Sweden experienced a severe banking crisis in the beginning of the 1990's. To manage the crisis, the authorities had to implement measures that averted a complete breakdown of the financial system while attempting to limit the costs to taxpayers. In addition, it was imperative for the authorities to consider whether or not their actions would contribute to future moral hazard problems. During the crisis, an effective division of labour developed between the authorities involved, namely the Ministry of Finance, the Riksbank, the central bank, Finansinspektionen, the financial supervisory authority and the Bank Support Authority. As the crisis progress, the Riksbank provided emergency liquidity assistance (ELA) twice, acting in its role as the "lender of last resort". The remarks in this note are largely drawn from the Riksbank's experiences in this crisis, and how these experiences can be used to assessing the role of ELA in a general crisis management infrastructure.
When discussing the infrastructure that is in place to handle financial crises, two aspects are of primary importance. First, ELA must be put in the context of the overall methods crisis management that the authorities would employ should a financial crisis occur. ELA is not a measure that can or should be used in every financial crisis. The way other crisis management procedures are set up also affects the scope for ELA. This is covered in the first part of this note.
Second, the way ELA issues are organised within the central bank is also very important for whether it is an effective measure and, in particular, whether it will reduce or increase moral hazard in the financial sector. This is discussed in the second part of the note.
Before discussing these two aspects, the importance of speed in granting ELA must be emphasized, as the speedy implementation of ELA is essential to its being effective in a crisis situation. . ELA is the only measure in the crisis management framework that normally can be used quickly. It should be seen as an acute measure that is used for avoiding a sudden disruption of the large value payment system, in particular when threats to the system emerge suddenly.
Incidents like the failure of Barings bank, LTCM and the Swedish currency and banking crisis show that threats to the financial system may indeed emerge suddenly, and that authorities must react very quickly - if not within minutes, at least within hours. The growth and increasing globalisation of financial markets, the high volatility in these markets and the increasing complexity of financial institutions, are all examples of factors that increase the risk for sudden crises and increase the probability that such crises will spread from one market to another.
ELA as a part of general crisis management and resolution
Should a financial crisis occur, there are a number of measures that may be used by authorities to limit the consequences of the crisis. In addition to ELA, examples of these measures are financial support to protect the solvency of institutions, brokering of private sector support to or acquisitions of troubled institutions, management of deposit insurance schemes and insolvency procedures for financial institutions. In general, these measures may be undertaken by different authorities. A consequence of this is that co-operation between authorities is an important element of crisis management. The authorities must have a common strategy for how the crisis should be managed in order to avoid excessive support to the financial sector. Excessive support may lead to undue costs to taxpayers and to unsound incentives for the financial sector. In this context, it is important to emphasise that these effects can arise from all kind of official support. Thus, the authorities must limit the support that is offered. Support may be warranted only when there is a considerable risk for a systemic crisis that may lead to a severe decline of real economic activity. Typically, real economic activity is affected by a breakdown of payments systems or restrictive granting of credit, a credit crunch.
Traditionally, ELA provided by central banks has been seen as a measure to deal with pure liquidity problems in banks that essentially are solvent. In practice, situations where pure liquidity problems arise seem to be rare. In most cases where banks face liquidity problems, there is some uncertainty about the solvency of the institutions involved. Typically, it takes time to make an assessment of the solvency of an institution. In the Swedish banking crisis, it took many months to fully assess the financial situation of the troubled institutions. Not even the institutions themselves had a good picture of how large their problems were. Recent developments in the financial system, such as consolidation, the emergence of geographically disperse financial conglomerates with activity in many product lines and a greater complexity of financial products, make this task even harder. It is unlikely that central banks or supervisory authorities will be able to make a valid assessment of the solvency of troubled institutions quickly enough. As mentioned above, requests for liquidity support may arise very rapidly. The conclusion of this is that it will in practice be very difficult to draw the dividing line between the central bank's responsibilities andthose of other authorities by distinguishing between institutions with solvency problems and those with pure liquidity problems, at least during a sudden crisis.
Another conclusion is that it is crucial to establish rules for information sharing between authorities, especially between the financial supervisory authority and the central bank in advance so that the central bank quickly can make an assessment. It is also important that the central bank, even if it does not have direct supervisory responsibilities, closely monitors general financial markets developments, including in the major institutions, to detect potential crisis situations that may call for ELA.
