Happy New Year everyone! Time to start fresh again in 2016. Here is the first tactical asset allocation update for 2016. As I mentioned last month, I am now using a new data source for the portfolio updates and also going to use a slightly different format for these monthly updates. I will also maintain the old portfolio formats, in Yahoo Finance, for a while. Here is the link to the Yahoo data. Lets dive right in.
Below are the updates for the AGG3, AGG6, and GTAA13 portfolios. The source data can be found here. The big change here is the use of FINVIZ data and more importantly that these signals are valid after every trading day. So, while I’ll maintain these month end updates this means that you can implement your portfolio changes on any day of the month, not just month end. FINVIZ will at times generate signals that are slightly different than Yahoo Finance (due to the use of month end vs weekly data).
The AGG3 portfolio is invested in VNQ and MTUM and so is the AGG6 portfolio. That means the AGG3 portfolio is 33% in cash and the AGG6 portfolio is 66% in cash.
For the Antonacci dual momentum GEM and GBM portfolios, GEM is in SPY and the bond portion of GBM is in cash. I’ve also made my Antonacci tracking sheet shareable so you can see the portfolio details for yourself. Here is the data.
Performance data for 2015 is presented below for the various portfolios I track. The only portfolio I don’t have data for is the GTAA13 portfolio. I’m working on that.
It was a tough year for any diversified portfolio. It was one of the worst in the last 30 years for any diversified portfolio. See here for a brief roundup. The 60/40 US centric portfolio was the best performer with just a slightly positive return of 0.37%, which was below the performance of just being in 100% cash. The best proxy for the Global portfolio, GAA, returned -4.1% for the year. I’ll have more on 2105 portfolio performance as more data becomes available but overall it was a tough year world wide for most asset classes.
On another note, I’ve commented here before on the performance difference of the tactical portfolios when using monthly vs 4 week rebalancing intervals. I’ve noted this difference in the table above for the AGG3, AGG6, and GTAA5 portfolios. For the AGG3 and AGG6 portfolios the difference is quite significant. Whether it is the use of lower volume ETFs, the increased popularity of these portfolios, or something else entirely, trading in some other date besides month end made a significant difference in 2015, for the better.
That’s it for this month. These portfolios signals are valid for the whole month of January. As always, post any questions you have in the comments. For those into trend following research there is a great beginning of the year research piece here. Enjoy!
36 Comments
B · January 3, 2016 at 8:30 am
Paul,
Happy New Year!
1. Thank you for all the TAA Updates for 2015 and this first one for 2016.
2. The new TAA Spreadsheet looks great.
3. It is nice that the spreadsheet is now valid daily so we can perform our updates on other days of the month (vs always using the 1st day of the month which can become a little hectic as you point out).
4. The info you provided about 4 week re-balancing (vs monthly) is very thought provoking.
5. Thanks for posting the link to the research piece … will take a look.
Thanks again and best wishes to all for a great 2016!
B
paul.novell@gmail.com · January 4, 2016 at 9:36 am
Thanks B. I think you’ll enjoy that research piece.
Shawn Murphy · January 3, 2016 at 4:59 pm
Thanks again Paul! I’ve been thinking about the slippage inherent in buying and selling ETF’s tactically, and I’ve been wondering if using transaction free custodial mutual funds (e.g. Schwab funds in a Schwab account, Fidelity in Fidelity, etc.) might do a better job at controlling costs. You wouldn’t be able to trade in the morning of course, and MTUM stands out as a unique ETF, but most of the 13 classes are effectively index funds…then you would get in and out at NAV. Is there a flaw in my thinking here?
paul.novell@gmail.com · January 4, 2016 at 9:21 am
No flaw. You avoid paying bid ask spreads as well. But the NAV you get out at is the close of the next day’s trading. So you have slippage from the previous close, which is what the ranking is based on, to the next day’s close which you buy at. You would have to ‘trade the close’ to get 0% slippage. But you can also do this with ETFs.
Kirk Doughty · January 3, 2016 at 6:44 pm
Paul,
I believe that the “Mebane Faber Version” of the AGG3 ranks the ETFs based on the average of their 1/3/6/12 month returns (and also being above the 200 day SMA cutoff). However, I seem to recall that you mentioned that you prefer to look at the simple 6 month returns (Half Year % Returns column) for your own portfolio ETF rankings. Is that the case and can you state your reason for doing so? Thanks for all you do on the blog, its a great resource for my understanding of TAA principals.
Kirk
paul.novell@gmail.com · January 4, 2016 at 9:18 am
In one of my models, I uses 6 month return ranking of the ETFs with a 3 month return ‘cutoff’. The reason is better historical performance. It is not major but it seems to help. For example, that type of AGG3 model returned 0.73% for 2015.
