After months spent analyzing empirical market data from three major exchanges—including 25 million-plus daily observations and 1 million-plus monthly observations, Abbott and Pratt found, as expected, that the relationship between size and liquidity is not statistically significant. “They are not substitutes for each other,” Abbott says. The data also confirmed the previously reported negative correlations between returns and both liquidity and size; that is, the larger market value firms exhibit lower returns, as do the more liquid firms. The ultimate query: Is the difference in returns for different size quintiles (using standard Ibbotson categories) statistically significant?
In a word—yes. “Significant differences exist between observed returns for different size quintiles, and for different liquidity quintiles,” Abbott says. Moreover, “liquidity has a stronger impact than size, significantly so.” Liquidity matters, he says, “much more” than expected. “That is the surprising contribution from the study,” Abbott adds. The practical takeaway for analysts: “They will first make a size adjustment to the rate of return, as they have always done,” he says, “and then they will have to look at liquidity and make an adjustment for that.” Catch both presenters and their complete findings at the ASA’s 2011 Advanced BV Conference, October 10-12 in Chicago.