Drill-Down to Security-Level Data in Sector Attribution

Drilling-Down to Security-Level Data in Sector Attribution

This page explains how to provide stock-level details in Sector Attribution.

The idea is best explained by working through an example. This example analyzes a single day of returns for an equity portfolio. The benchmark return was 1.39%, and the portfolio return was 1.71%. Exactly where did the 32 basis points of added value come from?

The attribution analysis decomposes the added value into asset allocation, stock selection, interaction, and transaction costs. It also calculates each of these management effects for each sector. Moreover, it works out the value added for stock selection and transaction costs right down to the stock level.

Table 1 shows the example:

Table 1: Sector Attribution with Stock-level Results for Stock Selection and Transaction Costs
	Benchmark Weight	Portfolio Weight	Benchmark Return	Portfolio Return	Asset Allocation	Stock Selection	Interaction	Transaction Costs	Total Value Added
Total	100.0%	100.0%	1.39%	1.71%	16 bps	26 bps	-3 bps	-6 bps	32 bps
Large Value	40%	33%	0.48%	0.88%	6 bps	21 bps	-4 bps	-4 bps	20 bps
Stock 1	8%	2%	0%	0%		3 bps		-
Stock 2	5%	3%	-1%	-1%		2 bps		-
Stock 3	9%	8%	1%	1%		0 bps		-
Stock 4	6%	8%	2%	1.5%		6 bps		-4 bps
Stock 5	7%	4%	-1%	-1%		3 bps		-
Stock 6	5%	8%	2%	2%		7 bps		-
Large Growth	40%	44%	1.50%	1.73%	0 bps	9 bps	1 bps	-	10 bps
Stock 7	5%	9%	4%	4%		8 bps		-
Stock 8	9%	3%	2%	2%		-3 bps		-
Stock 9	8%	6%	0%	0%		4 bps		-
Stock 10	7%	10%	1%	1%		-1 bps		-
Stock 11	4%	8%	2%	2%		2 bps		-
Stock 12	7%	8%	1%	1%		0 bps		-
Small Value	10%	9%	2.20%	2.00%	-1 bps	-2 bps	0 bps	-	-3 bps
Stock 13	1%	2%	2%	2%		0 bps		-
Stock 14	1%	0%	-1%	-1%		3 bps		-
Stock 15	1%	0%	3%	3%		-1 bps		-
Stock 16	2%	3%	2%	2%		0 bps		-
Stock 17	2%	0%	4%	4%		-4 bps		-
Stock 18	3%	4%	2%	2%		0 bps		-
Small Growth	10%	14%	3.80%	3.43%	10 bps	-2 bps	-1 bps	-2 bps	4 bps
Stock 19	3%	5%	3%	3%		0 bps		-
Stock 20	1%	2%	2%	2%		-1 bps		-
Stock 21	1%	0%	4%	4%		0 bps		-
Stock 22	2%	4%	6%	5.5%		2 bps		-2 bps
Stock 23	1%	2%	1%	1%		-1 bps		-
Stock 24	2%	1%	5%	5%		-2 bps		-

For a spreadsheet containing all the data and formulae, see AttribSectorDrillDown.xls. Later on this page, we provide a commentary on how the spreadsheet works.

The portfolio has been attributed against a sector scheme that divides the securities into four categories:

large-cap value;
large-cap growth;
small-cap value; and
small-cap growth.

The large-cap sectors each has a benchmark weight of 40%, while the small-cap sectors each have a benchmark weight of 10%.

Overview of the Attribution Results

The portfolio's asset allocation was tilted somewhat toward small growth stocks. As it turned out, this was exactly the right asset allocation on this particular day. The underweight of large value stocks added 6 basis points to performance, while the overweight in small growth stocks added 10 basis points.

Stock selection added 21 basis points altogether. Stock selection in the large cap stocks was particularly strong, while stock selection in the small cap stocks subtracted a small amount of value.

The interaction term was slightly negative, because the sectors where stock selection added the most value were also sectors that had been underweighted.

Transaction costs subtracted 4 basis points from performance. This was due entirely to transactions in Stock 4 and Stock 22 (the only stocks that the portfolio traded on this day).

Introducing the "Drill-down" Idea

It is very rare to see performance attribution reports that only calculate answers at the total level (i.e. the portfolio level). Everybody seems to agree that it's useful to have sector-level results so that one can see in more detail where the active return comes from.

It is a logical extension of this idea that one would like to see attribution results at the security level where possible.