Crisis Resolution and Financial Stability in Norway
Speech by Mr. Jarle Bergo, Deputy Governor of Norges Bank, at the 50th Anniversary of Bank Indonesia, 10 December, 2003.
Yours excellencies, ladies and gentlemen.
It is a great pleasure to attend this important conference in conjunction with the 50th anniversary of Bank Indonesia, and have the opportunity to share our experience of handling a banking crisis with colleagues from South East Asia.
The banking crisis in Norway in the early 1990s was not as severe as the banking crises some of the Asian countries faced in the late 1990s or the one our neighbouring countries Sweden and Finland experienced. In 1991 - at the peak of our crisis - bank loan losses amounted to 2.8 per cent of GDP. In the same year non-performing loans accounted for 9 per cent of loans outstanding for the entire banking industry1. The corresponding figure for Indonesia was as high as 70 per cent, 33 per cent in Thailand, and 18 per cent in Sweden2.
Nevertheless, there can be no doubt that the Norwegian crisis was systemic; in 1991 three of the four largest banks and a number of smaller ones had failed or were in deep financial trouble.
In this relatively short presentation, I will concentrate on the main issues of the crisis and its resolution. Thus, not every stage in the development of the crisis will be recapitulated, or every step and detail in its resolution. Nor have I tried to reflect every nuance in the public debate, in the reports from parliamentary commissions and in research papers on the subject3. The views I put forward are my own and have been deliberately sharpened to promote discussion.
The most important characteristics of the resolution of the Norwegian crisis can be summarized as follows:
- The banks' own collective guarantee funds handled the problems in the banking sector before the crisis became systemic.
- No blanket guarantee for the banks' debts was provided by the government.
- No regulatory forebearance.
- No liquidity support to banks whose solvency was in doubt.
- A clear and transparent division of responsibility between the political authorities, the supervisory authority and the central bank was established early on.
- Government support was contingent on strict requirements being met, e.g. existing shareholders accepting a write-down to cover losses to the extent possible.
- No micro-management of the banks.
- Measures taken to prevent supported banks exploiting the situation vis-à-vis non-supported banks.
- No asset management companies or "bad banks".
I shall comment on each of these issues, but before doing this let me briefly mention the background for the crisis:
During 1984 and 1985, the Norwegian banking system was deregulated. In this new environment, banks started to compete for market shares. We saw a huge increase in bank lending which, coupled with general optimism (oil!) and interest rates that were kept artificially low, created an economic boom. However, neither bank managers nor supervisors had much experience of banks operating in a competitive credit market, and did not pay sufficient attention to banks' risk exposure.
A sharp fall in prices for our main export commodities, oil and gas, and the need to consolidate the economy after the boom and bring down rampant inflation, led to a recession in the late 1980s.
In short, deregulation, a lending boom, and an ensuing recession set the stage for increasing bank losses.
The Norwegian banking crisis evolved in two stages:
In the first stage from 1988 till 1990, several small and some medium banks failed.
From 1990 - 1991, the crisis became systemic. The three largest commercial banks, accounting for half of the market for credit to the domestic non-financial sector, were in deep trouble.
During the first stage of the crisis, the problems were mostly handled by the banks' own guarantee funds. Membership in the funds is mandatory and the funds are financed ex ante by fees from the member banks.
These funds had a wider mandate than just paying out depositors at the closure of a failed bank. If liquidating the failed bank and paying out the depositors was considered a more costly alternative than recapitalizing the bank or providing guarantees or financial support to facilitate a take-over, the funds were free to do so.
However, by the end of 1990, bank-owned funds had effectively depleted almost all of their capital and it became apparent that the larger banks might also face problems. To make sure that there would still be a safety net present, the government established the Government Bank Insurance Fund - or GBIF - in early 1991. At the same time it was made clear that it was through this fund only that public money might be injected into failing banks. The central bank might provide liquidity support only once solvency was assured. The GBIF was given a specific grant by the Storting (parliament). The new fund could lend money to the banks' own guarantee funds to enable them to recapitalize failing banks. Later, as the crisis escalated, the GBIF was permitted to recapitalize the banks directly.
In both cases, strict conditions were attached.
Most often these included:
- the management and board of directors of the bank were replaced
- the existing share capital was written down to cover losses to the fullest extent possible
- the bank's operating costs were reduced and some of its activities downsized
- measures were taken to restrain growth in the bank's total assets.
An example of the conditions that were imposed can be found in chart 8. These conditions were of course rather unattractive to bank managers, board members and shareholders. And that was part of the objective; the management of a troubled bank should have a strong incentive to try other solutions before approaching the GBIF. For those that had to resort to the GBIF, an important aim was to force them onto a sounder track of business.
