A Variance Decomposition for Stock Returns

Unexpected stock returns are interpreted by breaking them into components attributable to both news about future returns and news about future dividends. The framework for the analysis is a log-linear approximation to the standard present value formula that is tractable even when expected returns vary though time. The analysis also uses a regression of the stock return, measured over a short period, onto variables known in advance. Using monthly data on the New York Stock Exchange value-weighted index over the period 1927-1988, it is estimated that a typical 1% increase in the expected return is associated with a capital loss of 4% or 5%. It is estimated that, over the full sample period, the variance of news of future cash flows accounts for only 1/3 to 1/2 of the variance of unexpected stock returns. The remainder of the stock return variance is due to news about future expected returns. This suggests that an economic explanation of stock market volatility must also explain short-term predictability in returns.