This is a standalone version of an article first published on quantstr.at. The Python and R code, the data files and the full tables are in the original.
TL;DR
- Strategy: Section 4.6 of 151 Trading Strategies holds the ETFs with a positive past return and weights them by return (Eq. 371), by return per unit of volatility (Eq. 372) or by return per unit of variance (Eq. 373).
- What we did: an attempt to replicate the rule on ten ETFs (US large and small stocks, developed and emerging markets, real estate, long and medium Treasuries, corporate bonds, gold, commodities) with SHV as cash, March 2007 to October 2026, net of 10 basis points on every dollar traded.
- Result: the three weightings returned 5.0% to 6.7% a year, against 10.9% for the S&P 500 ETF and 6.3% for holding all ten equally. Their worst falls were 26% to 27%, against 51% for the S&P 500 ETF.
- The weighting matters little: Sharpe ratios are 0.57 to 0.62 (zero risk-free rate), close to equal weighting (0.62). A 60/40 stock and bond mix had a higher Sharpe ratio (0.87).
- A 35% cap raised Eq. (371) to 7.7% and cut its worst fall to 17%, but it also holds more cash (8% of the portfolio on average, against 3%), and we chose it after seeing the results.
The rule
At each month-end, with a 12-month look-back, compute each fund's cumulative return Ri and keep the funds with a positive return and a price above their 200-day moving average. The weights are proportional to
and are scaled to sum to 1. Here σi is the volatility of daily returns over the previous 252 trading days. If no fund qualifies, the portfolio holds the Treasury bill ETF.
Method
- Data: daily Yahoo Finance prices adjusted for splits and distributions. The book gives no data or universe, so the ten ETFs are our choice, and the first month with a full 12-month return for all ten funds is February 2007.
- Timing and costs: we trade at the month-end close, earn the following month and charge 10 basis points on every dollar traded, including the SHV leg and the trades that restore drifted weights.
- Disclosed choices: 252-day volatility, the 200-day filter and a 35% cap are ours; the book gives ranges or no value.
Results
| Portfolio | Return, % a year | Max drawdown, % |
|---|---|---|
| Eq. (371) | 6.7 | -26.4 |
| Eq. (372) | 5.9 | -26.7 |
| Eq. (373) | 5.0 | -27.1 |
| Equal weight, ten ETFs | 6.3 | -33.2 |
| S&P 500 ETF | 10.9 | -50.8 |
| 60/40 stocks and bonds | 8.1 | -29.5 |
- Less risk, less return. Trend following roughly halved the worst fall against the S&P 500 ETF and gave up 4 to 6 points a year. US stocks outperformed the diversified mixes in this period.
- Not clearly better than holding everything. Only Eq. (371) beat equal weighting on return, by 0.4 points.
- Robustness. Across 18 settings (look-back 6, 9, 12 months; filter on or off; three weightings) returns ranged from 4.6% to 8.1%. The 12-month look-back did worse than 6 months in all 6 comparable pairings, so the main choice was not the best.
- Execution. Trading at the next day's close instead of the signal close gave 5.9%, 5.1% and 4.2%.
- The cap. Because fewer than three funds sometimes qualify, a 35% cap forces cash. Part of the lower drawdown comes from that lower exposure.
Limitations
- Ten ETFs of our choice and one 19-year history with one major crash; survivors only.
- Flat 10 basis point costs, no taxes; the R code reproduces the 18 main settings, while the cap, next-day and cost tests are Python only.
- This is an attempt to replicate; the book gives the rule but no results. Not investment advice.
Conclusion
Under our implementation, multi-asset trend following reduced the worst loss but also the return, and it was not clearly better than holding all ten funds or a 60/40 mix. The three weightings behaved similarly.
References: Kakushadze, Z. and Serur, J. A. (2018). 151 Trading Strategies. SSRN: https://ssrn.com/abstract=3247865, Chapter 4, Section 4.6.
The full article, with the Python and R code: https://quantstr.at/quant-strategies/multi-asset-trend/?utm_source=devto&utm_medium=syndication
If you want the next strategies from the 151 Trading Strategies series, follow @quantstrat here.
Educational content, not investment advice. Backtested results are not a forecast of future returns.
Top comments (2)
Here's the Python code: quantstr.at/quant-strategies/multi...
Here's the R code: quantstr.at/quant-strategies/multi...