Roundup: Tactical Asset Allocation
This is a roundup of our recent posts talking about Tactical Asset Allocation (TAA). Unfamiliar with TAA? Click for a primer.
I’m very satisfied with where this research into TAA led. I’ve committed a portion of my own capital to my TAA model, and I’ll be sharing the model’s trades and tracking performance each month here on the blog.
Defines TAA and why it’s an important component of a broader portfolio (even for active traders like us).
Backtested results of my own take on TAA. The benefit of these types of models has not necessarily been generating return, but managing volatility and drawdown.
Building the TAA Model: Step 1, Step 2, and Step 3
Walks through the model’s three-step approach: identifying uptrending asset classes, ranking and selecting asset classes to trade, and allocating to selected assets based on expected volatility.
Happy Trading,
ms
. . . . .
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Tactical Asset Allocation | 30 Comments
Just a quick note to say that I really enjoyed that series… (and by the count of number of comments on some posts, it feels like I’m far from being the only one!). Well done Mike and thanks for sharing most of the concepts behind the strategy!
Thanks for the kind words Jez!
it just looks too good to be true so it probably will be. too many people embracing this stuff. the markets have a way of slapping upside the head.
Hello Steve – my $0.02…
Will we be able to hit our numbers in real-time? I dunno’…I warn about the dangers of backtests and curve fitting here…
http://marketsci.wordpress.com/2010/10/13/notes-and-revisions-re-taa-backtest/
Having said that, the basic premise behind these type of systems, most importantly, trend following using long-term moving averages, is a concept that has consistently worked for well over a century. All while being popular and widely utilized for well over a century.
Trend-following is the meat of these strategies, everything else is gravy, so assuming TF continues to perform, these strategies should do just fine.
michael
i hear you michael and I’ve been trading a similar stategy with success for a while but I worry that everywhere i turn i read about more taa. if you are correct about expecting half the returns and double the dd’s then expected results quite frankly, suck ( at least as far as i’m concerned).
RE to Steve: to clarify, I’m NOT saying I EXPECT half the return and double the DD. I’m saying that investors have a bad habit of looking at backtests and getting googlie-eyed. As I tried to explain in the post, investors need to turn down excitement over backtests and turn up excitement about empirical concepts and real-time results. Put another way, I’m trying not to be one of these goofs who take people’s money by flashing a sexy (but utterly bullshit) backtest around.
TAA (in the Faber’esque sense) is no more than trend-following at it’s core. Has trend following become more popular in recent years? I think not. It’s always been a fundamental investment approach. And as we’ve demonstrated in the past, trend-following is as effective today (scratch that, MORE effective today) than at any point in a looong time.
I’m not defending any one investment approach because the future is always unknowable, but unlike some of the other very fleeting inefficiencies that we chase (ex. short-term MR) this is one I wouldn’t logically expect to suddenly stop working across the board.
Just my $0.02.
michael
I think you covered this, but I can’t find where it is. My question is: Why not go both long and short using the same strategy. Go long the best uptrending ETF and short the best downtrending ETF?
Thanks.
Hello Russ – historically, shorting naturally-appreciating assets (equities, treasuries, gold, real estate, etc) hasn’t been very successful in anything but the shortest timeframes (ex. RSI(2)), partially because of the generally appreciating nature of the assets and partially because of the assymmetry of bull vs bear markets (i.e. bull markets rise much slower than bear markets fall).
When asset class A goes through a long downtrend, I would rather rotate into asset class B that is in an uptrend, than try to capitalize on shorting asset class A.
Just my $0.02.
michael
I keep thinking about this issue. Being both long and short seems like a good way to add an additional hedge to the portfolio. Also, if there are no good short candidates, the short side could be kept in cash. Why not at least include the possibility of being simultaneously long and short as an option in your TAA strategy?
– Russ
RE to Russ: I would be open to it if I thought there was a way to predictably short asset classes in such a long timeframe (monthly). While this approach might have been successful over the last decade (because of the extended and well defined shifts between bull and bear markets) over the very long term, shorting in long timeframes has not been effective because of (a) the natural tendency for asset classes that are “stores of wealth” to appreciate, and (b) the fact that bear markets tend to strike (and finish) very quickly, while bull markets tend to draw out over a long period of time. Just my $0.02. michael
Michael, thanks for a great set of posts on TAA. You do an excellent job of boiling a complicated subject down into a few short, accessible posts.
Hi,
Excellent work Michael. Could pls clarify whether dividends are included for stock market indices in your backtest? Also I was wondering why you do not use corporate bonds (IG or HY) in your TAA model, it seems it would have been a good addition. Cheers.
Alex
Hello Alex – yes, I’m using div-adjusted data for equities. Corp and HY bonds are on my to do list to include, but I’m having trouble getting enough clean historical data for an index that closely replicates one of the liquid ETFs. Especially with HY bonds I’ve found that the different indexes produce radically different returns. More to follow on this. michael
I also really liked this series! I think there’s a huge demand for low volatility trading systems that trade infrequently – at least that’s my impression from discussions with friends & family. Everyone agrees that buy & hold is dead but no one has an idea what to do instead…
Hello psiegert: thanks for the kind words sir. Should be fun watching this one in real-time. Having a fair amount of capital invested in the model myself, I take some comfort in the asset class diversification that I don’t get with my other more aggressive strategies. michael
michael, for HY results best to use a MF. suggest PHDAX as it’s pretty plain vanilla and has been. also, a long history. not perfect but it gets the job done.
