What Alpha Means, and Why It Depends on the Benchmark

The idea

A fund returned some amount. Part of that is explained by the market having risen, and part is not. Alpha is the name for the part that is not.

The reasoning behind it is sound. If a fund holding Indian shares rises in a year when Indian shares rose, most of the result is the market, and crediting the manager with all of it would be absurd. So you subtract what the market would have delivered, adjusted for how much market exposure the fund actually carried, and whatever remains is attributed to the manager.

Positive alpha means the fund did better than its market exposure explains. Negative means worse. The concept is worth having, and everything difficult about it follows from one word in the definition.

The benchmark does the work

To say what the market would have delivered, you must first say which market.

That is a choice, and it decides the answer. A fund holding mostly mid-sized companies, measured against an index of large ones, will show positive alpha in every period when mid-sized companies outperform — and it will show it whether or not anyone at the fund made a good decision. The alpha is measuring the mismatch between the fund and the benchmark, not the skill of the manager.

Measure the same fund against an index that matches what it actually holds, and much of that alpha usually disappears. Nothing about the fund changed. The reference point did.

This is why the benchmark is the most important number in an alpha calculation and the one least often examined. It also explains a pattern worth recognising: benchmarks that flatter tend to persist while they flatter.

Three more reasons a positive alpha may not be skill

It may be exposure with a different name. Beyond company size, funds differ systematically in whether they favour cheaper companies, faster-growing ones, or steadier ones. Those tilts have their own long-run behaviour, and a fund tilted towards whatever did well will produce alpha against a benchmark that has no such tilt. Whether such tilts genuinely earn anything, and what they cost to implement, is how factor investing works and factor investing, liquidity and turnover.

It may be luck. Among enough funds, some will produce positive alpha over any period for no reason at all. Distinguishing that from skill needs far more evidence than one fund's record over one period, and the arithmetic of how easily chance produces impressive-looking results is in data mining and backtest overfitting.

It may be risk that had not yet arrived. A strategy that earns a little extra most of the time and loses heavily rarely will show steady positive alpha until the rare event happens. Measured over a window without the event, it looks like skill. This is not a hypothetical shape; it is what several categories of strategy look like by construction.

What alpha is genuinely useful for

Read carefully, it is a good diagnostic.

As a question rather than a score. A fund with persistent alpha against a well-matched benchmark raises a real question: what is it doing that the benchmark is not? If there is a clear answer — a specific tilt, a structural advantage, a cost advantage — you have learned something. If nobody can say, you have learned that too.

For spotting a mismatched benchmark. Large and consistent alpha is more often a sign that the benchmark is wrong than that the manager is exceptional. That is a useful thing to be able to detect.

And for what it costs. Alpha is normally quoted before the fund's charges, or against a benchmark you could have bought cheaply. The comparison that matters to you is against the fund you would otherwise have held, after both funds' costs — and a charge advantage is certain in a way alpha never is: how investment fees reduce wealth.

The honest gap

This page makes no claim about how much alpha Indian funds have delivered, or how many have delivered it consistently.

Answering that needs fund-level return history including funds that closed, matched to appropriate benchmarks, over several periods. This site holds index and inflation history only, so no figure appears above and no proportion of funds is quoted. The reasons that question is hard even with the data — survivorship, benchmark choice, the number of funds you were choosing from — are set out in why past performance is not enough to choose a fund.

What to take away

Alpha is return that a chosen benchmark cannot explain. The phrase "chosen benchmark" carries the whole of the difficulty.

When you are shown an alpha figure, ask what it was measured against, whether that reference matches what the fund actually holds, and whether the figure is before or after charges. If the benchmark does not match the holdings, you are looking at a description of the mismatch — which is information about the comparison, and not about anybody's skill.