Philip Sommer and Stefano Pasquali, in their recent paper Liquidity – How to capture a multidimensional beast, point to the highly dynamic nature of liquidity and the manifold factors that affect the liquidity of an asset from individual security characteristics (credit quality, maturity, amount issued, age, etc.) through market-related characteristics (volume, turnover, bid/ask spread, market breadth etc.) to market conditions, the business cycle, liquidity crises and even an individual trader’s level of sophistication and relationship with their counterparties.
It is common practice to use one or a combination of these factors as proxies to measure liquidity in the models commonly used for trading and portfolio manage
ment, yet their shortcomings as an adequate measure are increasingly pronounced since the financial crisis.
THE PROXY PROBLEM
One of the most commonly-relied upon proxies is the bid/ask spread, which has become popular because the data is readily accessible on-screen and because it is assumed to be highly correlated with other liquidity measures.
However, Sommer and Pasquali point to two obvious pitfalls in relying on the bid/ ask spread: firstly, quotes are often only indicative and, secondly, the bid/ask spread is only really useful up to the quoted volume. Beyond that, the spread widens as trade size increases and prices can deviate significantly from what the quoted bid/ask spread might suggest.
The volumes available to transact at on-screen bid/ask prices tend to be small and, thus, the spread has limited use as a gauge of liquidity for typical institutional trade sizes.
