Curriculum·S211 Portfolio Analytics and Performance Measurement·about 33 min
Attribution
By the end of this lesson you can
- →Decompose a result into benchmark, allocation, timing, selection, cost and tax lines
- →Compute how many independent observations your trading history actually contains
- →Distinguish the attribution lines you can estimate from the ones you only appear to
- →State the conclusion your sample supports, rather than the one it suggests
Sophomore · enrolled learners
This lesson opens with Luck versus skill, still unresolved after twelve years.
- What happened
- Fama and French published in the Journal of Finance in 2010, bootstrapping the cross-section of US equity mutual fund returns while retaining the correlation structure, and reported that the aggregate portfolio of actively managed funds is close to the market portfolio while the high costs of active management show up intact as lower returns to investors. Few funds produced benchmark-adjusted expected returns sufficient to cover their costs, and adding expense ratios back raised aggregate alpha to only about 0.1 percentage point a year. Harvey and Liu published in the same journal in 2022, reconciling that work with Kosowski and co-authors, and found that the Fama and French method suffers from an undersampling problem which fails to reject the null of zero alpha even when some funds generate economically large risk-adjusted returns, while the Kosowski method substantially over-rejects even when every fund has zero alpha. They proposed a new bootstrap.
- The decision point
- Two of the most cited researchers in finance, working with thousands of funds and decades of monthly returns, produced a result that a later paper in the same journal showed was an artifact of the sampling method. The unresolved question is not whether active managers add value; it is whether the statistical machinery can detect skill at all at that sample size. Your own record contains a few dozen decisions over twelve months.
What you will be able to answer
- →What are the attribution lines?
- →Which attribution line can you actually estimate?
- →Why is your effective observation count far below your trade count?
- →What conclusion does a year of trading support?
Orientation and Year One are open: anyone can read them without an account. From Year Two onward the lessons are for enrolled learners, because progress through the later years only means anything if it is tracked against a record.
It is free. We do not sell the list and there is nothing to buy at the end of it.
Sources and review
- https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1540-6261.2010.01598.x
- https://onlinelibrary.wiley.com/doi/10.1111/jofi.13123
- https://www.spglobal.com/spdji/en/research-insights/spiva/
- https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2308659
Confidence high·Volatility low·Reviewed 2026-08-06·Owner unassigned
Contested
This lesson deliberately presents a live methodological dispute rather than a settled finding, because the dispute is the teaching point. Harvey and Liu do not show that active managers add value; they show that a widely cited method could not have detected it either way.
The effective observation count formula used here assumes a constant average pairwise correlation, which is a simplification. Correlations in this asset class rise during stress, per S209, so the effective count is smallest exactly when the observations matter most.