This paper takes the recent experience of Indonesia in closing, and otherwise taking over, banks as an example for drawing conclusions about bank closures. The experience is highly relevant; there were four major sets of bank closures between November 1997 and March 1999, and several sets of open bank resolution through bank takeovers; the authorities have also sought to address bank insolvency in some cases through recapitalization. To some extent these approaches are complementary, but to some extent also substitutes. The process in Indonesia has been very controversial, with the initial closures in particular subject to much criticism. The more recent closures, on the other hand, have been viewed much more positively by observers.
Interventions in banks are often an integral element of a government’s program for addressing a systemic banking crisis. Interventions may be warranted because the banks concerned are deeply insolvent or riddled with fraud; they may be requiring substantial liquidity support from the central bank, thus putting strains on the operation of monetary policy. Interventions may be through “closures” or “open bank resolution”, i.e. where a bank remains open for business, but under new rules for the conduct of business (such as new owners) or maybe as part of another institution. In either case, the cost of the intervention arises not only at the time of intervention, but as the assets and liabilities of the banks are dealt with.
Closures may be more cost effective than seeking some form of open bank resolution, such as recapitalization or purchase of problem assets. A closure strategy mitigates the moral hazard problems that would arise with any “bail out” of the bank; it may bring about a necessary reduction in banking sector capacity, thus improving the viability of the remaining banks; it may enable the government to share the costs to the public deriving from banks’ insolvency, even in a situation where depositors are protected, by inflicting losses on the banks’ shareholders and subordinated debt holders; and it may demonstrate that the government is serious about addressing the banking system’s problems.
On the other hand, bank closures may be risky. Most significantly, where the extent of a crisis may not have been fully recognized by the public, the impact of the closures may lead to concerns about the remaining banks, and thus provoke bank runs; closures may disrupt credit or payment flows; and may inflict losses on vulnerable or, conversely, on powerful members of the community. Bank closures may be difficult to organize without provoking asset stripping by owners, managers, or workers. Closure will only be one element of the handling of a failed bank: loan collections may fall after a bank closure, thus increasing the ultimate cost of resolving the insolvency of the institution.
The focus on bank closures, and other forms of bank intervention, in this paper does not mean to suggest that they are the only element in a bank restructuring process; indeed, they may not even be the main element. Governments need to introduce a comprehensive program, and closures may be part of this program.2 Indeed, in Indonesia’s case, the start of the closure process coincided with agreement on a comprehensive adjustment program—including a range of measures to address weaknesses in the banking system and more widely—to be supported by the IMF; subsequent measures too have been embedded in programs supported by the IMF, the World Bank, and the Asian Development Bank.
This paper looks at the five principal examples of interventions into the banks in Indonesia (most of which included bank closures) during the banking crisis: in November 1997, February 1998, April 1998, August 1998, and March 1999. The main instruments employed were closure; “soft” open bank intervention by the Indonesian Bank Restructuring Agency (IBRA); full takeover by IBRA; and recapitalization. Lessons from the individual experiences are presented in each sector, and overall. Conclusions are drawn at the end.
II. November 1997
A. Bank Closures and their Aftermath
In late 1997 Indonesia began negotiations with the IMF on a comprehensive adjustment program. Since the unpegging of the Thai baht in July 1997, the rupiah had been under severe downward pressure. The authorities had abandoned the peg for the rupiah, and by October 1997 the currency had depreciated by almost 40 percent. At the same time, runs had been building up on some of the private banks, reflecting a “flight to quality” as depositors sought to move their funds out of the private banks that were believed to be in trouble into the state banks, which were widely thought to be more secure.
Data on the individual banks at this time were poor. Lack of adherence to international accounting standards, prudential regulations that were not commensurate with international best practices, and lack of monitoring and enforcement of data standards by Bank Indonesia (BI) concealed the extent of the banking system’s problems. Nevertheless, intensive work by Bank Indonesia during this period, assisted by experts from the international financial institutions, resulted in a view that although the bulk of the banking system seemed still to be solvent, there were a significant number of banks that clearly were not. Many of these also showed evidence of illegal practices; many were already subject to runs.
After prolonged discussions, as part of a comprehensive adjustment program with the IMF, the government adopted a bank resolution package covering 59 banks (66 percent of the assets of the banking sector) with graduated resolution measures. As part of the package, it was announced on November 1, that 16 banks, comprising around 2.5 percent of the assets of the banking sector would immediately be closed. All depositors would be fully protected up to Rp 20 million, equivalent to $6,000 at the then current exchange rate; this would cover 92.5 percent of depositors fully.
The immediate response to the announcement of the program was positive. The exchange rate rebounded slightly from its earlier steep falls. After a few days, the runs on the banks declined. Eligible deposits were transferred efficiently by BI from the closed banks to designated recipient banks.
