It’s the standard for reporting, but when it comes to identifying actual decision-making skill — whether you’re a portfolio manager or an allocator of capital — performance attribution leaves a lot to be desired. But there’s an alternative: decision attribution.
By Clare Flynn Levy
Clare Flynn Levy is CEO & Founder of Essentia Analytics. Prior to setting up Essentia, she spent ten years as a fund manager, in both active equity (running over $1B of pension funds for Deutsche Asset Management), and hedge (as founder and CIO of Avocet Capital Management, a specialist tech fund manager).
Everyone in the business of investment management is familiar with traditional performance attribution: it starts with portfolio returns and works backward to explain what happened. Did the manager outperform the index? Where did they get it right (or wrong) — stock picking, sector picking, country picking? Were they more or less exposed than their benchmark to growth, value or other factors?
There’s no question that performance attribution analysis is useful for constructing a narrative about why the portfolio performed the way it did. But because this type of analysis starts with the outcome, rather than the actual decisions the investor took that led to that outcome, it is heavily influenced by the random effects of luck. Moreover, it doesn’t give the manager any indication of how they might achieve better results in the future.
Decision attribution, on the other hand, takes a bottom-up approach. It looks at all of the different decisions (stock picking, sizing, timing, etc.) the manager made during a given period and measures the distinct value each decision generated or destroyed, identifying patterns of demonstrated skill — and specific areas for improvement.
At the end of the day, a fund manager’s job — indeed, their fiduciary duty — is to make the best decisions they can given the information available, at all times. Decision attribution tells us whether the manager is doing a good job on that basis or not — performance attribution cannot.
I discussed this topic in a recent seminar with the Portfolio Construction Forum, a professional development program for asset managers, allocators and fund buyers in Australia. This two-minute clip sums it up nicely:
Below is what decision attribution analysis looks like in the real world. The Essentia Behavioral Alpha® Frontier (EBAF) diagram illustrates where a manager is adding value (in this case, decisions around exit timing, scaling in/out and sizing — toward the top right in the diagram) and where there is room for improvement (entry timing, size adjusting and stock picking, toward the lower left). The dotted curve in the center represents what would be achieved by chance: decisions that are profitable 50% of the time (decision hit rate), and that are typically as profitable when they are right as they are unprofitable when they are wrong (decision payoff ratio). The investor’s most skilled decisions are those to the right of the dotted curve.
Essentia’s Behavioral Alpha Frontier considers decision hit rate (the frequency of value-accretive decisions) and payoff ratio (the magnitude of the value added or destroyed in each decision or decision type). The upper right quadrant — furthest away from 0,0 — represents the highest value-add.
Performance attribution has its place: it’s important that there is an industry standard for reporting. But the investment management industry is well aware that past performance is not predictive of future success — after all, it’s written in the fine print on every investment ad.
An investor who is ultimately trying to select managers who will perform well in the future should be looking deeper, at the manager’s actual decision-making. After all, an investor who makes profitable decisions more than 50% of the time, and whose profitable decisions tend to make more than their unprofitable decisions lose, will outperform (or make money, if you’re looking at it in absolute terms) — that’s just a mathematical fact.
If you’re an allocator of capital to equity managers, it’s worth asking any manager on your shortlist to show you their decision attribution analysis before you invest. They may not have it at the ready, but they do have the data required to do it — and the incentive, not only in the form of your potential investment but also the means to continuously improve their performance. And if they aren’t willing to look in the mirror at their own decision-making … enough said.
If you’re a manager, you already have the daily historical holdings data you need to do decision attribution analysis — just ask your custodian or fund administrator. And it doesn’t necessarily require a large financial investment — just ask us.
Essentia tracks manager decision-making and publishes the top five-ranked managers each quarter. To learn more — including which managers have shown the most decision-making skill over the latest three-year period — get the details here.
