moldova and Lithuania, where the central bank is experiencing major legal difficulties in closing insolvent banks, notwithstanding clear legal authority to do so.

Fewer than half of the countries have initiated formal deposit insurance schemes; however, in many, there seems to be an implicit government guarantee on deposits of the large banks. Household deposits at the state-owned banks, in particular, at the former Savings Bank, are by law guaranteed by the government in Belarus, Lithuania, Russia, Ukraine, and Uzbekistan. In the Kyrgyz Republic, the government has implemented payoff plans to compensate the depositors and other creditors of the majority state-owned banks.

Kazakstan and Lithuania have established an institutional framework to recover nonperforming loans. In Kazakstan since early 1995, the directed credits that are nonperforming are being carved out from bank balance sheets. Most of the credits were withdrawn from the commercial banks and converted to long-term state debt of the government vis-à-vis the national bank (domestic credit) or held as off-balance sheet items by the EximBank (foreign credits). In the Kyrgyz Republic, the government is creating an independent legal entity to take control over and dispose of some of the assets and liabilities of the former state-owned banks. The institution will be established as an autonomous unit within the Enterprise Reform and Resolution Agency framework, which deals with many of the state-owned enterprise borrowers of the problem banks. In addition, these two countries have established transitional arrangements to ensure the working of a minimum financial structure. In Kazakstan, the restructuring effort is staggered so that a core banking structure will continue to function at all times. Alternatively, in the Kyrgyz Republic, new nonbank service providers have been set up, such as a rural finance mechanism to fill the void after the proposed liquidation of the Agricultural Bank.

Another approach is that of Georgia, which is adopting a more “market-based” approach to restructure its banking system. A key element is the implementation of a bank-certification program objective that is to establish proper programs to bring banks into compliance with the prudential standards. Failure to gain certification will be accompanied by strict limitations on balance sheet growth. In Armenia, Azerbaijan, and Uzbekistan, since a systemic restructuring strategy will take time to design and implement, short-term stopgap measures have been invoked. These include curbing the activities of problem banks and limiting ad hoc injections of funds by the authorities. Azerbaijan, Belarus, Ukraine, and Uzbekistan have prohibited the banks that do not comply with prudential regulations from participating in the interbank market or credit auctions.

It is clear that bank restructuring in these countries will be an arduous and time-consuming task that, in most cases, is only just beginning. Most will have to tackle the problem of deciding the fate of large, state-owned banks more or less simultaneously with undertaking comprehensive enterprise restructuring. And, unlike the case in most other countries that have undertaken systemic bank restructuring, practically the entire operating framework for the banking system remains to be developed. Shortcomings in the legislation, prudential norms and accounting framework, and a general lack of adequate supervisory capacity are particularly acute. In general, the courts and legal system are not adequately developed to fully sustain market-based, commercial banking activity. Moreover, there is a serious lack of basic banking skills that will need to be addressed as a major component of operational restructuring.

Illustrative Case Studies of Fiscal Aspects of Systemic Bank Restructuring

The experiences of five countries, Poland, Latvia, Mauritania, the Philippines, and Sweden, have been chosen to illustrate the fiscal aspects of systemic bank restructuring.104

Linking Bank and Enterprise Restructuring in Poland

Bank-restructuring operations in Poland between 1993 and 1994, unlike in other transition economies, not only restored public banks to solvency, but also averted further losses by linking government assistance to a comprehensive program to restructure public banks and enterprises. The reasons for Poland’s comparative success be in the design and implementation of the government’s Enterprise and Bank Restructuring Program (EBRP). The major aspects of the EBRP were (1) emphasis on organizational restructuring, a clear statement of the ultimate objective and a thorough audit; (2) a decentralized approach; (3) improvement in the legal framework; and (4) the structure of the recapitalization program.

Organization and Framework

The aim of the EBRP was to privatize the major public banks. This aim gave purpose and commitment to the whole restructuring scheme. Banks to be restructured were first transformed into joint-stock companies, with the treasury as the sole shareholder. Independent supervisory boards were established, and long-term technical assistance contracts with reputable foreign banks (twinning arrangements) were signed. International auditing firms then performed a thorough audit of these banks (which revealed serious capital shortfalls).

Decentralized Approach

The EBRP did not adopt a centralized approach (e.g., a one-time swap of bad debts to a loan recovery institution for interest-bearing government debt) to recapitalization because the government doubted any loan recovery institution’s ability to collect; and considered the central problem to be the banks’ lack of market experience, which an asset swap would not remedy. Instead, a decentralized approach was taken where the banks were recapitalized while leaving the bad debt on their books to be collected by them. This kept the banks involved with enterprise monitoring and restructuring, and because the amount of recapitalization was not linked to the collection of bad debt (it was based on the earlier audit), banks had an incentive to make their bad debts perform.

Legal Framework

A potential problem of leaving the bad debt on banks’ books was that since the relationship between the banks and the debtors was not severed, banks might have continued to lend to enterprises, which were not servicing their debt. To guard against this, banks were obliged to properly provision for loans by the end of 1993 and to set up debt workout units that had to either sell or restructure substandard loans by April 1994. These safeguards were effected through the Restructuring Law (February 1993), which required that no new credit be extended to bad debtors unless in conjunction with a restructuring agreement and that one of the following events take place: (1) the loan is entirely recovered; (2) a restructuring agreement has been made; (3) the debtor is declared bankrupt by a court; (4) liquidation of the debtor has been initiated; or (5) the debtor has regained creditworthiness by servicing its debt for three months. If these requirements were not met, the bank would be obliged to sell the loan in the open market. However, for those enterprises that the government considered important from a socioeconomic perspective, a “safety net” mechanism was created to cushion the effects of restructuring or liquidation for certain firms that had not reached a restructuring agreement with their creditors. This mechanism largely insulated banks involved in the program from political pressure.

Recapitalization Program

The recapitalization program provided a substantial, one-time increase for the recipient banks to enable them to operate effectively and be suitable for privatization. It was effected by the transfer of 15-year bonds, nonnegotiable for 3 years, with biannual redemption starting 18 months from issuance. As a result of this recapitalization, capital-asset ratios improved to about 12 percent enabling the banks to adequately provision for the stock of nonperforming debt and to deal successfully with the flow of new credits. The nonnegotiability of the debt and delayed amortization also helped ensure that public funds were not misused while recipient banks were being restructured and learning to operate in a new environment.

Minimizing Public Sector Costs in Latvia

In the spring of 1995, the largest banking crisis in the Baltics, Russia, and other countries of the former Soviet Union to date involved Bank Baltija, by far the biggest bank in Latvia, as well as several mid-sized private banks, such as the Latvian Deposit Bank, Centra Bank, and Olympija Bank. Banking system assets and liabilities plunged about 40 percent, and by the time Bank Baltija’s operations were suspended in May 1995, most of its assets had been stripped, leaving negative net worth of around 8 percent of GDP. The government’s response provides an example of intervention that minimizes public sector costs while ensuring a viable banking system and political stability. In the wake of the banking crisis in the spring of 1995, the government initially announced before the election that it would generously compensate household depositors. After the election, however, the new government indicated that, along with claims from legal entities, household depositors would only be compensated to the extent assets are recovered. To date, asset recovery has been minor.

A wide range of concomitant measures were taken as part of the comprehensive restructuring strategy. Liquidation of Bank Baltija has been initiated and the mid-sized banks involved in the crisis have been declared insolvent. Efforts to strengthen banking supervision and tighten key prudential regulations culminated in a tenfold increase in the minimum capital requirement in 1996. Several types of restrictions were imposed on banking activities, and publication of banks’ quarterly accounts was made obligatory. As a result of the more aggressive supervisory stance, consolidation of the banking sector is under way. Fifteen banks were closed, and the activities of another 2 were suspended in 1995; the number of banks allowed to accept household deposits (“core” banks), had been reduced to 12 in June 1996. The Bank of Latvia began closely monitoring problem banks with the assistance of external auditors, revoking banking licenses in cases of violation of prudential regulations. One year after the onset of the banking crisis, the Latvian banking sector has achieved relative stability. The number of problem banks has declined steadily, compliance with prudential regulations has improved, profitability has been reestablished, bank capital has increased, and commercial banks’ asset growth has resumed. Public confidence in the financial system, however, is only returning slowly.

Two factors, particular to the Latvian experience, partly explain the authorities’ emphasis on minimizing public sector costs in responding to the banking crisis. First, public opinion did not support bailing out the comparatively well-off and politically well-connected depositors, who had been receiving high interest rates. Second, the authorities were strongly committed to stabilizing the economy by maintaining an SDR peg, which, because it involved tight fiscal and monetary policies, severely limited the scope for any bank assistance.