Within a few weeks, however, the positive sentiment was entirely reversed. Runs—which earlier had been largely a flight to perceived quality, that is, transfers of deposits from weak private to state banks—became pervasive across the system as concerns over banks’ safety merged into broader concerns over the currency and indeed the stance of economic policy overall. Between end-November 1997 and mid-January 1998 the rupiah lost 80 percent of its value. With depositors withdrawing dollar deposits from the banks, and banks seeking rupiah liquidity which they changed in the market to meet the demands from their customers for dollar liquidity, BI provided liquidity to the banking system amounting to 46 trillion rupiah (4.5 percent of GDP). As liquidity support and currency depreciation intensified, imminent financial “meltdown”—i.e., withdrawal of deposits from the banking system to such an extent that virtually all institutions would have to close their doors or rely on the central bank for constant support—seemed in prospect. Eventually, on January 26, 1998, as the liquidity support approached 60 trillion rupiah, the government announced a blanket guarantee for all depositors and creditors of all domestic banks, as well as the establishment of the IBRA, to take a coordinating role in handling the banking crisis.
Observations and lessons
Some commentators3 have blamed many of the economic problems that occurred after November 1997 on the closure of the 16 banks. Such a conclusion seems too strong, and it is not clear what counterfactual policy prescription at the time would have led to fewer problems. The sixteen banks were selected from an initial list of around 30 banks that aimed to resolve all banks that were clearly insolvent on the basis of data provided by the banks to BI; political pressures, however, ensured the removal of some well-connected banks from this list, although the condition of some of these banks was already well-known to the public. In addition, the closure of the 16 banks was one element of an overall bank resolution and macroeconomic program, and it was the failure to implement that program that is likely to have led to a loss of confidence in the banks, and in the economic management of the country more generally. A blanket guarantee had been introduced in Thailand a few months earlier, but because of the large moral hazard effect, it was considered preferable not to introduce such a guarantee at this time in Indonesia, particularly since the 16 banks being closed were all very small. Nevertheless, there are aspects of these closures that do provide lessons:
- (i) The actual process of closure was carried out efficiently. Despite having no significant earlier experience of bank closure, the central bank was able to carry out the logistics of bank closures effectively. Eligible deposits were transferred quickly.
- (ii) However, given the lack of experience with bank closures, no institutional arrangements were in existence for effective follow up. Remaining assets of these banks were therefore not secured. Moreover, deficiencies with regard to these closures were not corrected later,4 and there has been no orderly process of liquidation. It is estimated that there are now very few assets remaining in the banks, so ultimate recoveries are likely to be minimal.
- (iii) Absence of reliable information, and natural tendencies for denial, may well have led to underestimation of the extent of the crisis at this point.
- (iv) On depositor protection, although 92.5 percent of depositors’ accounts were fully protected, this only accounted for 20 percent of all deposits. Depositors who were not fully protected included some very powerful elements in Indonesia who were not prepared to accept their losses. There was serious intimidation of BI staff from these groups, leading to discouragement in actually inflicting the losses.
- (v) The criteria for banks to be closed was not sufficiently explained. Whilst the banks that were closed were all insolvent, the market was aware that some banks known to be also insolvent were kept open; this indeed was apparent from the announcements at the time.
- (vi) The authorities’ commitment to carry through a banking reform program was cast into severe doubt when a well-connected owner of a closed bank was permitted by BI in mid-November to take over a small bank, which was granted a foreign exchange license by BI, and to use it as a platform to resume the closed bank’s activities, thus effectively by-passing the closure process.
- (vii) The failure of the Indonesian authorities to carry out key elements of their adjustment program led to public disillusion with the economic management of the country. Dollar withdrawals from the banks led to concerns about the ability of the banks to continue to meet the demands for liquidity, prompting further withdrawals. At this point the crisis had become fully systemic, in the sense that depositors were fleeing from the currency and from the banking system as a whole rather than (as earlier) from particular banks perceived to be weak.
III. February 1998
In order to try to stop the bank runs, restore monetary control, and address the banking sector’s problems with the exchange rate appearing to head into free fall, the government announced in late-January 1998 a blanket guarantee for all depositors and creditors of Indonesian banks, as well as the establishment of IBRA. Thereafter, the authorities moved quickly to make the agency effective. A highly-regarded senior Ministry of Finance (MOF) official was appointed its head; additional MOF staff, and several hundred BI supervision staff were seconded to IBRA in order to provide immediate staffing. By mid-February 1998, IBRA was ready to take action. It was proposed that all banks that had borrowed from BI at least twice their capital and had a capital asset ratio (CAR) below 5 percent should be brought under the auspices of IBRA, with IBRA officials on-site at all head offices and principal branches, and all owners of “IBRA banks” working under memoranda of understanding agreed with IBRA restricting their activities.
On Saturday 14 February owners of 54 banks5 were summoned to BL were warned about their parlous financial condition, and were invited to apply to come under the auspices of IBRA. All the bankers agreed. They were informed that IBRA officials had already entered their banks that morning.
While the interventions were determined on a transparent and uniform basis, and were carried out smoothly, the government introduced a last-minute change to the plans that severely undermined the operation. While on this occasion there was no interference in the selection of institutions for intervention, there was some concern that the interventions could provoke renewed runs; it was therefore determined that no publicity should accompany the operation.
