Proponents of rational markets build on a common-sense foundation of reward for risk, with price variability (beta) as the fundamental risk. Since this single source of risk does not predict asset prices very well, rationalists have empirically appended to their models other sources of risk (proxied by size, value and momentum factors) in search of better predictions. Proponents of behavioral finance counter with innate cognitive and emotional biases (irrationality) as causes of rational model failures. Is there a way to prove one of these two views more correct? Should rationalists look for additional risk factors? Does some third perspective offer insight? In their January 2007 preliminary paper entitled “Failure of Asset Pricing Models: Transaction Cost, Irrationality, or Missing Factors” Joon Chae and Cheol-Won Yang tackle these questions. Using monthly stock return data for 700 Korean firms over the period December 1997 to November 2004 (84 months), along with associated measures for both potential degree of trader rationality (sophistication) and transaction costs, they conclude that:
- Both transaction costs and investor irrationality correlate negatively with the pricing forecast errors generated by rational market models.
- Rational models with additional risk factors such as market capitalization, book-to-market ratio, and recent returns (momentum) do not significantly change these correlations.
- Transaction costs and investor irrationality, not missing risk factors, therefore cause the poor performance of rational market pricing models.
- It appears that transaction costs are more important than investor irrationality in explaining rational market model shortcomings.
- Firm size correlates negatively with both transaction costs and investor irrationality, with the former relationship working against exploitation of the size effect.
In summary, the authors find that both market friction and investor irrationality play substantial roles in the pricing of stocks.
Confirmation of an irrationality effect offers support to traders looking for overreactions and underreactions. But, it seems that the only way to beat transaction costs is to be one who collects and not one who pays.
