RDP 2008-06: Promoting Liquidity: Why and How? 2. The First-best and the Real World

In thinking about how public policy should respond to asset illiquidity it is useful to step back and ask what the ‘first-best’ world would look like. This was done very nicely at this conference last year in a paper by Franklin Allen and Elena Carletti.[1]

They note that ‘if financial markets are complete, it is possible for intermediaries to hedge all aggregate risks in financial markets’ (p 207). In such a world, institutions could use securities, derivatives or trading strategies to ensure that liquidity is available when it is needed, with the price system ensuring that the liquidity was appropriately priced in every state of the world. In this perfect world, ‘market liquidity’ would be plentiful so that assets could be readily bought and sold at their fundamental value, and ample ‘funding liquidity’ would enable solvent institutions to easily borrow against their assets.

The real world falls well short of this first-best benchmark in at least two important ways. The first is that not all assets can be bought and sold in liquid markets, and where liquid markets do exist in normal times, they can disappear at short notice, just when they are most needed. The second is that the availability of funding can evaporate quickly, making it difficult for institutions to continue financing their assets. The effect of this can be particularly pronounced if it coincides (as is likely) with illiquid asset markets, as the institution experiencing the funding difficulties cannot simply downsize its balance sheet by selling assets in an orderly market. The reasons why these liquidity problems can emerge are discussed in the following section.

Given the limitations of the real world, distressed fire sales of assets can occur, and solvent institutions can find themselves unable to obtain funding, or sell assets on reasonable terms. As Allen and Carletti note, the result can be more volatility in asset prices than is socially optimal, and ‘costly and inefficient crises’ (2007, p 209).

While the real world clearly falls short of the first-best, many of the developments in the financial system over recent years can be seen as moving the system closer to this benchmark. One obvious example is the securitisation of loans on banks' balance sheets, with securitisation offering the promise that historically illiquid assets could be liquefied. Indeed, some financial institutions had included the possibility of securitisation in their contingency planning for a liquidity crisis. Another example is the widespread use of contingent credit lines, with the entity paying for such a line essentially insuring itself against the possibility of funding difficulties and/or being a forced seller of assets. There has also been very strong growth in the trading of a whole range of financial products, which has allowed various assets and risks that previously could not be traded in markets to now be traded; one example is the credit default swap (CDS) market which allows the trading of credit risk.

The paradox here, however, is that while these developments may have moved the system closer to the first-best world in normal times, they do not appear to have had the same effect under more turbulent conditions. Many of these developments assist with the management of idiosyncratic liquidity issues and aid the efficient functioning of the market under normal conditions. However, they have not proved particularly resilient under strain, and the comfort that they have provided to institutions under normal conditions may have increased aggregate liquidity risk by encouraging the belief that if things changed for the worse, the markets could be relied upon to manage both liquidity and asset positions.

As institutions have become more dependent upon financial markets for the management of their balance sheets, the importance of the smooth functioning of these markets has simultaneously increased. Not only are these markets used for managing many more risks than was once the case, they have also supported the increased use of mark-to-market accounting. One consequence of these developments is that if liquidity dries up, amplifying movements in the prices of financial assets, the potential systemic implications are much larger than they once might have been.[2]

Reflecting this, in the past decade there have been a number of cases in which concerns about market liquidity have been at the forefront of policy-makers' minds. The concerns have been most acute in situations in which the failure of an institution was considered a real possibility. In particular, in the cases of both Long-Term Capital Management (LTCM) and Bear Stearns, policy-makers in the United States were extremely concerned that markets could not deal with the closing-out of positions that would inevitably follow the failure of a major counterparty. As Bill McDonough, the then Head of the Federal Reserve Bank of New York, said in the wake of LTCM's problems, the closing out of these positions ‘… would have caused a vicious cycle: a loss of investor confidence, leading to a rush out of private credits, leading to a further widening of credit spreads, leading to further liquidations of positions, and so on’ (see McDonough 1998). Similarly, 10 years later, in explaining the Fed's actions in response to Bear Stearns' problems, the Head of the New York Fed, Tim Geithner, said that by agreeing to lend against a pool of assets, the Fed had ‘… reduced the risk that those assets would be liquidated quickly, exacerbating already fragile conditions in markets’ (Geithner 2008).

