New Evidence on the Time-Varying Correlation between Consumption Growth and Stock Returns for the G7 Countries

Campbell (1996) reports that, for most countries, the unconditional correlation between quarterly stock returns and consumption growth is small in magnitude and sometimes even negative. Using a bivariate GARCH framework, we examine whether the conditional correlation between stock returns and consumption is positive, even if the unconditional correlation is not. Consistent with this, we find strong evidence, both for U.S. monthly and most G7 quarterly data, that the conditional correlation between innovations in consumption growth and stock returns is positive and significant. Moreover, for six of the G7 countries, we reject the hypothesis that the correlation is constant. For three of the G7 countries (including the U.S.), the correlation is statistically higher for positive stock return shocks relative to negative stock return shocks. However, the correlation is unaffected by large movements in the stock returns for most countries. Our results support Campbell and Cochrane (1999b), who stress the importance of time-varying conditioning information for explaining asset prices. For policymakers concerned with the effect of the stock market on the real economy, our results suggest that the policy response may need to be stronger than normal when the stock market is performing better than expected. However, extreme market conditions, whether positive or negative, should not have additional effects on policy..