Intrinsic bubbles and Granger causality in the S&P 500: Evidence from long-term data

Results of research on whether changes in earnings can predict future stock returns are inconclusive. We add to this debate by using long-term data from 1871 to 2004 to examine the predictive power of changes in earnings in periods of intrinsic bubbles and in periods absent intrinsic bubbles. Our results show that accounting for bubbles is important in whether changes in earnings can predict future stock returns. In periods of no bubble, we find that changes in earnings Granger-cause future returns, whereas in periods of bubble, this Granger causality from changes in earnings to future returns cannot be found. We conclude that changes in earnings can predict future stock returns, but only in periods absent bubbles.