| dc.description.abstract | The purpose of this thesis is to investigate the effect of arbitrage activity on abnormal trading
profits based on the new measures of arbitrage proposed by Lou and Polk (2013) and Huang,
Lou and Polk (2014), called Comom and Cobar, respectively. First, I replicate the process of
Comom and Cobar construction and conduct an additional analysis of their specifications. I also
create a combined measure Comom/Cobar that measures arbitrage in both strategies
simultaneously. Second, I examine patterns of abnormal returns in momentum and beta strategy
conditional on the computed arbitrage measures. The study is conducted over the period January
1970 – December 2011.
The results of this paper indicate that such parameters as asset-pricing model and inclusion of
stocks below $5 into the sample do not affect the time series of the arbitrage measures, whereas
the choice of decile may significantly change the outcome. Consequently, I suggest using the
lowest decile for Comom and Cobar computation to avoid unrelated return comovements that
may arise in the highest deciles. I also find that Cobar and Comom cannot substitute each other
when used for abnormal return evaluation. After estimating abnormal returns through
constructed measures, I find that the effect of arbitrage activity does not create common patterns
in abnormal returns across beta and momentum strategies but rather produces specific price
reactions in each strategy. | nb_NO |