Transparently Recording Fiscal Costs in Mauritania

Following macroeconomic instability, poor central bank supervision, and weak risk management by banks, the insolvency of the banking system (four commercial banks and one development bank) erupted into a liquidity crisis in 1993. As the liquidity crisis mounted and deposit withdrawals increased, the authorities instituted a comprehensive bank-restructuring program. The financial restructuring program involved liquidation, recapitalization, and privatization. The insolvent development bank was assessed to be nonviable and was liquidated. In this process, full cash compensation was provided for small depositors, foreign embassies, and expatriate workers. Treasury debt was issued to the central bank for claims on this bank and to public enterprises in partial compensation for their deposits. Of the four commercial banks, one was recapitalized by equal capital injections of treasury debt and of capital from the Libyan partner. Another bank was recapitalized by an equity injection from private investors, and the government sold its 10 percent share. For the main public bank, the government relinquished its share (46 percent), canceled treasury claims on the bank, assumed the bank’s nonperforming claims on public enterprises, and swapped central bank claims for treasury debt. In return, private owners injected capital and assumed full ownership. The government’s share of another bank was also sold.

These bank assistance operations were fully recorded in the budget. Revenue fell by 0.6 percent of GDP as a result of losses from canceled treasury claims outweighing privatization proceeds. Cash outlays associated with deposit reimbursement (1.2 percent of GDP) and the principal component of the treasury debt issuance (5.6 percent of GDP) were recorded as expenditure under the category “restructuring and net lending.” Interest on the debt issue was recorded with other interest payments. In total, the restructuring outlays increased the deficit by 7.4 percent of GDP in 1993 (see Table 13). This transparent recording of costs, in addition to improving confidence and governance, contributed to the recognition of the need to take offsetting measures and to moderate aggregate demand.

Table 12.Cost Estimates of Systemic Bank Restructuring, by Performance Groups

(In percent of GDP)1

Substantial progress countries
Côte d’Ivoire13.0
Peru0.4
Philippines4.0
Spain15.0
Sweden4.3
Moderate progress countries
Chile33.0
Egyptn.a.
Finland9.9
Ghana6.0
Hungary12.2
Korean.a.
Poland5.7
Slow progress countries
Kuwait45.0
Mauritania15.0
Tanzania14.0
Countries with bank-restructuring programs after 1994
Argentina0.3
Indonesia2.0
Japann.a.
Kazakstann.a.
Latvian.a.
Mexico12–15
Moldovan.a.
Venezuela17.0
Zambia3.0
Note: n.a. denotes not available.

Calculated by expressing fiscal or quasi-fiscal outlays in each year as a percentage of that year’s GDP. The percentages are then added (e.g., if the costs amounted to 11 percent of GDP in 1981, in 1982, and in 1983, the total cost would be noted as 33 percent in this table). These estimates do not take into account cost recoveries achieved by governments.

Note: n.a. denotes not available.

Calculated by expressing fiscal or quasi-fiscal outlays in each year as a percentage of that year’s GDP. The percentages are then added (e.g., if the costs amounted to 11 percent of GDP in 1981, in 1982, and in 1983, the total cost would be noted as 33 percent in this table). These estimates do not take into account cost recoveries achieved by governments.

Table 13.Mauritania: Fiscal Balance and Bank Restructuring(In percent of GDP)
1992199319941995
Total revenue19.425.723.224.0
Total expenditure and net lending24.836.827.724.8
Overall balance−5.4−11.0−4.5−0.8
Bank-restructuring costs7.40.2
Overall balance excluding bank-restructuring cost−5.4−3.6−4.3−0.8
Source: IMF staff.
Source: IMF staff.

Achieving Fiscal Consolidation While Restructuring Banks in the Philippines

The expansionary, inward-looking, economic policies of the 1970s, coupled with preferential lending to priority sectors, built up problems for the banking sectors in the Philippines, especially the two major public banks that channeled the bulk of this credit. These problems came to a head following changes in external conditions, interest rate increases, and political uncertainty at the end of the decade and in the early 1980s. The financial restructuring of the banking system between 1983 and 1986 involved assistance to banks of about 25 percent of GNP, of which about 7 percentage points of GDP were recorded in the budget (see Table 14). Despite this level of assistance to the banking sector, the deficit of the consolidated public sector fell from 9 percent of GNP to 4.8 percent of GNP. Much of this reduction was due to lower capital expenditure, mainly investment by public enterprises. During this period, public sector gross domestic investment decreased by 4.7 percentage points of GDP.

Table 14.Philippines: Macroeconomic Indicators and Bank Restructuring
1983198419851986
(in percent)
Real GDP growth1.9−7.3−7.33.4
Inflation (GDP deflator)10.050.323.10.8
(in percent of GNP)
Gross domestic investment27.117.414.313.2
Of which: public7.74.53.73.0
Gross national savings19.013.414.316.5
Consolidated public; sector balance−9.0−8.2−5.5−4.8
Of which: recorded assistance to financial institutions0.31.41.93.3
Source: IMF staff.
Source: IMF staff.

This reduction in the fiscal deficit and in public investment occurred as private investment slowed, and the economy fell into a sharp recession. Growth between 1983 and 1986 averaged -2.3 percent annually, accompanied by inflation of 21 percent. Investment as a share of GNP fell from 27 percent to 13 percent of GNP during this period. The accompanying macroeconomic developments, however, can be seen as a result of the poor economic policies of the preceding years rather than the fiscal consolidation. Indeed, the fiscal consolidation possibly laid the foundation for the subsequent upturn in growth, investment, and price stability. The ability to borrow abroad was limited and greater recourse to domestic borrowing would have exacerbated both the banking crisis and worsened the prospects for recovery.

Recovering Public Sector Outlays in Sweden

The bulk of government assistance to banks was capital injections into two large banks, Nordbanken and Gota Bank, and their associated asset management companies, Securum and Retriva, respectively. Total commitments amounted to 5.9 percent of GDP, but because not all guarantees were called, the initial (1991–93) cost to the government was about 4.2 percent of GDP. After much of the assistance had already been provided, the government passed the Bank Support Act (1992), which explicitly aimed at minimizing long-term fiscal costs and at recovering as much as possible of any support. This Act also stated that the assistance is exceptional and temporary.

Guided by the Act, the government started recovering its outlays by privatizing Nordbanken. In October 1995, 35 percent of the shares were sold with the budget receiving 0.4 percent of GDP as a result. Due to a strong increase in Nordbanken’s values on the stock exchange, the remaining shares are valued at about 1.4 percent of GDP. Asset recovery from Securum and Retriva was also substantially higher than initially expected. Originally, the government expected its capital injection to be written off; instead, it is likely that 45 percent of the equity injection (0.7 percent of GDP) will be recovered. The eventual net cost to the government of the bank assistance operations may be as low as 15–20 percent of the initial gross cost, as Table 15 indicates.

Table 15.Sweden: Gross and Net Costs of Bank Assistance, 1991–96(in percent of GDP)
Gross cost4.2
Less:
Asset recovery0.7
Privatization revenue
Realized (sale of shares)0.4
Redemption (of shares)0.4
Potential1.4
Dividends to the slate0.4
Total recoveries3.3
Net costs0.9
Source: Sveriges Riksbank.
Source: Sveriges Riksbank.

The Swedish case highlights the large potential reduction in fiscal costs associated with bank assistance that can be achieved through asset recovery, while still implementing a comprehensive and timely bank-restructuring program. In assessing the Swedish experience with cost recovery, however, certain aspects should be considered. First, state intervention was rapid and occurred before major losses were made. The problem was weak solvency of otherwise viable banks rather than insolvency of nonviable banks, partly because the government intervened promptly. Other affected banks were able to attract private capital injections to restore their capital adequacy. Second, just as the problems were largely due to an economic downturn, the cost recovery was greatly helped by the subsequent economic upturn. These conditions may well not be present in other cases of systemic bank restructuring, especially for less-developed countries.

Empirical Analysis of Common Characteristics of Bank Restructuring

This section supplements the discussion of the country cases. In particular, it draws general conclusions about aspects of bank restructuring that appear to be broadly characteristic of the group of restructuring countries as a whole. It is important to emphasize that only a limited number of countries implemented systemic bank-restructuring programs and, therefore, the analysis is likely to be subject to small-sample bias.

A sample of 24 countries where systemic bank-restructuring has taken place—the broadest for which comparable data are available—was selected, with a view to providing evidence on best practice policies through a statistical analysis. The sample is a representative group of countries reflecting a broad coverage across regions and levels of development. Countries were included only in cases where the problems were judged to be systemic. For the analysis that follows, “systemic” is defined as a situation where problems affected banks that, in aggregate, held at least 20 percent of the total deposits of the banking system. The sample countries listed in Table 16 include countries that have completed bank-restructuring efforts (15) and countries where restructuring is more recent (post-1994) and is still ongoing (9).105 In this respect, the main criterion is that the restructuring efforts have been brought to a close; however, the fact that the process has been completed does not necessarily imply that the restructuring has been a complete success. In effect, some countries have experienced recurrent banking sector problems.106

Table 16.Sample Countries
Country (By region)1Onset of Restructuring Action (Year)
Africa
Côte d’Ivoire1991
Ghana1989
Tanzania1992
Zambia1995 (recent)
Asia
Indonesia1994 (recent)
Korea21993
Philippines1984
Japan1995 (recent)
Europe 1
Hungary1993
Poland1993
Finland1991
Spain1980
Sweden1991
Europe 2
Latvia1995 (recent)
Moldova1995 (recent)
Kazakstan1995 (recent)
Middle East
Egypt1991
Kuwait1992
Mauritania1993
Western Hemisphere
Argentina1994 (recent)
Chile1983
Mexico1994 (recent)
Peru1991
Venezuela1994 (recent)
Industrial4
Developing15Recent9
Transition5Other15
Total24Total24

Regions are defined according, to the IMF country departments.

