Folio Lab/Documentation

Quintile Momentum

A non-MVO factor strategy that ranks the universe on a per-asset score over a lookback window and equal-weights the top quintile. The default scoring function is trailing-window momentum, directly implementing the long-leg of the classical Jegadeesh-Titman momentum factor.

Overview

Cross-sectional momentum is one of the most extensively documented anomalies in equity markets. Jegadeesh and Titman (1993) showed that buying recent winners and shorting recent losers earned positive risk-adjusted returns over 3-12 month horizons in US equities; subsequent work by Carhart (1997) added momentum as a fourth factor on top of Fama-French, and Asness, Moskowitz and Pedersen (2013) documented the effect across asset classes and geographies. In a long-only setting, simply holding the top quintile of past performers has historically outperformed equal-weighting the universe.

FolioLab's Quintile Momentum optimizer computes a per-asset score over a configurable lookback window (default 126 trading days, roughly six months), ranks the universe on that score, selects the top fraction (default 20% — the top quintile), and equal-weights the survivors. Two alternative scoring functions are exposed: a low-volatility score (negative trailing standard deviation, supporting low-vol factor portfolios) and an in-sample Sharpe ratio score.

The strategy is intentionally simple: no covariance matrix, no expected-return estimation beyond the score, no solver. This makes it numerically stable and insulates it from the optimizer-amplification problems that plague MVO (Michaud, 1989; DeMiguel, Garlappi and Uppal, 2009).

Mathematical Formulation

Notation

  • — price of asset at time
  • — lookback length in trading days (default 126)
  • — top-fraction parameter (default 0.20)
  • — cross-sectional score for asset
  • — size of the eligible universe (assets with complete lookback data)

Default score: trailing momentum

For asset the trailing-window cumulative return over the lookback window is

where is the optimization date. The classical academic specification skips the most recent month to avoid one-month reversal (Jegadeesh and Titman, 1993; Asness et al., 2013); FolioLab's default implementation uses the simpler unskipped form for transparency, but the lookback length is exposed as a research parameter.

Alternative scores

  • Low volatility: where is the trailing-window sample standard deviation of returns. Selecting the top quintile by picks the lowest-volatility names, replicating the long-leg of the low-volatility anomaly (Haugen and Baker, 1991; Clarke, de Silva and Thorley, 2006).
  • In-sample Sharpe: where is the trailing-window sample mean of daily returns. This is a quality-tilted variant; it reuses MVO inputs but does not pass them to a solver, so it inherits less estimation noise.

Selection and weighting

Select the top assets by score. Each selected asset receives equal weight:

Weights are long-only and sum to one by construction. There is no covariance matrix, no quadratic programme and no solver call.

Why equal-weight the survivors?

Equal-weighting inside the selected quintile is a deliberate choice. The cross-sectional momentum literature consistently shows that the alpha is concentrated in the rank of the score, not in fine-grained weighting differences within a given quintile. Imposing optimization-based weights inside the basket generally injects estimation noise faster than it improves expected return (DeMiguel, Garlappi and Uppal, 2009). FolioLab therefore keeps the inner step naive, and exposes the score function and lookback as the research levers.

For investors who prefer a vol-aware basket, the recommended pattern is to generate the quintile selection here and then pass the surviving universe to Inverse Volatility, Risk Parity, or Min-Variance for the within-basket weights.

Advantages & Limitations

Advantages

  • Simple: No solver, no covariance matrix, no inversion.
  • Empirically robust: Momentum has been documented across markets and decades.
  • Estimation light: Equal-weighting inside the basket avoids over-fitting weights to noisy moments.
  • Configurable signal: Swapping the score function (momentum, low-vol, in-sample Sharpe) lets the same framework span several factor styles.
  • Composable: The selected basket can be passed downstream to a risk-aware optimizer for finer weighting.

Limitations

  • Crash risk: Momentum portfolios suffer severe drawdowns in trend reversals (Daniel and Moskowitz, 2016).
  • Turnover: Periodic re-ranking can be expensive in small caps and in fragmented Indian universes.
  • No risk control: Equal weighting inside the basket ignores volatility heterogeneity.
  • Lookback sensitivity: Short-term reversal versus medium-term continuation is regime-dependent; a single lookback is rarely optimal across regimes.
  • Long-only: The short leg of the classical momentum factor is not available; the strategy captures only one side of the anomaly.

References

  • Jegadeesh, N., & Titman, S. (1993). "Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency." The Journal of Finance, 48(1), 65-91.
  • Carhart, M. M. (1997). "On Persistence in Mutual Fund Performance." The Journal of Finance, 52(1), 57-82.
  • Asness, C. S., Moskowitz, T. J., & Pedersen, L. H. (2013). "Value and Momentum Everywhere." The Journal of Finance, 68(3), 929-985.
  • Daniel, K., & Moskowitz, T. J. (2016). "Momentum Crashes." Journal of Financial Economics, 122(2), 221-247.
  • Fama, E. F., & French, K. R. (2008). "Dissecting Anomalies." The Journal of Finance, 63(4), 1653-1678.
  • Haugen, R. A., & Baker, N. L. (1991). "The Efficient Market Inefficiency of Capitalization-Weighted Stock Portfolios." The Journal of Portfolio Management, 17(3), 35-40. (low-vol score)
  • Clarke, R., de Silva, H., & Thorley, S. (2006). "Minimum-Variance Portfolios in the U.S. Equity Market." The Journal of Portfolio Management, 33(1), 10-24. (low-vol score)
  • Michaud, R. O. (1989). "The Markowitz Optimization Enigma: Is ‘Optimized’ Optimal?" Financial Analysts Journal, 45(1), 31-42.
  • DeMiguel, V., Garlappi, L., & Uppal, R. (2009). "Optimal Versus Naive Diversification: How Inefficient Is the 1/N Portfolio Strategy?" The Review of Financial Studies, 22(5), 1915-1953.