I Poo-Poo TAA and Readers Respond

12Feb13

I wear it as a badge of honor that my recent posts poo-poo’ing the future of tactical asset allocation (and its infinite reincarnations) have been consistently met with loud disagreement.

Folks getting upset with you isn’t necessarily a sign that you’re doing the right thing, but never having folks get upset with you is a sure sign that you’re not. Such is the nature of challenging conventional wisdom.

These very healthy discussions have helped me to crystallize and simplify my logic:

1. Almost everything in today’s market is too well correlated with equities, with the exception of just a handful of asset classes.

2. Treasuries (and similar assets), the most significant of those few low-correlation asset classes, is mathematically guaranteed to underperform in the future. It may not happen tomorrow, but it will happen.

3. Ponder the implications of the combination of #1 and #2 and riddle me how the super smooth backtested equity curves that the investing world is chasing, which inherently require effective diversification, represent realistic expectations for the foreseeable future.

Some smart responses have been:

1. Trend-following and momentum, market anomalies that have existed for well…as long as markets have existed, will be enough to overcome #1 and #2. (I disagree: real trading is difficult enough without entirely foreseeable holes in your game plan)

2. Alternatively, all of this is going to hurt TAA, but it’s still the best game in town and certainly better than buy & hold. (I don’t think it’s the best game in town, but I do agree it’s superior to B&H)

3. There are other sources of yield with low correlation to equities (including during market shocks) that will not be negatively impacted by rising rates.

4. Or alternatively, other low correlation asset classes, like gold or currencies, are going to make up for the role Treasuries play in most TAA backtests.

Those last two are very interesting to me. If I were going to continue putting my dollars into TAA, my energy would be focused there. These problems may not be insurmountable.

To date I haven’t seen anyone make a real effort to be the devil’s advocate when tackling this subject. I’m not talking about describing solutions. I’m talking about modelling solutions.

I’m not saying that finding yet another super duper way to measure momentum isn’t useful, but I am saying that when an obvious hole in your strategy is staring you in the face today, you don’t keep referring back to the past as justification to ignore it.

Just my (entirely biased as a short-term trader) $0.02…

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.



21 Responses to “I Poo-Poo TAA and Readers Respond”

  1. 1 MarketSci

    P.S. Before I rub anyone too raw, know that whenever I come across a quantitative TAA test, I always share it via The Whole Street, meaning, I love what you guys are doing. I’m just hoping to nudge the asteroid a little off course, not blow it up entirely. michael

  2. 2 minttowin

    Bingo. You are dead on target to challenge the status quo with the mathematically obvious. Alternatively riding volatile markets with low returns or going to cash won’t work.

    Adversity creates opportunity. Rising rates & inflation, devaluation, falling markets, and untimely flights-to-safety will eventually happen. “TAA” needs to be more dynamic or adaptive and have a broader scope.

  3. 3 Mark

    Are the rules in your TAA system not such that an possibly not/underperforming asset (IEF) would make sure that it’s not selected any more in the future?

    • 4 MarketSci

      Hello Mark – if you look at most TAA backtests (at least “TAA” in the true sense of the word), one of the most important ways they achieve the smooth equity curves that are the norm for these types of strategies, is through diversification.

      In a world where, for a long time (a decade plus?), UST and similar assets are no longer useful, I posit that diversification (and thus, those smooth equity curves) will breakdown.

      My argument is not that there isn’t value in trend-following, momentum, etc, but that smooth returns of the last few decades are not representative of the foreseeable future.

      michael

  4. I’ve really enjoyed the last few posts you’ve made on this subject, as well as some of the really smart comments made by others.

    I agree with nearly everything you have said about TAA’s future. To be honest I was never on board with it since my personal trading is almost entirely based on option trading. However I do get a little uneasy about people saying that it’s mathematically certain that treasuries will underperform in the future. I will grant you that is a true statement and simply a fact. However something none of us can say with any real certainty is how much underperformance we can expect. Everything should be measured on a scale, and while it’s obvious to say that future returns will inherently be less than we’ve seen in the past, I don’t really have a guess as to how low on that scale returns will actually be.

    So if you are going to abandon TAA entirely, I guess the real statement that sends is that your own expectation of the future underperformance in treasuries is going to be quite significant.

    Have you considered the macro reasons why maybe this supposed inevitability everyone seems to be talking about may in fact not happen at all, or maybe might be minor, or at the very least may be delayed for a significantly long period of time? I realize you’re a quant guy and the word macro doesn’t usually have a good place in the discussion, but i’m just not so sure the future of treasuries is so one-sided and obvious. If everybody knew what the future held, there would be no reason for diversified portfolios in the first place.

    Anyway, great posts. I think you’ve definitely stirred the pot, but as always have done it in a respectful way.

