Irrational exuberance reconsidered. The cross section of stock returns. (Q2386497)

For a review of the 1st ed. see Zbl 1050.91045. The present monograph investigates the so-called Winner-Loser Effect (WLE) and the questions, whether it may occur in rational pricing theory or is due to irrational behavior (new paradigm): stocks which outperformed the market in a formation period (''winner''-stocks), underperform in the following test period (and vice versa, ''loser''-stocks underperformed the market during the formation period, but outperform the ''winner''-stocks in the test period). It appears that past information influences the future markets. Therefore the fundamental valuation of stock returns (as cross section) is investigated and compared for different models: on the one hand, the traditional Capital Asset Pricing Model (CAPM) beta, on the other hand, the predictive power of earnings. However, the ultimative conclusion on the market's efficiency remains unsolved, also in this monograph. The book has been structured in four main parts. In part one, the start is laid ''unorthodoxly'' quoting an interview of the author with two practitioners (one fund manager and an equity strategist) to extract the importance of fundamental valuation, risk premium estimation, behavioral factors and corporate control. The second part gives empirical evidence for the overshooting of stock returns and the long-term reversals (WLE) within the existing literature. There explanations are given in terms of explanatory variables or change of exposure to risk, which were so far either to small to account for the total observed excess return or evidence was very weak. Furthermore, the author explores and develops a new resort of evidence within the cross sectional returns of German stocks (within DAFOX) based on the years 1968--1986. In the following third part, the author gives a theoretical framework, which can explain the WLE within rational pricing and gives empirical evidence for a relation to model parameters: (i) If the WLE occurs in rational asset pricing, then it must be related to a change in systematic risk (beta parameter of CAPM). (ii) It is deduced that fundamentals such as earnings and their changes have explanatory power for the returns (some surprise effect related to some news is thus related to fundamentals). (iii) Fundamentals can also account for changes in the exposure to systematic risk (and thus be explained without irrational behavior). In the statistical test units of part three, it is revealed that the CAPM model offers explanations for at most small parts of the observed excess returns in the DAFOX samples, too. In the relation to the fundamentals, evidence for a parallel movement of fundamentals and stock returns is found (``by and large''): consistent but small for dividends; in terms of profits, only for long-term reversals and not for short-term persistence (formation period). Also supportive evidence is found that changes in earnings and dividends account for a change in systematic risk. However, this is based on an indirect analysis, and therefore it is not superior to rational interpretation. In addition, a comparison of fundamentals and rational pricing parameters (beta, size) completes the third part. Overall a third of the variation in stock returns could be explained (in an ex post regression), whereof one third is related to the pricing parameters and two thirds are related to fundamentals. The last part of this monograph relates observed reversals to temporary problems in corporate control (loose and tight control). This hypothesis is newly introduced by the author, and yields mixed results depending on the respective parameters in view. In my view, this book on finance has impact on the modeling with some strong parts. However, with the eyes of a mathematician, the evidence in terms of statistical tests, is not strong enough for decisive modeling conclusions. It seems there are in the models too few variables and too few `independent' model components to explain complex behavior without relating everything to everything somehow. Also the author demands for a multivariable analysis as further research and anticipates challenges in the setup of some unifying model. Especially the last part seems to me highly speculative. Furthermore, the introductory interview is from my understanding nice, pushing the general line of investigation; there could be more links drawn between the presented research and the practitioners view in the interview. The demand for more complex models could be indeed some outlook. Last, from the reading perspective, each unit within the monograph has an own introduction, outline and summary and can be read independently, which adds a good deal of repetition. In conclusion, an interesting start for further research.