Curbing the activities of banks receiving capital from the GBIF also implied that there was no competitive advantage for these banks over rival banks that were not receiving this kind of support.
In addition, the value of the existing share capital in the bank would be written down according to the bank's losses. In cases where the losses exceeded the existing share capital, the entire capital would be written down to zero. Such decisions would normally have to be made by the banks' General Meetings. In order to avoid a stalemate if a majority at the meeting objected to the decision, the Storting had one month earlier made an amendment to the Commercial Bank Act. This amendment entitled the government by Royal Decree to write down the share capital of a bank against losses in the audited interim accounts, if the shareholders' General Meeting did not do so. This authority was used in two instances where shareholders refused to write down a bank's shares as required by GBIF. Shareholders in one bank brought the case to the courts, but lost.
That shareholders in this way were forced to take a bank's losses before tax payers' money was put on the table, was to my mind imperative in order to muster the necessary political support and acceptance among the electorate for the rescue operations. It is hard to imagine that it would at all have been possible to support the banks if the old managers and shareholders were allowed to maintain their stakes in a failed bank.
The conditions attached to the capital injections from the Government Bank Insurance Fund were explicitly set out in a written agreement between the bank and the GBIF. This agreement was made public. The bank was required to report regularly to the fund on the fulfilment of the terms of the agreement.
The GBIF was headed by a decision-making board of experts and its operations were carried out at a distance from the political authorities. However, the GBIF kept in close contact with the financial supervisors and the central bank. The central bank also played a part as lender of last resort once the banks had been recapitalized. General assurances were given that the central bank would provide liquidity support to sound financial institutions (including non-banks). However, the banks generally managed to keep their funding, and did not have to resort to the central bank to any extraordinary extent.
Before GBIF injected capital into crisis-stricken banks, efforts were made to attract private investors. However, these efforts did not succeed. Experience shows that in times of recession and high uncertainty investors will be extremely reluctant to take on risk. Under these circumstances, the government was the only source available to support the financial system. In a way, the government became the "owner of last resort".
In light of the high level of uncertainty, the funding of the GBIF was regarded as expenses in the fiscal accounts, not as an investment. In retrospect, however, capital supplied by the government to the problem banks proved to yield a reasonable positive return. Hence, it might have been a profitable investment for private investors, although with a huge risk ex ante.
To counteract the generally low supply of equity capital during the banking crisis, a separate Government Bank Investment Fund was set up. The Fund was intended to take part in equity emissions at (non-problem) banks on commercial terms. Certain other forms of general support to the banking industry were also provided.
At its peak, the share of state ownership in the banking sector was close to 60 percent.
Unlike many other countries dealing with a banking crisis, Norway did not set up a separate asset management company - or "bad bank" - to handle problem loans. There were several reasons why we chose not to do this:
- In none of the problem banks in Norway was the ratio of non-performing loans considered to be of such magnitude that it would require so much attention from the management of the bank that it would distract them from their main goal - to bring the bank back to profitability.
- The banks themselves knew their borrowers best and had experience in working out defaulted loans. Handling such loans will always be part of a large bank's business, and transferring employees with this expertise to a separate company might have left the banks more vulnerable when they encountered new problem loans.
- Setting up an asset management company and transferring bad loans from the banks would have required complicated accounting and legal work. In particular it would have been very difficult to find a fair price at which the loans should be transferred.
- The responsibility of handling the problem loans should remain with those who had the most to gain from a successful handling - the banks.
- Last but not the least; such an asset management company would have had to be financed by government money. Thus, more taxpayers' money would have been put at risk.
However, there was nothing to prevent the banks themselves, alone or jointly, establishing subsidiaries to handle the bad loans. None chose to do so.
Nor did the Norwegian authorities issue a blanket guarantee for the banks' liabilities. Such a guarantee can create moral hazard problems and actually lead to more bank problems in the future4. In addition, it was not considered proper for the government to take on an open-ended and not budgeted liability. Nevertheless, when the crisis emerged, it was made clear both by the Minister of Finance and by Norges Bank that measures necessary to bolster confidence in our financial system would be taken. No assurances were given that individual banks would be rescued. In this respect, it was probably a blessing that there was no government ownership in commercial banks before the crisis. In the end however, only two small banks were shut down and depositors paid back through the banks' guarantee funds.
By late 1993 - one year after Norway left the peg to the ECU and started to lower the interest rate - the crisis was effectively over and banks had started to earn normal profits. Banks in general came out of the crisis much leaner and more efficient than before.
