Momentum trading. Buy what is already working.

Momentum is the most academically documented anomaly in finance. Jegadeesh and Titman's 1993 paper showed that stocks outperforming over the past 3 to 12 months continued to outperform over the next 3 to 12 months. The effect persists across countries, asset classes, and decades. Despite being widely known, it continues to work, possibly because it requires patience through occasional momentum crashes that cause many traders to abandon the strategy.

What momentum actually is

Momentum buys stocks that have been outperforming a universe over a recent period and avoids or shorts those that have been underperforming. The lookback period is typically 3 to 12 months, with 6 to 12 months being the most studied range.

The relative ranking matters more than the absolute performance. A stock up 10 percent in a flat market is showing strong momentum. The same stock up 10 percent in a market up 20 percent is showing weak momentum.

Most momentum strategies rebalance monthly. The top decile (or quintile) of performers gets bought. The bottom decile gets sold or shorted. The portfolio rotates as the relative rankings change month over month.

The academic evidence

Jegadeesh and Titman (1993). The original momentum paper. Showed that US stocks ranked by 6 month past returns continued to outperform over the next 6 months. The top decile beat the bottom decile by about 1 percent per month, before transaction costs.

Asness, Moskowitz, Pedersen (2013). "Value and Momentum Everywhere." Showed momentum effects across 8 different markets including US equities, international equities, country indexes, bonds, currencies, commodities. The effect is global.

Carhart (1997). Added momentum as the fourth factor to the Fama French three factor model. Momentum became a recognized risk factor in academic finance.

Antonacci (2014). "Dual Momentum Investing." Combined relative momentum (rank by past returns) with absolute momentum (above its own 12 month moving average) to produce a strategy that rotates to safe assets during bear markets.

The mechanics of a momentum portfolio

Universe.

Liquid stocks. S and P 500 for retail. Russell 1000 or 2000 for broader exposure. ETFs across asset classes for global momentum.

Lookback period.

6 to 12 months is the academic standard. Some practitioners use 3 to 6 months for faster rotation. Longer lookbacks are smoother but slower to adapt to regime changes.

Ranking.

Calculate total return for each stock over the lookback. Rank the universe from highest to lowest return. The top N stocks are the momentum portfolio.

Portfolio size.

10 to 30 stocks for diversification. Smaller portfolios have higher tracking error and more variance. Larger portfolios converge toward the index but smooth the variance.

Rebalancing.

Monthly is standard. Some implementations rebalance quarterly to reduce turnover. More frequent rebalancing increases transaction costs without meaningfully improving returns.

Weighting.

Equal weight across all positions. Or volatility weighted (smaller positions in higher volatility stocks). Or market cap weighted (closer to a tilt of the index).

The momentum crash

Momentum strategies suffer occasional crashes when market regimes change abruptly. The most extreme example was March 2009.

In late 2008 and early 2009, financial stocks had been the worst performers (bottom of the momentum ranking) and consumer staples had been the best (top of the ranking). The momentum portfolio was long staples and short financials.

When the market bottomed in March 2009 and financials ripped higher 100 to 300 percent over the next 6 months while staples lagged, the momentum portfolio took massive losses on both sides. Many momentum funds had double digit single month drawdowns.

Momentum crashes happen at regime turns. The trader who runs momentum has to accept that these drawdowns are part of the strategy. Avoiding them requires either the absolute momentum filter (Dual Momentum) or active discretion about regime changes, both of which have their own trade offs.

The retail implementations

Stock momentum portfolio.

Rank S and P 500 stocks by 6 month total return monthly. Hold the top 20. Rebalance monthly. Equal weighted.

Backtests show roughly 15 percent annualized returns with 20 percent volatility over multi decade windows. Outperforms the S and P 500 by a few percentage points per year before transaction costs.

Sector momentum rotation.

Rank the 11 GICS sector ETFs (XLK, XLF, XLE, XLV, XLP, XLY, XLI, XLU, XLB, XLRE, XLC) by 6 month return. Hold the top 3. Rebalance monthly.

Simpler than stock level momentum. Lower turnover. Lower commissions. Backtests roughly match the S and P 500 long term with slightly better risk adjusted returns.

Asset class momentum.

Rank a basket of asset class ETFs (SPY, EFA, EEM, GLD, TLT, IEF, DBC) by 6 to 12 month return. Hold the top 2 to 3. Rebalance monthly. Rotate to bonds or cash if the top ranked is below its 12 month moving average (Dual Momentum).

Lower correlation to any single asset class. Strong risk adjusted returns over multi decade backtests. The Antonacci dual momentum approach is the most widely studied variant.

Momentum ETFs.

For investors who do not want to manage the portfolio themselves. MTUM (iShares MSCI USA Momentum Factor ETF) is the largest momentum factor ETF. SPMO, FDMO, and others provide similar exposure.

The ETFs do the ranking and rebalancing automatically. Slightly lower returns than custom implementations due to the ETF structure, but eliminate the operational complexity.

Why momentum continues to work

Behavioral biases. Investors underreact to news initially (slow to update views) and then overreact (chase the trend). The combination produces momentum that takes time to fully price in.

Career risk for institutional managers. Underperforming a popular winner stock creates career risk for institutional managers. They eventually chase the winner to protect their job, extending the momentum.

Limited arbitrage. Momentum requires holding through periodic crashes. Many funds cannot tolerate the drawdowns and exit at the worst time, leaving the alpha for the patient.

Pure flow effects. ETF rebalancing and passive index reweighting create flows that amplify recent winners and pressure recent losers.

What kills momentum traders

Chasing without rebalancing. Buying recent winners but holding indefinitely. The strategy requires the discipline to sell positions as they drop out of the momentum ranking.

Selling during the crash. The momentum crash periods produce the worst drawdowns. Selling at the bottom and missing the recovery defeats the strategy.

Insufficient diversification. Holding 5 momentum stocks instead of 20 to 30. The variance is too high. The strategy needs the diversification to produce its long term returns.

Excessive turnover. Rebalancing weekly or daily instead of monthly. The transaction costs eat the alpha.

Combining momentum with trend following or mean reversion without rules. The strategies conflict on individual names. Without rules to manage the conflict, the trader produces noise instead of edge.

Where the audit fits

The audit reads the actual trades and identifies whether the trader is implementing momentum systematically or just chasing winners. For most retail attempts at momentum the pattern is buying winners without selling losers, which is buy and hope rather than momentum. The plan locks the systematic rules. Five to seven pages.