Kbg wrote: ↑Sat Nov 28, 2020 10:50 am
Readings on dual momentum...read both the original and the rebuttal
Original:
https://blog.thinknewfound.com/2019/01/ ... entum-gem/
Rebuttal:
https://dualmomentum.net/2019/01/17/whi ... entum-gem/
You will learn a TON by reading these two articles. I was persuaded by the original and made adjustments to my approach accordingly. In practice what I have found using segmented momentum ranging from 3-12 months is that transitions between assets are no longer binary. It's more gradual in and out and the way I do it I can be holding 1-3 of the assets at the same time. Sticking with classic dual momentum is much easier to execute. What I do requires programming and data to pull off.
Ok I read through the articles. Yes those were really good articles. I remember watching Ocho despair here back in Q4 2018 when GEM fell apart. This is part of why I never really considered GEM, even before Q4 2018 it just seemed too fragile to me. It's not the actual dual momentum piece between US and Intl that is the issue, that performs just fine. It's the "risk-on/risk-off" comparison of the S&P to T-Bills, as well as the only 1x per month trading that leaves a big window open for bad things to happen. I would never run a quant strategy that did not have some form of a daily risk-on/risk-off signal. This I think is where the vol curve shines, being that daily signal to get you out when the market is crashing at the beginning or middle of the month. The monthly economic and trend signal from SpyComp is just a backup in case vol fails. It makes for a much less fragile and much more robust system than GEM. You really have to fool 2 systems here in order to sit through a crash and sell a market bottom like GEM did in Q4 2018.
Putting the risk-on/risk-off part aside. If we actually look at momentum as it is in my portfolio, just from an alpha generation add-on you get some interesting results.
Doing just vanilla DM since 1998 (For reference SPY alone was 7.37% CAGR)
10 month: 7.99% CAGR
https://www.portfoliovisualizer.com/tes ... odWeight=0
12 month: 7.2% CAGR
https://www.portfoliovisualizer.com/tes ... odWeight=0
Average 3, 6, and 12 month: 8.32% CAGR
https://www.portfoliovisualizer.com/tes ... odWeight=0
So doing the average of 3, 6, and 12 actually out performed both of the single periods. I also think fundamentally it's more robust. Other ways one can diversify to make this more robust? Well lets use that 3, 6, 12 and add some extra stuff in. So we start with the benchmark of US/Intl dual-momentum, since 1998 with 8.32% CAGR and 27 total trades in that timeframe.
Tri-Momentum, pick 1 US/Intl/EM: 10.59% CAGR with 48 total trades. So significantly more trades, about 2 trades per year average, but much higher CAGR.
https://www.portfoliovisualizer.com/tes ... odWeight=0
Ok, so what if we want to make this even more robust by adding in some diversification? Like holding only 1 thing is easy, but it is admittedly fragile. You're all in on one thing.
IJH, IJR, IVV, QQQ, VTMGX, VGSIX, VIVAX VEIEX, IEF, GLD, FXI pick top 4:
since 2006 11.02% CAGR 117 total trades:
https://www.portfoliovisualizer.com/tes ... odWeight=0
Now to be fair, not every one of those 117 trades is a full 100% portfolio turnover like it is in the other ones. We are doing pick top 4, so each of those trades is at least 25%, at most 100%. We also have a bunch of added diversification to the tri-momentum including gold, tech, bonds, real estate, mid caps, small caps, china, etc. The benefit here is that we have diversity in 4 picks each month, so if #1 underperforms but #4 out performs, it helps balance things out.
So basically, we started with single time period look backs, then went to multiple time period look backs averaged out to eliminate anomalies. Then we added in emerging markets to add a bit more diversification. Then if we want further robustness and diversification, we can take that and do a top x across a diverse basket of assets. Each step reduces fragility, adds diversification, and improved returns. Now, you take something like the tri-momentum or the take top 4 momentum and you add the vol curve and SPYComp (or GTT Trend) on top, and boom you got a 19-20% CAGR portfolio. Then add in leverage to you reach your desired risk/return flavor. There's so much that can be done in the momentum layer to improve return, diversity, and robustness.