What the Information Ratio Measures
The information ratio (IR) answers a specific question: how much return did a portfolio manager add over the benchmark, per unit of the risk taken in deviating from that benchmark? It is the standard metric for evaluating active management skill, because both the numerator and the denominator are defined relative to the benchmark the manager was hired to beat.
That benchmark-relative framing is what separates the IR from total-risk metrics. The Sharpe ratio, for instance, measures excess return over the risk-free rate per unit of total portfolio volatility. It tells you how well the whole portfolio compensated you for its overall riskiness. The information ratio ignores the portion of return and risk that comes from simply holding the benchmark, and isolates the part the manager's active decisions contributed. A manager can run a portfolio with low total volatility and still have a poor information ratio if the deviations from the benchmark added risk without adding return.
This concept sits in the CFA Level I Portfolio Management readings on portfolio risk and return, alongside the Sharpe ratio, the Treynor ratio, and M-squared. You can review the full topic list on the Level I Portfolio Management unit page. The same machinery (active return and tracking error) also underpins the discussion of active versus passive management in Level I Equity, so time spent getting the definitions precise pays off in two topic areas.
The Formula, Unpacked
$$\text{IR} = \frac{R_p - R_b}{\sigma_{(R_p - R_b)}}$$
In words: the information ratio equals active return divided by tracking error. Each piece has a precise definition worth committing to memory.
Active return is the portfolio return minus the benchmark return. It is sometimes called the "active" or "benchmark-relative" excess return. Be careful with the word "excess" on the exam: excess return over the risk-free rate belongs to the Sharpe ratio, while excess return over the benchmark belongs to the information ratio. Questions are frequently written to punish candidates who blur that distinction.
Tracking error (also called tracking risk or active risk) is the standard deviation of active return, not the standard deviation of the portfolio's return. A portfolio that moves exactly with its benchmark has a tracking error of zero, however volatile both may be. Tracking error only grows when the portfolio's returns diverge from the benchmark's.
Annualisation conventions matter when the inputs are measured over shorter periods. If you compute active returns monthly, the annualised tracking error is the monthly standard deviation multiplied by the square root of 12, and the annualised active return is (approximately, using arithmetic scaling) the monthly mean multiplied by 12. Keep the numerator and denominator on the same time basis. An IR built from an annual active return and a monthly tracking error is meaningless, and mixing bases is a classic way to get a plausible-looking wrong answer.
A Worked Example
The figures below are invented for illustration. They are not real fund data, and they are chosen to keep the arithmetic clean so you can focus on the method.
Suppose a manager runs a portfolio against an equity index and, over the past year:
- Portfolio return: 9.2%
- Benchmark return: 8.0%
- Tracking error (annualised standard deviation of active return): 3.0%
Step 1: Compute the active return. 9.2% minus 8.0% gives an active return of 1.2%.
Step 2: Divide by the tracking error. 1.2% divided by 3.0% gives an information ratio of 0.40.
The interpretation: for every percentage point of active risk the manager took, they generated 0.40 percentage points of return above the benchmark.
Now compare a second (equally hypothetical) manager against the same benchmark:
- Portfolio return: 10.4%
- Benchmark return: 8.0%
- Tracking error: 8.0%
The active return is larger at 2.4%, but the IR is 2.4% divided by 8.0%, which is 0.30. The second manager beat the benchmark by twice as much, and still ranks lower on the information ratio, because they took proportionally more active risk to do it. That comparison is exactly the judgement the metric exists to make, and exam questions regularly present two managers in this configuration to test whether you rank by active return (wrong) or by IR (right).
IR vs Sharpe vs Treynor
The three headline ratios differ in what they subtract in the numerator and what they divide by in the denominator. Keeping the table straight is worth easy marks:
| Ratio | Numerator | Denominator | Risk concept |
|---|---|---|---|
| Information ratio | Portfolio return minus benchmark return | Tracking error | Active risk |
| Sharpe ratio | Portfolio return minus risk-free rate | Standard deviation of portfolio returns | Total risk |
| Treynor ratio | Portfolio return minus risk-free rate | Portfolio beta | Systematic risk |
A quick way to choose the right one in a question: ask what the scenario is comparing against and what risk is relevant. Evaluating an active manager against an assigned benchmark calls for the information ratio. Evaluating a portfolio that constitutes someone's entire wealth calls for the Sharpe ratio, because total volatility is what the investor experiences. Evaluating a portfolio that is one well-diversified sleeve among many calls for the Treynor ratio, because only systematic risk survives diversification.
