A return is a fact; attribution is an explanation

"The portfolio returned 8% last year." Fine — but is that good? Compared to what? Driven by skill or luck? Repeatable or a one-off? A raw return answers what happened; performance attribution answers why it happened, whether the comparison is fair, and what to review next. That distinction is the difference between a number on a statement and an insight you can act on.

TWR vs MWR: two returns, two questions

The first subtlety is that there are two legitimate ways to measure return, and they answer different questions:

• Time-weighted return (TWR) strips out the effect of cash flowing in and out, isolating the performance of the investment decisions themselves. It is the right measure for judging the manager or strategy, because it is not distorted by when the client added or withdrew money.
• Money-weighted return (MWR), essentially an internal rate of return, accounts for the size and timing of cash flows. It answers what the investor actually experienced in their specific situation.

A manager can have a strong TWR while a client has a weak MWR simply because the client added money right before a downturn. Knowing which question you are asking is the start of honest performance analysis.

Active return: results are relative

An 8% return means something completely different depending on the benchmark. If the benchmark returned 6%, the portfolio earned +2% of active return; if it returned 11%, the portfolio lagged by 3%. Performance only has meaning relative to the assigned benchmark — and the benchmark has to be a valid comparison. Comparing a conservative balanced portfolio to an all-equity index produces flattering or damning numbers that are simply not meaningful. Confirming the benchmark is appropriate is a prerequisite to any judgment.

Information ratio and tracking error: is the active return earned well?

Beating a benchmark by 2% can be impressive or reckless depending on how much risk was taken to do it. Tracking error measures how much the portfolio deviates from its benchmark; the information ratio divides active return by tracking error to show the consistency of outperformance per unit of active risk. A high information ratio means steady, reliable value-add; a low one means the outperformance came with wild swings that may not repeat. This reframes "did we beat the benchmark?" as "did we beat it efficiently?"

Brinson attribution: allocation, selection, interaction

The heart of attribution is decomposing active return into its sources. The Brinson framework splits it into:

• Allocation effect — value added (or lost) by over- or under-weighting sectors or asset classes relative to the benchmark. Did being overweight the right area help?
• Selection effect — value added by picking better-performing securities within each sector. Was it good stock-picking?
• Interaction effect — the combined impact of allocation and selection decisions together.

This decomposition answers the question every investor and committee actually cares about: was the result driven by where we invested or what we picked? A portfolio might beat its benchmark entirely through sector allocation while its security selection actually detracted — a completely different story than "good stock-picking," and one that changes what you do next.

Contributors and detractors

Beyond the structural decomposition, identifying the top contributors and bottom detractors grounds the analysis in specifics. Which positions drove the result, for better or worse? This is what turns an attribution report into a conversation — with a client, a committee, or an internal review — and points directly at what deserves scrutiny.

From explanation to action

Attribution is not an end in itself; it is a routing tool. A valid, well-decomposed performance story tells you whether the next step belongs in risk (an exposure that drove volatility), tax (gains to manage), reporting (a clear client explanation), or strategy review (a selection process that is not working). In a continuous, agentic platform, attribution becomes the shared context that lets performance, risk, tax, and reporting act on the same evidence rather than re-deriving it separately.

The takeaway

Performance attribution turns "what happened" into "why, and what now." Use TWR to judge the strategy and MWR to reflect the investor's experience; measure active return against a valid benchmark; use the information ratio to see whether outperformance was earned efficiently; and use Brinson decomposition to learn whether allocation or selection drove the result. Do that and every performance review stops being a scorecard and becomes a source of better decisions.