Clevency
The Full Fundamental Law formula
The Fundamental Law of Active Management has two forms. The basic form is:
IR = IC × √BR
Where IR is the Information Ratio, IC is the Information Coefficient (correlation between forecasts and realised returns), and BR is Breadth (the number of independent active decisions per year).
The Full Fundamental Law adds the transfer coefficient to reflect implementation friction:
IR = IC × √BR × TC
Where TC captures the correlation between the active-weight positions the manager would take if unconstrained and the positions actually held after constraints are applied.
What transfer coefficient actually measures
Formally, TC is the correlation between risk-adjusted active positions (optimal) and risk-adjusted actual active positions. Intuitively:
- TC = 1.0 — perfect implementation. Every forecast translated into the portfolio weight the optimiser recommends. Hedge funds with few constraints can approach this.
- TC = 0.8 — strong implementation. Mild constraints (soft position limits, moderate tracking error cap).
- TC = 0.5 — typical long-only equity mutual fund. Can go overweight, cannot short.
- TC = 0 — no implementation. The portfolio has no relationship to the forecasts.
Worked numerical example
Suppose a manager has:
- IC = 0.10 (modest but realistic forecasting skill)
- BR = 100 (100 independent active positions per year)
- TC = 0.80 (constrained long-only portfolio)
Step 1 — basic Fundamental Law:
IR (unconstrained) = 0.10 × √100 = 0.10 × 10 = 1.00
Step 2 — Full Fundamental Law:
IR (actual) = 0.10 × √100 × 0.80 = 0.80
The constraints have cost the manager 0.20 of information ratio — a 20% reduction in risk-adjusted active return.
Step 3 — Convert to active return. If the manager runs 4% active risk (tracking error):
Active return (actual) = IR × active risk = 0.80 × 4% = 3.2% per year
Versus 4.0% unconstrained — a 0.8 percentage-point annual drag from constraints alone.
Constrained vs unconstrained portfolios
| Fund type | Typical TC | IC × √BR = 1.0 translates to IR of | Why |
|---|---|---|---|
| Market-neutral hedge fund | 0.90–1.00 | 0.90–1.00 | Can short, flexible sizing, no benchmark deviation cap |
| Long-biased hedge fund | 0.70–0.90 | 0.70–0.90 | Some sizing flexibility, limited shorting |
| Active equity mutual fund | 0.40–0.70 | 0.40–0.70 | Long-only constraint binds; position limits in prospectus |
| Enhanced-index fund | 0.20–0.40 | 0.20–0.40 | Very tight tracking error band on top of long-only |
| Pure index fund | 0 (no active) | 0 | No active positioning by design |
Why TC matters in practice
Two managers with identical IC and BR can post very different IRs because of implementation. The TC framework is why the industry increasingly reports "active share" and portfolio-level deviation data: these are proxies for TC. A fund marketed as "high-conviction active" with low active share has — by definition — low TC and will underperform its unconstrained alpha estimates.
The lesson for allocators is uncomfortable: a 0.5-TC fund with brilliant forecasts can be beaten by a 0.9-TC fund with mediocre forecasts, if the Full Fundamental Law maths works out.
How this appears on the CFA exam
Transfer coefficient is a Level 3 Portfolio Management topic, under the Fundamental Law of Active Management LOS. The exam patterns are predictable:
- Calculation — given IC, BR, TC, compute IR; or given IR and two of IC/BR/TC, back out the third.
- Interpretation — identify which constraint (long-only, position limit, risk cap) would reduce TC the most.
- Comparison — two managers with different TC values: which produces higher IR and why?
- Alpha decomposition — attribute underperformance between IC drop, BR drop, and TC drop.
Practice problem with worked solution
Problem: James Whitfield, CFA, runs an active equity fund. His IC is 0.08, BR is 144 independent active decisions per year. His fund is currently long-only, giving TC = 0.60. His tracking error target is 5%. His board asks: if we relax the long-only constraint to permit modest shorting (lifting TC to 0.80), by how much does our expected active return increase, holding IC and BR constant?
Solution:
Current IR: 0.08 × √144 × 0.60 = 0.08 × 12 × 0.60 = 0.576
Current active return: 0.576 × 5% = 2.88% per year
Proposed IR: 0.08 × √144 × 0.80 = 0.08 × 12 × 0.80 = 0.768
Proposed active return: 0.768 × 5% = 3.84% per year
Expected uplift: 0.96 percentage points per year, or roughly a one-third increase in active return, from the single change of permitting some shorting. The IC and BR did not improve; only implementation did.
Frequently asked questions
What is a good transfer coefficient?
Context-dependent. For a long-only active equity fund, 0.6+ is considered strong. For a market-neutral hedge fund, below 0.9 would be unusual. The relevant comparison is TC versus the achievable TC under that fund's constraints, not an absolute number.
Can TC be negative?
Mathematically yes — if actual positions are systematically opposite to optimal positions. In practice this would be an operational failure, not a measured outcome. Treat negative TC as a signal that the forecast or the implementation data is wrong.
How is TC estimated in practice?
By correlating the manager's actual active weights with the active weights an unconstrained optimiser would produce from the same forecasts. Firms with a research-to-implementation workflow track this routinely.
Does TC apply to fixed-income managers?
Yes, with modifications. The constraints look different (duration bands, credit limits, issue-size minimums), but the implementation-friction concept is the same. TC values for fixed income are typically lower than for equity because the constraint set is denser.
Why isn't TC in the basic Fundamental Law?
The basic form (IR = IC × √BR) assumes perfect implementation. Grinold and Kahn introduced TC in the Full Fundamental Law to make the model realistic for real-world constrained portfolios.
How does leverage affect TC?
Leverage itself does not change TC — it scales the whole portfolio equally. What changes TC is the structure of constraints (long-only, position limits, risk caps), not the gross exposure.