Wow! What a month. After 3 years of a steady grind higher there is finally some significant volatility in the market. Several of the portfolios had a drawdown of about 6% during the month along with the major indices. If this is a surprise or it seems excessive, it’s not. Historically the portfolios have exhibited drawdowns up to 20%. With such big swings the portfolio risk management rules kicked in and now the allocations are mainly cash. Now comes the interesting part – to see if these signals are head fakes or the start of something significant.
Here are the tactical asset allocation updates for September 2015. All portfolio updates are online as part of Paul’s GTAA 13 Portfolio New sheet.
First, for the basic portfolios – the GTAA5 and the Permanent Portfolio. GTAA5 is now all cash. For the timing version of the Permanent Portfolio there were no changes this month. The TAA version of the Permanent Portfolio is also now all cash.
Now for the more aggressive GTAA AGG3 and AGG6 portfolios.
Big changes this month. AGG3 is 33% invested in VGIT, the rest in cash. AGG6 is 16.7% invested in VGIT, the rest in cash. This from basically a 100% equity allocation last month. Risk management is fully engaged.
Performance for the portfolios so far this year is in the table below. Numbers are for each month. The figures are estimates taken from a variety of sources. I don’t do detailed performance tracking until the end of the year.
If you’re a fan of the Antonacci dual momentum GEM and GBM portfolios, GEM portfolio also experienced a significant change and is now invested in bonds, BND or the AGG ETF, and the bond momentum option (DMFI) of the GBM portfolio is invested in VMBS. No changes from last month. I’ve also made my Antonacci tracking sheet shareable so you can see the portfolio details for yourself.
That’s it for this month. These portfolios signals are valid for the whole month of September. As always, post any questions you have in the comments.
55 Comments
Sergio Molina · September 1, 2015 at 2:22 am
If uses ETF BIL , GEM model go SPY
paul.novell@gmail.com · September 1, 2015 at 8:02 am
True. Depends on what you use for your cash proxy. I guess the purest form is just using the yield of the 3 month T-bill. That would also suggest staying in VTI.
Mike · September 1, 2015 at 4:50 pm
Purest form of GEM for domestic component would probably be and S&P 500 ETF not VTI, as per the dual momentum errata website (pg 112): http://www.optimalmomentum.com/page14.html
Gary prefers an S&P 500 security to VTI because momentum seems to work better with large caps. I think the difference is probably minimal but I still use VOO. However I use SHY as the cash proxy because it get me out of the market a little more early and into the market a little later which is just a personal preference and also probably won’t make much of a difference in the long run.
paul.novell@gmail.com · September 1, 2015 at 5:07 pm
Good points Mike. Thanks. I also like the use of SHY as the cash proxy. At zero rates the slippage from BIL is enough to incur losses just being in cash and I also have found that SHY is better in actual implementation.
Paul
Bryce · October 8, 2015 at 12:59 pm
Paul,
You’re never actually in cash though in the GEM portfolio (it’s either equities or bonds, never cash). Antonacci uses 1-3 month t-bills (BIL) to make that determination. What’s your rationale for using 1-3 year treasuries (SHY) to make that determination?
paul.novell@gmail.com · October 9, 2015 at 8:41 am
True.
I think SHY, or better said the 1 yr T-Bill, is a better fit for what the portfolio is trying to do based on 12 month returns. And as you said the portfolio is never invested in cash, its only used for a signal. SHY was as close as I could get in the spreadsheet to importing the 1 yr T-Bill. But I haven’t paid that much attention to it since I don’t implement the portfolio.
All in all don’t think it matters much unless the treasury curve would be unusually steep.
Paul
Florenzo · September 4, 2015 at 9:50 pm
Hi Mike & Paul.
My question is, for GBM, which “basket” of bond ETFs do you use?
At the moment I use the best of BND, TLT, JNK and SHY.
But I also read (https://investingforaliving.us/2014/12/06/a-trend-following-bond-portfolio-for-any-environment/) IGOV, TIP, LQD and MUB.
So, some doubt rises.
Thanks.
paul.novell@gmail.com · September 6, 2015 at 6:21 pm
Florenzo, the Antonacci GBM bond model has changed since the publishing of his book. He now uses a bond model called DMFI, dual momentum fixed income. See here. You can use the ETFs I list in the sheet I use to track the model or others of your preference.
Paul
Florenzo · September 7, 2015 at 7:34 am
Ok.
Thank you!!!
