Swedroe: Trend-Following Strategies Work

As an investment style, trend-following, also referred to as time-series momentum, has existed for quite some time. Time-series momentum examines the trend of an asset with respect to its own past performance. This is different from cross-sectional momentum, which compares the performance of an asset with respect to the performance of another asset.

Academic research has provided consistent, long-term evidence that trends have been a pervasive feature of global markets, not just in equities but also among bonds, commodities and currencies. Carl Hamill, Sandy Rattray and Otto Van Hemert contribute to the literature with their August 2016 study, “Trend Following: Equity and Bond Crisis Alpha.”

Their study covered the period 1960 through 2015 and four asset classes: stocks, bonds, commodities and currencies. Following is a summary of their findings:

• Depending on the lag period chosen, the Sharpe ratios for trend-following strategies in stocks, bonds, currencies and commodities were generally higher, and in some cases dramatically higher (as much as 1.18), than the historical 0.4 Sharpe ratio for domestic equities.
• Performance was not only strong in the worst equity and bond market environments, but also in the best, revealing a well-known “equity smile” as well as a lesser-known, but even more pronounced, “bond smile.”
• Returns to trend-following strategies also tend to display a considerable amount of positive skewness. This is important because investors dislike negative skewness, which creates the potential for large losses. Typically, investors are willing to pay a significant premium to avoid negative skewness.
• The robustness of these findings is confirmed by similar results in both the pre- and post-1985 time periods, and by also examining results that excluded 2008, which produced the most extreme results.

Consistent Results

The preceding findings are consistent with those of Ian D’Souza, Voraphat Srichanachaichok, George Wang and Chelsea Yao, authors of a January 2016 study titled, “The Enduring Effect of Time-Series Momentum on Stock Returns Over Nearly 100 Years.” They found that the time-series momentum premium was persistent, pervasive across countries and robust to various formation and holding periods.

Hamill, Rattray and Van Hemert’s results are also consistent with those found by researchers at AQR Capital Management in a 2014 paper, “A Century of Evidence on Trend-Following Investing.” They constructed an equal-weighted combination of one-month, three-month and 12-month time-series momentum strategies for 67 markets across four major asset classes (29 commodities, 11 equity indices, 15 bond markets and 12 currency pairs) for the period January 1880 to December 2013.

They found performance that was remarkably consistent over an extensive time horizon that included the Great Depression, multiple recessions and expansions, stagflation, several wars, the global financial crisis of 2008, and periods of rising and falling interest rates. Furthermore, returns were uncorrelated with both stocks and bonds.

Biases Can Cause Trends
Researchers at AQR observed that “a large body of research has shown that price trends exist in part due to long-standing behavioral biases exhibited by investors, such as anchoring and herding [and I would add to that list the disposition effect and confirmation bias], as well as the trading activity of non-profit-seeking participants, such as central banks and corporate hedging programs. For instance, when central banks intervene to reduce currency and interest-rate volatility, they slow down the rate at which information is incorporated into prices, thus creating trends.”

The authors continue: “The fact that trend-following strategies have performed well historically indicates that these behavioral biases and non-profit-seeking market participants have likely existed for a long time.”

They noted that trend-following has done particularly well in extreme up or down years for the stock market, including the most recent global financial crisis of 2008. In fact, they found that during the 10 largest drawdowns experienced by the traditional 60/40 portfolio over the past 135 years, the time-series momentum strategy experienced positive returns in eight of these stress periods, and delivered significant positive returns during a number of these events.

AQR also noted its results were achieved even with a “2-and-20” fee structure. Today there are funds that can be accessed with much lower—although still not exactly cheap—expenses, including AQR’s own Managed Futures Strategy, AQMIX, which has an expense ratio of 1.21%, as well as the R6 version of the fund, AQMRX, which has a lower expense ratio of 1.13%. (Full disclosure: My firm, Buckingham, recommends AQR funds in constructing client portfolios.)

Additionally, AQR has found that its actual trading costs have been only roughly one-sixth of the estimates used for much of the sample period (1880 through 1992) and about one-half of the estimates used for the more recent period (1993 through 2002).

Before concluding, we’ll take a look at one more recent paper that specifically examined the performance of trend-following strategies during crises.

