rolling-window method (Hardy[2003]) and style drift score(Idzorek and Bertsch [2004]) based on the return-based model (see Sharpe[1992]) which only requires a time-series of historical fund returns and is usually easy to get comparing to other models such as characteristic-based model(Daniel, Grinblatt, Titman and d aWermers[1997]). Because up-to-date holdings of mutual funds are often no available, characteristic-based style analysis leads to poor information. This makes return-based style analysis the more popular approach to pursue.
First, to examine the application with respect to the fund
miss-specification phenomenon, we consider 30 domestic-equity mutual funds that their investing periods exceed 10 years and using rolling window
regression. In the second part, we calculate style drift score and drawing the allocated weights maps over time. Finally, we conclude.
2. Literature Review
The number of mutual funds has increased fast over the last decade.
Despite the fact that general performance of mutual funds managers have poor performance, investors have increased their demand for investment management; see Gruber (1996). Different needs of investors are reflected by in the different investment objectives. In the foreign countries, a large part of the funds describe its style rather properly through its name. However, in Taiwan, a substantial part of funds have misleading names, obscure investment objectives, or pursue a different style than advertised. A tool introduced by Sharpe(1992) might be appropriate to obtain a first insight of historical exposures of mutual funds. Sharpe(1992) contributes a great model
‧
that investors can utilize return-based style analysis to track their portfolios at relatively low cost. Money managers are increasingly evaluated relative to a performance specific to their style, such as a growth or a value index. Each index represents a dimension of the behavior of returns corresponding to a particular style. The fund's estimated sensitivities or loadings are taken as measures of its style. Alternatively, the characteristics of the stocks held by a fund serve as another indicator of the types of firms in which the fund invests, and hence its style. Grinblatt and Titman (1989) use such an approach to evaluate fund performance. Sharpe’s econometric model involves a constrained regression that uses several asset classes to replicate the
historical return pattern of a portfolio. The constraints are imposed to improve an intuitive interpretation of the coefficients. First, the coefficients as weights within a portfolio the factor loadings are required to add up to one. Second, coefficients should be positive to reflect the short-selling constraint which is suitable to most fund managers.
Sirri &Tufano (1998) indicate that mutual funds with higher rank in the
performance lists of magazines attract more money from the investing public. It might be a lure for mutual manager to deviate his objective and invest in more risky assets. There are abundant evidences of misclassification of mutual funds. Both Brown & Goetzmann (1997) and Dibartolomeo & Witowski (1997) use the realized fund returns as inputs for their analysis. Their results suggest that up to 40% of mutual funds are in one way or another misclassified. Kim, Shukla & Tomas(2000) report classification up to 50% when also taking into account other fund attributes than risk and return measures. These studies do not consider style changes. Of great concern about style consistent, Louis K.C.
Chan, Hsiu-Lang Chen and Josef Lakonishok (1999) finds the results that
‧
funds applying consistent styles over time might outperform the funds with inconsistent investing style, and funds with poor past performances are more likely to change styles. In particular, in some situations a manager may have an incentive to deviate from his declared style, in hopes of recovering from past losses or simply to follow the crowd and adopt whichever style has been successful.
Brown and Harlow (2004) find that style consistent managers are less likely to make asset allocation errors than those that try to time the market. There is some evidence to suggest consistency is a more valuable talent within some style classes (e.g., large- and small-cap) than others (e.g., mid-cap). Also, although their results do not negate the possibility that managers who follow an explicit tactical style timing strategy can be successful, they do suggest that unintentional style drift can lead to inferior relative performance; indeed, the decision to remain style consistent may be more useful in helping managers avoid consistently poor performance than creating an environment that fosters persistent superior relative returns.
Buetow, Robert, Johnson and Runkle(2000) utilize historical returns to obtain the results that Growth Equity Fund and Aggressive Equity Fund significantly have dynamic drift than Balanced Funds, Asset Allocation fund, Growth and Income Fund, and Index Fund.
Tracking error relative to a market benchmark can be a reasonable measure of style consistency (Seigel 2003) ,however, it doesn’t measure the style drift of a manager directly. Tracking error is an evolution of the asset class coefficients over time that indicates style drift indirectly. Idzorek and
Bertsch(2004) provide a critique of tracking error and style benchmark
turnover as a measure of style rotation and propose a new statistic method to
‧
measure style drift, the style drift score, which measures the variability of style through time. The main advantage of the style drift score is that it makes an easier evaluation through numerous rolling asset allocation graphs
unnecessary by providing additional complicated information.
Holmes and Faff (2007) find there is some evidence that SDS is related to fund performance. In particular, when conditional performance models are used, style drift and selectivity skill are positively related, indicating that managers that are more successful at stock selection tend to be less consistent with respect to style.
Brown and Harlow (2002) demonstrate the results that more style-consistent funds to produce higher total and relative returns than less consistent funds, after controlling for past performance and portfolio turnover. These findings are robust across fund investment style classification, the return measurement period, and the model used to calculate expected returns.
Finally, Brown and Harlow (2009) also conclude that deciding to maintain a consistent investment style is an important aspect of the portfolio
management process.