Date of Award

Summer 2010

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Program/Concentration

Business Administration-Finance

Committee Director

Licheng Sun

Committee Member

Mohammad Najand

Committee Member

David Selover

Abstract

This dissertation, by employing different trading strategies, addresses the trading profitability issue in a broad scope of different markets.

In the equity market, I construct a group of BUY-SELL portfolios based on prior stock returns, and find that contrarian and momentum strategies are both significantly profitable. Investor sentiment, in addition to firm-specific risks, provides behavioral explanations to the profitability. Three popular sentiment measures are used for the purpose of study: two reduced-formed sentiment indexes that are constructed by Baker and Wurgler (2006) and the survey-based University of Michigan Consumer Sentiment Index. Several interesting findings are revealed: (1) extreme sentiment levels (either optimistic or pessimistic) tend to be followed by higher contrarian profits; (2) momentum profits appear to be negatively related with the lagged average 6-month sentiment levels; (3) former loser stocks are more important in determining the average contrarian profits, while momentum profits largely result from former winner stocks. The results are robust for all three sentiment proxies, and are consistent with the core implications of behavioral models. Specifically, contrarian profits are consistent with the overreaction hypothesis, and momentum profits can be explained by investor overconfidence and self-attribution.

In the foreign exchange market, I employ a different Weighted Relative Strength Strategy (a.k.a. WRSS). The WRSS strategies uncover similar profitability: eighteen among sixty-four basic strategies generate significant trading profits, and all of them are momentum. Contrarian profits mostly emerge in the second subperiod from 1999–2007, but none of them is statistically significant. Due to the difficulty of generalizing investor sentiment in the global context, the underlying autocorrelation structure of currency returns and the cross sectional dispersion in mean returns of individual currencies are responsible for the abnormal returns. It is found that the time serial predictability plays a critical role in determining trading profits and accounting for market inefficiency. The profits remain significant even when transaction costs come into effect.

DOI

10.25777/31jy-7873

ISBN

9781124291482

Share

COinS