Guide · Portfolio construction · Stage 4

Smart beta and factor investing in India, honestly

Most explainers of smart beta stop at the brochure: here are the factors, here is a chart of how well they backtested, invest now. That is exactly the framing that gets people hurt, because the interesting and honest part of factor investing is the part the marketing omits. The premia are real in the data, but they are averages measured over decades, they take turns leading and lagging, they shrink once a signal is crowded, and live results trail the fitted backtest. This guide builds the mechanism from the construction rule up, gives each factor its academic thesis and its risk story, and then spends its weight on the cyclicality, because that is what decides whether a factor tilt is a discipline you can hold or a fashion you will abandon at the worst moment.

What smart beta actually is: the weighting rule

Start with an ordinary index. A market-cap weighted index takes a universe of stocks and gives each one a weight proportional to its market value, the share price times the number of shares outstanding. The mechanism is passive in the strict sense: the index does not decide anything, it simply mirrors how the market has priced each company. As a company grows in value its weight rises automatically, and the biggest names dominate. That is both the strength, it is self-adjusting and cheap to run, and a quiet bias, because a cap-weighted index is by construction most exposed to whatever is already the largest and often the most expensive.

A smart-beta index keeps the rules-based, transparent, low-cost character of that machinery but changes one thing: the weighting rule. Instead of weighting by size, it ranks the same universe by a factor, a measurable characteristic that research links to a distinct return stream, and weights constituents by that factor score. A value index weights toward the cheapest stocks on fundamentals. A low-volatility index weights toward the stocks with the smallest historical price swings. Nothing about the process is discretionary: the factor definition, the selection cut-off, the weighting scheme and the rebalance calendar are all published in advance and applied mechanically. The judgement was made once, when the rule was written, not stock by stock.

That is why the label sits where it does. Smart beta is passive in method, rules-based and hands-off, but active in intent, because it deliberately deviates from the cap-weighted market in the hope of a different and, over long horizons, better-compensated return stream. It will diverge from the index most people call the market, sometimes for uncomfortably long.

The five factors, their thesis, and why the premium might exist

A factor earns the name only if two things hold: the return pattern shows up across long samples and different markets, and there is a credible reason it should persist rather than being a fluke of data mining. The reason is always one of two kinds. Either the factor loads on a real risk, so the extra average return is compensation for bearing something painful at the wrong times, or it exploits a persistent behavioural mistake that other investors keep making. Which explanation is true matters, because a risk premium should survive being widely known while a pure mistake can be arbitraged away. Here are the five best-documented factors.

Value

The value factor buys stocks that are cheap relative to fundamentals, low price to book or low price to earnings, and underweights the expensive. Its academic anchor is Fama and French (1993), whose three-factor model added a size factor and a value factor, High Minus Low or HML, to the market and found they explained returns that market risk alone could not. The risk explanation is that cheap firms are often distressed or out of favour, so their extra return compensates the holder for bearing that distress risk. The behavioural explanation is that investors over-extrapolate, paying too much for glamour and too little for the unloved. Both can be partly true at once.

Momentum

The momentum factor buys recent relative winners and underweights recent losers, on the finding that performance over roughly the past three to twelve months tends to persist over the next few months. The canonical source is Jegadeesh and Titman (1993), who showed that a strategy of buying past winners and selling past losers produced returns that were hard to reconcile with an efficient market. The usual explanation is behavioural: investors under-react to news at first and then herd once a trend is established. Momentum is the factor most exposed to sharp reversals, because the same crowding that drives the trend unwinds violently when it breaks.

Low volatility

The low-volatility factor tilts toward stocks with the smallest historical price swings. Its interest is that it contradicts the textbook: standard theory says more risk should mean more return, yet lower-volatility stocks have historically not delivered proportionately lower returns, and on some measures the reverse. Baker, Bradley and Wurgler (2011) documented this low-volatility anomaly and proposed a structural cause: many investors cannot or will not use leverage, so they chase return by overpaying for high-volatility stocks, while professional managers judged against a benchmark are discouraged from arbitraging the gap. The premium survives because the arbitrage is constrained, not because it is riskless.

Quality

The quality factor favours profitable, low-leverage, stable firms over weak, indebted, erratic ones. Novy-Marx (2013) showed that gross profitability predicts returns with power comparable to the classic value measure, and Asness, Frazzini and Pedersen built the Quality Minus Junk factor around profitability, growth, safety and payout. Quality entered mainstream asset pricing when Fama and French added profitability and investment factors to their five-factor model in 2015. The thesis is that the market does not fully price durable business quality, so higher-quality firms are mildly underpriced, and quality often cushions drawdowns because sound balance sheets survive stress.

Size

The size factor, Small Minus Big in the Fama and French framework, holds that smaller companies have historically outperformed larger ones over long horizons. The risk story is that small firms are more fragile, less liquid and more sensitive to the cycle, so their extra average return is payment for that fragility. Size is also the factor whose historical premium has proven least stable and most sensitive to how the sample and the small-cap cut-off are defined, which is a useful reminder that a factor is only as robust as its definition. In practice size is frequently used as an input to multi-factor blends rather than on its own.

