Tuesday, December 21, 2010

All assets are illiquid

Much of financial theory is built on the premise that markets are liquid for the most part and that illiquidity, if it exists, occurs in pockets: it shows up only with very small, lightly traded companies, emerging markets and privately owned businesses. In fact, almost the prescriptions we provide to both investors and corporate finance reflect this trust that both security and asset markets are liquid.

To see how unrealistic the assumption of liquidity is, consider what a liquid market would require: you should be instantaneously be able to sell any quantity of an asset at the prevailing market price with no transactions costs. Using that definition, no asset is liquid and the only question then becomes one of degree, with some assets being more liquid than others.

Given the premise that all assets are illiquid, here is a follow up question: how do you measure illiquidity? One obvious measure, especially in securities markets, is the total transactions costs, including not only the brokerage costs, but the bid-ask spread and the price impact from trading. The other is waiting time. In real estate, for instance, illiquidity manifests itself in properties staying on the market for longer periods. Days on market (DOM) is a widely reported statistic in real estate and is used to measure the health (and liquidity) of different markets.

There is significant empirical evidence that illiquidity varies across asset classes, within assets in a given asset class and across time.
  • Across asset classes, illiquidity is more of a problem in real asset markets (real estate, collectibles) than in financial asset markets. Within financial asset markets, the US treasury market is the most liquid, followed by highly rated bonds and developed market stocks, with low-rated (junk or high yield) bonds and emerging market stocks bringing up the rear.
  • Within each asset class, there are wide variations in liquidity. In the treasury market, the just-issued, standard maturity treasuries (3 month, 6 month, 10 year) are more liquid than the seasoned, non-standard maturity treasuries. Within the stock market, larger market cap and higher priced stocks are more liquid than smaller market cap, lower priced stocks.
  • Liquidity also varies widely over time. While the long term trend in liquidity in equity markets has been towards more liquidity, liquidity moves in cycles, increasing in bull markets and decreasing in bear markets. Punctuating the long term trend are crises, like the 1987 sell-off in the US and the 2008 banking crisis, where liquidity dries up even for the largest market cap companies. During these crises, illiquidity manifests itself in many ways: trading halts, higher bid-ask spreads and bigger price impact when trading.
None of this evidence would matter if investors did not care about illiquidity but there is clear evidence that they do: liquid treasuries have lower yields (and higher prices) than illiquid treasuries and investors demand higher returns on stocks with lower trading volume and higher bid-ask spreads. One study find that every 1% increase in bid-ask spreads increased expected returns by 0.25%, and these higher required returns push down asset prices. Adding to the problem, the price that investors charge for illiquidity also varies over time, spiking during periods of crises: illiquid assets get discounted even more during these periods.

If illiquidity varies across asset classes, across assets and across time, and investors price in this illiquidity, it seems prudent that both portfolio  and corporate financial theory be modified to reflect the potential for illiquidity. Alas, given the length of this post, that has to wait for the next one.


Radovan said...
This comment has been removed by the author.
stephanie.lee said...

"Within financial asset markets, the US treasury market is the most liquid, followed by highly rated bonds and developed market stocks, with low-rated (junk or high yield) bonds and emerging market stocks bringing up the rear."

What about currencies? Plain vanilla interest rate swaps? They are pretty darn liquid.

Manish Poddar said...

Hi sir,
According to you which has been the Asset class of the year 2010 and what are your views for the next year.

Anil Bains said...

Sir I have a question here..
In India we take the 10 yr Govt. Bond yield as Risk-Free rate for cost of equity calculation using CAPM.
Shoudn't we:

1. Subtract all the spreads such as credit, liquidity out of the yield to get true risk-free rate.(India is BBB rated in sovereign debt)

2. Ideally take 1 Yr T-Bill's yield for 1st year's cost of equity calculation, 2-yr Govt. bond's yield for 2nd year's cost of Equity or use some extrapolation technique(Nelson Seigel Svensson Eq) if 2 yr bond is not available. Then using these different Cost of Equities for DCF purpose.

Aswath Damodaran said...

Stephanie, nothing matches treasuries for liquidity. Currencies are not an asset class and plain vanilla swaps are a risk management tool.
Anil, Yes and Yes.

David Darten said...

There are a number of reasons why an asset can become illiquid. One common reason is uncertainty about the value of an asset, which can be caused by general financial instability, or issues specific to the asset.

Interest Rate SWAP