What Is the Information Ratio? Active Return per Unit of Tracking Error

The information ratio measures how much a portfolio returned above its benchmark for each unit of volatility in that margin. In plain terms, it does not just ask whether a manager beat the index. It asks whether they beat it steadily, which is a much harder test and a much more useful one.

It is the standard benchmark-relative measure in institutional performance review, and its appeal is that it is difficult to flatter with a single good year.

What it measures

The formula in words:

Information ratio = active return divided by tracking error.

Both terms need unpacking, because both are defined relative to a benchmark rather than to a risk-free rate.

Active return is the portfolio’s return minus the benchmark’s return over the same period. If a portfolio returned 15 percent and its benchmark returned 11 percent, the active return is 4 percentage points. Active return is often called excess return over benchmark, though “excess return” is also used for return over the risk-free rate, so the context matters. It is the same quantity people loosely call outperformance.

Tracking error is the standard deviation of that active return. To compute it, take the difference between portfolio and benchmark return in each period, giving a series of active returns, then measure how much that series varies around its own average. A portfolio that beat its benchmark by roughly the same margin every quarter has low tracking error. One that beat it by a lot in some quarters and lagged badly in others has high tracking error, even if the average margin is identical. See what is tracking error for the mechanics and for what levels are typical of different mandates.

So the ratio asks: for each unit of deviation from the index that the manager imposed, how much did they gain from that deviation.

An important consequence falls straight out of the arithmetic. A portfolio that simply tracks its index closely has near-zero active return and near-zero tracking error, and its information ratio is undefined or meaningless. The metric only has something to say about portfolios that actually take positions different from the benchmark.

How to read it

Sign first. A negative information ratio means the portfolio lagged its benchmark over the window, and the deviations from the index destroyed value rather than adding it.

Above zero, the number describes consistency. Some hypothetical arithmetic makes this concrete. Suppose two portfolios each beat their benchmark by an average of 3 percentage points a year over five years. The first did so by beating it in most quarters by a modest, similar margin, producing a tracking error of 3 percent, so its information ratio is 1.0. The second did so by lagging in four of the five years and then beating the index by a very large margin in one, producing a tracking error of 12 percent, so its information ratio is 0.25. Same average outperformance. Very different processes, and very different confidence that the result was repeatable.

Reading habits.

• The benchmark decides the answer. Change the index and both the numerator and the denominator change. A portfolio measured against a benchmark that does not match its actual investment universe will show large, meaningless active return and large, meaningless tracking error. Choosing the benchmark is not an administrative step; it is the measurement. See benchmark selection for portfolios.
• The window has to be long. Tracking error is a standard deviation, so estimating it from a handful of periods produces a noisy number. An information ratio from a single year is close to meaningless, and even three years is thin.
• Read it beside active share. Tracking error tells you how much the return differed from the index. Active share tells you how much the holdings differed. A fund can have low tracking error and high active share, which usually means the differences offset one another.
• Check whether the returns are net of fees and costs. An information ratio computed on gross returns can be respectable while the net figure, which is what an investor actually received, is negative.
• Compare like with like. A benchmark-hugging large-cap mandate and a high-conviction concentrated one are playing different games. Their information ratios are only comparable within the same mandate type, window and benchmark.

The information ratio rewards a manager for being reliably different, and penalises them for being erratically different. That is close to the right incentive.

Where it fits among the other ratios

It is worth being precise about what makes this metric distinct.

The Sharpe ratio asks whether the investor was paid for taking risk relative to a safe asset. The Treynor ratio asks whether they were paid for taking market exposure. The information ratio asks something narrower and, for an active manager, more pointed: were they paid for the specific decision to differ from the index.

That makes it the natural companion to alpha. Alpha estimates the return that cannot be explained by market exposure. The information ratio measures whether whatever return was earned beyond the benchmark showed up dependably enough to look like a process rather than an accident.

What it does not tell you

It says nothing about absolute outcomes. A portfolio can post a strong information ratio while losing money, because it lost less than its benchmark did. In a market that fell 30 percent, a fund that fell 24 percent has positive active return. Investors do not spend relative performance.

It is entirely hostage to the benchmark. This bears repeating because it is where most misuse occurs. An unrepresentative benchmark manufactures both the numerator and the denominator. A benchmark that is easy to beat produces an information ratio that measures the benchmark’s weakness, not the manager’s strength.

It treats good and bad deviation identically. Tracking error is a standard deviation, so a quarter of unusually large outperformance inflates the denominator exactly as a quarter of underperformance does. This is the same symmetry complaint that motivated the Sortino ratio, and the same fix applies in principle.

It is silent on drawdown. A fund with a good information ratio can still have suffered a severe peak-to-trough fall alongside its benchmark. Depth requires maximum drawdown and the Calmar ratio.

It does not identify the source of the difference. Active return may come from stock selection, from sector tilts, from a persistent factor exposure, or from holding cash. The ratio does not decompose it. Factor exposure analysis is the tool that does.

It is noisy and easily overinterpreted. Both inputs are estimates from a limited sample. Small differences in information ratio between two managers over a short window are rarely distinguishable from chance, however precisely the decimals are printed.

It can be inflated by a benchmark mismatch in style. A fund with a persistent small-cap tilt measured against a large-cap index will show high active return and high tracking error in a small-cap rally. The ratio may look like skill and be closer to a style exposure. Understanding this is why factor investing in India matters when interpreting benchmark-relative results.

It cannot repair a flawed measurement. If the return series is built from data that would not have been available at the time, the whole calculation is describing something that did not happen. See why point-in-time data matters.

Using it in practice

The information ratio does its best work in manager review and in strategy comparison, where the benchmark is fixed and agreed, the window is long, and the returns are net of everything an investor actually pays.

Report it with four things attached: the benchmark, the window, the return frequency, and whether returns are gross or net. Read it beside tracking error itself, active share, a drawdown measure and an absolute return figure. And treat it as evidence about consistency rather than proof of skill. Consistency over a long window is the best available signal that a result came from a process, but it is a signal, not a certificate.

Related reading

• Portfolio metrics explained: the hub that maps every risk and return metric to the question it answers.
• What is tracking error: the denominator on its own, and what levels are normal.
• What is alpha in investing: return beyond what the benchmark and risk explain.
• What is the Treynor ratio: reward per unit of market exposure rather than per unit of deviation.
• What is the Sharpe ratio: the total-risk starting point every other ratio adjusts.

This article is educational. Altys Labs is not a registered research analyst or investment adviser, and nothing here is investment advice or a recommendation to buy, sell, or hold any security.