Pure Benchmarks · Explainer
Decision-Based Attribution: An Alternative to Brinson Attribution
Short answer
Brinson attribution splits a portfolio’s return against a benchmark into an allocation effect, from weighting groups differently, and a selection effect, from the holdings picked inside each group. Decision-based attribution answers a different question. It freezes the portfolio the moment before each change, runs that untouched version forward next to the real one, and credits or charges the gap to the change itself. Brinson tells you which part of the portfolio drove the result. Decision-based attribution tells you whether each thing you did beat doing nothing.
Brinson attribution is the standard way professionals explain a portfolio’s return. It is rigorous and it answers a specific question well: compared with a benchmark, how much came from being overweight or underweight in a group, and how much came from picking better holdings inside it. What it does not answer is the question most individual investors actually have after making a change. Was that change worth making? Decision-based attribution is built for that question, and it works without a benchmark weight file at all.
Brinson attribution and decision-based attribution side by side
| Brinson attribution | Decision-based attribution | |
|---|---|---|
| Question answered | Which groups and holdings drove the return against a benchmark? | Did each buy and sell beat the portfolio held before it? |
| Baseline | A benchmark index or policy portfolio | Your own portfolio, frozen the moment before the change |
| Unit of analysis | Sector, asset class or security group over a period | One decision, scored from its date forward |
| Output | Allocation, selection and interaction effects | Dollar gain or loss of each change against doing nothing |
| Needs a benchmark weight file | Yes | No |
| Typical user | Asset managers and performance teams | Individual investors and anyone reviewing an advisor’s changes |
Brinson attribution follows the framework published by Brinson, Hood and Beebower in 1986 and the later Brinson-Fachler variant. Decision-based attribution here describes the method Pure Benchmarks uses in Decision Benchmark.
How Brinson attribution works
Brinson attribution compares your portfolio with a benchmark group by group. The allocation effect measures what you earned or lost by holding a different weight in a sector or asset class than the benchmark did. The selection effect measures what you earned or lost by holding different securities inside that group. An interaction term captures the overlap, and the Brinson-Fachler variant adjusts the allocation effect so it is measured relative to the total benchmark return. Every number is relative to the benchmark you chose.
Where Brinson attribution stops being useful to an individual investor
Three things get in the way. First, it needs a benchmark with group weights, which most personal portfolios do not have a natural match for. Second, it explains a period, not a decision: if you sold a fund in March and bought another in June, the effects of both are blended into the period’s allocation and selection numbers. Third, the comparison is always with an index, so it cannot tell you whether the portfolio you gave up would have done better than the one you replaced it with.
How decision-based attribution works
Decision-based attribution starts from the portfolio itself rather than from an index. At the moment of each change, the holdings as they stood are frozen into a buy-and-hold counterfactual, sometimes called a ghost portfolio. That counterfactual is priced every day using the same market data as the real account. The difference between the two, after removing new deposits and withdrawals, is the value of the decision. Each buy and sell carries its own result, and results can be added up across a year or across an advisor’s whole record.
Which one to use
Use Brinson attribution when the question is about style and skill relative to a mandate: did the manager beat the benchmark by allocating well or by picking well? Use decision-based attribution when the question is about the changes themselves: did the trades I or my advisor made improve on what I already owned? Many reviews benefit from both. They are complementary calculations, not competing ones.
Where Pure Benchmarks fits
Pure Benchmarks, our own product, runs decision-based attribution through Decision Benchmark on linked brokerage and retirement accounts. It does not produce a Brinson decomposition. If you need allocation and selection effects against a policy benchmark, the attribution tools compared on our attribution page cover that, and most of them are sold to advisors or institutions.
Your current platform won't show you how your portfolio ranks against real investors in the same risk category. Create your secure Pure Benchmarks account and see exactly where you stand.
See Your Free Benchmark ReportFrequently asked questions
What is an alternative to Brinson attribution?
Decision-based attribution. Instead of splitting a return into allocation and selection effects against a benchmark, it compares the real portfolio with the version you held the moment before each change, so every buy and sell carries its own result against doing nothing.
What is decision-based attribution?
A method that attributes portfolio results to the individual decisions that produced them. Each change is measured against a frozen buy-and-hold copy of the portfolio from just before the change, adjusted for later deposits and withdrawals.
Is Brinson attribution wrong?
No. It is the institutional standard and it answers its own question well. It is a poor fit for judging individual trades because it measures groups over a period against an index, rather than each decision against the portfolio it replaced.
What is the difference between Brinson-Fachler and Brinson-Hood-Beebower?
Both split a return into allocation and selection effects. Brinson-Fachler measures the allocation effect relative to the total benchmark return, so overweighting a group that beat the benchmark is rewarded even if the group’s return was negative in absolute terms. Brinson-Hood-Beebower measures it against zero.
Do I need a benchmark for decision-based attribution?
No. The baseline is your own previous portfolio. That is what makes it usable for a personal account that does not map neatly onto any index.
Can decision-based attribution evaluate my financial advisor?
Yes, at the level of the changes the advisor made. Each change in a linked account is scored against the portfolio it replaced, so you can see whether the advisor’s trades added value over simply holding what was already there.
Keep exploring
- Best Performance Attribution Tools for Investors in 2026
- What Is a Decision Benchmark?
- What Is a Buy-and-Hold Counterfactual Portfolio?
- How Do I Measure the Quality of My Investment Decisions?
- MSCI (Barra) Alternatives in 2026
- IBviz Alternatives in 2026
- Capitally Alternatives in 2026
- Stock Rover Alternatives in 2026
See how a specific firm's advisor decisions compare to other firms:
New here? Read the portfolio benchmarking FAQ or the Pure Benchmarks blog.
This page is an information baseline for comparison only. It is not investment advice and not a recommendation to buy, sell, replace, or transfer any specific asset, account, or firm. Past performance does not guarantee future results. All figures shown are illustrative.