This article evaluates the tracking efficiency of physical and synthetic Exchange Traded Funds (ETFs). Theoretically, synthetic ETFs should provide greater tracking efficiency. We study this using ETFs domiciled in Europe tracking the European, United States and Emerging Markets. We calculate the tracking error differences and use univariate analysis across both replication and regional strategies. Our contribution is in identifying the variables that influence ETF tracking efficiency. We find the total expense ratio remains the pre-eminent explanatory variable, but the type of replication strategy is a significant secondary factor. We observe that the counterparty risk in synthetic strategies is not rewarded relative to physical replication strategies, but the synthetic ETFs increase returns of tangency portfolios compared to physical ETFs when analyzed using mean-variance spanning techniques. This suggests investment in synthetic strategies is suited to risk tolerant investors and investment in physical strategies is more suited for risk averse investors. TOPICS:Exchange-traded funds and applications, analysis of individual factors/risk premia, VAR and use of alternative risk measures of trading risk, risk management, interest-rate and currency swaps Key Findings • The type of replication strategy employed by an ETF is a significant secondary factor in its tracking efficiency. • Synthetic ETFs’ replication increases returns of tangency portfolios compared to physical ETFs when analyzed using mean-variance spanning techniques. • Investment in synthetic strategies is suited to risk tolerant investors, while investment in physical strategies is more suited for risk averse investors.
Read full abstract