Trading on Deviations from Put-Call Parity and Mean-Reversion: Evidence based on the American market

Jesper Kjær Bahlke & Natalia Kornela Setlak

Student thesis: Master thesis

Abstract

This thesis builds on information regarding utilizing predictive signals embedded in put-call parity deviations and past stock price paths. Using daily option and stock price data from January 2011 until December 2015 and following Cremers & Weinbaum (2010) and Hong (2013), we set up trading strategies based on the open interest-weighted option-implied volatility spread and price to moving average ratio and subsequently measure portfolio performance using the 3-fama French (1993) factors and the Carhart (1997) momentum factor. It is demonstrated that stocks with relatively expensive calls relative to puts outperform equities characterized by relatively expensive puts relative to calls. A portfolio longing high volatility spread stocks and shorting low volatility spread stocks yielded on average an 4-factor abnormal risk adjusted return of 19 bps (t-stat 2.45) and 73 bps (t-stat 2.79) one and four weeks after the investment. Further, evidence indicates that recent loser stocks outperform recent winner stocks confirming the return reversal phenomenon. A portfolio longing past losers and shorting past winners yields on average an 4-factor abnormal risk adjusted return of 17 bps (t-stat 1.46) and 46 bps (t-stat 1.05) one and four weeks after the investment. A strategy utilizing both predictive measures in conjunction successfully yielded on average an 4-factor abnormal risk adjusted return of 64 bps (t-stat 2.47) and 151 bps (t-stat 2.83) one and four weeks after the investment. Moreover, as the long arm of the hedge portfolio was largely responsible for the alpha returns, doubts that short-sale restrictions are driving profits are evaporated. Fama-Macbeth two-step regression procedure fails to reconfirm the economic utility, which the initial analysis suggested, as none of the factor risk premiums were statistically significant. Adjusting for non-synchronicity bias between option and stock markets significantly distorts the strategies profitability. Single sorted hedge portfolios experienced an overall loss of statistical significance as well as reduced profitability, while the double-sorted hedge portfolio only experienced reduced profitability, but still enjoys 5% significance. Overnight returns from Friday to Monday were thus found to be positive, contrary to findings presented by Harris (1986). Finally, results indicate that predictability is improved at times when option liquidity is high, but provided mixed conclusions for stock liquidity level, hence only partially supporting the model presented by Easley, O’Hara, & Srinivas (1998). Our main contribution to the literature lies in improving the theoretical understanding of asset pricing and examining practical applications of it in the given investment strategies. Results can be generalized to the US equity market, but not necessarily other developed and emerging markets, as the US market is characterized by high liquidity and market capitalization of companies.

EducationsMSc in Finance and Accounting, (Graduate Programme) Final ThesisMSc in Applied Economics and Finance, (Graduate Programme) Final Thesis
LanguageEnglish
Publication date2017
Number of pages88