Dealing with bank system failure: Indonesia, 1997–2003

Abstract

Indonesia's crisis recovery program has failed badly in relation to the two key objectives of development economics policy making: efficiency and equity. The economy went into severe recession within a few months of the IMF appearing on the scene, and six years later output was little higher than before the crisis. The collapse of the banking system and the associated government bailout of depositors has imposed a loss on the public of at least 40% of GDP. This paper describes that collapse and the government's policy response to it, under advice from the IMF. It goes on to propose an alternative scheme that might have been followed—and that could be followed in future banking crises. Its twin objectives are to maintain the integrity of the payments system and to avoid inequitable wealth transfers that result from government bailouts of banks and their depositors.

Notes

^*This paper is drawn from a study of the letters of intent (LOIs) from the government of Indonesia to the IMF from October 1997 through June 2002. The author gratefully acknowledges financial support from the Australian Agency for International Development (AusAID) for the preparation of this study. The views expressed in the paper are those of the author and not necessarily those of the Commonwealth of Australia, which accepts no responsibility for any loss, damage or injury resulting from reliance on any of the information or views contained herein.

¹These included reductions in tariff protection in certain sectors; reduced spending on infrastructure projects; removal of many restrictions on direct foreign investment; tighter controls on procurement and contracting procedures; accelerated implementation of the ruling of the WTO (World Trade Organization) dispute panel on the national car project; the phasing out of a number of import and marketing monopolies; closure of a number of private banks (some controlled by members of the Soeharto extended family); and privatisation of state banks that had previously lent heavily to companies favoured by the president. For a discussion of the politico-economic system under Soeharto, focusing on how economic policy was used to enrich the first family and its business associates and to entrench Soeharto's hold on power, see McLeod (2000a, 2000b).

²The January 1998 LOI reported that it had been decided to discontinue immediately any budgetary and extra-budgetary support and credit privileges granted to projects of the state aircraft manufacturer, IPTN, headed by Habibie.

³This view confused the symptom (capital outflow) with the cause (the negative reassessment of risk).

⁴McLeod (1997: 103) warned of the danger to the economy of these inappropriate responses by the government to the sudden depreciation of the rupiah.

⁵This paper makes frequent references to the series of LOIs. All may be found at the Indonesia section of the IMF website: http://www.imf.org/external/country/idn/index.htm?type=9998#23. They are identified here by the date (month and year) they were sent.

⁶For detailed discussions of the impact of the crisis on the banking sector and the estimated cost to the government, see Fane and McLeod (2002: 287–8) and IMF (2000: 26–8). The Fane and McLeod estimate of the net cost to the government of the banking sector collapse is around 40% of GDP, while the IMF staff estimate, which uses slightly different methodology, is about 47% of GDP.

⁷Calomiris and Kahn (1991) note that studies in the US ‘indicate that fraud and conflicts of interest characterize the vast majority of bank failures’ (p. 499). They go on to present a model of banking in which the possibility of a bank run, given the existence of deposits able to be withdrawn on demand, helps ‘to prevent absconding [by bankers] from taking place’ (p. 510). In light of this, it is interesting to note that the provision of last resort loans or blanket guarantees to illiquid banks significantly weakens the capacity of depositors to prevent bankers from absconding, since it largely removes the threat of forced liquidation.

⁸Subordinated loans rank above shareholders' equity, but below all other liability categories, in the case of bankruptcy.

⁹A 20% share in Bank Mandiri was floated to the general public in July 2003, and a 49% share in Bank Rakyat Indonesia in November 2003, but this achieved little since it left control of the banks firmly in the government's hands.

¹⁰This refers to the risk that cheques drawn on one bank and payable to another might not clear because of a deficiency of funds in the paying bank's account at the central bank. It appears that in late 1997 and early 1998 this risk was being carried by Bank Indonesia, rather than being borne by the recipient bank.

¹¹Imposing high interest rates on last resort loans is rather pointless if the banks are already insolvent. The banks' owners have nothing more to lose, and will therefore be happy to take expensive loans in order to get the chance to ‘gamble on redemption’. Perhaps, therefore, the intention was to require banks' owners to put up better collateral for their banks' borrowings.

¹²‘With the [interest] cost assumed on the budget, it is demonstrated transparently with other elements of public expenditure …’ (Enoch et al. 2003).

¹³This new CAR requirement was relaxed even further in June 2000 by allowing banks to include their loan loss reserves in measured capital (Jakarta Post, 16/6/00), despite the fact that loss reserves are an estimate of losses the banks expect to incur—whereas ‘capital’ for the purpose of calculating the CAR is supposed to be available to meet unexpected future losses.

¹⁴Banks with CARs below –25% were to be closed; those above 4% needed no further immediate action.

¹⁵The regulations specified limits on the amount of loans by banks to affiliated companies, but these were routinely ignored both before and during the crisis. The banking law contains severe penalties, including imprisonment, for bank owners or managers who cause these limits to be exceeded.

¹⁶For example, in the October 1998 LOI it was stated that ‘negotiations with other former bank owners [were] expected to yield appropriate settlements in coming weeks’.

¹⁷‘… on March 7 the government decided on a strategy to resolve the long-standing problem of shareholder settlement agreements with former bank owners that have fallen into dispute and/or default’ (April 2002 LOI).

¹⁸In the event, fixed and floating rate bonds were issued to recapitalise the banks. ‘Hedge bonds’, effectively indexed to the Rp/$ exchange rate, were issued to two of the state banks (IMF 2000: 29).

¹⁹That is, to ensure that loans are carried in the books so as to reflect the best estimate of their true value, after taking into consideration the probability of default.

²⁰‘Obligor’ is the term given to groups of debtors to IBRA that have common ownership.

²¹For an alternative view on the policy of forced mergers, see Fane and McLeod (2002: 14–15).

²²A further 20% shareholding was divested in September 2003.

²³Unfortunately, the former shareholders were left in effective control of these assets. It is probably no coincidence that their market value was later estimated at only about one-third of their transfer value (IMF 2000: 31).

²⁴‘To have tried to force depositors to bear the costs of the banking failures could have led rapidly to the collapse of most, or all, banks in the country, turning Indonesia into a wasteland of financial intermediation and returning it to a cash or barter economy from which it would have taken many years to recover’ (Enoch et al. 2003).

²⁵According to the then governor of BI, deposit accounts up to Rp 20 million constituted more than 80% of the number of accounts, but less than 20% in terms of the rupiah value (J. Soedradjad Djiwandono, personal communication).

²⁶Chapter 7 procedures differ from those of Chapter 11 in that with the latter, incumbent owners and managers retain control of the firm, rather than this being handed over to a trustee of the creditors (Gertner and Scharfstein 1991: 1,209).

²⁷This is double the international standard of 8%, in line with the recommendation by Fane and McLeod (1999: 408–9) that banks in developing countries such as Indonesia should have to meet much stronger capital adequacy requirements than those in the more advanced economies, where uncertainty is less and the rule of law is more reliable.

²⁸This would have no implications for those members of the general public who owned no bank deposits at all.

²⁹A precedent for the government deceiving itself as to what is required to insure itself against banking losses is the loan insurance company, PT Askrindo, which it established in the 1970s to insure the state banks against non-repayment of subsidised loans given to small businesses (McLeod 1983: 85). Since the essential purpose of insurance is to spread risks, this was a total waste of resources: both the insurer and the insured belonged to the government, so the risks remained with the public sector—and ultimately with the general public.