RDP 2008-06: Promoting Liquidity: Why and How? 7. Policy Discussion

It is clear that liquidity problems can have significant effects on the financial system and the real economy. It is equally clear that there is no single solution to addressing these problems. Reducing information asymmetries and improving market infrastructure have an obvious role to play. An improvement in the way that institutions manage their own liquidity is also required. Further, at the supervisory level, attention needs to be paid to the potential for system-wide liquidity problems, and not just problems that are restricted to a single institution. Central banks (and possibly other public sector bodies) also have an important role to play. At issue is how extensive this role should be, and what conditions should apply.

Unfortunately, too often discussions of this issue are derailed by quick references to the dangers of ‘moral hazard’. It is sometimes argued that if the public sector provides any form of liquidity services to the private sector, the result will be more risk-taking, and ultimately either a more crisis-prone system, or higher costs to the taxpayer.

While not wishing to downplay the risks, this argument misses a key point, namely that, while the provision of liquidity services by the public sector will undoubtedly change the behaviour of the private sector, this change in behaviour need not be welfare-reducing. If some form of systemic liquidity services are not provided, private institutions need to provide their own insurance by holding more liquid assets than would otherwise be the case. The end result may be a higher cost of financial intermediation and, in turn, a lower capital stock. Institutions may also be less prepared to commit funding for longer-term projects and more likely to cut back credit lines when troubled conditions emerge (although presumably the emergence of such conditions would be less likely). Indeed, making a credible ex ante commitment to provide a certain degree of liquidity assistance may actually reduce moral hazard relative to a statement that the central bank will not provide liquidity assistance. If the private sector does not believe that such a statement is credible, then it is likely to condition its behaviour on the level of liquidity assistance that it thinks the central bank would provide.

None of this is to imply that institutions themselves should not have responsibility for managing their own liquidity. They clearly do. Moreover, they need to be prepared for significant dislocations in the key markets in which they operate and disruptions to their normal funding patterns. Over recent years, many institutions appear not to have done this adequately, undertaking too much maturity transformation, with too little capital, and on a funding base that was much less stable than widely assumed. It is important, though, that in responding to these shortcomings there is not an overreaction the other way which requires the private sector to fully self-insure against system-wide liquidity problems. Given that, to some extent, these problems arise from underlying distortions or market failures, full self-insurance is unlikely to be consistent with welfare maximisation.

In our view there is a strong case for the central bank to play the sort of liquidity smoothing role discussed in the previous section, increasing the supply of liquid assets at a time when the market places a very high value on these assets. It can do this by increasing the size of its own balance sheet and/or changing the composition of its assets during times of strain. While playing such a smoothing role will lead to an increase in the risk carried on the central bank's balance sheet, this increase can be limited by the use of appropriate haircuts, and the central bank will be compensated for this additional risk through higher expected returns on its asset holdings.

For this role to be played effectively, the central bank needs to have a considerable degree of flexibility in its market operations, including the frequency, maturity and scale of these operations. Many central banks have moved in this direction over the past year.

We also see a strong case for the central bank being prepared to purchase a wide range of third-party assets under repo. Doing so can reduce illiquidity premia that apply to these assets and reduce the possibility that solvent financial institutions find themselves needing to seek emergency support. One useful criterion to apply in considering where the boundary should be between acceptable and non-acceptable assets is the degree of information asymmetry, with the greater the asymmetry, the weaker the case for the central bank buying the asset under repo. In some situations, this criterion might rule out accepting assets that an institution has originated itself, or at least requiring greater protection through larger haircuts.

In extremis, there may also be grounds for the public authorities to purchase outright a very limited range of assets. However, the risk-return trade-off from such purchases is, in most cases, likely to be much less attractive than the actions discussed above. This means that the ‘burden of proof’ that the public sector needs to meet in justifying such intervention should be set very high. One variant of this approach is for the public sector to assist with the off-market transfer of assets of a troubled financial firm, including possibly, in extremis, taking assets directly onto the public-sector balance sheet and disposing of those assets gradually over time. One argument for doing this is that in some extreme situations it is in the public interest for the assets owned by a troubled institution to be sold in a measured way, rather than dumped onto markets when risk and illiquidity premia are at their highest.

