What Is the Momentum Factor? Lookback Windows and Crash Risk

The momentum factor ranks every stock in a universe by its own price return over a past lookback window, commonly six to twelve months, and holds the strongest performers. It is the most purely price based of the main factors: it uses no earnings, no book value and no view on the business, only the shape of the stock’s own trend.

It is also the factor with the sharpest personality. Momentum has the highest turnover of the standard factors, the most trading cost sensitivity, and the most violent short term reversals. Understanding how it is measured makes both the appeal and the danger obvious.

How momentum is defined and measured

A momentum score has three ingredients that must be stated explicitly.

The lookback window. How far back you measure the trend. Twelve months is the classic academic window. Six and nine month windows are also widely used, and Indian index methodologies often blend several windows so the score is not hostage to one arbitrary period. A shorter window reacts faster and trades more. A longer one is steadier and slower to exit a name that has turned.

The skip period. Almost every serious momentum definition ignores the most recent month. The reason is empirical: at very short horizons, returns tend to reverse rather than continue, so including last month mixes a reversal effect into a continuation signal. Writing the window as “twelve months, skipping the most recent one” is the standard convention.

The risk adjustment. Raw price return favours whatever is most volatile, because volatile stocks produce the biggest numbers in both directions. Many methodologies therefore divide the return by the stock’s volatility over the same window, producing a risk adjusted momentum score. This tends to reduce the portfolio’s tilt towards the wildest names in the market.

Two mechanical requirements sit underneath all of this.

Prices must be adjusted for corporate actions. A share split or a bonus issue cuts the raw price without changing anything about the investment. Feed unadjusted prices into a momentum screen and the stock will register a catastrophic decline on the ex date. Any credible calculation uses a series adjusted for splits, bonuses and dividends.

The universe must be liquidity screened first. An illiquid small company can post an enormous return on very little traded value. Momentum rankings computed without a liquidity filter fill up with names you cannot actually buy in size, which is a core theme in liquidity constraints in backtesting.

How to read a momentum score

A momentum score says one thing: this stock’s price has trended more strongly than its peers over the measured window. It is a statement about relative price behaviour, nothing more.

Read it with three qualifications.

It is relative, not absolute. A stock can rank in the top decile of a falling market while having lost money. This is the difference between relative strength and absolute momentum, and it matters a great deal for how a portfolio behaves in a downturn. Many practical implementations add an absolute filter, requiring the stock or the index to be above some trend level, precisely to avoid holding the best of a bad lot.

It decays quickly. A momentum ranking is a fresh measurement each rebalance. Names enter and leave frequently, and a stock that was top decile last quarter carries no privileged position this quarter. This is why momentum portfolios rebalance more often than value or quality portfolios, and why rebalancing frequency is a first order design decision here rather than a detail.

It is silent about the business. Momentum makes no claim that earnings are growing or that the company is well run. The most common explanation offered for why momentum persists is behavioural: information gets absorbed into prices gradually, and investors extrapolate recent performance. That explanation, if true, means the factor depends on collective behaviour continuing, which is not guaranteed.

Momentum is a bet that a trend, once established, tends to persist a while longer. It is not a bet that the trend is justified.

The risks

Momentum’s risks are unusually concrete, and they are the reason it is rarely run on its own.

Crash risk is the headline. Momentum’s worst episodes are not slow bleeds, they are sharp reversals. The classic pattern occurs when a market that has been falling turns abruptly upward. By that point the momentum portfolio has typically rotated into defensive names and away from the most beaten down, highest beta stocks. Those are exactly the stocks that rebound hardest in the turn, and the portfolio can lag the index badly over a few weeks. This is well documented in academic work on momentum in markets worldwide, and it is a structural feature of the strategy rather than bad luck.

Turnover is high and costs are real. Because the ranking refreshes constantly, momentum portfolios trade far more than other factor portfolios. Every rebalance incurs brokerage, statutory charges, slippage and impact cost, and for a taxable investor, capital gains events. A momentum result computed on closing prices with no cost assumptions can look meaningfully different once realistic frictions are applied. See transaction costs in backtests.

It buys after the move. By construction, a momentum portfolio buys stocks that have already risen. Some of those have risen because the market correctly repriced good news, and some because sentiment has run ahead of the business. The score cannot separate them, and the second group is where the pain concentrates when sentiment turns.

Parameter sensitivity invites overfitting. Lookback length, skip period, rebalance frequency, portfolio size and weighting scheme all move the historical result. It is very easy to search that space until the numbers look excellent, which is overfitting rather than discovery. A momentum rule that only works at one specific lookback is a warning sign, not a finding.

Crowding. Momentum is well known, widely published, and increasingly tracked by rule based products with predictable rebalance dates. That predictability can itself create trading pressure around those dates, a dynamic discussed in factor crowding.

What it does not tell you

A momentum score does not tell you why the price moved. Earnings upgrades, a sector rerating, a takeover rumour and a liquidity squeeze all produce the same input to the formula.

It does not tell you that the move will continue. Momentum is a statistical tendency observed across large groups of stocks over long periods, and it fails regularly at the level of individual names. Any single stock’s trend can end the day after it enters the portfolio.

It does not tell you anything about valuation. Momentum and value frequently point in opposite directions, which is one reason they are often combined. Neither corrects the other’s blind spot on its own.

It does not tell you about downside. A high momentum ranking says nothing about how far the stock could fall, which is why momentum portfolios are usually assessed alongside maximum drawdown rather than on return alone.

And it does not identify any stock as worth buying. Ranking a stock highly on a trend measure is a description of past prices, not a view on the company.

Related reading

• Portfolio metrics explained: the hub for the risk, return and factor measures referenced here.
• Factor investing in India: what a factor is, and how the main ones are defined.
• What is the value factor: the factor momentum is most often paired with, and most often at odds with.
• Relative strength explained: how relative strength differs from absolute momentum, and how it is computed.
• Factor cyclicality and drawdowns: why every factor, momentum included, goes through long uncomfortable stretches.

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.