A News-Based Model for Pricing Equity Exchange-Traded Fund Holdings

Abstract

A news-based model (NBM) in which stock prices are determined by three types of news is proposed. The first type is nondiversifiable macroeconomic and geopolitical news. Their impact on prices is accounted for using the total market return in the spirit of the capital asset pricing model. The second type is the equity sector/industry news whose impact on prices is modeled with returns of the relevant industry exchange-traded funds (ETFs). Finally, the company-specific news and its momentum are described using an optimized ARMA-GARCH model. It is found that, in some equity sectors, the industry impact on stock returns may exceed the total market impact. A comparison of the accuracy of NBM and the momentum-enhanced five-factor Fama–French model for a representative list of holdings of 10 major US equity sector ETFs demonstrates superiority of the former in most examples.