Price Action Lab Blog

Premium Market Analysis

Trading Strategies

Two Billion Dollar Trend-Following Strategy

We look at the trend-following strategy employed by PTLC ETF and try to identify its strength and weakness.

Net assets of the Pacer TrendPilot® US Large Cap ETF are close to $2B. The strategy is detailed in website of Pacer Financial, Inc and briefly it is as follows:

  • 100% in S&P 500 total return index, if the index closes above its 200-day moving average for five consecutive days.
  • 50% in S&P 500 total return index and 50% in 3-month US T-Bills if the index closes below its 200-day moving average for five consecutive days.
  • 100% in 3-month US T-Bills if while below its 200-day moving average the index also closes lower than its value from five days ago.

The first step involves writing code for the strategy and checking against PTLC ETF results. Then we extend the backtest with longer history and analyze the results.

Code check

Below are the results of our backtest and comparison to actual PTLC ETF performance. We use SPY total return and BIL ETF for T-Bills and there may be slight differences. The period of the backtest is from 06/12/2015 to 02/22/2019.

Results for CAGR, Maximum DD, Volatility, Sharpe, Sortino and MAR (CAGR/Max. DD) are very close. We attributed the minor differences to our use of SPY ETF.

Extended backtest

The main objective of the extended backtest is to compare the strategy performance to the performance of SPY total return. The period of the backtest is from 05/30/2007 to 02/22/2019.

Below is a performance comparison table.

Parameter Strategy SPY TR 50/200 cross
CAGR 7.81% 7.42% 7.41%
Max. DD -17.31% -55.2% -19.4%
Sharpe 0.75 0.37 0.63
MAR 0.45 0.13 0.38
Return 142% 132% 132%

We have also included the performance of the long-only 50/200 day cross strategy. It may be seen that the strategy used for PTLC outperforms both buy and hold and 50/200 cross on both absolute and risk-adjusted returns basis in the test period.

First conclusion: The PTLC strategy performance is attractive.

In fact, this looks like a good strategy. Note that it is not strange at all that the strategy is fully disclosed. A total expense of 0.60% is justified in context of handling all the practical issues from following such a strategy and actually transforming tactical investing to a higher level of passive investing. Note that for most individual investors the transition from strategic to tactical is not trivial. Tactical investing is not even trivial for traders since it may be affected by loss of discipline, especially in the case of trend-following strategies. In simpler words, you pay 0.60% someone else to do the dirty work and in my opinion the fee is low compared to what some hedge funds charge that may use even inferior strategies.

What about statistical significance?

Below is a simulation of 20K random traders (no commission for comparison purposes) in the extended backtest period. The random traders use a fair coin to go long SPY when head show up and exit the position when tails show up.

The net return of the PTLC strategy of 142% ranks at 8.6%, meaning that 8.6% of random traders realized higher returns than the strategy while ignoring other metrics. While this is higher than the 5% we usually expect for significance, it is nevertheless pretty close. Based on this test we cannot reject the strategy as insignificant. Note that we do not use t-stats or other naive metrics because we do not know how the strategy was developed in the first place, i.e., whether it was the outcome of multiple comparisons or a unique hypothesis.

Second conclusion: The PTLC strategy is not statistically non-significant.

What could go wrong?

Since the strategy shows good potential based on an extended backtest that included a major bear market and several corrections and at the same time statistical significance is not all that bad, what remains is to think of some scenarios that could cause it to break down. An example of an ominous development that is in general detrimental to all trend-following strategies is an extended sideways market, such as the one that occurred in emerging markets from 2011 to 2015. Below we repeat the backtest but we use EEM ETF in place of SPY.

As it may be seen, extended whipsaw triggers an unstable mode and results are terrible. If something similar occurs in US market, then not only PTLC but all trend-following strategies will encounter problems. The probability is small but still finite but we cannot know its magnitude.

Third conclusion: The PTLC ETF strategy could suffer from an extended period of whipsaw that usually affects all trend-following strategies.

Summary

As the above analysis showed the strategy employed for the PTLC ETF is sound. The only potential weakness could be a change in dynamics of the US stock market. If bear markets continue to exhibit similar dynamics as in the past, then the strategy may also continue to perform well.  In the low probability event that the dynamics change and extended whipsaw emerges, then the strategy in its current form may face difficulties depending on specific path the market follows in time domain.


If you found this article interesting, I invite you follow this blog via any of these methods below:

RSS or Email, or follow us on Twitter

If you have any questions or comments, happy to connect on Twitter: @mikeharrisNY

Disclaimer

Market signals from systematic strategies are offered in our premium Market Signals service. For all subscription options click here.

Copyright Notice

Leave a Reply

Theme by Anders Norén