If noise is a significant component of stock prices, does a portfolio that favors large market capitalization stocks automatically underperform? In the May 2006 draft of their paper entitled “Pricing Noise, Rejecting the CAPM and the Size and Value Effects”, Robert Arnott and Jason Hsu examine the implications of a very simple model that assumes stock prices deviate from fundamental value based on a single source of unknown risk (noise). They assume the deviations revert to a mean of zero, with no long-term effect on stock returns. Based on this model, they conclude that:
- A market capitalization-weighted stock portfolio generally underperforms similar unweighted portfolios because it overweights stocks likely to be overvalued and underweights stocks likely to be undervalued.
- Large capitalization stocks tend to underperform small capitalization stocks, and high price-to-book (growth) stocks tend to underperform low price-to-book (value) stocks.
- The size and value premiums derive entirely from equity market pricing noise (not specific risk factors). This same noise effect also drives mean reversion of stock prices and contrarian excess profits.
In summary, small capitalization value investing works by systematically buying negative noise and selling positive noise.