;o)
Pete · September 1, 2015 at 4:28 am
I’ve been a lurker for a while, so wanted to say thanks for all you do, Paul. According to finviz, vgit is below its sma 200, so wouldn’t that portion also be out of equities?
Lars · September 1, 2015 at 3:49 pm
Were you looking at the total return SMA, or the price SMA? Looking at Schwab, the price was just a cent or two below it’s 200 day average, so adjusted for dividends (which are paid monthly), the ETF was comfortably above.
By the way, I haven’t found a free charting tool yet that allows you to show the total return SMA. Any pointers?
paul.novell@gmail.com · September 1, 2015 at 5:08 pm
Always include dividends. I don’t use chart to base my signals on. I use the actual price data but stockcharts.com use dividend adjusted prices.
Paul
Bill R · September 1, 2015 at 7:44 am
Where can I find the YTD performance for Paul’s GTAA 13 portfolio?
Thanks
paul.novell@gmail.com · September 1, 2015 at 8:04 am
Bill, I don’t keep a YTD performance on GTAA13. It’s too much trouble to do on a month to month basis. I only update it at the end of the year.
Paul
Ben · September 1, 2015 at 8:00 am
Thank you for the update, and all the work you do! Quick question; in recent posts you have stated that the 10 SMA is used for the AGG3. I take it you are using the 200 day SMA for the current allocation? I seem to recall it might even be a variation of the 200 day SMA. I’m positive you have covered this before but I was unable to find the post.
Thanks!
Ben
paul.novell@gmail.com · September 1, 2015 at 8:03 am
Ben,
AGG3/6 uses the 10 month SMA (aka 200 day SMA) to decided on invested or not for each asset class.
Paul
Justin W · September 1, 2015 at 8:32 am
Hey Paul
Just want to thank you for the updates you do. I was digging through some of your old posts and noticed you were an active trader in selling option premium. Out of curiosity, do you still use options as part of your retirement strategy? If not, why?
I think supplementing this with the Timing portfolios you track can be very beneficial.
Thanks
Justin
paul.novell@gmail.com · September 3, 2015 at 7:58 am
I’m pretty much out of the trading game but every so often when the market presents good opportunities I jump back in for a bit. I don’t use options any more. When I do ‘trade’ I tend to just use stocks for individual companies and for indices I use futures. I have found the risk, reward better and execution much easier to control.
Paul
Andrew · September 1, 2015 at 8:37 am
Thank you for continuing to update monthly. It’s interesting how close even VGIT was to being cash. The GTAA 13 portfolio really illustrates how all asset classes can respond similarly during a pullback/market decline.
Bryan Katz · September 1, 2015 at 11:57 am
Paul,
How do you handle a day like today when we are supposed to rebalance our portfolios and wake up to a 2-3% drop in all the previously allocated AGG3 sectors. Do you hold off the rebalance for another day, or just take the hit? Thanks again for your excellent blog.
Bryan
paul.novell@gmail.com · September 1, 2015 at 5:17 pm
Unfortunately, that’s a long a complex answer Bryan. Gaps in markets seem to be a more and more frequent phenomenon. In general, with any system you’re going to get caught in days like today. As a rule of thumb, I never trade in the first 15-45 min on a day like today and I never use market orders. But today none of that would have helped much. It happens. It is part of ‘slippage’.
But I do try and have a system that reduces the probability of days like today. For example, a system that doesn’t use month end to make decisions, a system that uses more advanced order types (e.g. market on close orders). To give you a concrete example I use Portfolio123 to automatically generate the signals for my AGG3 portfolio (via email). The AGG3 portfolio I set up re-balances ever 4 weeks, not every month. This is a subtle difference but important difference. My signal for this “month” came last week, not at the official month end because that is when the 4 weeks were up.
This prob deserves a post in and of itself.
Paul
Chris · September 1, 2015 at 8:04 pm
Glad to see this point brought up because I was going to ask this as well. I just started doing some of these momentum trading systems today, and I have to say, it was a bit unsettling to open up my accounts to set the trades as the positions were already getting hammered. In one of my systems, RPG was on a buy signal at the end of August (but just barely). As I was putting my trade on, I couldn’t help but think that it had likely dropped below the ‘buy’ signal point.
But since one of the advantages of these systems is to have a mechanical system that takes out the human emotion and second-guessing, I went ahead and bought it. And the value side of me liked the fact that I had actually gotten it cheaper, but I couldn’t help but think about how it was going against the system.
Anyways, I’d love to see that future post, if you get the opportunity to dive into the topic.
