Table of Contents
There is a peculiar habit in finance of borrowing words from other fields and then quietly stripping them of their original meaning. Volatility no longer suggests something dangerous or unstable in any human sense. It just means a number wiggles. Risk has been demoted from the possibility of ruin to the standard deviation of a return series. And then there is information, perhaps the most abused word of them all, which sits inside one of the most respected ratios in active management and pretends to mean something profound.
The Information Ratio is everywhere. Pension consultants love it. Fund managers quote it on tear sheets. Allocators rank managers by it. Academics build careers around it. It is the supposed measure of skill, the cleanest available proof that a portfolio manager is doing something other than getting lucky in a rising market. And the very name implies a noble pursuit. The ratio of information. As if the manager were a quiet scholar in a candlelit room, gathering insights the rest of the market has missed, and converting those insights into returns at a steady and superior rate.
The reality is somewhat less romantic.
What the Ratio Actually Measures
Before getting to the philosophical mischief, it helps to understand what the Information Ratio is literally doing. It takes the excess return a manager earns over a benchmark and divides it by the volatility of that excess return. The numerator captures how much better the manager did. The denominator captures how bumpy the road was while doing it. A high Information Ratio means the manager consistently beat the benchmark by a meaningful margin without too much weaving across lanes.
So far, so reasonable. It is a measure of risk adjusted outperformance, in the same family as the Sharpe Ratio but more focused. Sharpe asks whether you were paid for taking risk in absolute terms. The Information Ratio asks whether you were paid for taking risk relative to a benchmark. It is the metric of choice for evaluating active managers who promise to beat an index.
The word information enters because, somewhere along the way, theorists decided that the only legitimate way to beat a benchmark consistently was to have some kind of edge. And edge, in turn, was assumed to be information. Better data, faster data, smarter interpretation of data, or some proprietary insight that nobody else had figured out yet. The ratio was christened to honor that assumption. If you have information, you should produce alpha. If you produce consistent alpha, you must have had information.
This is where the trouble begins.
Information Was Never the Right Word
The word information, in its older and richer sense, refers to something that reduces uncertainty. If I tell you the dice rolled a four, you now know something you did not know before. The information has content. It changes your understanding of the world. Claude Shannon built an entire branch of mathematics around this idea, and the elegance of his framework is that information is defined by surprise. The less expected the message, the more information it contains.
By that standard, most of what active managers act on is not information at all. It is interpretation. It is opinion. It is hunch dressed up as analysis. It is the same earnings report everyone else read, filtered through a slightly different valuation model. The market is not a quiet room where some people have access to messages others do not. It is a noisy auditorium where everyone has the same loudspeaker pointed at them, and the question is who can hear the signal through the static. Calling that information is generous. Calling it the basis for a ratio named after information is closer to branding than to science.
What the Information Ratio is really measuring is not the quality of someone’s information. It is the consistency of their deviation from a benchmark, scaled by how bumpy that deviation is. That is a real and useful thing to measure. But it has almost nothing to do with whether the manager actually knows something the market does not.
The Quiet Tautology
Here is the part that bothers me when I think about it too long. The Information Ratio assumes that consistent outperformance must come from information, and then uses outperformance as the evidence that information was present. It is a closed loop. You cannot tell from the ratio itself whether a manager has genuine insight or has simply been making a structural bet that happens to pay off across the measurement window.
Consider a manager who quietly overweights small cap value stocks against a broad market benchmark. For years at a time, this might generate steady excess returns with controlled tracking error. The Information Ratio looks beautiful. The tear sheet glows. Allocators write checks. But what the manager is actually selling is not information. It is a factor tilt. The manager has not discovered anything. They have simply chosen a corner of the market that has historically rewarded patience, and they have packaged it as skill.
Now consider another manager who genuinely does original research, talks to suppliers, builds models nobody else has built, and occasionally makes a brilliant call. This manager might have wild swings. They might be early. They might look wrong for two years before being spectacularly right. Their Information Ratio could be mediocre or even ugly, because the denominator punishes them for the very volatility that comes from holding contrarian positions long enough to be vindicated.
Which manager actually has more information? The honest answer is the second one. Which manager has the better Information Ratio? Probably the first. The ratio rewards the appearance of knowing, not the substance of it.
The Tracking Error Trap
The denominator of the Information Ratio is tracking error, which sounds technical but really just means how much the portfolio drifts from the benchmark. The lower the drift, the higher the ratio for any given level of outperformance. This creates a subtle and corrosive incentive. Managers learn that the safest way to a high Information Ratio is not to take bold positions and be right, but to hug the benchmark closely while adding small, persistent tilts.
This is the closet indexing problem dressed in a different costume. A manager who holds something very close to the benchmark with minor adjustments can produce a respectable Information Ratio with very little actual investment thinking. The math practically forces it. Tracking error is in the denominator, so reducing it makes the ratio look better even when the underlying skill is identical or worse.
The irony is that the ratio designed to identify skilled active managers ends up rewarding managers who behave passively while charging active fees. The more a manager hugs the index, the easier it is for them to look good by this metric. The more they actually try to be different, the harder it becomes. We have built a measurement system that penalizes conviction and praises caution, then we wonder why so much active management has quietly become expensive indexing.
Time, Sample Size, and the Illusion of Skill
There is another problem, which is that the Information Ratio is extraordinarily noisy over the time periods most investors actually use to evaluate managers. To have any real statistical confidence that a manager’s Information Ratio reflects skill rather than luck, you typically need decades of data. Most allocator decisions are made on three to five years of returns. Most career trajectories are decided on even shorter windows.
This means the ratio is regularly used as if it were a precise instrument when it is closer to a foggy weather vane. A manager with a great five year Information Ratio might be skilled, or might have been on the right side of a regime that is about to end. A manager with a poor five year ratio might be terrible, or might be a thoughtful contrarian whose thesis has not yet played out. The number itself cannot tell you which is which. It can only tell you what has happened, and even that comes with a margin of error wide enough to drive a truck through.
Yet we treat it as oracle. Investment committees pull up Information Ratios in PowerPoint slides and discuss them as if they were measurements of inherent quality, like a runner’s personal best or a violinist’s tone. They are nothing of the sort. They are summary statistics from a tiny sample of an extraordinarily complex system, dressed up in a name that suggests certainty.
The Deeper Lesson
The misnaming of the Information Ratio is a small example of a larger problem in how we talk about investing. We borrow words from rigorous fields, attach them to noisy measurements, and then act as if the rigor of the original word transferred to the new use. Information becomes a property of returns rather than a property of knowledge. Risk becomes a property of variance rather than a property of consequence. Alpha becomes a residual in a regression rather than a meaningful description of insight.
Each of these moves makes the field feel more scientific than it actually is. And in feeling more scientific, it becomes more dangerous, because we trust numbers more than we should and we trust the names attached to them most of all. The Information Ratio is a calculation. It is a useful one, in context, with caveats. It is not a measure of how much the manager knows. It never was.
The next time you see one quoted, remember what it actually measures. Remember what it cannot tell you. And remember that the word information, in its richest sense, refers to something that changes what you believe about the world. By that standard, the most informative thing about most Information Ratios is the gap between what they claim to measure and what they actually do.
That gap, more than any number, is worth paying attention to.