RE to Steve: thanks for the follow up, but PHDAX is not an accurate enough proxy for the ETF (obvious from just an eyeball of the charts) and I’d like to see a lot more data. michael
Michaael,
Will you be posting the re-allocations as and when they arise ?
Hello Ian – yes, on the last trading day of every month prior to the close I’ll be sharing the model’s allocation for the following month here on the blog. michael
Michael,
Do you have back test results of Faber’s original work comparable to your latest test results?
Thanks
Michael,
I have thoroughly enjoyed your blog , and in particular this thread on TAA. I have been following and doing testing based on Fabers paper and 10 month SMA model. I am excited to follow (and learn) from your work. By the way, in case you haven’t seen this from Faber last week…
http://www.mebanefaber.com/2010/10/18/gtaa-etf-launch-and-increasing-amounts-of-granularity/
Regards
Jim Satterfield
The Faber ETF (GTAA) launches today. According to Advisor Shares the holdings are as follows:
Sticker Name Weight
VWO VANGUARD EMERGING MARKETS ETF 5.03%
SCZ ISHARES MSCI EAFE SMALL CAP INDEX 5.00%
RWX SPDR DOW JONES INTERNATIONAL R 4.99%
TLT ISHARES BARCLAYS 20+ YEAR TREASURY BOND FUND 4.99%
IYR ISHARES DOW JONES US REAL ESTATE INDEX FUND 4.99%
RWR SPDR DOW JONES REIT ETF 4.99%
VB VANGUARD SMALL CAP ETF 4.01%
IWC ISHARES RUSSELL MICROCAP INDEX 4.00%
VO VANGUARD MID-CAP ETF 4.00%
TIP ISHARES LEHMAN TREASURY INFLAT 3.99%
SHY ISHARES BARCLAYS 1-3 YEAR TR 3.32%
BIL SPDR BARCLAYS CAPITAL 1-3 MONTH T-BILL ETF 3.32%
BSV VANGUARD SHORT-TERM BOND ETF 3.32%
DBP POWERSHARES DB PRECIOUS METALS 3.02%
DBA POWERSHARES DB AGRICULTURE FUN 3.02%
DBB POWERSHARES DB BASE METALS FUN 3.02%
EWX SPDR S&P EMERGING SMALL CAP ET 2.52%
VEU VANGUARD FTSE ALL-WORLD EX-US ETF 2.50%
LSC ELEMENTS LINKED TO THE S&P COM 2.01%
QQQQ POWERSHARES QQQ 2.00%
DBE POWERSHARES DB ENERGY FUND 2.00%
VTI VANGUARD TOTAL STOCK MARKET ET 2.00%
BWX SPDR BARCLAYS CAPITAL INTL D 1.99%
AGG ISHARES BARCLAYS AGGREGATE 1.99%
IEF ISHARES BARCLAYS 7-10 YEAR 1.99%
FXC CURRENCYSHARES CANADIAN DOLLAR 1.25%
FXA CURRENCYSHARES AUSTRALIAN DOLL 1.25%
FXY CURRENCYSHARES JAPANESE YEN TRUST 1.25%
FXE CURRENCYSHARES EURO TRUST 1.24%
PLW POWERSHARES 1-30 LADDERED TREA 1.04%
FXF CURRENCYSHARES SWISS FRANC TRU 1.00%
DBC POWERSHARES DB COMMODITY IND 1.00%
CEW WISDOMTREE DREYFUS EMERGING CU 1.00%
BZF WISDOMTREE DREYFUS BRAZILIAN R <1%
FXB CURRENCYSHARES BRITISH POUND STERLING TRUST <1%
HYG ISHARES IBOXX $ HIGH YIELD COR <1%
CYB WISDOMTREE DREYFUS CHINESE YUA <1%
LQD ISHARES IBOXX INV GR CORP BD <1%
MUB ISHARES S&P NATIONAL MUNICIPAL <1%
MBB ISHARES BARCLAYS MBS BOND FD <1%
Total 100%
There is a general point that gets my thinking: high correlation doesn’t imply low effectiveness of asset class picking. Correlation measures how often asset classes tend to move in the same direction but dispersion is the metric we may need – standard deviation of returns of all asset classes – since it measures how much returns deviate from each other.
We can have high correlation together with medium dispersion (like nowadays) which would make asset class picking still very effective in a long-short perspective.
example: all asset classes perfectly correlated – moving in the same direction – but with very different returns – high dispersion. We can still benefit from a correct long-short asset class exposure no matter how high the correlation will be.
I remember you are using a long-only strategy due to low rebalancing frequency and assymmetry of bull vs bear markets but a long-short strategy would be able to take advantage of dispersion (ideally being market neutral – zero-beta – and taking alpha only out of the market)
What’s the correct way to address it in a TAA perspective?