Home / Wiki / I / Information Ratio (IR)

Information Ratio (IR)

Investment and Capital MarketsUpdated 16 Nov 2026

The information ratio (IR) is a measure of risk-adjusted performance that compares a portfolio with a benchmark. It divides the portfolio's average return in excess of the benchmark, called the active return, by the variability of that excess, called the tracking error. A high ratio indicates that the manager has beaten the benchmark by a meaningful margin and done so steadily, instead of through a few large and possibly lucky bets.

The ratio is the standard tool for assessing active management. Pension funds, consultants and fund selectors use it to compare managers who follow the same benchmark, and investment firms use it internally to set risk budgets and judge portfolio managers. It is closely related to the Sharpe ratio, which uses the risk-free rate as the reference point in place of a market benchmark.

01

Formula

The calculation is:

IR=Rˉp−Rˉbσ(Rp−Rb)IR = \frac{\bar{R}_p - \bar{R}_b}{\sigma_{(R_p - R_b)}}
  • Rˉp\bar{R}_p is the average return of the portfolio over the measurement period.
  • Rˉb\bar{R}_b is the average return of the benchmark index over the same period.
  • σ(Rp−Rb)\sigma_{(R_p - R_b)} is the tracking error, the standard deviation of the period-by-period differences between the two returns.

Both parts are normally annualised. With monthly data the average active return is multiplied by 12 and the tracking error by the square root of 12, so a monthly ratio is multiplied by the square root of 12, about 3.46.

02

Worked example

Suppose a fund's returns minus its benchmark's returns in six successive months were 0.5, -0.3, 0.8, 0.2, -0.2 and 0.2 percentage points. The sum is 1.2, so the average monthly active return is 1.2÷6=0.21.2 \div 6 = 0.2.

The deviations from that average are 0.3, -0.5, 0.6, 0, -0.4 and 0. Their squares are 0.09, 0.25, 0.36, 0, 0.16 and 0, which sum to 0.86. Dividing by five, one less than the number of observations, gives a variance of 0.172, and the square root is a monthly tracking error of 0.415. The monthly information ratio is 0.2÷0.415=0.4820.2 \div 0.415 = 0.482, and the annualised figure is 0.482×3.464=1.670.482 \times 3.464 = 1.67.

Six months is far too short a record to be reliable and is used here only to show the arithmetic. Stated in annual terms, the same logic applies directly: a fund that returns 9 percent against a benchmark return of 7 percent with a tracking error of 4 percent has an information ratio of (9−7)÷4=0.5(9 - 7) \div 4 = 0.5.

03

Interpretation

The ratio answers the question of how much excess return was earned for each unit of risk taken relative to the benchmark. Two managers who both beat an index by 2 percentage points are not equally skilled if one deviated from it by 4 percent a year and the other by 10 percent. A passive management strategy that tracks its index closely has an active return and a tracking error near zero, and its ratio has little meaning.

A commonly quoted rule of thumb, taken from the work of Richard Grinold and Ronald Kahn, treats an annualised ratio of about 0.5 as good, 0.75 as very good and 1.0 as exceptional when sustained over many years. Few managers maintain a ratio above 0.5 after fees over long periods. The same authors popularised the fundamental law of active management:

IR≈IC×BRIR \approx IC \times \sqrt{BR}

Here ICIC, the information coefficient, is the correlation between a manager's forecasts and the outcomes, and BRBR, breadth, is the number of independent bets made in a year. A manager with a modest information coefficient of 0.05 who makes 100 independent decisions a year has an expected ratio of 0.05×100=0.50.05 \times \sqrt{100} = 0.5. The law implies that a small edge applied many times can be worth more than a large edge applied rarely.

  • Sharpe ratio. It measures total risk-adjusted return against cash and suits the assessment of a whole portfolio. The information ratio measures relative return and suits a manager hired to beat a specific index.
  • Alpha-based measures. The appraisal ratio divides alpha, the return left after adjusting for market exposure measured by beta, by the variability of that residual. It differs from the information ratio when the portfolio's beta is not one, since a manager can beat a rising benchmark simply by taking more market risk.
  • Downside measures. The Sortino ratio and the Calmar ratio penalise only losses or drawdowns, whereas tracking error treats outperformance and underperformance alike.
05

Limitations

The result depends heavily on the benchmark. A manager measured against an index that does not match the investment style, for example a small-company fund compared with a large-company index, can show a high ratio that reflects the mismatch and not skill. The choice of period matters as well, and ratios estimated over three years or less have wide margins of error.

Tracking error assumes that active returns are roughly symmetric. Strategies that earn small, regular gains and occasionally suffer a large loss can display a high ratio until the loss occurs, so the figure is often read alongside the maximum drawdown. The ratio also ignores the absolute level of return: a fund can have an excellent ratio while losing money, if the benchmark lost more. Finally, a manager can raise the reported ratio by holding illiquid assets whose prices are smoothed, which understates tracking error. For these reasons the ratio is used with contribution to return analysis and other attribution tools, not on its own.

Last updated 16 Nov 2026