Examples of Style-Based and Style-Specific Investment Performance Attribution
Ron Surz December, 1999
949-488-8277

In the November, 1999 issue of Senior Consultant I described a new attribution approach that separates style effects from skill in sector and stock selection. Until now, attribution analyses had used economic sectors rather than styles. In this follow up article I show examples of style-based and style-specific attributions and contrast these to the old-fashioned sector-based approach.

Exhibit 1

Exhibit 1 shows 3 components of return for each of the 3 attribution approaches. The leftmost panel is a summary of a 3rd quarter, 1999 style-based attribution. The middle panel is a summary for the same portfolio and time period using sector-based attribution. The rightmost panel uses style-specific attribution. Style-based and sector-based attributions break the portfolio into styles or sectors, respectively, and then within these segments evaluate stock selection. The performance effect of the style or sector bets is called the ÒAllocationÓ effect. The contribution of individual stock selection is called ÒSelectionÓ. Both Allocation and Selection are measured relative to the market, so the 3 components of return are Market, Allocation and Selection. The style-specific attribution shown on the far right is a little different in that the market is defined to be a style-specific subset of the broad market, and then the portfolio is segmented by economic sector within this custom ÒmarketÓ, or benchmark. As you can see the Market effect is the same Ð6.4% for both the style-based and stock-based attributions because the broad market is used in both cases. By contrast, the style-specific Market, which really is a custom benchmark, is down only 4.3% implying that this fund's style was in favor for the quarter. Indeed, we see the style-based Allocation effect is a positive 1.9% meaning this fund's style allocation added 1.9% to performance in the quarter. Furthermore, stock Selection added another 4%, so the total buy-and-hold return was Ð6.4% Market + 1.9% Allocation + 4% Selection = -.5%, which is the solid red line shown running across the exhibit. If we use sector-based attribution instead we would attribute most of the value added, 5.2%, to stock Selection. As stated in the article, both the style-based and the sector-based attributions can lead to faulty inferences. In this case it's the inference that Allocation contributed positively to performance. The style-specific attribution shows that Allocation actually subtracted 1.8% from performance and that stock Selection is what won the day. How can these analyses be so different?

Exhibit 2

Let's look at the details of the style-based and the style-specific attributions. We'll skip the sector-based attribution because it's similar to the style-based, and less meaningful because this fund is managed to a style not a sector. Exhibit 2 shows the style-based analysis. The shaded blue area shows the fund's commitment to each style, which contrasts to the market's commitment shown as the solid red line. As can be seen this is predominantly a large cap growth portfolio. The floating bars show the range of results that could have been earned in each style, with the median also shown below as the ÒMarket ReturnÓ. Large cap growth was in favor for this quarter, losing only 2.5% when the broad market was down 6.4%, hence the style Allocation is positive, as shown in exhibit 1. Also, this fund has above average return in every style, so stock Selection is positive as well, actually contributing more to performance than style Allocation.

Exhibit 3

Exhibit 3 uses a similar format to analyze performance within the style-specific market indicated at the bottom row of the exhibit. Here we see that the fund was over-weighted to Finance with a 21.03% allocation, versus the style-specific market's 7.29% weight. This over-weighting subtracted significant value because Finance was by far the worst performing sector in the quarter. Offsetting this deleterious effect is very good stock selection in Finance, and all of the other sectors. Stock Selection really did win the day. You may have figured out that the reason for the superior selection in Finance is that the stocks selected for this sector were not large cap growth stocks. This is easily seen in the Appendix to the analysis where characteristics are shown for each stock, portfolio segment and the market.

This may seem complex, but once you've used style-specific attribution for awhile you'll see just how easy it is to use, and how much better the insights are compared to those of the old-fashioned sector-based approach.