THE JOURNAL OF FINANCE • VOL. LXI, NO. 6 • DECEMBER 2006 Information, Trading, and Volatility: Evidence from Weather-Sensitive Markets

We find that trading- versus nontrading-period variance ratios in weather-sensitive markets are lower than those in the equity market and higher than those in the currency market. The variance ratios are also substantially lower during periods of the year when prices are most sensitive to the weather. Moreover, the comovement of returns and volatilities for related commodities is stronger during the weather-sensitive season, largely due to stronger comovement during nontrading periods. These results are consistent with a strong link between prices and public information flow and cannot be explained by pricing errors or changes in trading activity. A LARGE SEGMENT OF THE FINANCE LITERATURE investigates the link between information and prices. Theory suggests that prices are a function of public information and order flow (see, e.g., Grossman and Stiglitz (1980), Kyle (1985), and Glosten and Milgrom (1985)). Order flow is driven by both public and private information as well as investor shocks, which may be either rational (e.g., noninformation-based liquidity trades) or irrational (e.g., trades based on noise as described by Black (1986)). Prices can deviate from fundamental value due to market microstructure, liquidity, and hedging effects, and pricing errors can arise from noise trading and the systematic underreaction or overreaction to information. Much of the empirical literature attempts to discriminate among public information, private information, market frictions, and pricing errors as sources of return volatility. Some researchers argue that observed volatility is much higher than can be explained by fundamentals, by market microstructure o