The discounted value of the gross fiscal costs of the resolution of the Norwegian crisis resolution is estimated at 2.9 per cent of GDP as of year end 1993. At the same time, the net fiscal costs - adjusted for the value of the government's shares in the banks - were estimated at 0.8 per cent of GDP5. As referred to above, for the whole period up to the present a positive return was achieved on the capital invested.
It might be interesting to compare fiscal costs in Norway to fiscal costs in Sweden which had a crisis of similar size relative to the size of the economy. Looking at simple non-discounted sums, gross fiscal costs for Sweden were 3.6 per cent of GDP in 1997. For Norway, the comparable figure was only 2 per cent6. Part of this difference can probably be traced back to Sweden's use of government-financed asset management companies as part of their crisis resolution.
However, during a banking crisis there are important costs to society in addition to the pure fiscal costs. Some attempts have been made to estimate the cumulative effects of the banking crisis on Norway's GDP. These estimates vary considerably from a low of 9.8 per cent of GDP to a high of 27.1 per cent7. This wide variation illustrates the methodological difficulties involved in isolating the effects of the banking crisis per se.
No doubt the economic costs of the crisis - however measured - would have been far larger if the problem banks had been forced to close. Insured depositors might not have lost that much - although this is questionable - but the situation for borrowers might have been serious if they had been forced to repay their loans early. Given the state of the Norwegian economy at that time, finding a new bank willing to extend sufficient credit would no doubt have been difficult - we would most probably have experienced a severe credit crunch that deepened the recession8. The main purpose of the rescue operations was to prevent such a scenario from becoming reality.
Banks receiving a capital injection from the government were able to continue their normal bank lending and other banking operations. Empirical studies indicate that the credit conditions of firms borrowing from these troubled banks were no worse than for borrowers at non-crisis banks.9
In 1994, the government started to sell its shares in the banks. At present the government has a substantial share in only one of the three banks it took over during the crisis. The two others are now subsidiaries of a Danish and a Nordic banking group respectively. Fears of a foreign take-over, also of the largest commercial bank, has - at least so far - led the government to opt for remaining an important owner.
When the government as part of a crisis resolution steps in as an owner of last resort, the intention is usually to get out of this position once the bank has proved to be profitable again. This is probably good advice, but there is a delicate balance between restoring private ownership quickly and trying to recover as much as possible of the government's fiscal costs.
The Norwegian banking crisis was handled swiftly and transparently and at relatively low cost to the taxpayers. Furthermore, the banks that came out of the crisis had trimmed their operating costs and established much better systems to evaluate credit risks.
In one respect, we were rather lucky compared to others. The crisis built up gradually, so there was time to assess the situation and strengthen the defences of the financial system before the crisis became systemic. The banks' equity capital provided the first line of defence and the banks' collective guarantee funds the second when bank losses started to soar. The third line of defence - the Government Bank Insurance Fund - was only established when it was evident that the first and second lines of defence did not hold, but it was nevertheless established in time.
By and large most scholars seem to agree that the resolution of the crisis in Norway was fairly successful. But even if it was, right there and then, there is no guarantee that the same recipe will be successful in future or in other countries. In recognition of this and to avoid moral hazard, the law establishing the Government Bank Insurance Fund has now been repealed and the Fund has ceased its operations. Even if government capital injections in the large problem banks were the right resolution method in the early 1990s, it is far from evident that such a rescue operation would be the right method should a large bank fail in the future. However, I think it is possible to identify some principles that should be adhered to in order to increase the chances of success. The focus must be on saving the financial system, not the individual bank. With regard to the way failed banks are handled, the division of responsibilities between various parts of government must be clear, as it was ten years ago. As then, owners should be the first in line to shoulder losses. Similarly, the board and top management responsible for the failure of a bank should not be allowed to continue.
Hopefully we have all, bankers and authorities alike, learnt from the crisis. At the central bank, we have strengthened our surveillance of the soundness of our financial system (macro-prudential supervision). There is also an extensive exchange of information with the supervisors in the Banking Insurance and Securities Commission. Twice a year, Norges Bank publishes its financial stability report where the risks to the present and future solvency of financial institutions in Norway are assessed. The aim is to detect the first signs of banking problems more quickly and possibly prevent them from escalating into a crisis. However, the financial system is evolving rapidly. Tomorrow's problems could easily be different from the past.
Today, a large part of our banking industry is part of Nordic banking groups. This raises the question of whether a model for crisis resolution only based on national considerations would work. In light of this challenge, Nordic central banks have issued a memorandum of understanding setting out principles for the establishment of a structure for crisis management and the handling of information if a pan-Nordic bank should have problems. Nevertheless, the emergence of multinational banking groups raises the question of whether some banks may be "too big to save".