Andrew · September 4, 2015 at 10:35 am
This is very interesting to me, but in the P123 data it looks like first of month and end of month trading does about the same, and mid-month actually performs a bit worse. I’m not sure it’s consistent because it’s only been tested on the AGG 3, but the most I can make of it is that over the last 7 years it hasn’t made much of a difference yet.
Andrew · September 4, 2015 at 10:52 am
Probably should take that back. I went ahead and did another test on more positions and the 4 week rebalance outperformed the other two. So I don’t know, really!
Steve · September 1, 2015 at 6:26 pm
Paul
I agree on using limit orders and not trading in the 1st 15 minutes of the day; however I don’t believe the 1st of the month vs every 4 weeks makes much of a difference. It all depends on where the market is on the specific day.
paul.novell@gmail.com · September 3, 2015 at 7:54 am
Steve, in general I agree but there is a high propensity for market shenanigans around option expirations, end of month window dressing, especially since the rise of the algos, bots, and HFTs, post 2008. I’d rather trade any other time except end of month even if it doesn’t make a difference. Maybe its just a cop to an emotional bias but that’s part of the reality of investing, finding little things that help us overcome ourselves.
Paul
MaxD · September 3, 2015 at 11:32 am
Hi Paul- I’m with you on this. Month-end seems to be becoming crowded- qualitative assessment from blogosphere. Can we use the GTAA spreadsheet for timing other than month-end, as it is currently setup?
Thanks for sharing your expertise!
Andrew · September 4, 2015 at 10:21 am
Unfortunately it sources the data from Yahoo which gives end of month pricing. Yahoo doesn’t offer a way to easily alter these days (e.g. set second of month, etc.).
paul.novell@gmail.com · September 6, 2015 at 6:22 pm
You can use weekly pricing. When I get some time or the inclination I’ll try a weekly sheet out and see how it works.
Paul
paul.novell@gmail.com · September 6, 2015 at 7:27 pm
Well, wasn’t nearly as time consuming as I thought. Here is a test of a weekly sheet that is valid at the end of every trading week.
Since there are odd numbers of weeks in months, the rankings are now based on 4 week, 12 week, 24 week, and 48 week returns. Wonder if it should be 4 week, 13 week, 26 week, and 52 week returns to catch up to the calendar….
https://docs.google.com/spreadsheets/d/1e38XOSAPMmrjOIlLftOMLKCJ3HaDRVgFlvYQU_bO7p0/edit#gid=206915515
Paul
B · September 7, 2015 at 7:20 am
Paul,
Wow, Great job and so fast!
I would like to put a plug in for using the second option you list: “4 week, 13 week, 26 week, and 52 week returns to catch up to the calendar”.
It may make no difference in returns but it seems better to be “synced-up” with the calendar.
Thanks again for all you do!
paul.novell@gmail.com · September 7, 2015 at 9:08 am
Agree B. That’s how P123 does it also. I’ve changed to to 4/13/26/52 week returns.
Also added a calendar test function looking for the end of the week, Friday, to conditionally format to green (“update valid”) at the end of weeks. if this works I’ll add to the monthly updates as well.
Now, if I could only get google sheets to email a message based on that conditionally formatted cell…
B · September 27, 2015 at 2:06 pm
Paul,
Check the weekly GTAA13 you just created where you point out whether the results are valid. I noticed this week it was lighting up as “green” or “valid” on Friday and Saturday but not Sunday. It looks like there is a minor error in formula: Sunday is day 1 and Saturday is day 7. How about something like the following:
=if(weekday(N3)=1,”update valid”,if(weekday(N3)=7,”update valid”,”update not valid”))
PS – Thanks for this weekly spreadsheet!
paul.novell@gmail.com · September 28, 2015 at 7:47 am
Thanks for catching that B. Fixed.
P
Bill R · September 2, 2015 at 7:56 am
Where can I find the “Rules” for Paul’s GTAA 13 & AGG 6?
The rebalancing time periods?
Thank you
paul.novell@gmail.com · September 3, 2015 at 7:51 am
This post discusses the various portfolios. I recommend reading the source material in the links for the full details.
Paul
Gary Conley · September 5, 2015 at 6:07 pm
Hi Paul,
Just wanted to say thanks for your work on all this stuff. I’ve been following the GTAA3 for a year and a half now. I like to be a pretty hands-off investor and your efforts make it easier to do so. The blog posts and comments are always interesting reading – keep it up!
