Trend Is Your Friend, or Is It: Empirical Evidence from 60-Year US Stock Market Data

Abstract

The assumption of normality for the distribution of daily stock price returns is the core of the modern portfolio theory. This allows the correct estimation of Value at Risk. This paper performs a long term normality test for the market and finds that stock market returns are not in fact normally distributed, supported also by previous literature. We then analyze whether the distribution of daily returns for some period allows us to make predictions for a trend continuation by testing a simple strategy based on the distribution of previous daily returns. We find that by taking into account the previous trend one can achieve a much better risk-reward ratio while investing. This leads us to conclude that trends have much bigger importance in the market than it is currently recognized.