The honest core: factors are cyclical, not free money

This is the section the brochures skip, and it is the one that matters most. A factor premium is an average measured over a long sample, and an average is a summary of a wide distribution, not a promise about any particular year. Around that average sit long stretches, often several years, where the factor lags the plain market badly. That is not a bug. If a premium exists because the factor bears a genuine risk, then the periods when it loses are precisely the compensation mechanism at work: you are paid on average because sometimes it hurts. If the premium is behavioural, the mistake it exploits can persist or even widen for years before it corrects.

The clearest recent illustration is value. Having been one of the most celebrated factors for decades, value spent much of the 2010s lagging, a stretch widely called its lost decade. Research on factor performance over 2010 to 2019 found the classic value factor delivered a negative average over the period, as low interest rates and a technology-led market favoured growth, though the episode was not unprecedented and echoed a similar drought in the 1990s. An investor who bought value in 2010 on the strength of its long-run backtest had to endure roughly a decade of disappointment before any reversion. That is the emotional reality a factor chart never shows.

Two further honest points compound the cyclicality. First, premia are time-varying and can be crowded. Once a factor is documented and a wave of capital chases it, its future premium tends to shrink. McLean and Pontiff (2016) studied 97 published return predictors and found their returns were on average about 26 percent lower out of sample and about 58 percent lower after publication, consistent with investors learning the signal and arbitraging part of it away. The premium rarely disappears entirely, because limits to arbitrage remain, but the honest expectation for a famous factor is smaller and less certain than its backtest.

Second, live results lag backtests. A backtest is fitted to history with the benefit of hindsight and often ignores the frictions that eat real returns: trading costs at each rebalance, the market impact of moving size, the bid-offer on less liquid names, and the simple fact that the future sample is not the past one. A marketing chart that shows only the good years, on a curve that was optimised after the fact, tells you almost nothing about what you will actually earn. The correct response is not to abandon factors but to demand diversification across them and the patience to hold through the droughts, and to treat any single-factor, single-decade backtest with deep suspicion.

How it combines: single-factor, multi-factor, and timing

Given cyclicality, the natural question is whether to hold one factor or several. A single-factor index gives you the purest, fullest dose of one idea, and therefore the fullest dose of its long lean spells. A multi-factor index blends several factors so that when one lags another may lead, smoothing the combined ride. Indian index providers describe their multi-factor indices in exactly these terms, as a way to counter the cyclicality of a single-factor strategy by taking exposure to several factors through one product. The cost is that blending dilutes any one factor and introduces new construction choices, chiefly how the factors are combined and rebalanced.

The tempting alternative is to time factors, holding value when value is about to lead and momentum when momentum is. This is very hard, for the same reason timing anything is hard: the signals that tell you a factor is about to turn are weak and noisy, and by the time a rotation is obvious it is largely priced. The lesson is the same one that governs sector rotation in India, where the theory of moving between sectors ahead of the cycle collides with the practical difficulty of being early and right repeatedly. For most holders, diversifying across factors and rebalancing on a rule beats trying to predict which factor is next.

There is also an overlap caveat. Because factor indices are drawn from overlapping universes, holding several factor products, or a factor product alongside broad-market and thematic funds, can quietly concentrate you in the same names rather than diversifying you. A momentum tilt and a quality tilt can share many constituents in a given regime. The only way to know your true exposure is to look through to the holdings, the same discipline covered in the guide to mutual fund overlap analysis, which shows how apparently different funds can hide a single concentrated bet. Judging where a factor tilt fits in a whole portfolio, rather than admiring its backtest in isolation, is exactly the upstream work that the method we teach is built around.

The India context: judge the rules, not the backtest

Factor and smart-beta indices, and products that track them, exist on the Indian market across the familiar factors: value, momentum, low volatility, quality, an alpha construction, and multi-factor blends, alongside simpler equal-weight variants that reduce the mega-cap concentration of a cap-weighted index. Described generically, they work exactly as the mechanism above sets out: a published methodology defines the factor, ranks the eligible universe, selects the constituents and assigns weights, then rebalances on a set schedule such as semi-annually. Two indices carrying the same factor label can hold materially different stocks depending on how the factor is defined, how deep the selection goes and how weights are assigned, so the label alone tells you little.

That is why the right way to assess an Indian smart-beta index or the product that tracks it is on process, not on the performance chart. Four things carry almost all the signal. First, the rules: read the methodology and confirm you understand what it actually does at each rebalance. Second, the factor definition: how the factor is measured and scored, because that is where two same-named indices diverge. Third, the cost, since a factor's modest, uncertain premium has to clear the expense of capturing it before you see anything. Fourth, the liquidity of both the product and its underlying holdings, because a tilt you cannot enter or exit cleanly is a tilt you do not really own. Backtested return is the least reliable input of all, because it is fitted to the past and precedes the crowding that erodes it.

Where this sits in the curriculum

Factor investing, smart-beta construction and the honest treatment of factor cyclicality are covered in Stage 4, where the curriculum builds the quantitative research workflow, and are extended in the later portfolio-construction material into how factor tilts are combined and risk-budgeted inside a whole portfolio. The framing throughout is methodological: how the machinery works, how to read a methodology document, and how to size a tilt against its risk. The curriculum does not recommend any specific factor, product or allocation, and it treats every factor premium as historical and uncertain rather than promised.

Related reading