As discussed above, situations can also emerge where providing a loan directly to a troubled, but solvent, institution may also be in the public interest. Over time, however, with market operations becoming more flexible, the probability of such support being used to assist an institution over temporary funding difficulties has probably declined. It is more likely that such support provides bridging finance while new ownership and management are put in place.

In supporting a role for the public sector in providing a range of systemic liquidity services to the private sector, the moral hazard concerns discussed above need to be addressed. In doing so, it is important to recognise that the relative public versus private benefits of the various liquidity services differ across these services. In particular, there is a strong public good element in the central bank to play a contrarian role when liquid assets are in high demand, and in helping reduce illiquidity premiums in financial assets. While financial institutions benefit from these services, these benefits are spread widely and are not concentrated in a particular institution. In contrast, providing a direct loan to an institution can lead to significant benefits to those associated with that institution; of course there may be also benefits to the market more generally, particularly if, in the absence of the liquidity support, the troubled institution would be liquidated, causing widespread dislocation in financial markets.

How then can the moral hazard concerns be addressed? We see three not mutually exclusive possibilities.

  1. The first is a strengthening of the macro-prudential elements of supervision.

    While we have argued that there are strong grounds for the central bank to take a contrarian position in the sense we discuss above, and to assist more generally when system-wide liquidity problems emerge, there is a certain asymmetry if such actions occur only when conditions are unsettled. It is not implausible that this asymmetry itself could affect private-sector behaviour. One way of addressing this is for supervisory requirements to be tightened in the good times, when liquidity is judged to be ample and credit risk low. The case for this type of cyclical supervisory response is strengthened by the observation that system-wide liquidity problems invariably have their roots in the underestimation of risk in good times.[16] If the public sector is to provide some form of systemic liquidity insurance – and inevitably accept a higher level of risk in doing so – the trade-off may be a tightening of supervisory requirements in good times. In a sense, such a tightening could be thought of as part of the ‘insurance premium’ that the private sector pays for the liquidity services that the public sector provides. It would also assist in the building-up of the system's buffers in good times and reduce the probability of liquidity problems emerging when conditions eventually deteriorate.

  2. A second possibility is to ensure that institutions are subject to prudential regulation if there is any possibility that the public sector might need to offer some form of institution-specific support.

    Significant moral hazard issues arise if an institution is able to sit outside the regulatory net but obtain support when times are troubled. Again, submitting to prudential regulation can be part of the ‘insurance premium’ that institutions are required to pay if they are to ever obtain institution-specific assistance. There is a strong case for them to be required to pay this ‘premium’ if they are large and have complicated dealings in financial markets.

    A tangentially-related issue is who the central bank should be prepared to deal with in its daily market operations.[17] Where these operations are conducted in high-quality, third-party assets, the counterparty risk being run by the central bank is normally low. There is, therefore, a strong case for the eligibility requirements to be largely limited to operational issues related to the effective implementation of monetary policy. If this is the case, then a very wide range of institutions – including non-banks – can participate in market operations (as is the case in Australia). The situation is somewhat different when it comes to transactions in assets that have been originated or sponsored by the central bank's counterparty. Accepting related assets can involve significant additional risk, and the case for doing so in the course of normal market operations appears weak. This is particularly so for an institution that is not subject to prudential regulation.

  3. A third way of addressing moral hazard relates directly to the conditions that apply to liquidity assistance outside normal market operations.

    As discussed earlier, if despite the central bank having flexible operating procedures, an institution requires emergency liquidity assistance, then that institution is probably in very significant trouble. Extending support to such an institution may be in the public interest, but it also risks providing benefits to those directly associated with the institution, including its managers and shareholders. Given this, it may be better to think of emergency liquidity support as the public sector providing bridging finance, while new ownership of the institution is being arranged. This was what essentially happened in the cases of Northern Rock and Bear Stearns, with in one case the new owner being the government, and the other, a private bank. It seems likely that the days are gone (if indeed they ever existed) in which an institution could obtain emergency support, then repay that support after the funding problems resolve themselves, with the bank institution then continuing on as normal.


For a fuller discussion of this option see Borio (2007) and Borio, Furfine and Lowe (2001). [16]

See Hilton (2008) for a discussion of recent changes in the Fed's range of eligible counterparties. [17]