Time-Series Momentum During Crises

Mark Hutchinson and John O’Brien contribute to the literature on time-series momentum with their 2014 study, “Is This Time Different? Trend Following and Financial Crises.” Using almost a century of data on trend-following, they investigated what happened to the performance of the strategy subsequent to the U.S. subprime and eurozone crises, and whether what occurred was typical of what happens after a financial crisis.

Hutchinson and O’Brien observed that “identifying a list of global and regional financial crises is problematic.” Thus, they chose to use the list of crises from two highly cited studies on financial crises, “Manias, Panics, and Crashes: A History of Financial Crises” and “This Time Is Different: Eight Centuries of Financial Folly.”

The six global crises studied were: the Great Depression in 1929, the 1973 Oil Crisis, the Third World Debt Crisis of 1981, the Crash of October 1987, the bursting of the dot-com bubble in 2000 and the Sub-Prime/Euro Crisis beginning in 2007. The regional crises studied, with year of inception, were: Spain (1977), Norway (1987), Nordic (1989), Japan (1990), Mexico (1994), Asia (1997), Colombia (1997) and Argentina (2000).

The start date for each crisis was considered to be the month following the equity-market high preceding the crisis. Because neither of the two aforementioned studies provided guidance on the length or end date of each crisis, rather than attempting to define when every individual crisis ended, the authors instead focused on two fixed time periods: 24 months and 48 months after the prior equity-market high.

Their data set for the global analysis consisted of 21 commodities, 13 government bonds, 21 equity indexes and currency crosses derived from nine underlying exchange rates covering a sample period from January 1921 to June 2013.

Findings
The authors’ results include estimates of trading costs as well as the typical hedge fund fee of 2% of assets and 20% of profits. While they found that time-series momentum has been extremely successful over the long term, with their global portfolio from 1925 to 2013 posting an average net return of 12.1%, volatility of 11% and a Sharpe ratio of an impressive 1.1 (a finding consistent with that of other research), the predictability of future market returns tended to break down during crisis periods.

Specifically, Hutchinson and O’Brien found that, while in no-crisis periods, market returns exhibit strong serial correlation at lags of up to 12 months, subsequent to a global financial crisis, trend-following performance tends to be weak for four years, on average. This lack of time-series return predictability reduces the opportunity for trend-following to generate returns.

Comparing the performance of crisis and no-crisis periods, the average return in the first 24 months after the start of a crisis (4.0%) is less than one-third of the return earned in no-crisis periods (13.6%). Performance in the 48 months after a crisis starts (6.0%) was well under half the return (in that of no-crisis periods 14.9%).

They also found that the results were consistent across stocks, bonds and currencies. The sole exception was commodities, where returns were of similar magnitude in pre- and post-crisis periods. They found a similar effect when examining portfolios formed from local assets during regional financial crises.

Hutchinson and O’Brien noted that behavioral models link momentum to investor overconfidence and decreasing risk aversion, with both leading to return predictability in asset prices. Under these models, overconfidence should fall and risk aversion should increase after market declines, so it seems logical that return predictability would drop following a financial crisis.

Finally, it’s important to note, as the authors did, that “governments have an increased tendency to intervene in financial markets during crises, resulting in discontinuities in price patterns.” Such government interventions can lead to sharp reversals, with negative consequences for trend-following.

The authors concluded that the performance of trend-following strategies “is much weaker in crisis periods, where performance can be as little as one-third of that in normal market conditions.”

They write: “This result is supported by our evidence for regional crises, though the effect seems to be more short lived. In our analysis of the underlying markets, our empirical evidence indicates a breakdown in the time series predictability, pervasive in normal market conditions, on which trend following relies.”

Conclusion

The data from the aforementioned studies provide strong out-of-sample evidence, beyond the substantial evidence that already existed in the literature. It also provides consistent, long-term evidence that trends have been pervasive features of global markets.

In addressing the issue of whether we should expect trends to continue, researchers at AQR concluded: “The most likely candidates to explain why markets have tended to trend more often than not include investors’ behavioral biases, market frictions, hedging demands, and market interventions by central banks and governments. Such market interventions and hedging programs are still prevalent, and investors are likely to continue to suffer from the same behavioral biases that have influenced price behavior over the past century, setting the stage for trend-following investing going forward.”

Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.