    • 6 MarketSci

      Hello VT – thanks for the comment. One note however: the eventual underperformance in UST is a mathematical certainty caused by rates increasing from a low base (caused by insufficient coupon today to offset changes in bond prices). As you know, I (literally) never express opinions on future changes in any asset, but this one isn’t an opinion, it’s just simple math. michael

  5. 7 Vix-Trader

    Yes I completely agree on the “mathematical” certainty of a future slide in returns of treasuries. I guess my point was more to the scale of the underperformance, and what degree of underperformance would make TAA not worth trading anymore. In the end, that’s the real question.

    ( and of course my comments are all to be taken with a grain of salt, since I already said I have never traded TAA. I also think volatility trading like you have recently shifted to has a stronger future, so i’m not even advocating that you yourself take a closer look at continuing TAA. I guess i’m just babbling, but anyway…. )

    TAA includes several assets being traded in conjuction, preferrably with low correlation but as we know that does ebb and flow. So although treasuries may not perform as well as they have in the past, that doesn’t imply they won’t still be tradable, and it doesn’t imply that other asset classes couldn’t pick up the slack so to speak.

    That’s kind of what I was meaning when introducing that ugly word “macro” into the mix. Some would argue that the very fact that treasuries are destined to underperform would mean that other assets may necessarily outperform.

    You still have to trade the whole mix, including treasuries because of course none of us have a crystal ball, but just because one of the assets in the mix underperform, that doesn’t mean the whole model is doomed.

    Compared to option trading, i’ve always felt like TAA was a waste of time. Also compared to trading the recent volatility products, I would also say that TAA doesn’t stand much chance either to compete.

    However when strictly talking traditional portfolio management, I’m not so sure TAA is doomed. Until we see a better, and more importantly PROVEN alternative, TAA is just as viable today as it was over the last 30 years.

    Feel free to ignore my ramblings :)

  6. Greetings Michael! I have enjoyed your blog since before you launched the TAA model. You have started a good discussion so I will just contribute what I can feel from my end of the investing elephant :-) I of course have no concern if you put more or less money behind TAA in the future.

    I do asset allocation: holding multiple ETPs at once for the sake of the whole portfolio’s performance. I don’t know if others would recognize what I do as TAA or not. I also research other trading styles. I don’t calculate or care about historical correlation matrices, but I do care about historical draw down of the portfolio. I trade portfolios that use un-leveraged ETFs with modest absolute return goals but high risk-adjusted return goals, I also trade portfolios that use leveraged ETPs that have highish absolute return goals and modest risk-adjusted return goals.

    What I have found in the last year or two is that my portfolios that use leveraged ETFs have been flat-ish, and my conservative portfolios have been consistently positive. Long term bonds have played a relatively minor role in my portfolios except during periods where equities were struggling – during those times the algorithm rotated more heavily into long bonds.

    I seem to remember that Meb posted a brief piece showing that the long bond returns are more potent as the yield dwindles to zero. US 30 year yields are stil above 3% we may not have even begun to feel the bond bull market(!) – but I am no prognosticator. Just saying that if Japan bonds were still yielding above 3% they could milk a decades long bond bull market out of that LOL. It doesn’t matter much to me how bonds end up performing because if long bonds stop delivering returns or diversification my quant models will have me in other stuff.

    I have also fallen out of love with a different investment methodology: value stocks. After doing it for a while I realized that there were good returns to be had but there were also long holding periods, un-avoidable volatility and a lot of research work on my part. That is why I decided to study asset allocation in the hopes of finding good returns, with lower volatility and less week-to-week demand for my research time. I would say I made a good trade, but the end of the story is all that matters and I haven’t gotten there yet.

    If/when you ever decide to spin up more research into asset allocation I would recommend the following based on my limited findings: 1) Get rid of IEF and go for TLT. 2) Add short ETPs to the list of options that your optimizer can select – this opens up all kinds of crosses that can be trending and offer diversification. 3) Ignore correlations during “stable” periods.

  7. 9 m

    I still except smooth returns with a diversified basket of assets. The problem is I expect the slope to be lower…lower annual returns, since bonds will be a drag instead of a pull as they have been in the past several decades.
    I’m going with the following weights in equity heavy portfolio:
    US equity- small 15, low vol 15
    Emerging equity – all cap 12, low vol 12
    Europacific equity – small 8, low vol 8
    US high yield corporate – 5
    US long term corporate – 5
    US real estate – 10
    Foreign real estate – 5
    Gold – 4

    • 10 MarketSci

      Hello M – am I missing something here? Every asset class you mention, with the exception of long-term corporates, gold (and possibly foreign real estate, but I don’t know), is well correlated with equities. And those 3 asset classes I mention comprise just a small % of the total portfolio. I’m not judging your strategy b/c it’s your money, but that doesn’t strike me as a particularly diversified portfolio.

