 |  | |||||
| ||||||
 |  | |||||
| | |||||
| ||||||
| | |||||
 |
Style Primer
The Background of Style Analysis In the late 1980's Nobel Laureate William F. Sharpe introduced an analytical process for determining the composition of an investment portfolio solely based on knowing the performance history of the portfolio. The result of the analysis was expressed in terms of an "effective mix" - a combination of market indices whose return most closely approximated that of the portfolio. Effective mix tells investors objectively how an analyzed portfolio behaved in the past and how it is expected to behave in the future. To estimate effective mix, Sharpe used a mathematical process called quadratic optimization. His process made it possible to look at the return history of a portfolio, like a pension fund, and attribute its performance to its respective exposures to various domestic and international asset classes as represented by published market indices. The return of the weighted combination of indices represented the fund's market derived (systematic, or later also called style) return, and the portion of the fund's return not explained by the estimation process was attributed to the manager's selection skill (specific return). When indices became readily accessible that segmented the domestic equity market into value/growth and small cap/large cap spectra, they were used in estimations to characterize portfolios in terms of "style" and the estimation process came to be called returns-based style analysis. Sharpe's idea became very popular by the mid 1990's, partly because of its novelty and the stature of its creator, and importantly because of the success of its implementation in an elegant commercial software application called StyleAdvisor®. The original version of StyleAdvisor, now owned by Zephyr Associates, Inc., was built in 1992 by Michael Markov and Mik Kvitchko of Markov Processes (MPI). In 1996 the development team at MPI renewed its investigation of Sharpe's methodology with the objective of addressing some of its significant shortcomings. Working together, Michael Markov and Victor Zurakhinsky hit upon the idea of "locally weighted regression" or LWR, which was implemented by MPI the next year in a product called mpi Stylus. LWR represented a significant advance in style estimation technology. It overcame the sizable lag in mix estimation of Sharpe's methodology (dubbed the "rear view mirror effect" by critics) and significantly reduced statistical "noise." Thus the second-generation of returns-based style analysis came into being, and because of its improved accuracy brought with it a number of new uses for the concept, all of which are being employed with success by MPI's institutional clients. Style Metrics Returns-based style analysis has as its output measures that address exposure to classes of assets and the return implications of these measures. Here's how we see these measures in use in mpi Stylus. Asset Loadings (Exposure)
The Asset Loadings bar (or area) chart depicts the
periodic exposure of an investment product to various
asset classes that are represented by proxies or explanatory
indices, providing a quick and lasting impression
of the portfolio manager's movement among areas of
the market. The chart above shows the exposure
dynamic of Janus Fund over the last five years. If
we were to look at the same plot for an index or an
index fund, the lines separating the areas of color
almost would be parallel. Style Map
A Style Map is a scatter plot typically used to show where a product lies on the value-to-growth and small-to-large cap spectrums. In the visuals found in Stylus RIA, you will often see a style map on which four to six style reference indices are plotted, with the value oriented ones on the left and the growth oriented on the right, small cap at the bottom and large cap at the top. In the illustration above you can see the movement of a product (red disks) in the large growth area of the map. The smaller dots represent earlier points in time, the larger more recent. The design gives you a sense of the dynamic of a fund and allows you to form an opinion about its trueness to its style mandate, which is represented in our example by a benchmark index also plotted on the map in blue. Style maps also are used to show the dynamics of global products and work well with fixed-income portfolios too, in which case the axis parameters usually designated are duration and quality. Style Drift
The Style Drift line chart is yet another way to sense the trueness of a fund to its style mandate. The same two data series used to plot exposure coordinates on the style map are plotted linearly on the style drift chart. When the value-to-growth line (in blue) is in the positive (upper) part of the chart, it indicates growth exposure, and when in the lower half, value exposure. Similarly, when the small-to-large (yellow) line is in the upper (positive) part of the chart, it indicates a large cap exposure, and when in the lower part, small cap exposure. If the benchmark used in the study is represented on the chart, it would be by lines that are virtually horizontal. Style Return, Timing Return
Style Return represents the systematic portion of the product's return and is derived from the estimated exposures to the explanatory indices over time. It is depicted in the above illustration along with the total return of the product and the market benchmark return. If the returns of the style index and the benchmark index differ, then the fund manager has structured the fund in a way that is different from the structure of the benchmark index, in other words, has bet against the market. This return difference is the timing component of the fund's excess return and is shown in the lower part of the chart. Timing return can be refined by attributing the effect on total return of each weighting difference between the product and its benchmark index. Selection Return
Selection Return represents the security selection and trading effects of portfolio management. It is the difference between the return of the style index and the total return of the fund - the part of total return that cannot be attributed to exposure to the explanatory indices. It and timing return are the components of excess return. All of the above measures, when combined with the output of conventional performance and risk analysis, provide a rigorous and convenient way to examine, select and compare investment products. |
 |
|||||||||
 |
|
|
|
||||||
|
|||||||||
