Tactical Asset Allocation: It Really Is that Good

04Oct10

Inspired by the excellent work of Mebane Faber I’ve been tinkering obsessively with a Tactical Asset Allocation (TAA) model and I’ve come to the conclusion that TAA really is that good (as one component of a broader portfolio…more on this later).

This post is a primer for those unfamiliar with TAA, and in a follow up, I’ll take a more technical look.

What is TAA?

From Wikipedia:

Tactical asset allocation (TAA) is a dynamic investment strategy that actively adjusts a portfolio’s asset allocation…Unlike stock picking, in which the investor predicts which individual stocks will perform well, tactical asset allocation involves only judgments of the future return of complete markets or sectors.

Three big differences between TAA and what we do (active trading):

First, the focus is on asset classes (not individual stocks, funds, etc.) Second, though the definition doesn’t strictly call for it, TAA tends to trade less frequently and ignores more day-to-day market noise. And third, TAA relies more on diversification to provide for risk management (as opposed to entry/exit points).

When I talk about TAA I’m using Faber’s published model as a jumping off point (“standing on the shoulders of giants” and all that jazz), so the focus is on trend-following/momentum.

My end goal is a strategy that trades infrequently (once per month or less), selecting from a diversified basket of asset classes those that are exhibiting the strongest trend/momentum, and allocating the portfolio between them in a “smart” way (i.e. in a way that maximizes expected return versus volatility).

My basket includes: U.S. Stocks, China Stocks, Japan Stocks, Gold, Oil, Commodities, Real Estate, and U.S. 10-Year Treasuries.

Why Faber’esque TAA?

Active strategies (including our own) are based on market inefficiencies and anomalies that are by their nature fleeting.

Folks who trade the way we trade are in a perpetual quest to stay ahead of the curve because the doors of short-term opportunity are forever opening and closing.

Contrast that with tried-and-true trend-following, which works as well today as it did 100 years ago. The lowly 50/200-day moving average crossover, that’s been around since the dawn of ticker tape, was more effective over the last decade than at any point since at least 1930 (read more).

It’s that kind of reliability that makes sleeping at night easier on us active traders.

Why not just trade trend-following/momentum?

Because the successful among us will, at the end of the day, soundly outperform the best TAA strategy. The point isn’t to replace what we do now, but to provide a stable foundation for a portion of our portfolio that’s reliably rooted in the long-term.

What’s Next?

1. I’ll post a follow up this week taking a more technical look at Faber’s TAA model versus our own.

2. I’ve allocated a portion of my own portfolio to trading our TAA model (I eat my own cooking). I’ll be sharing our picks monthly and tracking performance here on the blog (a value-added service of being the best readers in the blogosphere).

[Edit: click for a summary of all posts in this series on TAA] 

Happy Trading,
ms

P.S. to learn more about Faber’s TAA model pick up his book The Ivy Portfolio.

. . . . .

To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.




21 Responses to “Tactical Asset Allocation: It Really Is that Good”

  1. 1 Jon

    Thanks so much for this series regarding TAA. Despite its history, I also have concerns about several aspects…its slowness in moving, it tendency toward whip-saw. I think this market will be a difficult one for this strategy.

  2. 2 eber terandst

    I am nervous about TAA. I did quite some work on it, and it is exquisitely dependent on the particular set of assets that you select for the rotation pool. The theory is that they should not be correlated, but correlations have been increasing with the years, so I am suspicious about historical results.
    In fact, the only fault that I find with Mebane paper is his selection of REITs, which certainly were not correlated when his study begins ( the 70’s) but by now are very correlated.
    In summary, a good part of the backtesting should be on sensitivity to the composition of the assets pool.
    eb

    • 3 MarketSci

      RE to EB: you bring up a very valid point about increasing correlation of many assets (especially anything equity-related). I think that there are ways to combat that though.

      That’s part of what I’m getting at when I talk about allocating the portfolio in a “smart” way. Simply cutting the portfolio into equal chunks (volatility and correlation be damned) is clearly not the best solution. Neither is allowing the portfolio to be overly concentrated in on asset group (equities, commodities, etc) in any given month.

      My point is that even though correlation between many asset classes has progressively increased, as long as the effectiveness of trend-following/momentum holds, there are ways to combat it.

      Just my $0.02…michael

  3. 4 Hector

    check this out, you can use this to diversify directly with Meb http://gtaa.advisorshares.com

  4. Looking forward to more posts on TAA/Trend Following (I would, wouldn’t I? ;-)

    btw – did you see that Faber’s coming out with his own ETF (GTAA): TAA applied to 50/100 ETFs representing various asset classes (as opposed to the 5/6 in his paper)

    • 6 MarketSci

      RE to Jez: I’m very interested to see how GTAA performs. I think the concept is outstanding.