Thus, instead of being able to demonstrate that they had started to take hold of the situation, IBRA officials had to work over the following weeks against a public perception that IBRA was still nonoperational. Also, with the lack of publicity accorded to the operation, IBRA officials in the banks appear to have carried little standing. It was not possible for officials to assert full control over the staff of the institutions, and to make clear to the staff the new direction in which the banks were now intended to be going. In any case, there seems to have been little change in the behavior of many of the banks. BI liquidity support continued at its earlier levels for several more weeks.6
Observations and lessons
- (i) Once again the logistics of the interventions were handled effectively, with several hundred IBRA officials securing banks representing 40 percent of the banking system over the weekend.
- (ii) The emphasis on uniformity of treatment was a marked advance over the experience of November 1997. The authorities were concerned to classify banks on the basis of simple transparent criteria. With information on banks’ solvency position limited, categorization was by necessity on the basis of liquidity indicators,7 in particular the size of borrowings from BI.8
- (iii) February 1998 provided an example of “soft” open bank resolution, as opposed to the “hard” open bank resolution (involving full exclusion of former owners and managers) that took place in April 1998. There may well be a role for such interventions. Effectiveness, however, depends upon the elaboration of new rules under which the management and staff of the banks will operate the banks, and upon management and staff of the banks understanding that these new rules will be enforced. In the case of Indonesia at this time, where there was great uncertainty as to what the new rules were, and how far they would be enforced, the absence of any publicity for the operation was a major impediment.
- (iv) The establishment of the new restructuring agency was an important step forward. Contrary to some earlier plans, however, IBRA was established as an agency of the MOF, rather than in an autonomous role. This meant that over the following months its effectiveness was compromised by needing to obtain political authority even for its technical operations. Whilst such political oversight may well have been inevitable, some political interventions seemed to indicate a lack of commitment to IBRA’s objectives. The refusal of the authorities to allow IBRA to publicize its February interventions was an early example.
- (v) While in theory an IBRA presence in the banks was expected to bring the banks under government control, this objective was seriously impaired by the nonremoval of shareholders and managers.
- (vi) The February 1998 interventions were logistically effective because of very close cooperation between IBRA and BI. However, the creation of IBRA was seen by many in BI as an indictment of their record, and relations were not always cooperative. As noted above, BI staff were subsequently progressively withdrawn from IBRA. There were reportedly also ongoing problems with the transfer of documents on the individual banks between BI and IBRA. While such interagency rivalry may be inevitable, it is necessary to have a clear conflict resolution mechanism.
IV. April 1998
By April 1998 it was apparent that forceful intervention into the banks was necessary in order to establish the credibility of the restructuring strategy and of IBRA, and to halt the continuing liquidity emissions from BI. A new IBRA team, appointed in March 1998, recognized the need for (a) transparency and uniformity of treatment of the banks; and (b) publicity for their actions. Considerable efforts were made first to ensure that, once criteria were established, they were applied entirely and consistently, and second, to prepare the public relations aspect of the exercise; a professional public relations firm was employed, and fully involved in all stages of the process.
With reliable information on the solvency of the banks still not available, and with an urgent need to curtail the liquidity emissions from BI, focus was again on liquidity support from BI. At this point, over 75 percent of total liquidity support to the banking system (comprising 222 banks) was accounted for by only seven banks, representing 16 percent of the liabilities of the banking system. Each of these banks had borrowed at least 2 trillion rupiah ($240 million); four of the banks had borrowed 5 trillion rupiah ($600 million). IBRA determined “hard” open bank intervention into these banks, i.e., the suspension of the shareholders’ rights, and the assumption of ownership control by IBRA; and the replacement of the managers by management teams from designated state banks.9 In addition, seven small banks, comprising 0.4 percent of total banking sector assets, which had each borrowed more than the equivalent of 500 percent of their capital, and more than 75 percent of their assets, from BI were closed. The action was to be validated through the conversion of past BI liquidity support into an IBRA stake in each bank, with the government recognizing its responsibility for the financing of bank restructuring by committing to issue bonds to BI to compensate it for the liquidity support.
Once again the operation was carried out smoothly, with IBRA staff intervening in the banks on the weekend of Friday, April 3. Frequent television press conferences took place over the weekend in which the Finance Minister, and the Head and Deputy Head of IBRA repeatedly explained what was happening, and that all depositors were totally protected. Given that these were the first bank closures since the runs following the closure of the 16 banks in November 1997, there was some nervousness among the authorities. In the event the closures were handled successfully and all depositors were able to have immediate access to the deposits from the designated state bank at the opening on Monday morning.10
While the closures were handled successfully, there was at this stage also some uncertainty among the public about the implications of the bank takeovers. This was a new concept in Indonesia, and some depositors were uneasy at banking at a bank that had been taken over and was being operated by IBRA (these became known as the BTO banks). There were some runs on one of the BTO banks which had been subject to protracted runs before. However, the safe transfer of deposits from the closed banks and the continuation of all banking services at the BTO banks ended the runs within about three weeks. Deposit withdrawals were also held in check by premium interest rates offered by most of the BTO banks.11
While the closed banks’ liabilities were transferred efficiently, there were problems handling the banks’ assets. Although IBRA was able to secure the premises of these banks, inordinate delays in the passage of the necessary legislation and subsequent implementing regulations and in the resolution of complex negotiations with BI over the valuation of the liquidity support, which was the basis for IBRA’s equity claims on the banks,12 meant that IBRA was unable to take control of the banks’ assets until February 1999. This undoubtedly delayed the start of asset recoveries and led to severe erosion of the asset values.