Data on the extent of problems in Korea are limited; Korea was nonetheless included because there was a rapid increase of nonperforming loans requiring major action by the authorities.

Regions are defined according, to the IMF country departments.

Data on the extent of problems in Korea are limited; Korea was nonetheless included because there was a rapid increase of nonperforming loans requiring major action by the authorities.

Methodology

A questionnaire was sent to country authorities and, in some cases, to IMF or World Bank staff with special expertise on banking sector problems, The study considers changes over a nine-year period for countries where the restructuring began before 1991.107 Data were requested for three points in time: the year when bank-restructuring action started (onset of action), and four years before and four years after that date (or most recent). This information and data covered five broad areas: banking structure; bank performance; banking sector institutional framework (regulatory, legal, and accounting environment); instruments of bank restructuring used; and costs and budgetary implications. Corresponding macroeconomic data were also collected for the entire period under consideration. Only a truncated period is available for episodes of bank restructuring that began after 1992. The information from the questionnaire was complemented with other studies and published materials.

In the following analysis, the data are first classified by broad country groups: industrial countries, developing countries, and transition countries. The division reflects the plausible conjecture that there would be substantial differences in the experience of these countries with respect to their initial conditions. Other important factors that may characterize country groups are the restructuring approach and the particular instruments and combinations of instruments used, and the speed and urgency with which banking sector problems were tackled. For instance, given the fundamental nature of their problems, as well as their undeveloped state of market-oriented institutions, the policy responses of transition countries might be expected to be radically different from those of other country groups. In contrast, given the presence of developed asset markets and infrastructure, it might be presumed that their experience could be sharply differentiated from other countries. The countries in the sample were then ranked by relative progress in resolving banking sector problems; that is, data on banking performance and changes in financial system intermediation capacity were used to group countries into three broad categories, ranging from “substantial” to “slow” progress.108

The next step was to relate the performance rankings obtained to the institutional and regulatory measures that the countries used in their restructuring operations, assess the impact of accompanying macroeconomic policies, and examine the extent to which the use of particular restructuring instruments contributed to success. This facilitates the assessment of whether the presence or absence of particular factors contributes to the degree of success of a bank-restructuring program and allows the empirical analysis to identify best practices that seem to be effective across a wide range of individual country experiences.

Then, to highlight the explicit or implied incentive structures of various bank-restructuring techniques, some aspects of bank-restructuring instruments were studied in more depth. The working hypothesis was that the choice and design of instruments provide information about implied incentives and determine the outcome of the restructuring operations. The role of existing or newly created institutions (central bank, ministry of finance, restructuring agencies, and so forth) in the bank-restructuring process was surveyed for each country to determine the type and extent of involvement and the distribution of the associated costs among the central bank, the government, and the banks.

Assessing the Effects of Bank-Restructuring Operations

Bank-restructuring operations have two main objectives: to restore the financial viability of the banking system (restore solvency and sustainable profitability) and to restore the system’s intermediation capacity and an appropriate level of banking services relative to aggregate economic activity. This section assesses the extent to which countries succeeded in meeting each of these objectives and then obtains measures of overall success for each country (a performance ranking), taking into account the extent to which both objectives were met. In later sections, comparative performance rankings are analyzed to identify the policies and instruments that underpin best practice bank-restructuring policies and instruments.

The technique used to establish the overall performance ranking for each country is a procedure based on the direction of the change (improvement) in each of the selected performance indicators following bank restructuring. It is constructed for each country by adding the number of indicators where improvement actually occurred. The procedure has some advantages over other procedures that might seek to quantity the magnitude of change in each indicator, including avoiding scaling problems in comparing movements in a particular indicator across countries, and avoiding a weighting problem in combining changes in disparate indicators to provide an overall ranking for a particular country. It also facilitates use of both quantitative and qualitative data and minimizes data quality problems that inevitably plague unsound banking systems.

Bank Performance

Bank performance involves the two aspects of solvency and sustainable profitability. As solvency-improving measures primarily affect banks’ balance sheets while profitability-improving measures affect banks’ income, they are referred to as “stock” measures and “flow” measures, respectively. Stock improvements in banking system performance emanate chiefly from financial restructuring operations, while sustainable flow improvements result from operational restructuring measures.

The indicators of stock improvement used here comprise the ratios of nonperforming loans to total loans; loan-loss provisions to total loans; and capital to assets. Generally speaking, banks with large holdings of troubled assets have high provisioning costs and must provide for losses on a significant portion of those assets. This reduces net earnings and, ultimately, capital. Improvements in stock effects require a reduction of the ratio of nonperforming loans to total loans, a reduction in loan-loss provisions, and an increase in capital. The flow improvement indicators comprise the ratios of operating expenses to assets; interest income to assets; and profits to assets. Unless improvements in the income position of the bank occur, an ongoing need for future restructuring efforts is likely to occur. In general, reducing expenses and increasing levels of interest income and profitability will enable banks to boost capital and improve their economic viability.

The results presented in Table 17 emphasize the differences in the extent to which particular groups of countries were able to improve bank performance. However, there is an important and striking similarity in that all groups of countries—industrial, developing, and transition—were substantially more successful in addressing stock problems than flow problems. For the “all other countries” category, the success index in solving stock problems is 71 compared with 53 for the index of flow improvement. One reason is that stock indicators can be improved more quickly. Swaps of bonds for nonperforming loans, for example, instantly improve all three stock indicators; they do not necessarily have an effect on costs, earnings, or profits. Achieving positive flow effects requires operational restructuring, which is more difficult and takes more time. Another reason appears to be that, in practice, the design of restructuring packages has been somewhat unbalanced, focusing more on financial restructuring measures at the expense of operational restructuring measures. The evidence of the relatively disappointing performance with respect to resolving flow problems may have important implications for the future of the banking sector in that it suggests the probability of recurrent banking problems and possibly the need for further bank restructuring. Second, our analysis of the experiences of the sample countries following the initiation of bank restructuring suggests that establishing or significantly improving the soundness of the banking sector is a relatively long-term process, and improvements are not always steady. In particular, countries that did not address the flow problems decisively have experienced recurrent problems in the banking sector. In some, such as Hungary and Mauritania, repeat bank restructurings were necessary.

Table 17.Improvement of Bank Performance After the Onset of Bank-Restructuring Programs1(in percent of countries in each subgroup)
Progress in Addressing Stock EffectsProgress in Addressing Flow Effects
Decline in nonperforming loans/loansDecline in loan-loss provisions/loansIncrease in capital/assetsIndex of stock success2Decline in operating expenses/assetsIncrease in interest income/assetsIncrease in profits/assetsIndex of flow success2
Recent experiences(9)675667636701126
All other countries(15)8760677171535353
Industrial countries (3)10010010010067333344
Developing countries (10)8050606350606057
Transition countries (2)10050506750505050

Measures changes between average performance in the four years prior and four years following the onset of restructuring. Figures in parentheses indicate number of countries.

Calculated as simple averages of the three indicators in each category.

Measures changes between average performance in the four years prior and four years following the onset of restructuring. Figures in parentheses indicate number of countries.

Calculated as simple averages of the three indicators in each category.

Intermediation Capacity of the Banking System

Six indicators were selected to measure the improvement in the financial intermediation capacity of the banking system following the bank-restructuring process (Table 18). These were divided into three subcategories. The first measures the scale of intermediation and includes the ratio of the growth of credit extended to the private sector relative to GDP growth and the ratio of broad money to GDP. Credit to the private sector would be expected to expand when banks are successfully restructured. Furthermore, as public confidence in the banking system rises, it can be expected that the demand for deposits will increase as depositors return to the banking system. On the other hand, an ongoing sharp increase relative to GDP growth may imply continuation of bad lending practices. A sharp decline in the ratios might also be indicative of a credit crunch. The interpretation of these figures is thus not entirely unambiguous.

Table 18.Improvement in Intermediation Ability of Banks After the Onset of Bank-Restructuring Programs1(In percent of countries in each subgroup)
Scale of IntermediationEfficiencyRisks
Increase of credit to private Sector/GDP2Increase in M2/ GDPIndex of intermediation3Decline in interest spreadsDecline in central bank credit to banks/GDPIndex of efficiency3Decline in real interest rate4No recurrence of banking problemsIndex of risk3
Recent experiences (9)11221733133344022
All other countries (15)536760535353875370
Industrial countries (3)06733676767100100100
Developing countries (10)707070404040804060
Transition countries (2)5050501001001001005075

Measures changes between average performance in the four years prior and four years following the onset of restructuring. The higher the number, the more pronounced the improvement. Figures in parentheses indicate number of countries.

