Font Size: a A A

Essays on fundamental analysis

Posted on:2004-12-27Degree:Ph.DType:Dissertation
University:University of California, BerkeleyCandidate:Briginshaw, John Anthony JeffreyFull Text:PDF
GTID:1466390011977156Subject:Business Administration
Abstract/Summary:
Previous literature on fundamental analysis is reviewed, revealing a conflict between the fundamental analysis school and the efficient markets school. The fundamental analysis school has demonstrated significant abnormal returns to fundamental analysis. The efficient markets school has countered by showing that these returns either do not exist out of sample, by arguing that the returns are rewards to risk, and by arguing that the returns are too large not to have been traded away by market participants.; Fundamental analysis research is vulnerable to data-snooping bias or spurious correlation between valuation indicators and future returns. We classify prior research based on the ordering method by which preferred and less-preferred stocks are identified for portfolio construction. We argue that the research least vulnerable to spurious correlation will be that which attempts to predict intrinsic value, which we term a type 1 ordering method.; The effect of varying parameters within a simple valuation model is examined. The parameters of the valuation model are current residual earnings, current book value, persistence of residual earnings and cost of capital. A residual income valuation framework is used to estimate the intrinsic value of firms, while varying the persistence of residual earnings and cost of capital. These firms are sorted by calculated intrinsic value over market capitalization of equity, and deciles are formed, so that the hedge returns between deciles can be calculated. The results are that abnormal return from a high minus low decile portfolio varies between -2 to +14%. The probability density function of persistence and cost of capital is derived from historical and survey data respectively. This is used to weight the abnormal returns and predict an abnormal return based on the valuer's forward looking view of the parameters. The resulting weighted abnormal return is 8.3%, which is significantly different from the midpoint return (which does not allow for uncertainty of valuation parameters) of 10.4%, showing a valuation parameter bias of 2.1%, which is significant at the 1% level. This indicates that rewards to fundamental analysis are present, but are not as large as have previously been claimed.
Keywords/Search Tags:Fundamental analysis, School
Related items