Stock Prices Matter
49 Pages Posted: 18 Feb 2015 Last revised: 25 Jan 2018
Date Written: January 2018
Explicitly controlling for size, we find that nominal stock prices matter in predicting future returns with high price stocks outperforming low price stocks by an average raw return of 0.51% per month (or 6.29% annually) and by a 5-factor alpha of 0.44% per month (or 5.41% annually) with a significantly higher Sharpe ratio. These results are based on the entire U.S. stock market instead of just low price stocks. Since size and price are positively related (correlation of 0.70), and size and returns are negatively related (Banz, 1981), a control for size is needed in any analysis of price and returns to remove its confounding effect (Berk, 1995). In the absence of a control for size, there is no statistically significant difference in returns for stocks at varying price levels.
Furthermore, returns to high price stocks are less sensitive to market movements and investor sentiment. These results are robust to samples constructed with different screens: price cut-off of $5, sub-periods, and arbitrage-related constraints imposed by low institutional ownership, high idiosyncratic volatility, high illiquidity, and high idiosyncratic skewness. Similar results are obtained in longitudinal settings following stock splits and earnings announcements.
Keywords: investor behavior, nominal prices, market capitalization, idiosyncratic volatility, institutional ownership, sentiment
JEL Classification: G11, G14
Suggested Citation: Suggested Citation