Examining the relationship between time and factor anomaly returns

• Main Author: Kokkonen, Joni
• Publication Date: 2019
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10400.14/29043
• topic: Rebalancing; Factor investing; Anomalies; Abnormal returns; Market efficiency; Rebalanceamento; Investimento factores; Anomalias; Retornos anormais; Eficiência de mercado

By examining the return characteristics of 10 different, long minus short, accounting-based anomaly portfolios, we find that a portfolio that rebalances daily, based on either 10-K or 10-Q releases, is optimal when transaction costs are low (around 2% or less) and the anomalies are robust, with a 4.5% annualized return differential being created versus a yearly rebalance. However, when that is not the case, we find that a yearly rebalancing is best. Furthermore, we determine that no other common rebalancing periods (such as weekly) should be used as they consistently fail to outperform either the daily or yearly ones. We also find a significant clustering of anomaly returns around the release of 10-K and 10-Q financial statements, primarily up to 15 days, corroborating theories connecting news releases to anomaly returns. Nonetheless, we find anomaly returns to still be consistently present up to 365 days after a financial statement release, opposing the idea of fast market responses to information and large clustering. Overall, there appears to be a connection between anomaly returns and time considerations that warrants further study.