Kirk Doughty · January 4, 2016 at 3:09 pm
Thanks for the rely Paul. So if I understand correctly, you would use the 6 month return for ETF ranking purposes, but would disqualify any ETFs if their 3 month return was less than 0.0%?
paul.novell@gmail.com · January 5, 2016 at 12:04 pm
Yep. exactly.
Mandar · January 3, 2016 at 6:57 pm
Thank you thank you thank you thank you!!!! Can’t thank you enough for the daily FINVIZ updated portfolio google spreadsheet … it will be very very helpful!! Thanks for finally implementing it!!!! And thank you for all your wonderful updates as always and a happy new year to you Paul.
paul.novell@gmail.com · January 4, 2016 at 9:37 am
Thanks Mandar.
Matt · January 4, 2016 at 12:11 am
Many thanks as always, Paul. Your updates and tracking are greatly appreciated.
As for the study you linked to on Philosophical Economics, that’s an incredible effort. One “concern” I have with that is the imposition of 0.6% slippage on each round trip trade. I understand why the author used that number (he had to use a consistent slippage figure across all asset classes, and had to look to historical numbers to get a number) but it seems unduly harsh when applied to trend-following strategies on modern, highly-liquid instruments like SPY.
paul.novell@gmail.com · January 4, 2016 at 9:13 am
True but its very very easy with a careless implementation of the portfolios to go way over 0.6% slippage. And anyway, if it works with 0.6% it will work with less slippage.
Sergio Molina · January 4, 2016 at 2:44 am
Finviz not adjusted dividend.
See VMBS in ETF Replay or stockcharts
paul.novell@gmail.com · January 4, 2016 at 9:11 am
FINVIZ does adjust for dividends. E.g. SPY. In general. I have notice some discrepancies vs lets say Stockcharts but they are usually fixed quickly.
Steve · January 4, 2016 at 6:58 am
Thanks Paul for your work. Indeed 2015 was a tough year, and out of the gate in 2016 does not look good.
paul.novell@gmail.com · January 4, 2016 at 9:37 am
Looks to be interesting for sure.
Greg · January 4, 2016 at 10:05 am
Is there no January 2016 allocation recommendation for the GTAA5 portfolio?
paul.novell@gmail.com · January 4, 2016 at 2:54 pm
Yes, its in the spreadsheet I linked to. The old one that uses the Yahoo Finance data. I just didn’t include the update in the post.
Greg · January 4, 2016 at 3:01 pm
Sorry Paul, I just looked at your new spreadsheet, and didn’t even explore the yahoo version
Peter Harrison · January 4, 2016 at 1:31 pm
A happy new year to you and a big thanks for the continued updates Paul. The Philosophical Economics research piece is awesome. And the Finviz spreadsheet is an education in itself…it’s amazingly easy to plug in any ETF symbol and get the same info, which is useful to me in other ways. I’ve been considering a change in my investment approach for several weeks now and your website and it’s links has been a major resource in helping me think about this. Thank you.
Peter Harrison · January 4, 2016 at 1:58 pm
Regarding Matt’s comment about the Philosophical Economics author using a 0.6% slippage, I think the author is just trying to prove that the basic trend-following concept works across almost all asset classes with the most basic and conservative rules and assumptions, rather than being his view on what performance is possible with a real life model. It’s also the first of a series of posts on the subject…his comment about having identified a way to improve the strategy based on an understanding of the underlying math (and its geometric properties) is intriguing and has me looking forward to the sequel posts.
Another blogger has also promised a future post on how to improve the performance of trend-following models in bull markets…here. http://econompicdata.blogspot.com/2015/09/momentum-applied-to-mutual-funds.html?m=0
Aung · January 4, 2016 at 8:12 pm
Hi Paul,
Happy New Year!
I’ve been following your website for the past few months. Thank you very much for your posts and tests.
There was an article you wrote on how to improve value portfolio by screening out companies with a lot of change in debt.
I also read your article on how to do Trending Value on micro caps.
I was wondering if you ever tested your Micro Cap Trending Value strategy with the Change in Debt?
Thank you so much Paul
paul.novell@gmail.com · January 5, 2016 at 12:06 pm
Hi Aung, yes I’ve tested both trending value and micro cap trending value with change in debt. It reduces returns. Both those strategies are more about momentum than value.
Paul
Richard Wilkes · January 6, 2016 at 3:33 am
Hi Paul,
I just started looking at your material and wonder how to replicate the filtering on the FINVIZ site? I currently use two other similar sources ETFreplay.com and another similar but very unique site called SectorSurfer by Sumgrowth.com. Does FINVIZ allow for other filters other than the standard 20/50/200 sma? Doing some backtesting on ETFreplay I found that a 2 ma over under using a 90 day long and 9 day short worked somewhat better.
