| Summary: | 碩士 === 義守大學 === 財務金融學系碩士班 === 95 === It’s a regret that most mutual funds underperform their respective benchmarks. For decades, scholars have provided empirical evidences, indicating that managers on average could not beat their benchmarks. Significantly, Roll (1992) argued that, since investors pay for sound performance, the least managers can do is earning returns in excess of their respective benchmarks to justify the fees collected from investors.
Jorion (2003) proposed constructing portfolios with constrained tracking error volatility and portfolio variance. We adopt this approach and analyze the performance of portfolios, so constructed, under different market conditions.
Choosing Dow Jones Industrial Average (DJIA) as the benchmark, we construct our initial portfolios based on the daily trading data of index constituents from year 1996 to 1997. Taking trading costs into account, we set up portfolios with five different levels of tracking error volatilities to compare their performances. We then observe the performance of these portfolios through the period of 1998 to 2006. The managing portfolios are actively rebalanced only when the benchmark index is reconstituted. Furthermore, the benchmark samples are categorized as bullish or bearish, in order to observe the wax and wane of portfolios performance against DJIA.
Based on their annualized compounded rates of return, we find that these tracking error portfolios significantly outperform DJIA under almost all market cycles, except when the market was bearish. However, when we limit the portfolio’s tracking error volatility to within 1% of the benchmark, the extent of underperformance may be tightly controlled to within the same amount.
|
|---|
