← Q&A archive
05-information-ratio May 2, 2026

Why is IR a better metric than alpha alone?

Why IR Beats Raw Alpha as a Yardstick

alpha" title="Alpha — definition">Alpha tells you how much a manager beat their benchmark. IR tells you how efficiently they did it — alpha per unit of active risk taken. That distinction matters for two reasons.


1. Alpha ignores the risk taken to earn it

Consider two funds, both beating the Nifty Midcap 150 by 3% per year over five years:

Fund Alpha Tracking Error IR
A 3% 12% 0.25
B 3% 4% 0.75

Same alpha. Fund B achieved it at one-third the active risk. Fund A may have simply concentrated in a handful of stocks that happened to rally — statistically fragile. Fund B is generating consistent, spread-out outperformance — the signature of genuine skill. [1]


2. Alpha is structurally identical to a Sharpe ratio — but applied to the active bet

$$IR = \frac{\overline{rr}}{TE}$$

where $\overline{rr}$ is the mean active return and $TE$ is its Volatility) — definition">Standard Deviation (Volatility) — definition">standard deviation. This isolates the quality of the active management decision, stripped of the fund's market exposure. The Sharpe Ratio — definition">Sharpe ratio blends both systematic and active risk together; IR looks only at the active layer. [3] [5]


3. IR connects directly to the fee decision

An active fund earns its higher Expense Ratio — definition">expense ratio only if:

$$IR \times TE > \text{fee differential vs. passive}$$

Raw alpha doesn't give you this test cleanly, because it doesn't tell you whether that alpha is reliable or the product of a lucky concentrated bet. [4]


The Fundamental Law says:

$$IR \approx IC \times \sqrt{B}$$

Alpha can be high for one lucky year. A sustained IR above 0.5 over 5+ years implies the manager has genuine information advantage ($IC > 0$) deployed across many independent bets ($B$ is meaningful). A high alpha with low IR is consistent with luck; a sustained IR is not. [6]


Practical rule: Use alpha to know the size of outperformance. Use IR to know whether it's worth paying for. [5]


Apply this → Sort funds by 5-year IR at Explore Funds and cross-check whether the IR × TE clears the fee hurdle for each one.

Sources cited