Non-Linearity in the Dividend Yield: A Comparison of the US and Japan

Abstract

In order to examine non-linear predictability of the US and Japanese dividend-yield ratio, smooth transition regression model analysis is applied to an extended time period of data. The theoretical basis for investigating non-linear behaviour in stock returns can be based on the interaction between noise traders and arbitrageurs or behavioural finance theories of non-linear risk aversion. Our findings support non-linearity in the US and Japanese dividend yield that might be linked to differences in the market structure of the Japanese stock market compared to the US. Specifically, there is evidence of an inner momentum and an outer mean reversion regime in both countries. However, the momentum regime appears to be larger in the US compared to Japan.