Not logged in.
Quick Search - Contribution
|Title||Explaining Stock-Market Factor Rotation|
|Institution||University of Zurich|
|Faculty||Faculty of Business, Economics and Informatics|
|Number of Pages||90|
|Zusammenfassung||This thesis examines the returns of the most common equity factor strategies and tries to explain, based on the evolutionary finance model by Evstigneev et al. (2016), the time-series variability in their risk premia. Despite the increasing interest in so-called smart beta strategies and the growing number of suggested factors, only few academics examine the impact of capital transfers to factor strategies (Harvey et al. (2016), Cong and Xu (2016), Ratcliffe et al. (2017)). Similar to prior studies by, amongst others, Arshanapalli et al. (1998), Ahmed et al. (2002), Amenc et al. (2003) and Asness et al. (2013) I show that factor returns are time-varying and thus diversifying across several factors or timing individual factor exposures could enhance portfolio returns. Furthermore, I present evidence for momentum in factor returns over the next one to twelve months which is consistent with Ehsani and Linnainmaa (2017), Arnott et al. (2018) and Gupta and Kelly (2018) who document strong and significant profitability of time-series and cross-sectional factor momentum strategies. However, most of the factor rotation strategies neglect potential offsetting effects of the capacity and cross-impact of the underlying factors (Doskov et al. (2018)). In this thesis, I present a new approach how to endogenize the population weight in the context of the evolutionary finance model by Evstigneev et al. (2016) by aggregating market value-weighted factor betas across stocks. The population weight can be regarded as a proxy of total capital or share of the total market capitalization allocated to a specific strategy. I find that population weights are time-varying but less than the corresponding factor risk premia. Furthermore, my results demonstrate that future population weight changes can be predicted by their past changes and hence suggesting the presence of momentum in population weights. Lastly, I detect a negative relationship between an increase in a factor's population weight and its corresponding risk premia. These findings are mostly consistent with Doskov et al. (2018). Finally, I demonstrate several out-of-sample time-series and cross-sectional multi-factor strategies which are based on the findings of momentum in population weights and the impact matrix by Doskov et al. (2018). I show that the time-series and cross-sectional strategy (long-only as well as long-short) generate attractive risk-adjusted returns and statistically significant alphas relative to a passive equally-weighted strategy over various holding and look-back periods.|