Market Efficiency and Return Predictability: A Dynamic Perspective

We view the state of aggregate equity market efficiency as an unobserved, time-varying variable, and propose to use relative predictive gains as its proxy. Given the difficulties in meaningfully forecasting the equity premium in the presence of structural breaks and instabilities, we employ the novel methodology of robust forecast combination to obtain predictive gains, thus dynamically tracking the changing magnitude of market efficiency. The robust combinations alleviate the impact of over-penalizing an otherwise outperforming model for the occurrence of outliers owing to instability, thus providing a theoretical foundation for the benefits of combining forecasts in unstable environments. Our empirical results reveal an increasing degree of equity market efficiency, particularly since the 2000s. Attempting to explain the elusive nature of return predictability and rising market efficiency, we explore the impact of events such as the dot-com bubble, the Sarbanes-Oxley Act, and the advent of new information-sharing technology.