Banks whose owners were politically connected were not protected from these interventions. One bank taken over by IBRA was part-owned by one of the President’s closest associates, who at the time of the closure was in the cabinet. This nondiscrimination added significantly to public perceptions that the authorities were finally getting a firm grip on the banking sector.
Like the November 1997 experience, the announcement of the bank closures was followed rapidly by the announcement of agreement on an IMF-supported program. Unlike in November 1997, however, the authorities demonstrated commitment in the following weeks to carry out the other elements of the program.13
In the BTO banks, new managements were brought in from “twinned” state banks. Over the coming months, however, these managements were unable to undertake significant operational restructuring of these banks. Subsequently, in all cases, the managements were again replaced or the banks were closed.
Observations and lessons
- (i) Bank closures, if carried out effectively and on uniform principles, and provided that the cut-off point is explained and accepted by the public, and if this is set within a broader plan that is announced and credible, can have a positive impact on sentiment in assuring the public that the government has a coherent strategy for tackling the banking crisis.
- (ii) Investment in public relations is likely to be particularly rewarding. This is likely to involve both external consultants and a heavy involvement of time by the principal policy makers themselves.
- (iii) In the immediate term, efficient transfers of deposits to recipient banks will be the major operational exercise.
- (iv) The authorities need to have resources and legal powers to minimize the deterioration of the assets that have come under their control.
- (v) The costs resulting from the closures will be higher if the required legal powers for the supervisory agency are not fully in place. These include powers for the agency rapidly to take full control of the assets of the closed banks.
- (vi) Open bank resolution is not necessarily simpler to carry out, or cheaper, than bank closures. Keeping a bank open while under the control of the restructuring agency requires pre-identification of competent management to safeguard the agency’s interest, and to provide confidence to the public that it is still desirable to do business with that bank. The managements brought in by IBRA necessarily had to be recruited at very short notice, and by and large did not prove successful in turning the banks around. Most of the taken-over banks offered very high deposit rates subsequently to try to regain lost deposits. The resultant negative interest margins added to the ultimate costs of resolving the banks.
- (vii) It is critical for the process that there is at least one “good” bank in which the public has confidence and which can be used as the recipient of the liabilities (and in some cases the assets) of the closed banks. In Indonesia, with limited time, capacity, and sound banks, options were very limited. All deposits from banks closed at this time were transferred to a single bank, state-owned BNI, widely regarded as the strongest bank in the country, although at the time still unrestructured.
V. August 1998
Over the course of the spring and summer of 1998 international accounting firms conducted portfolio reviews on the banks taken over by IBRA in April 1998,14 as well as on 16 of the larger of the remaining private banks,15 and on the state banks. Similar reviews were conducted for the remaining banks over the following months.
The first results for the banks taken over in April 1998 were devastating, showing levels of nonperforming loans ranging from 55 percent to over 90 percent of the banks’ portfolios. Although there was some question as to whether the accountants had been excessively diligent in marking down the portfolios, there could be no dispute of their finding that the examined banks were deeply insolvent.
In June 1998, very soon after their completion, the results of the audits on the IBRA banks were leaked to the press. There was an immediate shock that the state of the banks was so bad; beyond that, however, the leaks prevented any further denial of the seriousness of the crisis, and forced the authorities to recognize that drastic further action was urgently needed.
The two largest banks taken over in April (BDNI and Danamon) were listed on the Jakarta stock exchange. Under stock exchange rules, shareholders’ stakes could only be written down after a declaration of insolvency of a bank; this in turn required the calling of an extraordinary shareholders’ meeting. In July 1998, IBRA obtained a declaration of insolvency for Danamon; the shareholders of BDNI, however, used technical procedures to prevent such a declaration for that bank.
By this time the authorities had begun to develop an overall plan for the eight banks they had taken over but not closed in April and May 1998. The one bank taken over in May (Bank Central Asia) was considered inherently sound, and IBRA began discussions as to its disposal. The one state bank (Exim) was essentially moved out of IBRA, pending merger with other state banks. Three of the remaining banks had been found to have almost no performing loans, and were therefore to be closed. The authorities’ strategy was to use the largest remaining bank (Bank Danamon) as a sort of “bridge bank”—i.e., a repository for assets and liabilities of other banks, and possibly a source for specified banking services—and therefore to keep it open under IBRA control. The other two banks (PDFCI and Tiara Asia) were in marginally better condition, and—unlike the others—not riddled with prudential violations, in particular connected lending. These were to be merged into the bridge bank in the event that IBRA could not sell them within a specified period of time.
