How can trend following (intrinsic or absolute or time series momentum) beat the market, while ostensibly similar return chasing transfers wealth from naive to smart investors? In their January 2016 paper entitled “Return Chasing and Trend Following: Superficial Similarities Mask Fundamental Differences”, Victor Haghani and Samantha McBride offer a plausible and testable definition of return chasing and explore its differences from trend following. They characterize trend followers as mechanical and decisive and return chasers as discretionary and slow moving. For quantitative comparison, they consider three long-only, no-leverage strategies:
- 50-50 (benchmark): 50% equities and 50% U.S. Treasury bills (T-bills), rebalanced monthly.
- Trend following: 100% stocks (T-bills) when real stock market return over the past year is greater than (less than) 2.5%.
- Return chasing: increase (decrease) exposure to stocks each month by 20% of however much real stock market return exceeds (falls short of) 2.5% over the past year, holding the balance in T-bills.
They test these strategies with Robert Shiller’s long-run U.S. stock market data spanning 1871 through 2015 and with separately specified Monte Carlo simulation (5,000 runs of 20 years based on weekly simulated prices). Using these two approaches, they find that:
- For the Shiller data:
- Over the entire 145-year sample period, trend following (return chasing) outperforms (underperforms) the 50-50 benchmark by a gross 3.6% (1.2%) per year.
- Results are similar for the most recent 90-year and 40-year subperiods (though weaker for the last 40 years). Results are also similar using the MSCI World developed index during 1975 through 2015.
- Return chasing also performs poorly for look-back intervals or moving average windows of six months and two, three, four and five years.
- One interpretation is that trend followers front-run the slow adjustments of return chasers.
- However, turnover is ten times higher for trend followers than return chasers.
- Monte Carlo simulation indicates that trend following (return chasing) outperforms (underperforms) the 50-50 benchmark by a gross 7.3% (2.0%) per year, generating annual gross Sharpe ratio 0.62 (-0.18). Results are generally robust to varying simulation parameters.
In summary, evidence suggests that, despite superficial similarity, return chasing badly underperforms trend following.
Cautions regarding findings include:
- Per Robert Shiller’s description: “Stock price data are monthly averages of daily closing prices…” These data therefore blur past fixed-interval return calculations. Results for some of the above calculations based on end-of-month data may therefore differ. The long sample period and corroboration from a recent sample of end-of-month index readings mitigate this concern.
- Performance results are gross, not net. As indicated by different turnovers, accounting for trading frictions would narrow the gap between trend following and return chasing.
- A 60%-40% stocks-U.S. Treasury notes (10-year T-notes) benchmark may be more challenging for trend following than the benchmark used above. T-notes generally offer higher yields than T-bills and may contribute capital gains via the specified market timing.
- As noted in the paper, investors may adapt such that the “rent” paid by return chasers to other strategies dissipates in the future.
Frequently Asked Questions
What is the difference between trend following and return chasing in investing?
Trend following (also called time series or absolute momentum) is a mechanical, decisive strategy that moves fully into stocks or T-bills based on whether past-year real market return clears a threshold. Return chasing is a discretionary, slow-moving approach that gradually raises or lowers stock exposure after good or bad returns. In the Haghani and McBride study, trend following went 100% stocks when the prior-year real return exceeded 2.5%, while return chasing adjusted exposure by 20% of the excess or shortfall each month.
Does trend following beat a 50-50 stock and bond portfolio while return chasing does not?
Yes, in Shiller U.S. stock market data for 1871-2015, trend following outperformed a 50-50 stock/T-bill benchmark by a gross 3.6% per year, while return chasing underperformed it by 1.2% per year. Results were similar in the most recent 90-year and 40-year subperiods, though weaker over the last 40 years, and in MSCI World data for 1975-2015. Return chasing also performed poorly across look-back windows from six months to five years.
How are trend following and return chasing strategies tested against each other?
The study compares three long-only, no-leverage strategies using monthly rebalancing: a 50% stocks/50% T-bills benchmark, trend following, and return chasing, all keyed to a 2.5% past-year real return threshold. They were tested on Robert Shiller's U.S. stock data from 1871 through 2015 and on a separate Monte Carlo simulation of 5,000 runs of 20 years built from weekly simulated prices. The simulation showed gross outperformance of 7.3% per year for trend following and gross underperformance of 2.0% per year for return chasing, with Sharpe ratios of 0.62 and -0.18.
Why does trend following make money while return chasing loses money if both follow past returns?
One interpretation is that trend followers front-run the slow adjustments of return chasers, so wealth transfers from the slow, discretionary investors to the fast, mechanical ones. The two strategies are only superficially similar, because trend following acts decisively on a threshold while return chasing adjusts gradually. Trend following also has about ten times the turnover of return chasing, which matters for the cost of the strategy.
What are the limitations of the evidence that trend following beats return chasing?
The results are gross of trading costs, and trend following's turnover is about ten times that of return chasing, so accounting for frictions would narrow the gap. Shiller's stock prices are monthly averages of daily closes, which blurs fixed-interval return calculations, though the long sample and a recent end-of-month check mitigate this. A 60-40 stocks/10-year T-note benchmark may be tougher for trend following than the T-bill benchmark used, and investors may adapt so that the wealth transfer from return chasers dissipates.