eThesis - School of Business electronic theses

PURPOSE OF THE STUDY This paper examines the relation of systematic cashflow risk and cross-section of stock returns. Specifically, I study whether the systematic cashflow risk determined using analysts’ earnings forecast revisions, i.e. the earnings beta, explains the possible value and size anomalies, as well as cross-sectional return variation in these value and size portfolios. I replicate the methodology used by Da and Warachka (2009), and apply it in European setting. This is the first study to investigate the systematic cashflow risk and return relation with a European sample. In addition I create earnings beta estimates also for country and industry portfolios, which is unstudied with this cashflow risk measure.

DATA The sample used in this study consists of companies from 15 European stock exchanges between 1990 and 2009, and the total number of firm years amount to 9,463. The analyst earnings forecasts are obtained from Institutional Broker’s Estimate System (I/B/E/S) database, and my sample includes 340,704 forecast observations. The fundamental company data is collected from Thomson One Banker and Datastream databases.

RESULTS My results indicate a significant value effect in Europe during my sample period, but earnings beta estimate is negatively related to the returns. This could be explained by adverse portfolio performance during the sample period. Additionally, I find a reverse size effect that is best explained by the inadequacy of my sample to be used for size comparison, since it is heavily biased towards large and liquid companies. The earnings beta estimate of the size sorted portfolios shows to support the return variation, but the difference in the earnings betas of small and large stocks is not statistically significant and thus no holding conclusions can be made whether it explains the reverse size premium.

The economic importance of my earnings beta estimates seem to be rather insignificant. Due to the inconsistencies in the results above, such as difference in the relation between the earnings beta and return as well as the insignificant difference between the earnings betas of small and large portfolios, the betas explain only 3.2 % of the cross-sectional return variation. Thus, one can conclude that under this dataset, the earnings beta model does not hold in the European setting.