Guidelines for asset pricing research using international equity data from Thomson Reuters Datastream

Abstract

Abstract We provide detailed guidelines and code to derive high quality international equity data from Thomson Reuters Datastream (TDS) data. Our approach increases stock and country coverage (to 91 countries), improves data accuracy, filters problematic data and reduces survivorship bias and data staleness. We validate our approach by demonstrating that our U.S. TDS factors are statistically and economically indistinguishable to standard Fama-French CRSP factors. On the other hand, when we compare our international factors to other publicly available international factors, differences are significant, so we justify and detail every aspect of our proposed guidelines. Our guidelines and accompanying code and data should be especially useful for international research focused on wide coverage, equal weighted portfolios, small stocks and countries with a limited number of stocks and for researchers wishing to analyze the US market with access to only TDS but not CRSP-Compustat data.