Another difference from the early 1990s is the government's one-third ownership in the largest bank. In terms of crisis resolution, this is most probably a further complication.
Banking crises have occurred at irregular intervals as long as banks have existed. And it could happen again. We need to be vigilant and alert in our macro prudential supervision, and willing to consider new approaches in crisis resolution to lessen the impact of any new financial crisis.
Thank you for your attention.
Drees, Burkhard and Ceyla Pazarbasioglu (1998). The Nordic Banking Crisis: Pitfalls in Financial Liberalisation. IMF Occasional Paper.
Hoggarth, Glenn, Ricardo Reis and Victoria Saporta (2002). Costs of banking system instability: Some empirical evidence. Journal of Banking and Finance, 26, pp. 825 - 855.
IMF (2001). Thailand: Selected Issues. IMF Country Report 01/147, International Monetary Fund, Washington D.C.
Jennergren, Peter and Bertil Näslund (1998). Efter Bankkrisen: Vad blev notan för skattebetalarna? Ekonomisk Debatt 1998/1, pp. 69 - 76.
Kane, Edward J. and Daniela Klingebiel (2002) Alternatives to blanket guarantees for containing a systemic crisis. At: www1.worldbank.org/finance/assets/images/ Kane_Klingebiel--alterns_to_blnkt_guars-doc.pdf,
Moe, Thorvald G., Jon A. Solheim and Bent Vale (2004). The Norwegian Banking Crisis, Norges Bank Occasional Paper No. 3X. Forthcoming.
Moen, Harald (2003). Nåverdien av statens investeringer i og støtte til norske banker. Working Paper ANO 2003/6, Norges Bank, Oslo
Norwegian Official Reports (1992). Report by the Commision on the Banking Crisis. NOU 1992: 30E, Royal Ministry of Finance and Customs, Oslo.
OECD (2001). OECD Bank Profitability 2000.
Ongena, Steven, David C. Smith and Dag Michalsen (2003). Firms and their distressed banks: lessons from the Norwegian banking crisis. Journal of Financial Economics, 67, pp. 81 - 112.
Peek, Joe and Eric S. Rosengren (2000). Collateral damage: Effects of the Japanese Bank Crisis on Real Activity in the United States. American Economic Review, 90, pp. 30 - 45.
Sandal, Knut (2004). The Nordic banking crises in the early 1990s - resolution methods and fiscal costs. In Thorvald G. Moe, Jon A. Solheim and Bent Vale (eds.), The Norwegian Banking Crisis. Norges Bank Occasional Paper no. 3X. Forthcoming.
Schwiertz, Christoph (2003). Economic costs of the Nordic banking crises. Working Paper. ANO 2003/X forthcoming, Norges Bank, Oslo.
Steigum, Erling (2004). Financial Deregulation with a Fixed Exchange Rate: Lessons from Norway's Boom-Bust Cycle and Banking Crisis. In Thorvald G. Moe, Jon A. Solheim and Bent Vale (eds.), The Norwegian Banking Crisis. Norges Bank Occasional Paper no. 3X. Forthcoming.
Stortinget (1998). Rapport til Stortinget fra kommisjonen som ble nedsatt av Stortinget for å gjennomgå ulike årsaksforhold knyttet til bankkrisen. Dokument nr. 17 (1997-98), Oslo.
Vale, Bent. (2002). Firms' inventory investments, financial conditions and banking crisis. In Iftekhar Hasan and William C. Hunter (eds.), Research in Banking and Finance, vol. 2. JAI Press, pp. 179 - 198.
Wilse, Hans Petter (1995). Management of the banking crisis and state ownership of commercial banks. Norges Bank Economic Bulletin, 66, pp. 217 - 229.
1Source: Sandal (2004)
2Source: Kane and Klingebiel (2002) and Norges Bank
3Interested readers are referred to Moe, Solheim and Vale (2004), Norwegian Official Reports (1992), Wilse (1995), and Stortinget (1998). Steigum (2004) sets out the macroeconomic background and implications of the banking crisis.
4 Cf. Kane and Klingebiel (2002)
5 Source: St.meld. 39 (1993-94)
6 Source: For Sweden: Jennergren and Näslund (1998). For Norway: Norges Bank, Moen (2003)
7 See Hoggarth, Reis and Saporta (2002).
8 Peek and Rosengren (2000) look at the loan supply shock facing U.S. firms borrowing from Japanese banks during the banking crisis in Japan. They identify a substantial impact on U.S. real estate activity from this supply shock.
9See Ongena, Smith and Michalsen (2003) and Vale (2002)