Beta-arbitrage: Institutional Implementation: Size, Time, and Short-Selling

Søren Funch Adamsen & Mette Schmidt

Student thesis: Master thesis

Abstract

It is well-documented that the Capital Asset Pricing Model does not adequately capture the expected return of securities. Empirically, the Security Market Line almost consistently overestimates the performance of high beta securities while it underestimates the performance of low beta securities. The line is too flat, the intercept is too high, and consequently beta may be a valuable tool for arbitrageurs. Using more than 30 million data observations across the US and international markets, in equities and fixed income, we construct 16 betting against beta portfolios and confirm the persistence of this apparent disequilibrium. We investigate the effect that time, size, and short-selling frictions may have on the performance of these portfolios. We find that beta-arbitrage performs well on average, but with significant time-variation of returns. Return variability seems positively correlated with implied leverage, as measured by the beta spread. Next, we find that the adverse effects of short-selling costs are negligible and that longshort is mostly preferable to long-only implementation, gross and net of lending fees. Moreover, in the US sample, alpha contribution is roughly equally distributed across size deciles and we find that excluding the smallest equities has little impact on overall performance of the portfolio. Our findings suggest that beta-arbitrage portfolios are implementable and robust when adjusted for size and shorting frictions and present an attractive investment opportunity for sophisticated institutional investors.

EducationsMSc in Finance and Accounting, (Graduate Programme) Final Thesis
LanguageEnglish
Publication date2015
Number of pages116