On the weekend of August 20, IBRA therefore closed three banks, representing 5 percent of the liabilities of the banking sector. Resource capacity was severely stretched, in part because the earlier mass secondments of staff from BI to IBRA had already been severely cut back. Deposits of the closed banks were transferred over the weekend to BNI. IBRA also intensified its control over its banks that remained open.
In the period thereafter, the IBRA management team at Danamon (and to a lesser extent in the other IBRA banks) made considerable efforts to restore their earlier deposit base. Deposit interest rates were among the highest of major banks in the country. While this may have looked good from the point of short-term stabilization of the banking sector, the large number of additional deposits generated on the basis of a substantial negative interest margin added considerably to the ultimate cost of recapitalizing the bank.
Observations and lessons
- (i) Open bank resolution procedures may be a prelude to bank closures later. The initial open bank resolution may provide time for the authorities to obtain a full picture of the bank, so as to decide what to do with it. In such cases the authorities are literally “buying time,” since if a bank ends up being closed it is likely to be more expensive to do it later than if the bank had been closed at the outset, since losses are likely to continue to accumulate during the interim. Also, if the public realizes that open bank resolution may be a prelude to closure, confidence that there will be full and prompt protection of all depositors becomes even more critical.
- (ii) Even large amounts of banking capacity can be taken out of the system without significant impact on loss of banking services, if properly handled. Depositor services were immediately made available in other banks. Few performing loans still remained in the closed banks. Where they did, IBRA subsequently sought to provide services to these borrowers, either directly or through the banks it had taken over.
- (iii) A resolution strategy may involve a mixture of closures and open bank resolutions, so that the authorities maintain some core banking services and can use some institutions(s) for “bridge bank” purposes. Fraudulent banks and those with few performing loans are likely to be easiest, and cheapest, to resolve through closure. Where franchise value remains, and significant banking services are still provided, there may be grounds for seeking open bank resolution. The exact distribution between the two approaches will require careful consideration.
VI. March 1999
On September 26, 1998 the Governor of BI, on behalf of the government, announced a plan for the restructuring of the private banks. The objective was the retention of a residual private banking sector in an environment where the state sector had already increased its share substantially over previous months, and where there was evidence that most of the remaining significant private banks were also insolvent.
Under this plan, banks were to be subject to a threefold categorization. Those with estimated capital asset ratios greater than 4 percent (the minimum requirement for all Indonesian banks at end-1998) were deemed sufficiently strong not to warrant government support (the “A” banks). Those banks with CARs worse than –25 percent (the “C” banks) were too weak to warrant government support. Those with CARs between –25 percent and 4 percent (the “B” banks) were invited to submit business plans which would be assessed by independent advisors. If the plans were passed, and if the owners and managers passed a “fit-and-proper” test applied by BI, they would be eligible for joint recapitalization with the government. If the owners supplied unborrowed resources equivalent to 20 percent of the shortfall from the 4 percent CAR requirement, the government would supply the remaining 80 percent of the shortfall. The government would obtain a commensurate shareholding in the bank, but would leave the owners in day-to-day control of the bank, and provide the owners with the first option to buy back their shareholdings at the end of three years.
Implementation of the plan—which depended on the conclusions of a series of four interdepartmental committees—was to have been finalized by December 1998, but the need to achieve political consensus prevented the process from beginning until that month, and expected implementation was pushed back to February 1999. By early February, government ministers—wishing to avoid the risks of a sudden shock in expectations—were hinting strongly to the public that there would be a number of bank closures before the end of the month. By the last week of February, it was common knowledge that there were to be bank closings on the weekend of February 26. Two days before that date, however, the government announced a postponement. Political consensus had not been established on the treatment of a few of the banks. Public reaction was extremely negative, with a widespread perception—acknowledged by some officials—that there had been outside influence on the closure decisions.
Over the following two weeks, the various technical committees refined their estimates of the condition of the banks, and political consensus was established on the implementation of the restructuring plan according to the agreed principles. Also, owners of “C” banks were given a final opportunity to provide additional capital to raise themselves to “A” status; “B” banks without acceptable business plans were invited to resubmit plans.
On March 13 it was announced that 73 banks (representing 5.7 percent of the assets of the banking sector) had CARs above 4 percent (“A” category)16 and could continue to function without government support. Nineteen banks (representing 3 percent of the assets of banking sector) had CARs below –25 percent (“C” category) and were to be closed immediately. Of the remaining (“B” category) banks, nine (representing 10 percent of the banking sector) had passed the various tests and were eligible for recapitalization; seven banks (with 2.5 percent of the assets of the banking sector) had failed one or more of the other tests mentioned above, but were to be taken over by IBRA on the basis that they had at least 80,000 depositors; and nineteen smaller banks (with 2 percent of the banking sector) would be closed with the “C” banks.
