How Factor Investing Works

What a factor is

Kabir has been offered a fund that holds shares selected by a rule rather than by a manager's judgement or by size alone.

A factor is a measurable characteristic of a company or its shares that has been associated with different returns from the market as a whole. The characteristics usually named are things like how cheap a share is relative to what the business earns, how much its price has recently risen, how profitable the business is, how large it is, and how much it moves about.

A factor fund holds more of whatever scores highly on its chosen characteristic and less of what does not. The rule is written down and applied mechanically, which makes it closer to indexing than to active management — a different index, with a different rulebook — how an index fund works.

The two explanations, and why the difference matters

Where the extra return is supposed to come from has two competing accounts, and they carry completely different implications for whether it should persist.

The risk explanation. The factor identifies companies that are genuinely riskier in some way that ordinary risk measures do not capture, and the extra return is compensation for bearing that risk. If this is right, the premium should persist — because the risk persists — and it should arrive with the risk attached. You will be paid, and you will occasionally be paid for a reason.

The behavioural explanation. Investors systematically misprice these companies, through overreaction, inattention or preference, and the factor collects the mistake. If this is right, the premium is not compensation for anything and it should shrink once enough people know about it, because they will trade it away.

Both accounts are held by serious people and the evidence does not cleanly separate them. What follows for a buyer is the useful part:

Under the risk story, you should expect long stretches where the factor loses, and those stretches are the reason the premium exists. Selling during one is the mistake.

Under the behavioural story, you should expect the premium to decay as the strategy becomes crowded, and a factor that has been packaged into widely sold funds is by definition no longer obscure.

So the answer to "will this keep working" is not available, and anybody offering you one with confidence is telling you which story they believe rather than what is known.

Three problems that apply either way

The premium is not reliable over the periods people actually hold things. Factor returns are described over decades. Over five or ten years a factor can underperform badly, and the investor who bought it after reading about the long-run figure is the one who leaves during that stretch. A strategy you will abandon has no expected return worth quoting.

They are found by searching. A characteristic associated with higher returns is discovered by looking at data, and looking hard enough at any dataset produces characteristics that worked by chance. Distinguishing a real effect from a discovered coincidence is the central difficulty and it is the subject of data mining and backtest overfitting. The important consequence: a factor's published record is generally a backtest, and backtests are run by people who kept the versions that worked.

Implementation costs are real and are not in the backtest. A factor rule requires trading, and trading costs money that the historical simulation did not pay. The more turnover the rule demands, and the smaller and less liquid the companies it selects, the larger this gap — and it is measured on Indian data in factor investing, liquidity and turnover, which is the article to read before buying one of these.

What to ask before buying one

Which factor, defined how? The same-named factor is defined differently by different providers, and the definition decides what you hold.

What is the story? If the seller cannot say why the extra return should exist, the fund is a rule somebody found in data.

What did it cost to run? The turnover, and what the fund actually paid to trade — not what the backtest assumed.

What is the charge? A factor fund costs more than a plain index fund, and the extra is certain while the premium is not. That gap is the hurdle the factor must clear before you are ahead — how investment fees reduce wealth.

And how long is the record? A rule with a short live record and a long simulated one has mostly been tested against the period it was designed on.

The honest gap

This page quotes no factor premium for India, no historical return for any factor, and no proportion of periods in which any factor beat the market here.

Doing so needs return histories for factor-sorted portfolios of Indian shares over several cycles, which needs constituent-level data. This repository holds one index series and inflation. Premiums measured in other markets are not transferable — different companies, different investors, different costs — and borrowing them would be the sort of imported evidence this site refuses.

What to take away

Factor investing is systematic tilting: a different rulebook, applied mechanically, at a higher cost than holding the market as it comes.

Whether the extra return persists depends on which explanation is right, and that is genuinely unsettled. What is not unsettled is that the additional charge is certain, the trading costs are real, and the periods of underperformance are long enough that most buyers will not sit through them. If you want one, size it so that being wrong does not matter much, and decide in advance that a bad decade is the deal rather than a reason to leave.