Just a quick post today on the TAA bond strategy that I’ve posted on and keep track of here. Starting yesterday, the strategy is on AllocateSmartly. See the blog post discussing the details of the strategy and it’s historical performance. A few important items that I’d like to highlight follow.
First, the strategy is slightly different from the original that I posted here. The details are in the AllocateSmartly blog post. The ETFs are slightly different, the absolute momentum filter is applied against T-bills instead of using zero like I did, and the backtest results are much more robust.
The great thing about the strategy now is that there is a historical proxy from which to get better performance data. Here is the historical performance of the TAA bond strategy vs an equal weighted benchmark of bonds. Basically, higher returns, lower drawdowns.
And the portfolio statistics.
And even better now that the strategy is among many of the TAA strategies we can look at forming portfolios with it using it as a diversifying asset like bonds usually are. The strategy has low correlation with stocks, 0.19, and low correlation with other TAA strategies. For example, it has correlations of 0.27 and 0.34 with Antonacci’s GEM and Faber’s AGG3 respectively. Now lets stick those 3 into a portfolio and see what we get. This portfolio is 85% AGG3 and GEM equally split, and 15% TAA bond. What would the performance of that portfolio look like. See below.
And the portfolio statistics.
I chose this portfolio to have about a 10% max monthly drawdown. Again, benchmark beating performance, higher risk adjusted returns, and lower drawdowns. Of course, this is just one of many allocations and permutations you can create with the TAA bond strategy and other TAA strategies or even include quant strategies like I do.
That’s about it. The TAA bond strategy now has a proper home. It will be quite interesting to see how it performs if and when we do see a sustained period of higher interest rates. My expectations are that we will see lower returns to the strategy, how can it not from a lower absolute level of rates, but that it will also maintain it’s portfolio diversifying characteristics, and lower drawdowns than the buy and hold approach.
P.S. All data in the historical portfolio returns database has been updated to include the TAA bond strategy.
6 Comments
Steve · January 18, 2017 at 8:11 am
Paul
You also posted the bond TAA strategy using the top 1 performer, which, offers higher returns with a slightly higher risk. Will that also be tracked on Allocate Smartly?
paul.novell@gmail.com · January 19, 2017 at 6:11 am
Don’t know Steve. Don’t think so.
Paul
David E. McNamara · January 19, 2017 at 11:59 am
Paul,
In the case of a lump sum for investment would you favor a dollar cost average entry into the TAA or GEM strategies, or would you be all in from day one. My sense is that given valuation markers (CAPE, and others) as well as 8 positive years in the equities market it might be prudent at this time to dollar cost average in tenths over the next year or so. I would be interested in your thoughts on this issue. Thanks.
paul.novell@gmail.com · January 19, 2017 at 12:05 pm
David, historically the best strategy has been to invest 100% from day 1. Of course, that ignores the human component of investing. I think tenths is way too little at a time. I prefer 25% or 33% chunks but that’s up to you. CAPE and other valuation metrics are horrible indicators short term performance and the TAA strategies have built-in risk management anyway.
Paul
Matt · January 19, 2017 at 6:25 pm
Glad to see the bond strategy getting some much-deserved publicity, Paul. A couple (perhaps dumb, but not satisfied by me digging through a lot of posts) questions on your tactical approaches:
1. Is it safe to say that in your own TAA implementations, you’re now going exclusively with abs momentum instead of SMAs?
2. In implementing GTAA AGG3, are you always in the top 3 assets as long as they have positive results over the average of 1, 3, 6, and 12 months (i.e., you don’t have any other “filter”)? Seems that Faber’s original method was to calculate the average of 1, 3, 6, and 12-month returns and then to be in the top 3 as long as they were above their 10-month MA. Likely very similar but not quite the same thing, so I’m just trying to get it right. Many thanks.
paul.novell@gmail.com · January 20, 2017 at 6:49 am
Thanks Matt. As for your questions,
1. No. I’m using GEM and AGG3 as my two primary TAA strategies. GEM uses only abs momentum but AGG3 uses relative momentum plus the 10mo SMA as the ‘crash’ filter.
2. No. AGG3 ranks the ETFs by the avg performance over the last (1,3,6,12 mos.). I then looks at the top 3 by that ranking and only invests in them if they are above their 10 mo SMAs.
AllocateSmartly has both of these implemented which makes it super easy to follow.
Paul
Comments are closed.