Gary
paul.novell@gmail.com · September 6, 2015 at 6:17 pm
No problem Gary.
Matt Jerina · September 8, 2015 at 9:36 am
Hi Paul,
You have mentioned in the past how the Permanent portfolio is one of the best at reducing overall risk. Is the posted YTD return for the Permanent portfolio the fixed or TAA version.
paul.novell@gmail.com · September 8, 2015 at 9:45 am
Fixed
Jon S · September 9, 2015 at 4:51 pm
Well, I’ve been lurking for 2 years and finally decided to start AGG3’ing this month…a good point I thought as there was only a single buy signal! So $33k on VGIT…let’s see what happens!
Your blog and spreadsheets are both informative and inspirational! I would recommend followers to read the Ivy Portfolio book if new to this, as it gave me really good background on the fundamentals of the why/who/when and how of momentum portfolios.
Thanks again!
GLTA !
Jon
Pat · September 12, 2015 at 7:42 am
Paul, the frustrating thing for me is that since adopting the tactical allocation program in May my account is down 7% from what it would have been had I just stayed in stocks. Any words of wisdom would be appreciated.
paul.novell@gmail.com · September 12, 2015 at 8:22 am
Well, my first answer would be that any system will underperform over certain periods of time. That’s why I don’t like to pick just one system. Also, with any TAA system you should expect a mac drawdown of about 20% on a monthly basis – that’s what history shows has happened.
On specifics, your situation seems a bit odd. If you look at the YTD or even monthly performance numbers of the TAA systems none have underperformed stocks by that much. The worst one this year is AGG6 which has underperformed the SPY by 5% as of the end of Aug.
Paul
Alex Garcia · September 14, 2015 at 2:09 pm
Where are you getting the open, high, low, and close data?
paul.novell@gmail.com · September 14, 2015 at 2:37 pm
Price data comes from Yahoo Finance. I only use the adjusted close prices.
Alex Garcia · September 14, 2015 at 3:25 pm
Thanks Paul. I’m getting the same data for most, but can you double check your 9/1/2015 data for VTV, MTUM, and VBR? I’m not seeing the same on Yahoo Historical Quotes. Please correct me if I’m missing something.
Lars · September 15, 2015 at 7:57 pm
Alex,
The monthly historical quotes screen on Yahoo has a quirk with the dates. Even though it’s showing the date of the first trading day of the month, it’s actually showing the price at close of the last trading day of that month. For example for MTUM Yahoo is showing $69.70 for the close of 8/3/2015, but in actuality that was the price at the close of 8/31/2015. For September, it’s showing the close of the most recent trading day, even though the date says 9/1/2015. It becomes obvious when you compare the daily historical quotes to the monthlies.
-Lars
Alex Garcia · September 15, 2015 at 8:09 pm
Ahhhh gotcha… makes sense now! Thanks a lot.
Bill R · September 19, 2015 at 2:47 pm
If Yahoo Finance has this “quirk”, how can I calculate the 10m SMA for each of the ETF’s in the GTAA-13 at the end of the month?
Can I use StockCharts.com’s 40wk SMA number?
Or do I have to become a member of StockCharts.com to get the monthly chart?
Thanks
Kevin J · September 22, 2015 at 2:24 am
Hello – have been investing all year and experimenting with advice at another site. Making 6.5% YTD so far but still want to understand more. Any crash course reading suggested for a beginner ETF investor? Really feeling kind of lost with returns this year.
paul.novell@gmail.com · September 28, 2015 at 7:51 am
What kind of knowledge are you looking for?
Paul
John Power · September 23, 2015 at 6:44 pm
Any chance of indicating the last update on the GTAA spread sheet?
paul.novell@gmail.com · September 28, 2015 at 7:50 am
John, the spreadsheet updates every day.
Paul
B · September 26, 2015 at 8:23 am
Paul,
Is your tracking sheet for Antonacci Dual Momentum based on daily, weekly or monthly data?
On your Paul’s GTAA 13 Portfolio’s you post a note up top saying when the signals are valid (i.e. week-end or month-end). I was wondering what the Antonacci spreadsheet is based on (you might want to put a note when the signals are valid on the spreadsheet also).
Thanks for all you do.
B
paul.novell@gmail.com · September 28, 2015 at 7:49 am
The Antonacci pulls data from finviz which uses month, qtr, half year, and year returns and those figures are accurate daily. So, the Antonacci sheet is valid every day. Of course, the decision would still be on a monthly, in this case a 4 week, basis.
Paul
Comments are closed.