Research note: This material is for educational and strategy-design purposes. It is not investment advice or a promise of future returns.
Executive summary
Cross-sectional momentum ranks stocks against one another and buys relative winners while avoiding or shorting relative losers. The classic academic evidence comes from Jegadeesh and Titman and has been replicated internationally, while Nifty Indices’ momentum families show that volatility-adjusted intermediate-horizon momentum is also operationally relevant in India.
For India, the cleanest long-only version is usually preferable for most funds: select liquid NSE names from Nifty 200 or Nifty 500, rank them using six- and twelve-month returns excluding the most recent month, adjust for volatility, then hold the top decile or top 20–40 names with diversification and liquidity caps. A short-leg version is feasible mainly through a derivatives basket or SLBS-supported short book.
Description
This strategy seeks stock-selection alpha rather than market-timing alpha. In contrast with trend-following on the index, it can be fully invested even when the market itself is directionless, provided some stocks continue to outperform others. The trade-off is higher turnover and stronger exposure to crowdedness and reversals around market stress.
In India, the Nifty200 Momentum 30 methodology is a strong benchmark for practical design because it uses six-month and twelve-month returns adjusted for daily-return volatility, then tilt-weights by both factor score and free-float market capitalisation. That gives an official, investable template that is superior to naïve pure-return ranking for many allocators.
Key attribute table
The expectation range below is a practical inference from the academic momentum literature, Nifty momentum histories and implementation-cost research.
| Time horizon | Turnover | Typical Sharpe or return expectation | Data needs | Complexity | |---|---|---|---|---| | Medium term | Medium to high | Sharpe roughly 0.6–1.0; annualised return target roughly 12–20% gross | Stock-level prices, liquidity, corporate actions | Medium |
Details
Universe: default to Nifty 200, or Nifty 500 if capacity and execution allow. Exclude names below a minimum median traded-value threshold, for example ₹10–20 crore daily median value traded over the last three months. Signal: 0.5 × six-month return + 0.5 × twelve-month return, measured skipping the most recent one month to reduce short-term reversal contamination; divide by realised daily-return volatility over the prior year for a volatility-adjusted score, mirroring the spirit of Nifty200 Momentum 30. Rebalance monthly or quarterly; monthly is stronger on paper but quarterly often gives better net efficiency in India. Apply 5–7% max weight per name and 25% sector cap.
Risk controls should include beta monitoring, sector neutrality bands if the mandate is benchmark-aware, liquidity caps based on days-to-trade and one-name gap-risk limits around scheduled earnings. Position sizing can be score-proportional with volatility scaling, or equal-weight within the selected bucket. Data should come from official NSE price-volume archives and corporate-action files, with fundamentals not strictly required unless you add secondary filters.
Backtests must neutralise survivorship bias and corporate-action errors. Momentum signals are especially sensitive to stale-delisting assumptions, split adjustments and constituent-history mistakes. Edge cases include sharp regime reversals, upper and lower circuits, and momentum crashes following policy or liquidity shocks. Keep a turnover budget and a liquidity-constrained optimiser.
Implementation guide
- Build a liquid investable universe from NSE cash-market history and, if needed, F&O eligibility.
- Compute six-month and twelve-month total returns, skipping the latest month.
- Compute one-year realised volatility and convert raw momentum into a normalised score.
- Rank stocks, retain the top bucket and apply concentration and liquidity caps.
- Rebalance monthly or quarterly using next-day execution.
- Review turnover, crowding, sector bets and earnings-calendar gaps before each rebalance.
India-specific example
Suppose the eligible universe is Nifty 200 and the model is run on the last trading day of June. Sample names after screening might include TRENT, BHARTIARTL, SUNPHARMA, SRF, LT and DIVISLAB. Imagine TRENT has a six-month return of 28%, a twelve-month return of 46% and one-year realised volatility of 32%; the raw momentum blend is 37%, so the volatility-adjusted score is about 1.16. If LT has 8% and 15% returns with 20% volatility, its score is about 0.58. TRENT ranks above LT and receives a larger portfolio weight, subject to the name cap. The examples are illustrative ranking inputs, not live recommendations. Official market data, not vendor snapshots, should drive the production signal.
For a ₹50 crore long-only book holding 25 names, equal-weight sizing would start near ₹2 crore per stock before liquidity scaling. If TRENT trades with sufficient daily value, the PM can buy in cash during the regular session and hold until the next monthly rebalance. For a long-short overlay, any short in single names should be routed either through futures where available or through borrow made available under SLBS. Taxes and borrow fees must be included in net-return tests.