The creator of SectorSurfer uses electrical digital filtering techniques (yes a former electrical engineer) which is unique instead of the typical TA in the finance world to find the best of a group of stocks, funds, elf’s. You might enjoy a look there as it is very usable by technical and non technical folks alike. His current efforts are on making extended versions of common ETF’s for back testing and implementing strategies/portfolios to use when a strategy would be in cash otherwise. I would enjoy the conversation and or feedback in any event. Thanks for your contribution!
paul.novell@gmail.com · January 7, 2016 at 11:01 am
Can’t do it on FINVIZ. I know of SectorSurfer. Not a user. I don’t use anything that is proprietary. Everything I use has to be ‘open source’ and subject to verification going back to at least 1973.
Paul
Richard Wilkes · January 7, 2016 at 11:53 am
Thanks Paul,
Are you taking data from FINVIZ and putting into excel yourself to get your signals? For the same reason you don’t like to use a proprietary service that may or may not be in business next year, I’m trying to replicate the signals analytically using what’s available. Obviously I can create alerts on my trading platform for Crossing Moving averages. Thanks for your replies. My signals went out of everything but treasuries and bonds Jan 1. As of this writing SPX is down over 2%. Great time to be sitting out and looking at RV’s instead 🙂 Those will be less expensive too if this rout continues this year. As you have mentioned elsewhere Momentum strategies don’t perform well in years like 2015 where oscillation and whipsaw is around. Maybe we’ll get a true washout and be able to start anew again.
Dale · January 6, 2016 at 4:56 pm
Paul,
Thanks for the great blog on investing and for the great new lookup from Finviz. Pretty slick to parse the quote page, very neat, and very simple compared to the earlier sheet. And easy to plug in any other symbols.
Maybe the first column could be a link to bring up the current chart: ie: hyperlink(“http://finviz.com/quote.ashx?t=”&B8,”US Real Estate”)
Is there a price deadband around a SMA crossing that you use to enter or exit, looking now at data more often? With daily and intraday updates available, strictly following a SMA crossing, people might get more spurious trades? Whereas looking at the closing only at end of each month has filtered out some price movement. Or maybe the deadband is still time based, only look at it once a week at most.
Have you talked about how you do backtesting or tracking a given portfolio? What tools do you use? Quantopian is interesting, can even set up to trade automatically.
Thanks for the other blog link. There’s a lot more to understand there, and I like this part: ” why do the most vocal practitioners of market timing tend to perform poorly as investors? The answer is not that they are poor market timers per se. Rather, the answer is that they tend to always be underinvested. By nature, they’re usually more risk-averse ”
Congrats on what you’ve been able to do with finances and travel. We also watch the Wheeling it blog, we are taking the year to travel by camper for this year. One of my side goals for the year off is get a little more figured out with investing, and your blog is great resource. Thanks!
paul.novell@gmail.com · January 7, 2016 at 11:03 am
Thanks.
The models work best with monthly, i.e. 4 week, update intervals. This best avoids spurious signals.
Yes, I’ve talked about backtesting. For long term backtesting of asset classes I just use excel. For more recent stuff and for individual stocks I use Portfolio123.com.
Orlando · January 7, 2016 at 9:37 am
Hi Paul,
I just looked at the spreadsheet and I see that rank 5 is “Invested” while 2, 3, and 4 are “Cash”. Is that normal or is there a “bug” in the formulas?
Thanks as always!
paul.novell@gmail.com · January 7, 2016 at 10:57 am
No bug. Perfectly normal.
Jens · January 7, 2016 at 7:12 pm
Paul,
I’ve been reading your posts with interest over the last couple of days. Great work.
I noticed today that the GTAA13 spreadsheet didn’t update correctly for VNQ.
I’m not sure if there is an issue with the data source?
paul.novell@gmail.com · January 30, 2016 at 10:25 am
It happens sometimes. Just need to check back again every so often.
Paul
Jens · January 7, 2016 at 8:19 pm
Hi Paul,
Another question. I’m considering implementing the GTAA13 portfolio. With your new daily updated spreadsheet I guess I can start any day of the month, as long as I wait another calendar month (or 4 weeks) to react on the next signals? Or is it better to wait until February 1st to implement?
Also, to make sure I understand. As a new investor in the portfolio, I assume I only would purchase the positions that are listed as invested for the start date? The rest of the allocations will be held in cash until they are signaled as invested?
Thanks!
paul.novell@gmail.com · January 30, 2016 at 10:26 am
Yes, you can start any day. And yes, you would only ever have positions that are ‘invested’.
Mike · January 8, 2016 at 3:00 pm
Thanks for the Finviz version Paul! Super useful to be able to do this mid-month. I’m a corporate accountant so the end and beginning of the month are crazy for me.
paul.novell@gmail.com · January 30, 2016 at 10:27 am
Great. Glad it helps. It’s so nice to able to implement portfolios whenever you want as opposed to only at month end.
I have found it to be a huge help as well.
Comments are closed.