Similar concerns arose when the US hedge fund, Amaranth Advisors, got into trouble in 2006. In particular, its counterparties were concerned that if its positions had to be closed out on-market, there would be very large movements in prices with potentially destabilising effects. In that case, the situation was resolved by one of Amaranth's bankers eventually taking over its positions off-market at a substantial discount to their apparent market value. One view on why the situation with Amaranth was more easily resolved than LTCM's is that its positions were exchange-traded rather than being over-the-counter (OTC), an issue we discuss in the next section.

Liquidity issues have also been at the forefront of concerns arising from the sub-prime problem in the United States. A sharp fall in the demand for assets with unfavourable liquidity characteristics has seen a marked fall in the price of these assets relative to those whose liquidity is more assured, with many markets having essentially closed. Many financial institutions have also become much less willing to tie up their balance sheets in assets that cannot be sold easily, including term bank loans. This, combined with concerns about the ability to tap various funding markets on an ongoing basis, has resulted in a substantial increase in term spreads and a significant tightening of credit conditions. In some countries, there have also been runs on financial institutions, something that in the past has been quite rare in developed financial systems.

These various liquidity problems have not just affected a small group of institutions, but have been global in nature, and have had significant effects on economic activity. Indeed, the swing from a situation in which liquidity was unusually high, to one in which it is unusually tight, has been the major driver of the current business cycle in many countries.

Given the potential for adverse impacts of liquidity problems on the financial system and the real economy, a relevant question is: how should policy-makers respond? This question has taken on additional importance over time, particularly given that many developments may have moved the financial system further away from the first-best in troubled times. The arrangements for dealing with system-wide liquidity problems and, more broadly, disruptions to markets have become particularly important.

Here there are at least three (not necessarily mutually exclusive) perspectives, which we have stylised to make the views as clear as possible.

The first is that further financial innovation is required, so that the real world looks more like the first-best, not just in normal times, but also in troubled times. According to this perspective, the main problem with current arrangements is that there are still too many missing markets and too many impediments to state-contingent contracts, and that key parts of the financial infrastructure are underdeveloped. The key to a more stable system is to develop these markets, remove these impediments, and shore up the existing markets by improving the financial infrastructure so that participants can transact on reasonable terms in both good and bad times.

A second perspective is that financial markets will never be complete, and that realistically the various forces that periodically cause liquidity problems can never be completely overcome. In response, financial institutions need to hold more liquid assets than they have become accustomed to, and to be more realistic about their true potential liquidity needs (reflecting both explicit and implicit commitments). In doing so they need to take into account the possibility that normally liquid asset markets and reliable funding sources can evaporate in times of stress. According to this perspective, liquidity insurance has been underpriced for too long and many financial institutions have undertaken too much maturity transformation. Reflecting this, institutions need to either voluntarily hold more liquidity, or be forced to do so by regulators. The case for addressing this issue through regulation is strengthened by the idea that the benefits to the system of an institution holding more liquid assets are not fully internalised, with regulation potentially solving the distortion caused by this externality.

A third perspective is that while private financial institutions need to be responsible for ensuring that they can deal with idiosyncratic liquidity problems, they should not have to shoulder alone the burden of ensuring themselves against system-wide disruptions. According to this view, overall social welfare can be improved by the public sector providing systemic liquidity services to the private sector. In some situations it may be able to do this at little cost and with little risk to the taxpayer. In other cases, the risks may be significant, but so too may be the benefits; in particular, by playing this role, the public sector may be able to reduce the costs that society pays for financial intermediation.

We return to these various perspectives in the following sections. Before this, however, it is useful to discuss the reasons why asset markets are not always liquid.


See Allen and Carletti (2007), and also Allen and Gale (2004) and Holmstrom and Tirole (1988). [1]

Gai et al (2008) present a model which explains why financial innovation may have made financial crises less likely, but more severe if they occur. [2]