- Trend following is highly popular.
- It's not magical.
- As a method, its use effectively exchanges one risk factor for another.
It is almost impossible to peruse a list of investment bestsellers without noticing that they seem to break down into two main categories -- value investing and trend following. The value investing approach pioneered by Graham & Dodd is widely emulated and needs no further introduction.
However, trend following, while popular, is not often objectively studied. Its adherents often take on the wide-eyed devotion of cult members, and its detractors dismiss it out of hand without coldly examining objective evidence. Where does the truth lie?
As always, the market provides a laboratory to test ideas. A thoroughly exhaustive study of the issue is impossible within the confines of an article. However, we can easily apply the most popular trend following approach to ETFs and examine if they have currency.
The most popular trend following system is the Golden Cross/Death Cross. The Golden Cross, as popularly followed, calls for buying on the following day's open when the 50-day simple moving average of closing prices crosses above the 200-day simple moving average of closing prices.
The security is held until a Death Cross occurs, which calls for selling and moving to cash on the following day's open when the 50-day simple moving average of closing prices crosses below the 200-day simple moving average of closing prices.
How has this popular rule worked recently in various markets? First, let us examine the SPDR S&P 500 Trust ETF (NYSEARCA:SPY), which tracks the S&P 500.
Since 2000, it's been a wonderful period for trend following the SPY. A higher return, with lower drawdowns than buy-and-hold. Applying the same rules to the iShares MSCI Japan ETF (NYSEARCA:EWJ) has been interesting.
The same pattern of "higher" returns and lower drawdowns occurs, but the return is pitiful, even though it beats the negative return of buy-and-hold on the EWJ ETF.
But we need to examine the rule applied to multiple markets to come up with even a tentative conclusion. Next up, we have the iShares MSCI Emerging Markets ETF (NYSEARCA:EEM).
A much lower return to risk when compared to buying and holding the EEM. What is going on here? The rule looked magical when applied to SPY and EWJ. We need to examine more markets. Remember, always collect more data. Looking at the Vanguard FTSE All-World ex-US ETF (NYSEARCA:VEU) paints an even more murky picture (apologies that we are using data from 2008 instead of 2000, VEU has less data).
Again, the popular trend following rule fails to beat buy-and-hold. So far, we have two instruments where the rule outperforms, and two instruments where the rule underperforms. We need more data. How does the rule do on the Vanguard REIT Index ETF (NYSEARCA:VNQ)?
The rule gives higher performance and a lower drawdown than buying and holding VNQ for the limited data we are examining.
Moving on to the long duration government bond market, iShares 20+ Year Treasury Bond ETF (NYSEARCA:TLT), is even more perplexing.
The rules, applied to the TLT ETF, gives us lower returns and a slighter higher drawdown. And so we need to gather more data in another market. Let's examine the rule applied to high-yield (junk) bonds, the iShares iBoxx $ High Yield Corporate Bond ETF (NYSEARCA:HYG).
I am shocked by how well the rule works on HYG. It achieves a slightly higher return, but massively lowers the drawdown of buying and holding HYG. Very odd.
Thus far, we have perused the rule with stocks, domestic REITS, and certain bond markets, but we have not yet examined commodity ETFs. Let's take a look at the rule applied to the popular gold ETF, SPDR Gold Trust ETF (NYSEARCA:GLD).
The rule gives us lower returns and lower drawdowns.
But what are we to make of this mixed bag of performance. What is the truth about trend following? To say that markets sometimes chop sideways and sometimes exhibit serial-autocorrelation (trend) is almost meaningless. It's akin to commenting, "Sometimes it's Winter and sometimes it's Summer." We are trying to navigate a harsh market landscape and need answers to help us deal with the future.
I would posit that a priori, it is impossible to tell if a market or asset class's returns will trend or chop sideways for any given decade. But counterintuitively, that does not make trend following or buy-and-hold methods useless.
Each strategy provides a conscious, explicit method to trade one set of risks for another. For the traditional buy-and-hold investor, tail risk is embodied by a multi-year secular bear market, such as Japan's multi-decade bear market. However, for the trend-follower, tail risk occurs not because of bear markets, but because of trendless markets which result in trend methods experiencing whipsaws, whereby each whipsaw results in ever-greater losses.
Secular bear market tail risk is a widely embedded, or "held" tail risk by buy-and-hold investors. Whipsaw is a tail risk held by trend followers. It is important to remember that strategies do not merely contain assets. They contain statistical or probabilistic risks (risk factors) which are just as important as the assets themselves.
When examined in this light, it is not actually statistically interesting to determine if trend following is a holy grail. It is statistically interesting to understand that by engaging in trend following vs. buy-and-hold, one is quite literally exchanging one set of risks for another. Strategies are not about trading assets -- they are about trading risk factors. And it appears that buy-and-hold risk factors are more widely held in portfolios than trend following risk factors. Markets have cruel methods for correcting such imbalances.