When Does Trend Following “Kick In”?

BY DR. RUFUS RANKIN, DIRECTOR OF RESEARCH, EQUINOX FUNDS


October 2018 was a very difficult month, and managed futures* investors were severely disappointed as many trend-followers lost money along with equities. The increasing popularity of terms such as “crisis alpha” and “crisis risk offset” may have exacerbated confusion as to how managed futures may be beneficial during market downturns, leading some investors to think of managed futures as an equity market hedge.


In this note, we will show that historically, trend-following becomes increasingly negatively correlated to equities as the observation window lengthens (i.e. daily, monthly, quarterly) and that trend-following exhibits the strongest convexity (defined later) to equities at the quarterly horizon. This underlines what we have been saying at Equinox for years: managed futures are not a hedge, per se, but do provide some level of risk management during equity bear markets, when they often experience positive returns.

Correlation

Let’s first look at the basic or unconditional correlation between the S&P 500® (“S&P”) and the SG Trend Index (“Trend”) over various horizons:

The correlation between the two indices decreases monotonically as we go from daily to annual observations. This presents us with a clue as to what kind of behavior to expect from trend-following: daily and monthly correlations are close to zero, suggesting that Trend returns during negative days and months for stocks are almost equally likely to be positive or negative. For quarterly and annual observations, the correlation is much more meaningful and negative, suggesting that Trend is often, but not always, positive when the S&P has a losing quarter or year. 

If we zoom in on conditional correlations when the S&P has a losing month or quarter, we see an even stronger negative relationship. This translates into positive expected return for Trend when the S&P has a negative quarter. On the other hand, the correlation is positive when the S&P has a positive quarter, indicating that Trend returns are more likely to be positive than negative during positive S&P quarters.

Since January 2000, S&P returns for 24 out of 76 calendar quarters have been negative (approximately 1/3 of the time) with an average return of -7.8%. Trend was positive 14 of those 24 quarters (about 58% of the time) and had an average return of 2.9% when the S&P had a negative quarter. On average, when the S&P has had a negative quarter, Trend has made money. Trend also made money, albeit much less (0.3% vs 2.9%) on average when the S&P was positive. While this is not necessarily what a “hedge” would look like, it certainly shows that Trend has portfolio benefits and a tendency to provide some level of risk management when stocks have protracted losses. However, this analysis does not take the magnitude of loss into account; we discuss that in the next section.

Convexity

Artur Sepp defines convexity as “the beta coefficient of strategy returns to the square of returns on the benchmark. In this way, the convexity measures the dynamic risk of strategy performance in tails of the performance of the index.” Strategies with positive convexity tend to have strong positive performance when the benchmark has extreme returns. Trend exhibits almost no convexity relative to the benchmark when we look at daily returns but has strong positive convexity for quarterly observations. When we examine the scatterplot of quarterly returns for S&P and Trend, we see an asymmetric pattern that is sometimes called a “smirk” rather than a “smile:” Trend performs more strongly during large negative quarters for the S&P. This is exactly when investors would prefer that it “kicked in.”

S&P vs. TREND (Daily Observations)

Stocks and Trend are effectively independent at the daily scale.

S&P vs TREND (Calendar Quarters)

Trend kicking in when most desired (S&P suffers a large negative quarter).

Rationale

Why does Trend behave in this fashion? A very simple explanation is that most trend-following strategies, by definition, need time to change the direction of their positions. Depending on how things play out, a trend-follower could take several weeks to flip from long to short. Trend-followers will often lose money on long equity positions in the early stages of a bear market. If the bear market continues to develop, trend-followers will reduce their longs and ultimately go short, and make money during the downturn.

As most readers know, trend-followers also trade bonds, currencies and commodities.  A slightly more nuanced explanation (explored in our previous piece: “Managed Futures During Equity “Crises” – An Update”) is that bear markets do not play out in a vacuum – investors withdraw their funds from equities and move them elsewhere. As a bear market deepens, capital flows tend to accelerate from “risk-on” assets into “risk-off” assets, and trend-followers can also earn profits from being long “risk-off” assets. These profits may even exceed those from short equity positions, as trend-followers tend to size their positions inversely to volatility.  In a volatile equity bear market, short equity positions may be smaller than long “risk-off” positions whose volatilities may be lower.   

Conclusion

While managed futures should not be viewed as a reliable hedge for equity risk, they have tended to offer material protection during equity bear markets or “crises,” as we can see from the performance of Trend during negative quarters for the S&P, with higher returns “kicking in” during large negative quarters.

The Sharpe ratio was calculated using the Annualize Rate of Return on 3-month T-Bills, which was 0.78%.



Definitions of Terms

No amount of diversification or correlation can ensure profits or prevent losses. An investment in managed futures is speculative and involves a high degree of risk. You can lose money in a managed futures program. There is no guarantee that an investment in managed futures will achieve its objectives, goals, generate positive returns, or avoid losses.

Past performance data quoted here represents past performance. current performance may be lower or higher than the performance quoted above. Past performance does not guarantee future results.