Growth in credit to the private sector exceeding growth in GDP by no more than 100 percent.

Calculated as simple average of the two indicators in each category.

A decline in real interest rate or a shift to a positive real interest rate. (In most cases the deposit rate was used.)

Measures changes between average performance in the four years prior and four years following the onset of restructuring. The higher the number, the more pronounced the improvement. Figures in parentheses indicate number of countries.

Growth in credit to the private sector exceeding growth in GDP by no more than 100 percent.

Calculated as simple average of the two indicators in each category.

A decline in real interest rate or a shift to a positive real interest rate. (In most cases the deposit rate was used.)

The second subcategory measures efficiency of intermediation and includes indicators of interest spreads and the reliance of the banking system on the central bank (measured by central bank credit to banks as a percent of GDP). A decline of interest spreads and of central bank credit was interpreted as an improvement. The third subcategory measures the riskiness of the banking sector and includes changes in the real interest rate and the experiences with recurrent banking problems. An unsound banking system is likely to offer higher interest rates (to attract deposits and pay operating expenses), which may lead to higher risk through adverse selection. A decline in real interest rates (or no change) was interpreted as improvement. Repeated occurrences of systemic bank restructuring were interpreted as deterioration of risk.

The results presented in Table 18 indicate substantial variation across country groups. While countries usually were able to increase the scale of financial intermediation and reduce systemic risk following bank restructuring, less progress was typically made in improving the efficiency of financial intermediation. The latter result, in particular, is suggestive of the need for greater attention to operational restructuring measures.

Rank Ordering of Overall Country Performance in Bank Restructuring

In establishing the overall performance ranking for each country’s bank restructuring program, the six indicators of improvement in bank performance and the six indicators of improvement in financial intermediation capacity were all given equal weight.109

The resultant rank ordering permits the classification of the sample into three main groups: those that made substantial progress in restructuring their banking systems (countries that received a total score of 9 or more), those with moderate progress (countries with a total score of 6 to 9), and those with slow progress (countries with a total score of 5 or less). By construction, any one of the performance groups may contain a mixture of industrial, developing, and transition countries.

According to this grouping, five countries are included in the substantial progress category (Côte d’Ivoire, Peru, the Philippines, Spain, and Sweden); seven in the moderate progress category (Chile, Egypt, Finland, Ghana, Hungary, Korea, and Poland); and three in the slow progress category (Kuwait, Mauritania, and Tanzania). As the remaining nine countries in the sample initiated bank-restructuring measures in 1994 or later, they were categorized as recent and were not included in the performance analysis.

The following sections analyze the specific policy measures that carry the best prospects for a successful bank-restructuring program. In particular, the specific institutional reforms, the nature and design of bank-restructuring instruments, and accompanying macroeconomic policies that are characteristics of each performance group are analyzed. The analysis makes it possible to draw inferences on best practices from country experiences.

Causes of and Responses to Banking Sector Problems

Table 19 lists the principal causes of systemic banking problems and indicates to what extent the authorities took measures to address these. In addition to those that originated outside of the banking sector (exogenous, mainly macroeconomic shocks), problems could be attributed to deficient bank management and poor operational control, serious shortcomings in regulatory and accounting frameworks (the latter in part related to deficient management control), a concentration of problems in state-owned banks, and the application of excessive and distorted taxation schemes to financial institutions, for example, treatment of loan-loss provisions. Of these, management and control problems and deficiency in the regulatory framework afflicted all countries, followed in frequency of occurrence by problems with state banks.

Table 19.Diagnosis of Banking Problems and Measures Taken to Address Them(In percent of countries)
Substantial ProgressModerate ProgressSlow ProgressRecent
Deficient bank management and control
Diagnosed as a problem100100100100
Measures taken100713378
Shortcomings in regulatory and accounting framework
Diagnosed as a problem100100100100
Measures taken1001006789
State bank problems
Diagnosed as a problem608610078
Measures taken60863356
Excessive and distorted taxation
Diagnosed as a problem20296711
Measures taken201400
Exogenous shocks
Diagnosed as a problem100713389

The results for all country groups suggest that banking sector problems were never due to a single cause. In general, countries that have exhibited slow progress have had a greater number of problems to deal with than substantial progress countries. The latter have also been mostly countries where an exogenous shock triggered the need for action. But significantly, the weak performers have addressed the full range of their difficulties with a substantially lower frequency than have the moderate or substantial progress countries. In particular, the less successful performers have shown a failure or perhaps unwillingness to deal with problems in state-owned banks and, in some cases, nonfinancial public enterprises, or to tackle taxation problems that distort the incentive structure in the banking sector. By definition, substantial progress countries have dealt with all major problem areas.

The differentiation of experience among weak and strong performers can be taken as confirmation of several of the elements of what has been identified as best practice policy. First, the widespread incidence of multiple causes of banking sector problems confirms that best practices always need to begin with a diagnosis. It can be presumed from the experience of poorer performers that the failure to diagnose problems effectively leads to the design of restructuring programs that are less than fully comprehensive. Comprehensiveness, of course, is a key element of best practices. Moreover, the pervasiveness of deficient management and internal control problems further stress the need for a heavy focus on operational restructuring measures as a key element of best practice restructuring strategy. A failure to address internal management and control problems in 67 percent of the slow progress countries is correlated also with the low frequency with which they address problems in state-owned banks (67 percent) and problems of excessive and distorted taxation (zero percent). These linkages may be symptomatic of an inability to establish the strong political consensus that would be needed to deal with banking problems in a comprehensive way.

The final observation in explaining differences in countries’ degree of success in dealing with systemic banking problems appears to be the speed with which restructuring measures are undertaken. Best practices require rapid action to contain problems and minimize costs. As can be seen in Figure 2, there is a strong positive relationship between quick action and better performance.

Figure 2.Average Time Delay in Taking Bank-Restructuring Measures After Systemic Problems Arise

Source: An informal IMF survey of central banks.

Role and Effectiveness of Individual Bank-Restructuring Instruments

Many diverse instruments have been employed in bank-restructuring packages. Most of the instruments and techniques used in bank restructuring are modified versions of normal bank-management tools and strategies. Examples of commonly used business tools that are adapted for bank restructuring are the formation of specialized units to handle the problem of loan collection (“asset management”); merger with other banks; reconfigurations of core business, for example, by selling off certain product lines or branches (“splits”); and use of advisory and consulting services to improve specific aspects of bank operations (“twinning”). Where banking problems involve state banks, privatization is also a standard approach to improve the efficiency of banking. Also part of the standard repertoire of prudent banking strategies are central bank liquidity management, recourse to markets for new equity issues, as well as incentive structures to promote the effective exercise of ownership rights and ensure good management.

Table 20 lists the most frequently used instruments in bank-restructuring packages. Countries on average used eight instruments. As shown in Table 21, the industrial countries used a relatively small number of instruments (3–6). Central bank liquidity loans, bond swaps, and instruments to shift part or all of the costs to managers and owners are among those most frequently used. Contrary to what is usually viewed as best practices, instruments that place part of the burden on depositors are somewhat less widely used. Deposit insurance was in place for most of the industrial countries, while all developing countries and some transition countries introduced a blanket deposit insurance scheme in the aftermath of the banking problems. Depositors were fully compensated, in all of the sample countries, with the exception of Côte d’Ivoire, Latvia, and Spain.

Table 20.Most Frequently Used Instruments in Bank-Restructuring Packages1
Instruments Used in More Than

75 Percent of Countries
Additional Instruments Used in 60–75 Percent of Countries
Central bank liquidity loansCentral bank medium-term support
Bond swapsClosure and liquidation
Loan workoutPrivatization
Instruments aimed at owners and managers

In all 24 countries.

In all 24 countries.

Evidence of the effectiveness of particular instruments is presented in Table 22. There are significant differences among performance groups regarding the choice and frequency of use of instruments for bank restructuring. Best practices policies can be identified by examining these differences. In particular, the variation across groups is sufficiently broad to permit inferences on best practice policies regarding the role of the central bank; importance of loan-workout schemes; firm exit policy; privatization; enterprise restructuring; and incentive correction schemes. As the sample is based on a wide range of countries, it can be assumed that these conclusions on best practices of the use of instruments are robust to a wide range of particular circumstances and initial conditions.