      Side note: based on my own research, I think there are diminishing benefits of adding more and more asset classes to the mix. As Eber’s comment above talks about, it’s very difficult to find asset classes that provide real diversification in today’s market. I think the optimal number is certainly higher than the 5 Meb talks about in his paper, but probably far, far south of 50/100.

      michael

      • 7 marko

        So far the GTAA looks just like the S%P.

      • 8 MarketSci

        RE to marko: over the last month or so just about everything has looked like the S&P 500. Over the long-term (for better or worse) you’re going to see significant deviation between GTAA (or our own TAA model) versus the stock market. michael

      • 9 Bill

        Agree with posts about heavy correlation. Thus GTAA = SPY, but the fee on GTAA is much greater than owning SPY. I’m not sure why anyone created an ‘actively’ managed ETF when it only trades on multi-month swings (he just trades off of the simple mov avg – that’s kind of a joke). This is so simple of a model, why pay someone to do it? I just don’t get the value add here. To me it’s a lazy portfolio not an actively managed fund.

      • 10 MarketSci

        RE to Bill: I couldn’t disagree more.

        As Faber has repeated ad infinitum, the simple approach he takes on his blog and book are intended to be instructive, and it’s pretty clear from the underlying allocation of the ETF so far that he’s not remotely following either.

        I’m not a sycophant. I have no idea whether Faber will hit his targets. And in some ways I look at GTAA as my benchmark to beat, which puts us in competition (uni-directionally =). But to call TAA (when done right) a “lazy portfolio” very much understates the complexities involved.

        Just my $0.02.

        michael

  5. 11 codes

    Looking forward to the next post.

    Unlike Jon I’m of the impression that TAA will continue to work very well in this environment. Especailly helps sleeping at night – clear exit strategies provides piece of mind that is damn near priceless in times like ours.

    Really have no idea how the buy and hold crowd gets any sleep at all -given the history of bear markets (and the decimation of most buy and holders) if they’re sleeping they likely shouldn’t be.

    RL

  6. 12 Freeman

    Michael-While I do agree with much that has been said (including many elements in Faber’s paper), something is missing in your discussion. Faber mentions that much of the success of tactical assest allocation is due to the strategy’s move to cash when the potential for upside is not there.

    Your thoughts on the matter are appreciated.

    ~Freeman

    • 13 MarketSci

      RE to Freeman: I completely agree. That’s an implicit benefit of trend-following (the flavor of TAA I talk about in this post). When the asset class isn’t uptrending it’s either (a) in another asset class that is uptrending, or (b) in cash. michael

  7. 14 Lars Bech

    Hi Michael

    I cannot wait to read more about your findings in this very interesting matter. With no desire to be an active trader I am looking for a strategy that will help investing my money. Being Danish I would be concerned about the USD vs. DKK and would look to either Scandinavian ETF’s etc. and/or European (nominated in EUR).

    Best wishes, Lars Bech

  8. 15 steve

    michael, your etf choices are kind of odd. you have two foreign stocks etf’s that are pac rim but avoid the rest of the world completely (except US) . you have one bond etf that is really a “safe haven” choice but avoid riskier bonds.you have exposure to three commodities markets when oil and commodities are highly correlated. feedback?

  9. 17 fred

    For the neophytes among us….. can someone explain the subtleties/differences between TAA and trend following/momentum…….
    thanks.

    • 18 MarketSci

      Hello Fred – TAA doesn’t necessarily have anything to do with TF/momentum.

      TAA simply means dynamically allocating between various asset classes (as opposed to individual stocks, etc.) There are an infinite ways you could do that.

      In my case I’m following in Faber’s footstpes and using TF/momentum to decide WHEN to go long each aset classes.

      michael

  10. 19 Jan

    I do something similar to all this but I do not try to predict the market at all. I do not care which way it goes.

    1) You decide on an initial asset allocation It will get out of whack over time as some assets are more volatile.

    2) You gradually sell some of each asset as it goes up and buy more as it goes down.

    It is too complicated to explain here but basically if you sell 100 shares you can buy 130 shares when the security goes back down. You accumulate more shares with the same initial investment with each swing. It is very ZEN.

    see http://www.aim-users.com


  1. 1 Growing Popularity of Tactical Asset Allocation | Sage Investment Strategies
  2. 2 Ed Mamula.com » Automatic 7 TAA Results through January 10, 2011

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s