Momentum rotation
Each month, hold the few stocks that have risen the most over the last several months, and switch as the leaders change.
4 min readLesson 53 of 57
In short
It bets that recent winners keep winning for a while, which has often held on average but fails sharply when the market turns. An "only hold stocks that rose" rule, called dual momentum, moves to cash when nothing is rising.
Why it matters
Stocks that have risen strongly over the last several months have, on average, tended to keep doing a little better than the rest for a while afterward. This pattern, called momentum, is one of the most studied in finance, and it has shown up across many decades and many countries' markets.
A momentum rotation turns that pattern into a simple rule: keep switching into whatever has been rising most. It's also one of the most painful strategies to hold. It can lag badly when the market's leadership suddenly changes, and knowing why helps you judge a backtest that looks too good.
How it works
You choose a list of stocks and how many to hold, N. At the start of each month, measure each stock's trailing return over a lookback of 3, 6 or 12 months. MarketLab counts a month as 21 trading days, so a 6-month lookback compares today's close with the close 126 sessions ago.
Rank the stocks by that return and hold the top N, putting 1/N of the money in each. Everything else is sold. Next month, rank again: a stock that drops out of the top N is sold, and the new leader is bought. The weights you enter for the stocks are ignored.
An optional rule, "only hold stocks that rose", adds absolute momentum to the relative ranking; together they're called dual momentum. A stock in the top N is skipped if its own trailing return is zero or below, and its share of the money waits in cash. When everything is falling, the strategy is fully in cash, which can soften long bear markets, at the cost of missing sharp rebounds.
Researchers often use 12-1 momentum instead: the return from 12 months ago to 1 month ago, skipping the latest month, because the very latest month has tended to partly reverse. MarketLab's Screener uses 12-1 momentum for its Momentum leaders preset; the backtest strategy uses the plain trailing return.
Score = close today ÷ close (21 × months) sessions ago − 1; hold the top N, 1/N each
Rank by recent return, and own an equal slice of each of the leaders.
A worked example
You run a 6-month momentum rotation on four stocks, holding the top 2, with $10,000 and "only hold stocks that rose" turned on. Today's closes, and the closes 126 sessions ago, are: A $120 and $100, B $45 and $50, C $66 and $60, D $82 and $80.
- Trailing returns: A 120 ÷ 100 − 1 = +20%; B 45 ÷ 50 − 1 = −10%; C 66 ÷ 60 − 1 = +10%; D 82 ÷ 80 − 1 = +2.5%.
- Ranked: A (+20%), C (+10%), D (+2.5%), B (−10%). The top 2 are A and C, and both rose.
- Each gets 1/2 of the money: $10,000 ÷ 2 = $5,000 in A and $5,000 in C.
- A month later, after a sell-off, the 6-month returns are A −4%, B −12%, C +3% and D −1%. The top 2 are now C and D.
- D's own return is below zero, so with the rule on it's skipped. C gets its 1/2 share and the other 1/2 stays in cash. A is sold.
Result: The strategy rotated out of A as it fell, and moved half the money to cash when only one stock was still rising. Turned off, the rule would have bought D anyway, putting all the money back in stocks during the sell-off.
How to read it
- Short lookback (3 months)
- Reacts quickly to new leaders, but trades more and is fooled by short bursts.
- Long lookback (12 months)
- Steadier rankings and fewer trades, but slower to leave a stock after its run ends.
- Small N (1 or 2)
- Concentrated: the result hangs on a few stocks, so it can soar or slump far more than the group.
- Lots of time in cash
- With "only hold stocks that rose" on, a long spell in cash means the chosen stocks were mostly falling. Check time invested next to the return.
In the real world
- Momentum was documented in a widely cited 1993 study by Narasimhan Jegadeesh and Sheridan Titman, and it has since been found in stocks, bonds, currencies and commodities. Many funds now offer momentum strategies, often built on 12-1 momentum and rebalanced monthly or quarterly.
- Its big weakness is the momentum crash. When a long fall suddenly reverses, the beaten-down stocks rebound hardest while the strategy still holds last year's safe leaders. In 2009, momentum strategies in US stocks lost heavily in a few months as the market rebounded.
- Dual momentum, popularized by Gary Antonacci, uses the absolute rule as a crude bear-market switch. Like other timing rules, it has sometimes cut drawdowns and sometimes just missed rebounds, depending on the period tested.
Common mistakes
- Testing on today's winners. Choosing stocks you already know soared, then showing momentum picked them, is hindsight, not evidence.
- Ignoring trading costs. Monthly rotation can trade a lot, and each switch costs a little.
- Trying every lookback and N until one wins. That's Overfitting; the gaps between 3, 6 and 12 months should be modest if the idea is sound.
- Expecting it to work every year. Momentum has had long dry spells and sudden crashes even when it worked over decades.
See it in MarketLab
Backtest: momentum rotation →
Opens a 6-month rotation among Apple, Microsoft, NVIDIA and Exxon Mobil, holding the top 2 that rose, compared with holding all four equally.
Screener: Momentum leaders →
The Momentum leaders preset lists the top fifth of stocks by 12-1 momentum that are still above their 200-day average.
Test yourself
Optional: 4 quick questions on the words, using the idea, and how it plays out in real markets.
Loading your answers…