Explaining Changing “Investor Exuberance"

Abstract

We analyze the determinants of changes in S&P Composite Index Total Return Cyclically Adjusted Price-to-Earnings ratio (TR CAPE), to better understand changing “investor exuberance”. We use three different methods - linear regression using PCA, Lasso, and Ridge regression techniques, as well as ElasticNet method – and a large number of explanatory variables, to explain changing investor exuberance. Across all methods, we find that monthly changes in Michigan sentiment index is significantly associated with monthly changes in TR CAPE. When we crosscheck the results using annual changes (rather than monthly changes), across all methods, annual changes in Michigan sentiment index and changes in core inflation are significantly associated with annual changes in TR CAPE.

Keywords: TR CAPE, Cyclically Adjusted Price-to-Earnings ratio, investor exuberance, PCA, Principal Components Analysis, Lasso regression, Ridge regression, determinants, stock market variables, economy-wide variables, fixed income variables, commodity variables