On this page
FLOAM helps test whether an active manager’s claimed information ratio could plausibly arise from the way the portfolio is built. Start with the quality of the manager’s forecasts, count genuinely independent opportunities, then ask how faithfully the mandate turns those forecasts into positions. A strong-looking backtest is less persuasive if its bets repeat one factor or the actual portfolio cannot express the forecasts.
What the law is trying to explain
The fundamental law of active management links information ratio, or active return per unit of active risk, to two underlying ingredients: information coefficient and breadth. Information coefficient describes the relationship between forecasts and subsequent realised returns. Breadth is the effective number of independent decisions, not the count of securities held or the number of trades printed on a report. CAIA Association’s historical Level II workbook discusses these components and a transfer coefficient; the assigned Omni reading uses them as a due-diligence framework. This article explains that framework rather than presenting an old workbook as the current exam syllabus. CAIA Association says candidates should use the official curriculum for their own exam year, because editions change. The CAIA programme page gives the broader preparation context.
The square root is the first exam trap. More decisions can improve the opportunity to use skill, but an increase in the raw count of bets does not translate one-for-one into information ratio. If forecast skill is unchanged and breadth becomes four times as large, the simplified model makes the information ratio twice as large, because the square root of four is two. That result assumes the added decisions are independent and the same forecasting skill applies to them. It is a conditional model statement, not evidence that a real manager can simply trade more often to become better. If the new trades mostly repeat the original exposures, effective breadth barely changes.
Information coefficient is a correlation, not a percentage return. A positive figure means forecasts tend to rank later outcomes in the intended direction, but it does not specify the amount earned after costs. Estimating it requires a consistent definition of forecast, horizon, universe and realised outcome. A manager might show a high correlation in a narrow research sample yet struggle when the universe changes. A candidate should therefore ask whether the measured skill survives a different period, an out-of-sample test and a comparison with a sensible benchmark. The law organises these questions; it cannot verify the underlying data on its own.
Why portfolio constraints need their own term
The simple expression assumes that useful forecasts can be translated into appropriate active weights. Real mandates impose limits: some positions cannot be shorted, concentrations may be capped, liquidity may be thin and benchmark risk must be controlled. The extended form introduces the transfer coefficient, a measure of how closely the implemented active portfolio follows the signals. A transfer coefficient near one indicates strong implementation alignment in the model; a smaller one means useful forecasts are being blocked or diluted. CAIA Association’s workbook defines this term through the relationship between forecast active returns and active weights. It does not say every constrained manager is unskilled. A constraint changes implementation, not the quality of a forecast.
A long-only portfolio can express a negative view by owning less of an asset than its benchmark weight, but it cannot necessarily make the same position as an unrestricted long-short strategy. A narrow sector mandate may have many forecasts yet little room to scale the strongest signals without concentration. In each case the right due-diligence question is which forecasts are least well represented in the final portfolio, and why. The answer can point to legal limits, liquidity, trading costs, risk controls or discretionary overrides. The transfer coefficient is a compact diagnostic of the gap, not a licence to remove sensible safeguards merely to raise a modelled ratio.
An illustrative manager comparison
Take two invented managers with the same illustrative information coefficient of 0.08 and one hundred genuinely independent decisions over the measurement period. The square root of one hundred is ten, so the unconstrained basic model gives each a potential information ratio of 0.80. Manager A has an illustrative transfer coefficient of 0.90, while Manager B’s tighter mandate yields 0.50. Under the extended relationship, A’s modelled information ratio is 0.72 and B’s is 0.40. The arithmetic isolates implementation: neither manager is assigned more forecast skill or more independent opportunities. The assumptions are deliberately simple and the outputs are model estimates, not observed fund performance or promises.
Now suppose B claims that adding one hundred more holdings will restore the information ratio while its forecasts and constraints remain unchanged. The raw holding count is not enough. If the additional holdings reflect the same sector or factor views, there may be little new independent breadth. Even if effective breadth doubled from one hundred to two hundred, the model would give approximately 0.08 times the square root of two hundred times 0.50, or about 0.57. It would not reach A’s 0.72 under the stated assumptions. To match 0.72 at the same skill and transfer coefficient, effective breadth would need to be 324, since the square root would need to be eighteen. The implied requirement may be a useful challenge to B’s process.
Translate the formula into evidence requests
Begin with the forecast record. Ask how a signal was defined before outcomes were known, whether failed ideas are included, and whether the calculation uses the same horizon as the portfolio’s decisions. If the manager publishes many revisions to one forecast, count the underlying decision appropriately instead of treating each revision as a fresh independent opportunity. Then examine overlap. Positions spread across many securities can all be exposed to one macro driver. A manager can show breadth using a trade count, but the relevant object for the law is independent decision content. Correlations, common factor exposures and repeated catalysts can expose double counting.
Next compare forecasts with implemented weights. A manager who correctly ranks opportunities but applies nearly equal active weights may give weak signals too much risk and strong signals too little. A portfolio that systematically excludes the strongest negative forecasts may reflect a no-short constraint. A portfolio with large positions unrelated to stated forecasts suggests discretionary overrides or other drivers. These patterns make the transfer question concrete. Active risk, sometimes described as aggressiveness, also needs context: scaling positions more aggressively can change active return and active risk together. A higher gross active return is not automatically a higher information ratio. Good sizing must be tied to the strength and reliability of the signal.
Costs and capacity create another boundary around a promising calculation. A modelled information ratio before implementation does not guarantee the same ratio after spreads, market impact or fees. Increasing trade frequency can raise an apparent breadth count while eroding net returns. An opportunity set that worked with a small portfolio may become less accessible at larger scale. The law can help an allocator identify where a claimed edge ought to originate, but the final judgement also requires a credible trading and risk process. The Omni methodology page explains how its study material is anchored to official curricula.
How to answer a changed exam scenario
Identify which variable changed before recalculating. A fall in transfer coefficient because of new position limits reduces the modelled information ratio even when skill and breadth stay the same. More independent decisions raise it through a square root, provided the same information coefficient remains credible. If forecasts become weaker as the manager expands into unfamiliar markets, a breadth increase can be offset by a lower information coefficient. A question may give a target ratio and ask what skill, breadth or transfer would be required; rearrange the formula, then judge whether the implied input is plausible. The number is an implication of assumptions, not a verdict about an actual manager.
The most common conceptual mistake is to call every trade an independent bet and infer that a busier manager must have a higher ratio. A second mistake is to explain weak implementation as a forecasting failure when the portfolio is constrained. A third is to compare a historical ratio with a modelled target without checking the same period, benchmark, active-risk definition and cost basis. FLOAM is useful because it keeps these questions separate. Skill is about forecasts; breadth is about independent opportunity; transfer is about implementation; observed performance remains the outcome to test. Use the framework to interrogate a claim, not to certify it.
Sources
- CAIA Level II Workbook, September 2021 · caia.org · Retrieved
- Curriculum and Study Tools | CAIA Association · caia.org · Retrieved
- Explore the Curriculum | CAIA Association · caia.org · Retrieved