Table 21.Instrument Mixes for Bank Restructuring1
Central

Bank

Takes

Lead
Loan-Workout

Units
ClosureMergerSplitsPrivateTwinBondCentral

Bank

Loans
Central

Bank

Liquidity

Support
New

Equity
Deposit

Instrument
Owners

and

Management2
Enterprise

Restructuring
(Sum)
Argentina101001010101107
Chile1111010111011111
Côte d’Ivoire0110010111111110
Egypt000101010100005
Finland1111010010111110
Ghana010101111100119
Hungary11111100111101112
Indonesia100000001100003
Japan011110001110108
Kazakstan1111111111001112
Korea110000011111007
Kuwait110000011110006
Latvia1110010101111110
Mauritania1111010111011010
Mexico1101001101111110
Moldova101101000000016
Peru011101110001108
Philippines1111010111011111
Poland0111011111011111
Spain011000011010106
Sweden010100010010105
Tanzania10000001110105
Venezuela110001011111109
Zambia101000001100105
Sum16181514315519172012131811Average
Percentage67756358136321797183505475468

A “1” indicates occurrence.

Provision of appropriate incentives for managers and owners.

A “1” indicates occurrence.

Provision of appropriate incentives for managers and owners.

Table 22.Instrument Mixes for Bank Restructuring(in percent)
Central Bank as Only Agency for RestructuringCentral Bank Liquidity SupportLoan-Workout UnitsClosureMergerSplitsPrivatization Where Applicable1Enterprise Restructuring2TwinBondNew EquityDeposit InstrumentsOwners and ManagementAverage Number of Instruments Used
Substantial progress204010080600100402010060601008
Moderate progress5786865786141007129865757719
Slow progress100100673333033001003367337
Recent8910056674422714422564444788

Applicable only for countries that experienced problems specific to state-owned banks or state enterprises.

May not be applicable for some countries.

Applicable only for countries that experienced problems specific to state-owned banks or state enterprises.

May not be applicable for some countries.

The designation of the central bank as the sole agency for restructuring and provider of liquidity support was limited by the countries that were most successful in their systemic restructuring operations. This may partly reflect the fact that where there was a broad political consensus for comprehensive restructuring, it was carried out by specialized agencies to allow the central bank to continue to focus on its main function of implementing monetary policy. In particular, the authorities that achieved the best results determined at an early stage that the problem was bank insolvency, not lack of liquidity, and they precluded extensive use of lender-of-last-resort facilities. In contrast, all of the slow progress countries made extensive use of central bank instruments; in all of these countries, the central bank was the only agency responsible for bank restructuring. This may be a sign of lack of coordination and consensus between different institutions. Thus, it can be inferred that best practice policy is to minimize reliance on the central bank as a source of protracted liquidity support.

By contrast, the sample results also suggest that it was necessary for central banks to take the lead in most of the transition countries (Table 21). It appears that this choice has been strongly influenced by the limited availability of skilled human resources. Given the scarcity of banking expertise in the public sector, the central bank may be the only agency capable of addressing technical details of bank restructuring.

Table 22 also suggests that loan-workout units (central or bank-based) played an important role in all countries that made substantial progress in resolving systemic banking problems, while only about 70 percent of the slow progress countries established loan-workout schemes. It can be inferred here, too, that the use of distinct loan-workout units appears to be an important element of best practices (see Table 23).

Table 23.Loan-Workout Arrangements, by Performance Groups
CountryType of AgencyEstablished (Funded) byPerformance Indicators
Substantial progress countries
Côte d’IvoireBank liquidation agency.Government (1993).Strictly monitored. 45 percent of targeted recoveries achieved (1993–96).
PeruCOFIDE (Trust) managed by a special commission.Government.No performance indicators reported.
PhilippinesAsset Privatization Trust for assets of two state banks.Government (1986). Government assumed corresponding amount of bank liabilities.No information available.
SpainDeposit Insurance Company in charge of bank resolution.Central bank and commercial banks.Resolution completed.
SwedenMinistry of Finance helped establish 2 bank-based government owned AMCs capitalized with public funds. (AMCs later merged.) (Other banks also established AMCs without government assistance.)Government.Proceeds from asset sales are likely to cover initial government outlays.
Moderate progress countries
ChileCentral bank held nonperforming assets.Central bank.Repurchase obligations for old bank owners linked to dividend payments of banks. (Two classes of bank stock were created, for old and for new owners.)
FinlandCentral bank established asset-management unit.Central bank (1991).Asset management was not closely monitored and not given a high
Government established asset-management unitParliamentary Act (1993).priority until 1991.
GhanaAsset Recovery Trust.Government.No information on performance available
HungaryHungarian Investment and Development Bank manages assets for the government.Bank designated by government. No new agency created (1992).No information on performance available.
KoreaMultiyear loan write-off plan for each problem bank.Plans worked out with central bank.Nonperforming loans written off. No information on loan
No government assistance.collection.
PolandState-owned banks formed workout units.Directive by the Ministry of Finance.Banks to follow outlined plan for loan recovery and capital adequacy.
Slow progress countries
MauritaniaLoan Recovery Agency.Government (1993).No information on performance available,
Countries with ongoing restructuring progams
KazakstanGovernment (1994–95).Approach considered successful by some experts, but no detailed information is available.
Government assumes responsibility for foreign exchange or trade-related credit with government guarantees.
Rehabilitation Bank for loans of 30 most problematic enterprises (1995).
Agricultural Support Fund for nonperforming agricultural credits (1994).
Banks established workout units (1995).
MexicoTrust Fund holds assets but does not perform loan workout.Established by the government.Loan workout is performed by banks. (Repurchase obligations.)
MoldovaBanks required to establish loan-workout units at their own expense.Government rules but no funding.No information on performance.

Most of the countries exhibiting substantial and moderate progress made extensive use of mergers and closure of insolvent banks. This confirms the importance of firm exit policies. Furthermore, as shown above in Table 19, in about 60 percent of the countries with substantial progress and in most of the countries having moderate progress, problems with state banks and state enterprises contributed to banking system distress. All of the countries dealt with the problems of insolvent state banks and state enterprises particularly through privatization, enterprise restructuring, and closure. However, these policies were consistently avoided by countries exhibiting slow progress, which may imply that insolvent banks were allowed to operate, leading to a further deterioration in the conditions of these banks.

Bond instruments (such as an exchange of bonds for nonperforming loans) and issuance of new equity (e.g., equity purchased by the government) were widely used by all countries. However, such expenditures did not always seem to be disclosed in the budget. Splits and twinning with foreign banks as instruments for bank restructuring were methods mainly used in transition countries.

Similarly, provision of appropriate incentives for managers and owners is a key element of best practices. This is shown in the column labeled “Owners and Management” in Table 22, which suggests that all of the substantial progress countries emphasized the use of incentive corrective schemes, which further strengthened the market-based approach taken by these countries. Banks receiving support were almost always downsized. Only a minority of the countries with slow progress took measures to sanction management and owners, and little evidence was provided for stringent use of incentive compatibility within state-owned banks.

Importance of Incentives in Instrument Design and Use

The instruments and strategies that are typically employed in bank restructuring are also used in a functioning market by bank managers, whose ultimate aim is to increase profits. However, the same techniques can be used by the authorities to help alleviate systemic problems, for example, to prevent bank failure, to bail out depositors, or to transfer banking skills to poorly performing banks at below market costs. In the context of bank restructuring, the objectives of such instruments are, thus, likely to be at variance with normal business objectives and serious moral hazard problems may arise. Therefore, it is necessary to explicitly state the objective of any given instrument as well as the “corrective” for moral hazard problems.

Table 24 provides information on conditions and costs of instruments of bank restructuring that have been used in the sample countries. The second column provides some examples, drawn from the sample, of conditions attached to these instruments in order to introduce a corrective for moral hazard. The third column states the fiscal implications and other costs that might result from each instrument. “Other costs” include quasi-fiscal costs that are not directly shown in the budget but may, at a later point, have budgetary implications. For instance, to the extent that the central bank assumes part of the costs, the budgetary implication can be an indirect one (reduced remittances by the central bank). (Table 12 provides estimated costs of bank restructuring for individual countries.)