The announcement was well-prepared, and included major public relations efforts both from specialist consultants and the principal participants themselves. Market reaction was generally positive, with recognition that this was a careful and comprehensive resolution of the problems of the banking sector. Unfortunately, the closures and transfers of deposits did not occur as smoothly as in the previous closure operations. Although the total number of depositors in the closed banks was less than in the August 1998 operation, the sheer number of banks, their dispersion across the country, and the lack of prior IBRA access into them, made this operation more complicated than the earlier one. Even worse, the protracted period during which the closures had been expected provided the opportunity for workers facing redundancy to organize themselves, and to maximize their leverage in forthcoming redundancy negotiations by denying IBRA staff access to banks’ premises. Thus, while the majority of deposits were successfully transferred by the middle of the following week to five designated recipient banks, in around 20 branches workers prevented IBRA from obtaining access to depositors’ records and transfers could not be made.17 For several weeks IBRA became involved in redundancy negotiations with these workers; senior management were therefore distracted from their core functions. IBRA gradually overcame the problem, obtaining alternative sources for the needed documentation, and obtaining settlements with the workers.
While the public basically maintained confidence in the banks taken over by IBRA, there was—perhaps surprisingly—a loss of confidence in some of the banks deemed eligible for recapitalization. Apparent confusion by the authorities in the remaining steps necessary for the recapitalization to take place, together with punitive comments by some officials about the requirements that would be put on these banks, distorted the story that these “B-pass” banks were to be the core of the private banking sector of the future. Significant runs on some of these banks led to liquidity squeezes. Confusion as to the distinction between the “B-pass” banks and the BTO banks was heightened by public statements from the well-connected owner of one of the BTO banks saying that he was only temporarily out of his bank and that he expected that the government would be helping him with his bank’s recapitalization.
An additional issue arose from the operation of the interbank guarantee. While the depositor guarantee had been honored meticulously, the authorities had so far failed to make payments under the interbank guarantee, in part to try to contain fiscal costs.18 Several of the prospective “B-pass” banks had substantial claims under this guarantee. Their uncertain status seemed to jeopardize the banks’ eligibility for recapitalization.
By end-April most of these concerns had been largely resolved. IBRA had excluded the former owners from all the BTO banks, and changed all the managements.19 IBRA officials had progressively taken control of the assets of the banks. The owners of seven of the nine banks declared eligible for joint recapitalization with the government had provided their share of the capital. One of the other banks was provisionally sold to a major international bank;20 the other bank was taken over by IBRA. The government committed itself to honor claims under the interbank guarantee for all “B-pass” banks. In June 1999, the Head of IBRA announced that all the taken over banks (except the bank taken over in April) would be merged into Bank Danamon (the “bridge bank”) over the coming months.
The interventions of March 13, and the subsequent implementation of the measures announced then, were well-received by the markets, and by outside commentators, even those who had been critical of earlier interventions. The general feeling was that finally the authorities had a full grip on the banking situation.
Observations and lessons
- (i) The absence of surprise by the time the interventions occurred gave those affected—staff, owners, and managers—the opportunity to seek to protect themselves from the consequences of the interventions. The occupation of bank premises by the staff added to the costs of the closures by delaying IBRA’s achievement of full control of the banks’ assets. Owners and managers are likely to have transferred assets from the banks in anticipation of the interventions.
- (ii) The comprehensive nature of the action, involving a resolution decision based on uniform and transparent criteria for each of the private banks, added to the credibility of the action, and to its favorable reception.
- (iii) The “too-big-to-fail” argument of the authorities, used to justify the takeover rather than closure of the banks with at least 80,000 depositors, was subject to criticism. Such an argument is frequently made by authorities handling systemic banking issues, but it is hard to spell out, and seems to contradict the principle of uniformity of treatment. In Indonesia the authorities made it at an early stage regarding the seven state banks, which together comprise 50 percent of the banking sector: all these banks were to be recapitalized, even though on the basis of the criteria applied to the private banks all would have fallen into category “C” and have been closed.
- (iv) The March 13 interventions, comprising the 38 bank closures, the takeover of 7 banks, and the announcement of eligibility of the 9 banks for joint recapitalization had a major positive impact on the credibility of the restructuring program, and hence on the government’s economic management overall. At least partly as a result, with bank restructuring seen to be at the core of economic revival, market interest rates began falling rapidly from their crisis levels. With this fall in interest rates, prospects for the economy and for the banking system in particular, were dramatically improved.
VII. The closure process
The Indonesian experience since late-1997 has provided a number of lessons as regards the role of bank closures in a bank restructuring program.
- (i) In a case of massive insolvency and restructuring, such as Indonesia over this period, no single strategy with regard to closures could be put in place which would have worked successfully throughout the period. In practice, there had to be several phases as the true magnitude of the problem became apparent and the authorities over time moved to address it.