      (this is NOT personalized investment advice. it’s your ship cap’n =)

      michael

      • 11 m

        You’re right its not diversified, and I didn’t mean for it it to sound diversified. When I said, “I still except smooth returns with a diversified basket of assets.” I wasn’t referring to my portfolio.
        I am expecting high volatility, but greater returns than a portfolio with a large weight toward bonds. I guess, its not even a tactical portfolio since I am rebalancing yearly.
        Apologies, my comment didn’t fit with this blog.

  8. 12 RC

    When I back test my TAA strategy without bonds it goes to cash. The returns are lower, the sharpe ratio is decreased but it is still a better way for me to manage risk and return than any other method I am comfortable with.

    If investors are selling both stocks and bonds.The money would go into cash or some other asset class causing it to rise. In theory, the other asset class could be added to the TAA model as replacement for bonds.

    • 13 MarketSci

      Hello RC – I responded to another commenter who made a similar point below, so I’ve cut and paste my response:

      I agree with everything you wrote. And if I wanted to sit on my porch all day and not think about markets, I definitely think the trend-following/momentum components of TAA are superior to straight buy & hold for all the reasons you mention (as I talked about in the post). I think you’ve missed the point of my rant.

      The point is that the super smooth backtested equity curves that have become so commonplace may not be realistic given all the reasons I talk about in the post. Is TAA better than B&H or sticking your money in a coffee can? Probably, but that’s not the point.

      TAA has a few (possibly totally surmountable) obstacles at the moment and nobody is focusing on them. That’s the point.

      michael

  9. 14 RC

    According to Pragmatic Capitalism, 3 of the worst draw downs since 1970 in 10 year treasuries were in 2009, 2003, and 1999 ranging from -12% to -9%. During the same years Pimco Total Return Fund had returns ranging from 13.8% to zero. I use Pimco in my TAA.

    • I have the utmost respect for PIMCO, but if you allocate to them you are making a VERY different allocation than Treasuries. You are taking on leverage (they overlay tons of CDS), counterparty risk (all the derivatives), and risk asset macro calls. They have been able to turn all three of the above into “alpha” over the past 40 years (a LONG time), but that doesn’t mean there isn’t risk. If banks didn’t get bailed out in 2008, things would look a LOT different.

  10. 16 SSV

    I think there’s a misconception in how TAA works. TAA increases returns (and outperforms in backtests) by keeping you out of falling markets, (like in 2000 and 2008). Not by outperforming the stock market when it’s rising.

    The goal, when one uses TAA, should be to avoid large drawdowns. I agree that treasuries may have a limited upside but the point of including US treasuries into a TAA model would be to reduce volatility and drawdowns. It is the one asset that has a large negative correlation with equities when stocks are declining.

    I have used TAA for the last 2.5 years and while it has significantly underperformed the US stock market, it’s done better than developed market ex US, emerging markets and the GSCI commodity index. But 2.5 years ago, I could not have predicted all of this. My main goal is to preserve my capital and I’ve learnt that I can live with sub-par returns (compared to US equities) but cannot stomach drawdowns greater than 10-15%.

    • 17 MarketSci

      Hello SSV – I agree with everything you wrote. And if I wanted to sit on my porch all day and not think about markets, I definitely think the trend-following/momentum components of TAA are superior to straight buy & hold for all the reasons you mention (as I talked about in the post). I think you’ve missed the point of my rant.

      The point is that the super smooth backtested equity curves that have become so commonplace may not be realistic given all the reasons I talk about in the post. Is TAA better than B&H or sticking your money in a coffee can? Probably, but that’s not the point.

      TAA has a few (possibly totally surmountable) obstacles at the moment and nobody is focusing on them. That’s the point.

      michael

      • 18 SSV

        Michael – Thanks for your reply. I do think I missed quite what you ranted on about TAA. I got more into the reasons why I (and maybe others) use it – sitting on the porch and not worrying about the market or long-term performance of the strategy :)

        One note about the possibilities of using currencies in TAA. If one uses country ETFs (other than US), it gives you currency exposure as well. For better or worse. Just as an example, look at the returns for the MSCI Japan Index recently, compared to those for the ETF EWJ. Wonder if you ever looked into this?

        Thanks for a great blog!

  11. I think the info at this link is a thoughtful way to use the fixed income side in a strategic way:http://advisorperspectives.com/newsletters10/47-betabond4.php

    • 20 MarketSci

      Hello Lance – I haven’t looked at the specific strategy in that post (yet), but conceptually I agree that that’s one possible solution to the issues I raised in this post. Thanks for the contribution. michael


  1. 1 Pseudo Random News and Comment | Mortality Sucks

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