Dominance of External Features in Stock Price Prediction in a Predictable Macroeconomic Environment

Abstract

Understanding the factors affecting future stock prices has been of prime importance across the globe, as accurate stock price prediction is directly related to financial gains. Its interest has been reflected by a large and growing literature trying to investigate stock price prediction with an effort to gain higher prediction accuracy. Recent literature has identified relevant external features, such as current and anticipated future macroeconomic environment-related information, and has incorporated such external features along with historical data on stock prices into the prediction models to gain improved accuracy. However, the current literature fails to quantify the relative importance of those external features for a better understanding of their relevancy. In this article, we bridge this gap and quantify the relative importance of those external features in stock price prediction by combining macroeconomic data with historical stock price data and by utilizing dominance analysis. Our results demonstrate that external features are highly dominant in the prediction of future stock prices.