- (ii) The initial stages of a bank restructuring program are bound to be particularly difficult and messy. The authorities will always be acting on limited information. They are likely to have to operate with an inadequate institutional infrastructure and legal framework. They may well be working in a difficult political environment in which there are strong forces resisting change. While it will be critical to start to establish the infrastructure and legal framework immediately, it is likely that the authorities will not be able to wait until these reforms are completed before they have to intervene in the banking system.
- (iii) While every banking crisis is likely to be different, there may be a strong case for closures of some banks at the outset of a crisis. Some banks will be clearly deeply insolvent and racked with fraud. An open bank solution when there may be limited capacity to control further losses, and where the bank itself has lost credibility, may not be cost effective, either as regards the direct losses from the bank or in terms of demonstrating the government’s commitment to serious adjustment. In bank closures gross costs are realized upfront, and can be more easily identifiable (most immediately the transfers of deposits), while in open bank resolutions the costs are realized over time, and increase dramatically if the intervened bank is not properly managed.
- (iv) One of the most tricky issues for the authorities to determine is whether they will introduce a blanket guarantee at this stage. Partial protection when further closures are in prospect is unlikely to be effective. On the other hand, providing a blanket guarantee to depositors and creditors may be costly, and—unless handled carefully—raise well-known moral hazard concerns (for instance, banks may be able to offer unviable interest rates to attract depositors). Nevertheless, once a banking crisis has become systemic, such a guarantee often becomes an integral part of the resolution strategy, notwithstanding its cost. The guarantee may not be immediately credible; credibility may be achieved only once the authorities have demonstrated the smooth working of the guarantee in practice.
- (v) Large and protracted liquidity support to a set of banks will quickly threaten monetary control. Bank closures may be an integral part of a package of measures to address this. With limited information on banks’ solvency position, interventions at this stage are likely to be based on liquidity criteria. Evidence indicates that those banks requiring substantial liquidity support will also be deeply insolvent.
- (vi) Bank closures must be based on transparent, uniform, simple, and defensible criteria. There should be no exceptions to the specified rule, since the credibility of the entire operation will be only as strong as its weakest link—even one or two banks treated differently than indicated by the criteria can indicate political involvement, and thus discredit the criteria and the entire operation.
- (vii) In most cases, the authorities will have to identify one or more banks that will be in a position to immediately receive the deposits of the banks being closed. In the absence of a clear alternative, they are likely to select a state bank. While this may be appropriate in some cases (and in Indonesia there was probably little alternative), state banks may themselves also suffer from serious weaknesses, and there could be advantages in using a private, or possibly even a foreign, bank for this purpose if this does not serve to undermine further the public’s confidence in the domestic banks.
- (viii) The authorities will need to ascertain the true financial condition of the banks. Banks’ reported figures are unlikely to give an accurate picture of the banks’ health, so additional audits—by the banks’ auditors, by international accountants, or by other professional experts—will be needed. The results of these audits are likely to show a much worse situation in the banks than hitherto believed. These results will serve to confront any tendency among the authorities to deny the depth of their banking crisis.
- (ix) Upon receipt of the results, the authorities will need to conduct a “triage” of the banks. The best banks can be left alone; the middle tier of banks will be eligible for some form of assistance based on quantitative and qualitative selection criteria; the worst set will need resolution. This resolution may involve closures, open bank resolution, or a mixture of the two. The choice will depend upon the number and size of the banks in this category, the state of confidence in the economy, capacity to close the banks, or conversely to operate the banks economically in an open bank resolution strategy. Overall, the chosen strategy should be that which restores the system to soundness while minimizing public expenditures, but reliable estimates will be very difficult to make.
- (x) Bank closures at this stage should be based on estimates of insolvency, together with absence of any owners or outsiders willing to recapitalize the bank, or of a business plan showing how the bank can restore itself to viability. The critical solvency level may vary depending on the circumstances in the country and could be anywhere substantially below the minimum capital ratio required in the country at the time. Thus, in some cases it could be above zero. In other cases, for instance where the entire banking system is deeply insolvent and there is a desire to keep at least a residual element of the sector in operation, it may be below zero. In the case of Indonesia, CAR no worse than –25 percent was established as a criterion for possible eligibility for joint recapitalization with the government. This is likely to have been an extreme case. It reflected the depth of the banks’ insolvency, the recognition that some part of the insolvency was due to the unprecedented macroeconomic shock that had occurred, the desire to maintain at least some private banks in the system, and the belief (confirmed in the event) that a number of the owners had significant resources elsewhere which they would be prepared to use to share the recapitalization costs with the government.
- (xi) The closure process will, naturally, lead to a reduction in the number of banks, hence boosting the potential for profitability among those remaining. Given that most, or all, of the bank closures will occur in the private sector—and that open bank resolution will likely involve a takeover of private banks by the government—the share of state-owned institutions will rise during a bank restructuring. In Indonesia the domestic private banking sector shrank from 45 percent of the total in late 1997 to 15 percent in mid-1999. An important corollary—to which the Indonesian government has committed itself—is a program for the privatization of a large part of the private banking sector.
