No abstract

No abstract

There is considerable evidence showing that both mean reversion and momentum exist in stock prices, especially in financially-developed countries. We analyze these phenomena for two Central and Eastern European countries with very different transitions from centrally-planned to market economies: Poland and Romania. Although Poland’s stock market cannot be considered well-developed, its capitalization increased from 3 percent of GDP in 1995 to about 30 percent in 2017, while Romania’s stayed under 6 percent of GDP in the 1990s and early 2000s, and only recently has increased to about 21 percent. Examining how mean reversion and momentum affect stock prices, we find very similar results for the two countries. The speed at which stocks converge back to their fundamentals (i.e., mean reversion) is much faster than that of the developed markets, with half-lives of just over 9 months for both countries (similar to the results obtained in the literature for the Chinese market, but much shorter than the 3-4 years for  the well-developed markets). We also find that the momentum effect lasts less than in the developed countries. Therefore, in most cases, strategies combining mean reversion and momentum generate abnormal excess returns only for holding periods of less than 12 months.

This paper analyzes the components of the influential Baker and Wurgler (BW) indices of investor sentiment and their contribution to the volatility of U.S. stock portfolios. Using the Klein -Chow (2013) symmetric method of variance decomposition, we orthogonalize the five proxies of sentiment (dividend premium, first-day return on IPOs, number of IPOs, closed-end fund discount, and equity share in new issues) and determine the proportions of systematic risk contributed by these measures. We find that even portfolios that show no economically significant contemporaneous or lagged sensitivity to the BW indicators may still exhibit important sentiment risk, which raises serious questions about the validity of some of the proxies or the use of Principal Component Analysis to compute the sentiment indices. For instance, after controlling for the Fama -French three factors and for momentum, for the time interval 07/1965 -12/2014, the first-day return on IPOs explains roughly 9.8 and 19.2 percent of the systematic variance of high 10 -low 10 (long-short) Fama -French portfolios formed on variance and on accruals, respectively. For the same interval and portfolios, the overall BW sentiment-levels indicator explains, respectively, 1.7 and 0.0 percent of the systematic risk.