Glossary

Information Ratio — Definition, Formula, and Use by Allocators

Information ratio is excess return over a benchmark divided by tracking error. Definition, formula, worked example, and how institutional allocators use it.

By NakedPnL Research·May 7, 2026·4 min read

TL;DR

Information ratio (IR) is mean active return divided by the standard deviation of active return (tracking error).
It measures the consistency, not just the magnitude, of outperformance against a benchmark.
IR is a standard underwriting input for institutional allocators evaluating active managers.

Definition

The information ratio (IR) is the ratio of a portfolio's mean active return — the difference between portfolio return and benchmark return in each period — to the standard deviation of those active returns, known as the tracking error. A high IR indicates that the manager outperformed the benchmark not just on average but consistently, with low period-to-period dispersion of the active return.

Formula

IR = mean(R_p - R_b) / stdev(R_p - R_b)

where:
  R_p = portfolio return in each period
  R_b = benchmark return in the same period
  R_p - R_b = active return

The denominator stdev(R_p - R_b) is the tracking error.

IR is dimensionless and is conventionally annualised by multiplying by the square root of the number of periods per year.

Worked example

Over 36 monthly observations a portfolio outperforms its benchmark by an average of 0.30% per month, with a standard deviation of active returns of 1.10%. Monthly IR is 0.30 / 1.10 = 0.273. Annualised at √12, IR ≈ 0.94. Grinold's 'fundamental law of active management' (1989) treats this as a function of the manager's information coefficient and the breadth of independent bets they take per year.

How allocators use it

Institutional allocators — pensions, endowments, fund-of-funds — typically require an annualised IR of 0.5 or higher over a multi-year window before considering an active strategy worth its fees. An IR above 1.0 over five years is rare and is treated as a strong signal of either genuine skill or undisclosed factor exposure that the chosen benchmark fails to capture. Allocators usually run the calculation themselves against their own benchmark choice rather than relying on the manager's reported figure, because the result is highly sensitive to that choice.

Tracking error — denominator of IR
Sharpe ratio — IR's risk-free analogue
Active return — numerator of IR
Alpha — closely related but estimated via regression

Frequently asked questions

How does IR differ from the Sharpe ratio?

Sharpe measures excess return over the risk-free rate per unit of total return volatility. IR measures excess return over a chosen benchmark per unit of active-return volatility. Sharpe answers 'how rewarded was I for taking total risk?'; IR answers 'how rewarded was I for deviating from the benchmark?'

What is a good information ratio?

Grinold and Kahn classify 0.5 as 'good', 0.75 as 'very good', and 1.0 as 'exceptional', measured over a multi-year window. These shorthands assume the IR is computed against an honestly chosen benchmark and a sufficiently long window.

Does NakedPnL display IR?

No. IR is benchmark-dependent, and there is no canonical benchmark for the multi-asset, multi-venue population on the registry. Allocators can compute IR themselves against their own benchmark using the daily-return series exposed via the API.

References