Trend following is the simple idea of buying investments that have risen in value and either selling or avoiding those that haven't. However, the implementation details differ among specific strategies.
There are broadly two popular types of implementations. One type, dubbed the 'time-series momentum' strategy, leans on an investment's absolute returns over some time frame. For example, the strategy with a one-year time frame that trades stocks would buy or hold stocks if prices were higher than a year ago. Funds can differ in how they react when prices are down, however. More aggressive funds would short stocks - that is, bet against them such that the investor makes money when stocks decline. But shorting seems risky to many investors because stock prices usually go up, so it's more common to take a softer approach and hold cash instead.
# Time-series momentum if investment_performance > 0: buy_or_hold() else: sell_or_stay_in_cash()
The other popular type of trend following strategy uses simple moving averages. If we again apply the strategy to stocks, it would calculate the average price of stocks over, say, the past year and buy (or continue to hold) if today's prices are above it.
# Moving average crossover avg = average(stock_prices[one_year_ago:]) if price_today > avg: buy_or_hold() else: sell_or_stay_in_cash()
Despite their simplicity, trend following would have generated impressive returns over the past century. The more aggressive form of time-series momentum strategy that shorts stocks would have generated 11% per year after subtracting trading fees, which is about 2% per year higher than what investors would have earned from holding stocks at all times. More importantly, the strategy would have delivered these higher returns while assuming significantly lower risk. The strategy's volatility would have been 9.7% annualized - nearly half of the roughly 18% volatility a buy-and-hold investor would have endured.
The strategy's performance applied to other asset classes would have been equally impressive, with studies supporting its efficacy in bonds, currencies, and commodities.
Such was the strength of evidence that trend following became the predominant strategy among a type of funds called commodity trading advisors (CTAs), who, as the name implies, used trend following to trade commodities such as wheat but also used it to trade financial instruments as well.
But as with many strategies that became popular, trend following began to lose its lustre. The following chart shows the average performance of 10 CTAs that use trend following.
Source: SG Trend Indicator
The CTAs gained 120% from October 2002 (when the index began tracking) until the end of 2010 for an annualized return of roughly 10% per year. But since 2010, the annualized returns have dropped by more than half to approximately 4.5% per year. Looking back at the 'Exhibit 1' figure, you'll notice that 2010-2016 period had the worst performance, at least when we look at the column that excludes fees from performance calculations. It's fair to look at the performance gross of fees because the fees were more onerous when trading was done in pen and paper.
Because of this decline in performance, it's fair to ask whether trend following is worth investing in going forward. To answer that, we need to understand why trend following worked in the first place and judge whether those reasons will continue to apply.
Researchers have posited two types of forces that cause markets to trend, and those are behavioural and economic.
Behavioural forces consist of psychological biases, including anchoring and herding. Anchoring refers to the tendency of humans to under-react to news. Suppose news breaks out such that it should raise a stock's value by 50%. Investors may realize the magnitude of the news slowly, raising the stocks' prices by only 20% on the first day and gradually increasing it after that. The 'herding' bias, on the other hand, needs no explanation. Herding supports trends because when prices go up, people interpret that as a sign of popularity and feel the temptation to join in.
Trends also emerge because of the tendency of economic agents to implement their policies gradually. Central bank actions are a prominent example of this. If Canada were to enter a recession, the Bank of Canada would try to put the economy on a growth path by encouraging commercial banks to lend. Lowering interest rates, for example, is one avenue of such encouragement. However, central banks tend to spread out rate decreases over several meetings, and commercial banks tend to respond to such incentives slowly as well. Economies thus tend to recover only gradually, and assets such as stocks and commodities also rise in small steps.
While I think the economic forces shaping trends remain potent, I'm less optimistic about whether the behavioural forces are as relevant today as they were two decades ago. Take anchoring, for example. Stock prices move far more on earnings news today than they used to back then. Systematic strategies are also much more popular today, and computers don't succumb to psychological biases.
Trend following must also contend with the reality that it has become too well-known. Other traders can anticipate what the strategies will do and front-run them - for example, traders can sell stocks when they're close to, but not quite, crossing below the moving average.
Still, the economic forces should support the formation of trends. Is that enough? No one knows. But if I had to guess, I would think that trend following will work in the future, but perhaps not as well as what data from the past century would have you believe.
EDDY WEALTH