The credit crisis has highlighted three serious problems in current liquidity arrangements.
First, during the worst of the crisis, even collateralized borrowing against investment-grade securities may not be available to a large institution facing actual or perceived solvency and liquidity pressures, as Bear Stearns, the US bank, learned too late. The trend towards lower liquid-asset ratios and the shift from "owned" in favor of borrowed, just-in-time liquidity has exacerbated this vulnerability.
Second, banks may hoard liquidity in a crisis-because they are uncertain of their own future liquidity needs or because they are increasingly anxious about the creditworthiness of their clients. Central bank injections of liquidity into the system do not guarantee that this added liquidity will go to those that need it most.
What is required is a clearer picture of what constitutes regulatory liquidity, along with greater incentives for holding adequate amounts of it and sharing it with others.
Third, while central banks can compensate for the first two problems by offering large-scale liquidity assistance to a broad range of market participants against a wide range of collateral, the larger, the more frequent, and the longer-lasting such official support is, the greater the risk that this official lifeline will undermine incentives for market participants to self-insure against liquidity risks.
Contrary to recent official and industry reports, these problems will not be solved by further calls for more stress testing and contingency planning or by a revised set of principles on liquidity risk management. What is required is a clearer picture of what constitutes regulatory liquidity, along with greater incentives for holding adequate amounts of it and sharing it with others. I offer the following proposal.
One: regulators should define regulatory liquidity narrowly. This should give a dominant role to Treasury securities that will retain their unquestioned liquidity in a crisis and it should penalize very short-term borrowing relative to longer-term financing. Regulators should then set a minimum quantitative benchmark for regulatory liquidity. The algorithm for calculating minimum regulatory liquidity for commercial and investment banks would be subject to a variety of compromises and distortions, but these would be less damaging than the ambiguities and biases that have produced significant under-insurance against liquidity risk under current arrangements.
Two: there is a place for a liquidity pool among large and entangled commercial and investment banks of systemic importance. Each member would deposit with its central bank an agreed quota of Treasuries that it could draw instantaneously when needed and without challenge. Each member would also be able to overdraw by several times the size of its initial deposit-if needed to meet unusually large liquidity strains. Such overdrafts would have to be fully collateralized by investment-grade securities (from the member's holdings outside the pool) and would carry a market rate of interest. As a condition of membership, all pool members would agree to allow their deposits to be lent to other members. Since the pool would include some large banks with insured deposits (outside the pool) that typically do not decline-and may even rise-during a crisis, it is unlikely that too few members would be in a condition to lend liquidity to others. The central bank-or other relevant regulatory authority-would lend its support by determining eligibility for membership of the pool and by overseeing its operation. Market and default risks would, however, be borne exclusively by members of the pool.
Three: when liquidity needs went beyond the pool's capabilities, large commercial and investment banks would turn to their central bank to act as a lender of last resort, drawing, if necessary, on the full set of liquidity facilities employed during this crisis. Access to these facilities, however, would carry a higher cost of borrowing than in the pool and there would be a strong presumption that official liquidity assistance would come only after private sources had been exhausted.
Under this proposal, systemically important players would have added incentives to hold a minimum amount of owned liquidity to use in a crisis. Hoarding of liquidity would be ruled out by the commitment to lend to other members of the pool. Pool members would also have assurance that they could engage in collateralized borrowing even during the worst of the crisis.
And central banks, as the third line of defense against liquidity strains, could not only take less credit and market risk on to their balance sheets but also guard against becoming lenders of first-rather than last-resort.