Table 24.Instruments and Costs of Systemic Bank Restructuring
InstrumentsConditions Attached as Corrective for Moral Hazard (Examples)Fiscal Impact and Other Costs
Structural
Bank Rehabilitation Agency to take lead,1 nationalization.Conservatorship (management and ownership rights are suspended or revoked in case of nationalization).Budget allocation for the lead agency or for nationalization.
Central Asset Management Company (AMC).Bank-client relation is discontinued when asset is transferred to AMC. AMC has limited statute and continuous performance reviews.Budget allocation for AMC.
Bank-led loan workout (includes bank-based AMCs).Any subsidies tied to workout are subject to specified conditions.Direct support from the government.
Closure and liquidation.Outline liquidation plan and resources needed to liquidate bank.With courts overloaded, human resource cost is often borne by central bank that administers liquidation process.
Merger.Specification of conditionality for any payments or transfers made to the absorbing institution.Government may pay fee to acquire bank. Bank may absorb the costs.
Split-offs (turning branches or a part of a bank into an independent bank/ investment agency).Commercial viability of new bank must be proven. Fit and proper owners must be available.Start-up capital to new bank.
Privatization.Ensure efficient governance structure, fit and proper new owners. Anticipate “Bank Slaski problem” where equity prices shoot up when foreign bank buys shares. Government may be blamed for selling state property too cheaply.Administration of privatization, bank audits, investment banking services, and lower profit transfer.
Twinning (technical assistance usually provided by an international bank).Determine why the foreign bank is interested (e.g., foreign bank gets subsidy from its technical assistance donor; seeks to purchase market share in target country. Ensure competitive bidding where applicable).Fees paid to foreign banks.
Financial
Bond transfers to banks; subsidized government loans to banks.Bank may try to sell bonds for cash, take the money, and run. Correctives must be built in, e.g., by accordion style arrangements where bond transfer is contingent upon improvement in bank operations; nonnegotiable bonds.Bank audit. Bond issues. Cost is a function of terms and conditions (interest rate, currency, maturity and so forth). Cash transfers.
Central bank loans to banks; and other central bank support measures.If not fully collateralized, may need to be coupled with conservatorship, liquidation, or restructuring strategy.Shortfall of central bank’s profit remittance to budget; central bank reduces its demand for treasury bills; government costs of funding rise.
Equity injection (government buys equity shares).Ownership stakes would need to be calculated on the basis of market equity prices. (If zero, government may become sole owner.)Cash outlays for equity.
Depositor targeted instruments. (Blanket guarantees promised by the government; when banks fail depositors lose money; depositors receive bonds for deposits.)Incentive compatibility of deposit remuneration. When depositors lose money, contagion effect may follow. Specify marketability of bonds in case of swaps.Transfer to deposit insurance; bond issue.
Other instruments targeting owners, management, employees.Owner liability; merit-based management contracts; waivers for payroll tax arrears, severance pay, and so forth.Unemployment compensations; budgetary loss.
Enterprise Rehabilitation (support may be given to enterprises to pay off banks).Define operational restructuring of the enterprises and the role to be played by banks.Guarantees, bonds, budget allocation.
Inaction (time prior to restructuring).Not applicable.In the short term: government revenue shortfalls; arrears in payroll taxes and social security contributions from banks: bank owners or managers operating under perverse incentives may make more bad loans and loot banks’ assets, foreign exchange accounts, and so forth. In the long run: losses increase, requiring higher fiscal support.

To denote the lead agency (if any) during the restructuring process (deposit insurance fund; bank support authority; agency exercising state ownership rights; conservator-liquidator for banks).

To denote the lead agency (if any) during the restructuring process (deposit insurance fund; bank support authority; agency exercising state ownership rights; conservator-liquidator for banks).

The instruments were divided into structural and financial instruments. Structural instruments are those that directly affect the structure of the banking sector. Financial instruments are those that directly affect the banks’ balance sheets and income.

Role of the Central Bank in Restructuring

The preceding discussion of the use of bank-restructuring instruments shows that a good predictor of country performance is the extent to which the restructuring program emphasizes instruments other than central bank liquidity support. As shown in Table 25, most of the substantial progress countries refrained from using them and countries that made extensive use of central bank instruments typically made less progress in bank restructuring.

Table 25.Central Bank Leadership, Liquidity, and Other Support in Bank Restructuring(In percent)
Central Bank Acts

as Lead Agency
Central Bank

Liquidity Support
Medium-Term Central

Bank Support1
Substantial progress (5)22040603
Moderate progress (7)578686
Slow progress (3)100100100
All countries (24)678371

Loans and other forms of assistance.

Figures in parentheses indicate number of countries.

Although three out of five countries initially used medium-tern support, it was subsequently phased out in two of them.

Loans and other forms of assistance.

Figures in parentheses indicate number of countries.

Although three out of five countries initially used medium-tern support, it was subsequently phased out in two of them.

In some countries, the authorities have chosen to have the central bank act as lead agency in the bank-restructuring process and to assume extensive responsibilities in addition to its core monetary policy functions, including financial support, bank management, and asset (nonperforming loans) management (Chile and Kuwait). This can create certain difficulties because liquidity support to insolvent banks provides perverse incentives to banks and fails to address the underlying problem; direct ownership in banks and medium-term lending by the central bank produces conflicts of interest, especially when the central bank has supervisory responsibilities; and central banks are left with large structural positions.

Table 26 provides a review of the central bank instruments used in the bank-restructuring process in the three performance groups. Several countries (e.g., Peru and Sweden) have placed strict limitations on central bank short- and long-term finance when systemic banking problems arose. These countries were able to make progress in implementing their bank-restructuring strategies. Indeed, firm restrictions on the active involvement of the central bank appear to be an ingredient for successful bank restructuring.110

Table 26.Central Bank Instruments and Bank Restructuring by Performance Groups
CountryRole of the Central Bank
Substantial progress countries
Peru

Sweden
The central bank played a limited role in the bank-restructuring process. In Peru, the central banking law was changed restricting remaining central bank activity in banking.
SpainThe central bank played a lead role; however, its lending was done indirectly through a deposit insurance agency jointly owned by the banks. While bank restructuring was successful, it entailed very high costs (15 percent of GDP).
Côte d’Ivoire

Philippines
The central bank initially lent to government to help reduce government arrears on bank loans. The central bank role was phased out as part of concomitant reforms of bank-restructuring activity.
Moderate progress countries
Finland

Korea
In both countries, the central bank played a limited role. Nonetheless, in Finland, the central bank assumed equity stakes in one bank and organized a loan-workout agency for another bank.
Hungary

Poland
Reserve requirements and remuneration were adjusted to increase bank liquidity. In Hungary, the central bank provided bridge loans, some direct credit to banks and consolidation bonds. In Poland, the central bank discounted bonds of recovered banks.
Chile

Ghana
In Chile, the central bank assumed a major role in the restructuring process, including extensive direct lending to enterprises through commercial banks (pass-through loans). Bank restructuring was extremely costly (33 percent of GDP).
Slow progress countries
KuwaitThe central bank played a key role in the bank-restructuring process by providing liquidity through a broad application of discounting, repurchases, special facilities, as well as through “special deposits” at negative real rates held with the commercial banks. The central bank also became involved in debt collection and other aspects of the debt-restructuring process.
Mauritania

Tanzania
The central bank engaged in significant long-term lending to insolvent banks.

Moreover, restrictions should apply not only to the use of central bank financing, but also to ancillary activities that have little to do with core central bank activities. In some of the less successful restructuring experiences, for example, central bank involvement has extended to commercial bank management and ownership, loan workout, and credit allocation.

In the transition countries where bank restructuring progressed rapidly (Hungary and Poland), the central bank played an active role. However, it appears that the central bank reduced its involvement over time and placed great emphasis on appropriate incentives for banks. To better understand the importance of how central bank instruments have been used by the sample countries, a more detailed analysis of the various instruments, their costs, and incentives is presented in Table 27.

Table 27.Central Bank Instruments Supporting Bank Restructuring
Instrument (Countries Where It Was Used)Central Bank Balance Sheet Effects and Income EffectsBudgetary CostCost for Banks (Incentive Aspect)
Central bank organizes liquidity support among the banks; does not commit its own resources. (Côte d’Ivoire)NoneNoneAll costs borne by banks. Incentive:

Positive
Central bank organizes liquidity support among the banks; reduces reserve requirements explicitly for that purpose. (Venezuela)Liabilities fall; central bank may need to sell treasury bills or other assets.Yes. Central bank income will fall, reducing profit remittances to the government. The central bank may incurr losses requiring transfers from the government budget. In the case of foreign exchange support the central bank may incur capital losses when reserves are drawn down.Cost for banks reduced. Incentive:

Positive
“Credibility” strategy, announcing stabilization measures and benchmarks accompanied with more frequent and better disclosure (weekly publication of central bank balance sheet). (Mexico)Can be very substantial, for example, if currency stabilization measures become necessary, which absorb substantial portions of central bank’s foreign reserves.Costs increased. Funding costs rise and loan defaults may increase as interest rates rise. Tiering may result, strong versus weak banks.
There are also second round fiscal effects.Incentive:

Positive
Temporary or permanent reduction of reserve requirements in response to liquidity problems. (Venezuela, Spain, Argentina, Hungary)Liabilities fall; central bank may need to sell treasury bills or other assets.When central bank liabilities fall; the central bank may reduce its stock of or future demand for treasury bills, increasing the government’s cost of borrowing.Costs reduced. Incentive:

Positive
Increase in reserve requirements to counteract budgetary expansion coupled with increase in remuneration to preserve bank profitability. (Hungary, 1994)Revenue rises but so do expenses. Net effect depends on the magnitudes.Costs reduced. Incentive:

In determined
Short-term loans to banks fully collateralized, quasi-market rates. (Hungary, Indonesia)Short-term assets expand; earnings rise.NoneCosts reduced. Incentive:

Positive
Broad application of discounting, widening the range of acceptable collateral to lower quality paper. May include bank stock. (Argentina, Kuwait)Reduces the quality of Central bank assets; reduces income.Central bank income falls (see above for further details).Costs reduced. Incentive:

