The Transfer Coefficient

The fundamental law of active management says an unconstrained manager's information ratio depends on skill and breadth — how good the manager's forecasts are, and how many independent bets they make. But almost no real portfolio is unconstrained. A long-only fund can't short the stocks it dislikes. A benchmark-relative fund has position limits, sector limits, and turnover budgets. Every one of those constraints stops some of the manager's actual views from being fully expressed in the portfolio, and the transfer coefficient measures how much of that leakage happens.

Where it sits in the fundamental law

The original fundamental law states $IR \approx IC \times \sqrt{BR}$, information ratio approximately equals the manager's forecasting skill (the information coefficient) times the square root of how many independent bets they make (breadth). Adding the transfer coefficient $TC$ extends this to:

$$IR \approx IC \times TC \times \sqrt{BR}$$

In words: realized performance depends not just on how good your forecasts are and how many independent ones you make, but on how faithfully those forecasts actually get turned into portfolio positions. A brilliant forecaster stuck with a long-only mandate on a stock they desperately want to short is contributing zero transfer on that view, no matter how correct the view turns out to be.

Worked example

A manager has an information coefficient of 0.10 and 100 independent bets per year, giving an unconstrained information ratio estimate of $0.10 \times \sqrt{100} = 0.10 \times 10 = 1.00$.

1. Long-only constraint. The manager wants to underweight or short roughly 30% of the names in the universe based on negative views, but a long-only mandate caps the most negative active weight at simply not holding the stock (zero weight), not a true short. Studies of this specific constraint find it typically caps the transfer coefficient around 0.7-0.8 for a reasonably diversified long-only fund.
2. Applying $TC = 0.75$. Realized information ratio becomes $1.00 \times 0.75 = 0.75$.
3. The cost of the constraint. A quarter of the manager's theoretical edge, in information-ratio terms, is lost purely to the inability to express negative views as fully as positive ones — not because the forecasts were wrong, but because the portfolio couldn't act on them.

What this means in practice

The transfer coefficient is why relaxing constraints, moving from long-only to 130/30 (30% short, 130% long), or loosening position limits, can raise a fund's realized performance without the manager's forecasting skill changing at all. It's also why comparing two managers' realized information ratios can be misleading if one runs an unconstrained hedge fund and the other runs a heavily constrained long-only mandate: the gap might be entirely a transfer-coefficient effect, not a skill difference.