The traditional view that ELA never should be given to insolvent banks can also be discussed from another perspective. A prerequisite for all kinds of support, including ELA, should be that there is a considerable risk for a systemic crisis. Given that, it can be discussed whether it is worse to provide ELA to an insolvent institution, or to refrain from an action that could avoid a systemic crisis. If there is a risk for a systemic crisis, it is normally determined that public support even to insolvent institutions may be warranted, but this is normally given by the government through the fiscal budget, not by the central bank. The moral hazard aspects, that are the common arguments for not providing ELA to insolvent institutions, will be no different if the central bank provides support than if the government does it.
The central bank may have better capability than other authorities to provide financial support if the need arises very quickly. The main reason for this is that the central bank has the instruments for lending available and it has resources in its balance sheet or the credibility to get resources if they are needed (for example in foreign currency). It also has an established capacity to judge whether a systemic crisis may occur as a result of an institution's financial problems. The Riksbank is able to make this judgement by virtue of its role as provider of the central payment system and the responsibilities attached to this, its presence in financial markets and its general knowledge of the state of the economic and financial system. Other authorities typically have greater obstacles to reacting fast, as they do not in most cases have instruments or resources readily available for these sorts of operations. There may also be difficult political considerations that have to be taken. For example, a minority government may need to seek support from other parties to use budgetary resources, a support that may be hard to get or at least take some time. However, this is not to be taken as an argument for governments to use the central bank as a provider of support because it does not want to use budgetary resources. The central bank should be independent and control its own resources.
The conclusion of this is that ELA operations should have the role in the general crisis management infrastructure of dealing with quickly emerging problems. Where problems arise gradually, other measures available to supervisory authorities and governments are more likely to be used. ELA should be seen as a means for handling acute problems, so that the authorities gain time to decide whether other crisis management measures are needed, such as restructuring of troubled institutions. If private sector support can be organised, such as in the case of LTCM, that is of course to be preferred.
If the view is accepted that ELA should function as a "bridge" measure to prevent acute crises from occurring, it is very important that the support is given in a way that minimises the credit risk for the central bank, in order to avoid moral hazard effects. The central bank must have a well-considered exit strategy for its lending operations. The most obvious way to do this is to lend against collateral. In most cases where a need for support occurs, collateral that is normally accepted by the central bank will probably not be available. The central bank may in these cases be required to accept lower quality collateral. It should in these cases apply a larger haircut on the collateral than is used for normal collateral. Another part of a proper exit strategy is that financial support should not be given to shareholders or other holders of risk capital.
Internal procedures for ELA
In the previous section, two main points have been suggested; (1) that ELA is a measure that predominantly should be used for problems that arise without warning and (2) that ELA should only be provided if there is a risk that a systemic problem would occur if the central bank did not act. These two features bring about requirements for the way central banks should organise the ELA function internally. The possible need for speedy assistance necessitates that a central bank must have internal procedures for managing requests for liquidity support and doing so very rapidly. To accomplish this, central banks need to have a detailed crisis plan that can be put in place within very short notice. The crisis plan should include the following types of information:
Organisational structure
:
- A division of responsibilities for analysis, information, decision making etc.
- Staffing requirements for different tasks, with subsititutes
Analysis plan
- A description of the information necessary to make a decision about whether or nor a systemic crisis may occur.
- Necessary information about systemically important individual banks
Information strategy
- A plan for informing and co-ordinating information and decisions with the other domestic agencies involved in the crisis
- A plan for informing authorities in other countries about the situation
- A plan for communicating with the financial markets
- A plan for spreading information to the public and media (ex ante and ex post transparency)
Plan for implementing ELA
- A clear description of the routines which must be followed and the legal requirements which must be met for ELA
- Legal procedures, for example contracts, for lending against other collateral than normally eligible
- An inventory of the instruments available for implementing ELA
- Guidelines for acceptable collateral and appropriate haircuts.
Internal preparedness is important in order to be able to consider requests for ELA, as the timeframe for reaction normally is very limited. It is important to note that preparation is crucial also if a central bank is to deny a request for ELA. If the central bank has no organisation for analysing systemic consequences and the availability of measures, it is very likely that it will not dare not to lend.
No central bank wants to give financial institutions and markets the slightest reason to depend on central bank support rather than their own efforts to avoid financial meltdown. Clearly, ELA is a last resort measure that central banks use only when the stability of the financial system is at stake. A well-planned and clear LOLR function need not encourage moral hazard. On the contrary, a central bank with clear guidelines and a solid analysis capability is in the best position to pinpoint those situations in which ELA is necessary. In all other situations, clear guidelines help the central bank to stand firm against the institutions' demands for help and to force markets to solve financial problems by themselves.