Adverse
Long-term loan to the deposit insurance agency (or other bank-restructuring agency) replacing short-term bank loans (banks might otherwise have defaulted). (Venezuela)Loans to banks fall; low quality assets are transformed into high quality assets.Government pays interest/amortization and guarantees central bank loan).Costs reduced in the short run, may rise later if deposit insurance pricing is adjusted. Incentive:

Positive
Long-term loans to banks at below-market rates, also subordinated debt (Kuwait, Chile, Hungary, Indonesia)Central bank establishes structural position, incurs interest costs.Central bank income falls (see above for more details).Costs reduced. Incentive:

Adverse
Foreign exchange guarantees given to banks (e.g., for liabilities denominated in foreign exchange when foreign exchange creditors are defaulting due to major devaluation). (Mexico)Foreign reserves reduced.Central bank reserves and income from capital fall (see above for more details).Costs reduced. Incentive:

Adverse
Preferential foreign exchange rates either for bank borrowers or for banks to honor foreign exchange liabilities. (Chile, Mexico)Foreign reserves sold at a loss.Costs reduced. Incentive:

Adverse
Central bank issues bills to banks in exchange for nonperforming loans, assumes workout responsibility. (Finland, Kuwait, Chile, Hungary)Asset quality falls (depending on pricing of nonperforming loans); earnings fall. Central bank acquires structural position. Activities of central bank expand substantially, administration, loan-management and loan-workout activity requires substantial resources.Central bank income falls (see above for more details).Costs reduced. Incentive:

depending on pricing and design
Central bank takes over management of banks temporarily. (Chile, Kazakstan)Operating costs increase as Central bank engages in active monitoring and sometimes management of banks.Incentives:

Positive
Central bank purchases equity stakes in banks. (Finland, Indonesia)Long-term commitment of funds, with uncertain return, possibly further costs.Costs reduced. Incentives:

Adverse

In using instruments in support of bank restructuring, some central banks have limited themselves to providing temporary (mostly short-term) support, which was replaced by other sources (government budget) when the bank-restructuring strategy was put in place (Argentina, Kazakstan, Latvia, Mauritania). In Mexico, the central bank provided some of the support to banks via a government agency, thus protecting its own asset quality and drawing on government guarantees. In these countries, bank-restructuring strategies are ongoing.

There are some central bank instruments that may have no budgetary implications but can have strong positive incentive effects. These include liquidity support measures that are arranged by the central bank from within the banking community. Credibility policy, where the central bank attempts to exert a stabilizing influence on financial markets by pronouncing “once-and-for-all” policy guidelines and goals can also have no consequences for the budget. However, such a policy can also be very costly if the central bank fails to establish its credibility. This policy was used in Mexico, but as bank restructuring is ongoing, there is no empirical evidence for the success or failure of this strategy.

In virtually all other cases, central bank support to banks has indirect budgetary implications (fiscal costs or revenue shortfalls). The two main channels are reduced revenue resulting from lower bank income or higher costs and reduced demand for treasury bills when the bank’s liabilities fall or when the central bank engages in asset substitution to absorb nonperforming loans, leading to higher costs of government debt.

Macroeconomic Developments During the Restructuring Process

The economic background in the sample countries, against which the bank-restructuring operations took place, was analyzed to determine whether outcomes and best practice policies were sensitive to underlying economic conditions.111

It was possible to discern three broad patterns for GDP growth, inflation and the fiscal balance (Figure 3 and Table 28). One pattern is U-shaped, that is, in some countries macroeconomic conditions deteriorated slowly in the four years preceding bank restructuring and worsened significantly at the onset of bank restructuring, but they recovered in the following years. The fiscal balance improved with a considerable time lag, often reflecting the cost of bank restructuring.112 This pattern is best represented by Sweden. Average GDP growth in the four years prior to the onset of bank restructuring (1991) was about 2 percent, growth turned negative in 1991 and continued to be negative both in 1992 and 1993 by some 2 percent annually as the banking crisis culminated toward the end of 1992 and beginning of 1993, recovering to growth rates of about 3½ percent in both 1994 and 1995. Inflation showed a somewhat similar trend, although the pattern is a result of the temporary upturn in inflation due to the large depreciation of the currency. The fiscal balance was positive in the four years before the banking problems, but deteriorated sharply in 1991, as well as in the following two years and then began to recover in the third year after the onset of bank restructuring. A similar evolution of macroeconomic conditions can be observed for Chile, Finland, the Philippines, and Spain.

Table 28.Three Patterns of Macroeconomic Effects During Bank-Restructuring Process1
Average of Four Years Prior to OnsetYear of Onset of Restructuring (N)Average of Four Years After N
Pattern 1
Chile (N = 1983)
Real GDP growth2.33−3.495.44
Fiscal balance2.74−2.80−2.18
Inflation24.5427.2622.48
Finland (N = 1991)
Real GDP growth3.67−7.070.96
Fiscal balance1.42−4.48−10.01
Inflation5.474.161.79
Philippines (N = 1984)
Real GDP growth3.48−7.321.79
Fiscal balance−3.07−1.87−3.44
Inflation12.6550.329.20
Spain (N = 1980)
Real GDP growth2.081.221.17
Fiscal balance−3.17−2.38−4.69
Inflation18.6615.6113.11
Sweden (N = 1991)
Real GDP growth2.28−1.120.50
Fiscal balance2.02−4.00−11.87
Inflation6.739.342.91
Pattern 2
Côte d’Ivoire (N = 1991)
Real GDP growth0.660.002.07
Fiscal balance−12.29−9.87−7.85
Inflation3.561.5711.65
Hungary (N = 1993)
Real GDP growth−4.44−0.601.87
Fiscal balance−2.89−7.43−3.53
Inflation25.7722.4823.47
Mauritania (N = 1993)
Real GDP growth1.194.864.30
Fiscal balance−5.27−7.742.21
Inflation7.789.334.54
Peru (N = 1991)
Real GDP growth−3.962.876.18
Fiscal balance−6.11−2.42−2.57
Inflation2,908.27409.5039.22
Poland (N = 1993)
Real GDP growth−3.943.796.00
Fiscal balance−4.24−2.82−2.64
Inflation237.5535.3026.43
Pattern 3
Ghana (N = 1989)
Real GDP growth5.185.094.39
Fiscal balance−2.89−2.12−4.94
Inflation26.4725.2622.55
Tanzania (N = 1992)
Real GDP growth4.464.584.12
Fiscal balance−5.64−2.33−6.83
Inflation30.8224.8028.25
Source: World Economic Outlook database.

The fiscal balance is calculated as the central government balance as a percentage of nominal GDP; and inflation is calculated as the percentage change of consumer prices.

Source: World Economic Outlook database.

The fiscal balance is calculated as the central government balance as a percentage of nominal GDP; and inflation is calculated as the percentage change of consumer prices.

Figure 3.Patterns of Macroeconomic Effects During the Bank-Restructuring Process1

(In percent)
(In percent)

Source: World Economic Outlook database

1Covers four years before and four years after the onset of bank restructuring (a total of nine years). The fiscal balance is calculated as the central government balance as a percentage of nominal GDP; inflation is calculated as the percentage change of consumer prices.

The second pattern shows a steady improvement of macroeconomic conditions throughout the nine years. Countries that fit this pattern had experienced significant economic deterioration during the four-year period prior to undertaking bank restructuring; they adopted stabilization policies along with measures to stabilize the banking sector. Thus, bank restructuring does not appear to have been incompatible with economic recovery or with rapid economic growth. This pattern is most accentuated in Peru, where GDP growth rose from an average of about -5 percent during the pre-bank-restructuring years to an average of about 7 percent in the four years thereafter. Inflation fell from close to 4000 percent to 23 percent during the same period, and the government balance rose from -5 percent to about -2 percent. A similar pattern can be observed for Côte d’Ivoire, Hungary, Mauritania, and Poland.

The third pattern shows a slow but steady deterioration of certain macroeconomic indicators. An example is Ghana, where real GDP growth fell from an average of 5 percent to an average of 4 percent during the nine-year time period. Inflation fell from an average of 30 percent to an average of 22 percent, while the fiscal deficit rose from about 3 percent to almost 5 percent of GDP. A similar pattern can be observed in Tanzania. This pattern may be a variant of the first observed pattern, with a less pronounced deterioration at the onset of bank-restructuring action and with a much slower recovery.

While these patterns are interesting in their own right, they do not support the view that there is a strong link between underlying economic conditions and the success of restructuring operations. An environment of strong economic growth is conducive to successful bank-restructuring operations. Since bank profitability and retained earnings, and the underlying health of bank borrowers respond positively to economic growth, the empirical results indicate cases where measures have succeeded even where the macro situation remained weak. This is consistent with the best practice view that action should be taken promptly, without waiting for a serendipitous upturn in economic conditions to undertake otherwise difficult and unpalatable measures.

