The Role of "Other Information" in the Valuation of Foreign Income for U.S. Multinationals

In this paper, we examine investors' valuation of the domestic and foreign components of total earnings after controlling for information beyond current earnings. Our sample consists of U.S. multinationals during the 1985-2002 period. In a prior study, Bodnar and Weintrop (1997) find that investors place a higher weight on foreign earnings than on domestic earnings in valuing securities, and that this finding can be explained in part by the higher growth opportunities in foreign markets. While this explanation is intuitively appealing, other possible explanations include the varying importance of information other than current accounting earnings in pricing securities and the possible misspecification of their model. One potentially important source of other information is information contained in revisions of analysts' forecasts of future (abnormal) earnings and terminal values. Excluding this information from the regression specification potentially leads to a correlated omitted variables problem. In this paper, we use the Liu and Thomas (2000) proxy for "other information,"which is derived from analysts' revisions of near-term and long-term earnings forecasts and discount rate changes. Including the "other information" variable greatly improves the explanatory power of the returns-earnings regression. Consistent with our predictions, we find that the bias resulting from excluding other value-relevant information has a greater effect on foreign earnings than on domestic earnings. Foreign earnings are no longer incrementally value relevant when we control for "other information."