- (xii) “Too big to fail” may be a valid criterion—if not over-used—for open bank resolution rather than closing banks that play a significant role in the payments system, or in providing services to parts of the market. Again, the authorities should specify transparent criteria. Preserving the banks should not generally mean retaining the owners or managers. Except in the content of an overall recapitalization plan such as the one described above, loss minimization for the public sector must mean the elimination of the rights of shareholders for banks formerly in the private sector. For banks in the public sector, careful attention must be paid to avoid increased capitalization costs—following negative interest margins or weakened loan recovery efforts—if management feel relieved of the discipline of profit maximization objectives.
I am grateful for comments from Philippe Beaugrand, Thanos Catsambas, Eric Clifton, Olivier Frécaut, Lorenzo Giorgianni, Luc Jacolin, Michael Keen, Carl-Johan Lindgren, Hassanali Mehran, Elizabeth Milne, Anoop Singh, and Timo Välilä.
The Indonesian bank restructuring program as a whole, together with related issues, is to be discussed in “Indonesia: Anatomy of a Banking Crisis” by Charles Enoch, Barbara Baldwin, Olivier Frécaut, and Arto Kovanen (IMF Working Paper forthcoming).
See, for instance, “The East Asian Financial Crisis: Diagnosis, Remedies, Prospects,” Steven Radelet, Jeffrey D. Sachs, Brookings Papers on Economic Activity, Vol. 01, pp. 1–74, 1998, Harvard Institute for International Development.
Legislation passed subsequently enhanced the powers of IBRA in closing banks (see below). Since the 1997 closures predated the establishment of IBRA and the follow-up work continued to be handled outside IBRA, it continues under the pre-existing legislation.
These were all private banks. A state bank, Bank Exim, also had borrowings from BI of more than twice its capital—due to trading losses—and was the 55th bank to be brought under IBRA. Together these banks comprised 40 percent of the banking sector.
Liquidity support tapered off during the following month when the liquidity facility was redesigned to provide for nonfinancial sanctions—including takeover of the bank by IBRA—in the event of prolonged use of the facility.
It has been generally found that major liquidity problems arise after a bank has become insolvent. Thus, while reliance on liquidity criteria probably means that some insolvent banks will escape intervention, it is unlikely that those banks in which the authorities have intervened will, in the event, be found to have been solvent.
Reliance on this particular indicator was warranted also by the urgency of restoring monetary control.
Except for the one state bank among the seven banks, Bank Exim, where the treasury department managers (who had been responsible for the bulk of the bank’s losses) were replaced, but the remainder of the bank’s management were left in place.
Newspaper adverts were run across double pages with IBRA on the left hand page announcing the closure of the seven banks, and the designated recipient bank, Bank Negara Indonesia (BNI), advertising on the right hand page welcoming its new customers to its branches and summarizing the services it could provide.
In order to mitigate the moral hazard effects of the blanket guarantee, the authorities placed a cap on the interest rates that banks could offer depositors. At this time the cap was set at 500 basis points above the deposit rates offered by the “JIBOR banks” (the group of banks whose interbank rates were used as the benchmark for calculating JIBOR). Several of the BTO banks used the maximum premium afforded by the cap in order to safeguard, and later recover, deposits.
Amendments to the banking law were needed to give IBRA powers to take control of assets from former shareholders. After considerable and unexplained delays, the law was passed and signed by the President in October 1998. To be effective, enabling regulations from the amendments were also needed. These eventually received presidential signature in February 1999.
In May 1998 implementation of the program was interrupted as a result of rioting and looting and the further collapse of the exchange rate; President Soeharto was replaced by Vice-President Habibie. There were concerted runs on the largest private bank, as part of widespread targeting of Chinese-owned businesses. Following massive BI liquidity support, the bank was taken over by IBRA on May 16. IBRA at this point had taken over banks with 23 percent of the deposits of the system.
The operational importance of the February 1998 takeover of 54 banks decreased markedly during this period, no doubt due in part to lobbying on the part of the owners of some of these banks. IBRA inspectors were progressively removed from these banks, and in July 1998 responsibility for supervising these banks—which had been transferred to IBRA—was passed back to BI.
These 16 banks were selected on the basis of initial assessment by the authorities that these might be the “core” of a revised banking sector, given their relative size and good reputation.
For one additional bank a full determination had yet to be made. It subsequently passed the tests to achieve “A” bank classification.
The closed banks had employed 15,600 workers. The banks taken over by IBRA at this time employed an additional 11,900 workers—a significant factor in the decision to take them over rather than close them.
Foreign banks consistently maintained that this failure to honor the interbank guarantee was a severe impediment to the restoration of confidence in Indonesia.
Except in one bank, which had been taken over by IBRA only because there were no shareholders able to put up their share of the bank’s recapitalization requirement.
This takeover led to the discovery of side payments by the bank, Bank Bali, in order to obtain payments under the interbank guarantee in order to close the capital shortfall sufficiently that the bank would be able to participate in the joint recapitalization program. The “Bank Bali” scandal is beyond the scope of this particular paper.