In contrast to the differentiated experience regarding the evolution of the economic cycle before, during, and after bank restructuring, inflation followed a very definite pattern, declining in nearly all countries in the survey during the years after the onset of bank-restructuring action. One possible reason is that countries recognized that best practices do not involve inflating one’s way out of banking system problems. Another is that systemic banking problems often involve a large negative demand shock, for example, as a result of the associated wealth losses, which dominated other incipient inflation pressures, particularly in effects on aggregate demand.113 The lesson would appear to be that the probabilities strongly favor that restructuring will occur in a disinflationary environment. If so, the monetary policy asymmetry problem (i.e., there are limits on the extent to which monetary conditions can be tightened during restructuring, as discussed elsewhere in this book) may not be a binding problem in practice.

Most other macroeconomic indicators showed mixed results during the nine years for the sample countries. No clear patterns are visible for either of the two groupings (initial grouping and performance grouping). Changes in private consumption and savings vary widely across countries and no clear trends can be found. Current account deficits in the balance of payments relative to GDP also do not reveal clear trends. Some countries experience exchange rate shocks in the year(s) prior to the banking problems. In some countries the exchange rate continues to deteriorate in the following years, while in others the exchange rate stabilizes in die years after the onset of bank-restructuring action. Similarly, there is no regularity in developments in gross external reserves.

Lessons from Experience

Based on the case studies of Chile; Côte d’Ivoire; Mauritania; the Philippines; Poland; Spain; Sweden; the United States; the Baltic states, Russia, and other countries of the former Soviet Union; as well as the statistical analysis of a broader group of countries, this section provides a summary of policies judged to be successful and sufficiently robust to be useful in a wide range of circumstances and countries.

Diagnosis of the nature and extent of systemic banking problems proved to be an important component of the restructuring programs. In all countries, multiple causes contributed to the systemic problems. The statistical survey indicates that countries with substantial progress identified the underlying causes and designed a bank-restructuring strategy aimed at systematically addressing each one.

Thus, successful bank restructuring implies a comprehensive approach addressing not only the immediate stock and flow problems of weak and insolvent banks but also correcting shortcomings in the accounting, legal, and regulatory framework while improving supervision and compliance. Structural factors that stand in the way of efficient financial intermediation, such as exceedingly high reserve or liquidity requirements, interest rate controls, and distortions in the tax system, such as tax exemptions for state banks, may need to be removed. The case studies show that Chile, Côte d’Ivoire, the Philippines, Poland, Spain, and Sweden implemented far-reaching reforms in the banking sector as part of the bank-restructuring strategy. They also illustrate that the bank supervisory agency and the central bank have important roles to play in addressing and monitoring these aspects of the bank-restructuring strategy.

Prompt action is an important ingredient of success. Sweden and the United States emphasized and engaged in prompt corrective action and concluded that this was a crucial component of successful bank restructuring. The survey confirms that success is positively correlated with prompt action (Figure 2). Substantial progress countries took action within one year of problems emerging. The case studies of Chile and Côte d’Ivoire illustrate that action is sometimes delayed by several months because the authorities may not have the legal powers to intervene or because time is needed to determine the causes and the most appropriate action. However, both countries took speedy and comprehensive action once these difficulties had been overcome.

Operational restructuring is a necessary condition for banks to return to profitability and sustained solvency. As illustrated in the case studies of Côte d’Ivoire, Chile, Spain, Sweden, and the United States, management deficiencies were an important cause of the banking problems. This was recognized by the authorities and action was taken to address these problems, including the strengthening of banks’ risk-management systems and the replacement of management and owners. The experience of Mauritania illustrates that when financial support is given to banks without restructuring the banks’ internal operations, problems will recur as management continues to mismanage funds. The survey confirms that management deficiencies were identified as a cause of the banking problems in all sample countries and that progress in bank restructuring is highly correlated with whether or not these were addressed. All substantial progress and most moderate progress countries placed appropriate emphasis on operational restructuring, while the weaker performers generally neglected it.

Systemic bank restructuring should be coordinated and implemented by a designated lead agency. The case studies of Sweden and the United States illustrate that while the cooperation of the government, the central bank, and the bank supervisory authorities is necessary, the lead agency should have some degree of autonomy backed by a firm and unambiguous commitment to reform at the highest levels of government. Sweden formed a separate agency; in Spain and the United States, the deposit insurance agencies acted as lead agency; in Côte d’Ivoire, external donors played an important role in comanaging the bank-restructuring process. The survey shows that, when the central bank is the lead agency, frequently it is drawn into financing the bank-restructuring measures, exceeding its resources and conflicting with its other responsibilities.

Continuous monitoring of the bank-restructuring policies and of individual bank-restructuring operations is necessary. Côte d’Ivoire, Spain, Sweden, and the United States placed great emphasis on this aspect. The importance of monitoring is further supported by the finding of the survey that bank restructuring is a multiyear process, including significant public expenditure.

The central bank must stand ready to provide liquidity support during restructuring to viable banks. Many countries used temporary or permanent reduction of reserve requirements, and broad application of discounting facilities or short-term loans as a means of providing liquidity. The central bank should not provide long-term financing to banks, nor should it be involved in commercial banking activities, as this exceeds its financial resources and leads to quasi-fiscal costs. It also creates conflicts with its monetary policy objectives. The survey shows that very few countries refrained from using short-term liquidity support; however, countries exhibiting substantial progress took a conscious decision to minimize the use of central bank financing and avoid central bank lending to insolvent banks. The experience of the United States suggests that bank restructuring becomes more costly when the central bank lends to insolvent banks. This is supported by the case of Chile where extensive central bank lending to insolvent banks was associated with exceedingly high costs of bank restructuring.

Firm exit policies are an integral part of best practices. Closure was emphasized in Chile, Côte d’Ivoire, the United States, and in transition countries. Transition countries often made use of bank closure involving small private banks. Côte d’Ivoire and Latvia demonstrated to their banking communities that no bank would be protected exclusively because it was “too large to fail.” The survey confirms that most of the substantial progress countries used firm exit policies.

Government financial support of insolvent banks is unavoidable in most instances. As shown in the survey, bond transfers and other financial instruments were widely used but were not always associated with success. The country cases show that financial instruments are useful to improve the banks’ financial condition provided that they are designed in incentive-compatible ways and are used in conjunction with operational restructuring. In the Philippines, Poland, Spain, and Sweden, financial support was accompanied by detailed plans outlining operational restructuring targets and change in management.

The principle of loss-sharing between the state, the banks, and the public is an integral part of successful bank restructuring. One way of incorporating loss-sharing arrangements into the overall strategy is to designate a deposit insurance agency funded by contributions from banks as lead agency, as was the case in Spain and in the United States. Although the authorities avoid imposing outright losses on depositors in most countries, Côte d’Ivoire, Latvia, and Spain have successfully imposed limited losses on depositors and other creditors without causing a panic or run on banks.

Removing nonperforming loans from the banks’ balance sheets and transferring them to a separate loan recovery agency is an effective way of addressing the banks’ stock problem. The survey shows that most substantial and moderate progress countries made use of this technique. Carving out nonperforming loans immediately improves the banks’ balance sheet and helps banks focus attention on their core business. It does not, however, solve the banks’ flow problems. This result is confirmed in the survey, which indicates that most countries found it easier to address the stock problems than the flow problems.

Loan workout (foreclosure or asset sales) is important to recover some of the costs of bank restructuring and to send signals to delinquent borrowers. Loan workout can be done in a central organization, usually operated by the state, or in special loan collection agencies tied to individual banks. The case studies and the survey suggest that the institutional setting does not appear to matter. However, the cases of the United States and Sweden show that close monitoring of results from the workout process can be a key ingredient of efficient loan resolution. Some countries, including Chile, the Philippines, and the transition countries, approached the loan-workout issue indirectly by providing debt relief to borrowers or by engaging in enterprise restructuring.

While bank-restructuring programs may be initiated during a time of economic stagnation, positive economic growth helps banks to resume lending and return to profitability. Cote d’Ivoire is an example of a country where successful bank restructuring was started prior to the economic recovery. The survey shows that restructuring programs typically occur in an environment of low or moderate levels of inflation while the fiscal balance often deteriorates immediately following the onset of bank restructuring, but as the case studies of the Philippines and Mauritania show, fiscal adjustment can be achieved while restructuring banks.

Problems that are specific to state-owned banks or to development banks may require special attention. Privatization or closure of such banks worked well in many countries. The design of privatization is very important in determining the future profitability and viability of the banking sector. The experience of Chile in the early 1980s and of Mexico in 1995 demonstrates that a rapid and ill-designed process of bank privatization can contain the seeds of subsequent banking crises. Chile and Mexico went through an intensive process of bank privatization in 1974 and 1991, respectively. In both cases, preferential access to credit given to some bidders, overpricing of bank assets, and weak legislation against concentration of ownership allowed a few large business conglomerates to acquire a large portion of the financial system. In both cases, all of these banks were later intervened by the government, for either being insolvent or having a high lending concentration in affiliated companies.

    Other Resources Citing This Publication