What Counts as a "Good" Information Ratio?
Candidates often want a threshold, and the honest answer is that any threshold is a rule of thumb rather than a law. The most commonly cited benchmarks come from Richard Grinold and Ronald Kahn's book Active Portfolio Management, which describes an information ratio of 0.5 as good, 0.75 as very good, and 1.0 as exceptional. The same book presents empirical percentile data suggesting that a top-quartile manager achieves an IR of about 0.5, the 90th percentile sits near 1.0, and the median active manager comes in around zero.
That last figure deserves emphasis. An IR near zero for the median manager is consistent with the broader evidence on active management: before fees, active return across all managers is roughly a zero-sum game against the benchmark. So an IR of 0.5, which can sound unimpressive next to Sharpe ratios quoted for whole markets, actually describes a genuinely strong active manager. Treat the 0.5/0.75/1.0 ladder as attributed context you can cite, and not as an official CFA Institute grading scale, because the curriculum does not bless any specific cutoff.
A negative IR means the manager underperformed the benchmark over the measurement period. It does not necessarily mean the manager destroyed value in absolute terms (the portfolio may still have earned a healthy positive return), and it says nothing by itself about future performance. Interpreting the sign correctly, without over-reading it, is itself a tested skill.
Ex-Ante vs Ex-Post
The IR comes in two flavours, and the exam expects you to know which is which.
The ex-post (historical) information ratio uses realised active returns: you measure what the portfolio and benchmark actually did, compute the realised active return and its standard deviation, and divide. This is the version used to evaluate past performance, and it is the version in the worked example above.
The ex-ante (expected) information ratio uses forecast active return and forecast tracking error. This is the version a manager or allocator uses when deciding how much active risk to take going forward. Ex-ante IRs are generally positive, because no manager deliberately positions a portfolio expecting to underperform, while ex-post IRs are frequently negative, because expectations are often not met. If a question mixes forecasts and realised numbers, sort them into the right bucket before calculating anything.
At Level II, the ex-ante IR reappears inside the fundamental law of active management, which decomposes it into the manager's skill per decision and the number of independent decisions taken. For Level I purposes you need the definition, the calculation, and the interpretation.
Common Exam Traps
Subtracting the risk-free rate instead of the benchmark. The single most common error. If you catch yourself computing portfolio return minus the risk-free rate and dividing by tracking error, you have built a hybrid that is neither the Sharpe ratio nor the information ratio. The IR is benchmark-relative in both the numerator and the denominator.
Using portfolio standard deviation as the denominator. Tracking error is the standard deviation of the difference between portfolio and benchmark returns. Vignettes often supply the portfolio's own standard deviation as a distractor, and one of the wrong answer choices will be the result of using it.
Confusing ex-ante and ex-post. Read whether the question gives you forecasts or realised history, and match the version of the ratio accordingly.
Ranking managers by active return alone. As the worked example shows, the larger active return can come with the lower IR. When a question asks which manager displayed more skill per unit of active risk, compute both IRs rather than eyeballing the numerators.
Mixing arithmetic and geometric active return, or mismatched periods. Active return can be quoted as a simple difference (arithmetic) or as a compounded relative (geometric), and tracking error can be quoted monthly or annually. Use consistent definitions and consistent time bases throughout one calculation, and annualise with the square root of time for the risk figure.
Making It Stick
The information ratio is a small, self-contained piece of the Portfolio Management syllabus, and questions about it are highly patterned: define it, calculate it, compare it with the Sharpe and Treynor ratios, or rank two managers. That makes it ideal material to drill until the answers are automatic. Practise Level I Portfolio Management questions covering the risk-adjusted performance measures alongside the rest of the Portfolio Risk and Return topic, and work them until choosing the right ratio for a scenario takes seconds rather than a re-derivation.
As with everything on PopCFA, this post is exam revision material and not investment advice. The illustrative figures above are invented for teaching purposes, and real manager evaluation involves far more context than any single ratio.