Blog - Investing Notes

October 16, 2006 - Emergent Size-Value Patterns of Noise?

Are there investing/trading strategies that can turn stock price noise into alpha? More specifically, are there types of stocks for which the noise has a systematic effect on price? In the October 2006 draft of their paper entitled "Does Noise Create the Size and Value Effects?", Robert Arnott, Jason Hsu, Jun Liu and Harry Markowitz model the cross-sectional effects of mean-reverting noise on randomly walking stock values. Noise (for example, from overreacting, informationally challenged and/or liquidity-driven investors/traders) introduces random transients of inefficiency. Based on this model, they conclude that:

• Because they tend to have negative (positive) price noise, stocks with low (high) prices or low (high) price-fundamental ratios are more likely to be undervalued (overvalued).
• Small-capitalization and value stocks tend to have slightly higher betas and idiosyncratic volatilities than do large-capitalization and growth stocks. However, the primary driver of small-value excess returns is noise bias, not riskiness.
• These excess returns may disappear, or they may persist due to practical limits of arbitrage.

In summary, the value premium and the size effect are real manifestations of emergent patterns of noise across large groups of stocks.

For related research, see Blog Synthesis: The Value Premium and Blog Synthesis: The Size Effect. See especially our blog entry of 9/7/06 regarding a similar paper.