Regular readers will know that we often measure the influence of certain fundamental and macro
factors on stocks. The way we do this is to measure the correlation between the factor in question (with the individual readings broken into deciles) and the market performance over various time periods.

Over the last three months the most significant factors driving stock returns has been the one and three month change in EPS estimates. We’ve written a lot recently about earnings estimates and how the significant move we’ve seen in expectations has driven stock prices. Here is more evidence that changes in earnings estimates have in fact dominated investors votes.

The first table below shows the most significant factors sorted from highest to lowest over the previous three months. Since we are measuring r-squared in the first table – which only measures the magnitude of the explanatory power of the independent variable, not the direction – we show the correlation breakout for both the one month (2nd table) and three month (3rd table) change in EPS estimates.

From the second and third tables below we can see that companies in the top decile of EPS revisions outperformed companies in the 10th decile by a huge margin. We take this to imply that stock price performance is becoming much more sensitive to changes in bottom line growth than what we’ve seen previously in this cycle.  


In the last table we show the change in EPS estimates by industry group. All but three industry groups have experienced negative EPS revisions over the last three months. The energy area is the standout, but weakness has been widespread.