Just Curious


Has anyone seen an empirical response to the limited upside remaining in US Treasury funds (read more and more) and the impact to tactical asset allocation models and all their many variations?

I’m sure that there have been some smart things written on the subject only because it’s not a terribly big secret, but I follow the quantitative blogosphere pretty closely and I haven’t seen anything; lots of models, no thoughts on Treasuries.

Just curious what readers may have seen on the subject.

Related posts:

Happy Trading,

. . . . .

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

12 Responses to “Just Curious”

    • With regard to continued lower yields, I can respond to the implications in the alternative investment, macrotrend space, where a long-term trend can be seen as similar to a tactical asset approach:

      1. U.S. Treasuries, but especially the Eurobund, Euribor and related rates , were the best performing sector in 2012, even at much higher prices. This after a nasty downturn a year ago.
      2. Anyone who deleveraged undert the assumption of less potential or a turn down would have regretted that decision.
      3. Even declining returns in the interest rate/bond sector would provide essential diversification for a portfolio of rates, FX, index, and various commodities, although the commodity sector was the biggest drag on performance in 2012.
      4. On the negative side, much larger positions are needed in Treasuries because of declining absolute volatility. This presents an serious liquidity and event risk, which is then a dilemma for portfolio managers.

    • 3 MarketSci

      Hello Perry – thanks for the smart comments.

      One note re #2: I absolutely agree that preemptive rotation out of UST doesn’t make sense. I’ve shown in previous posts that there is a healthy bit of potential still left in UST.


  1. It’s an issue, but: Treasuries are not the only fixed income out there. Home bias isn’t just an equity thing, and there are ETFs that allow easy access to emerging markets sovereign debt, developed markets sovereign debt, emerging/developed markets corporate debt, emerging/developed markets high yield debt, etc. Just gotta broaden your horizons.

    I found this neat little graph I made some time ago that applied a GTAA approach (picking only the top asset within each group) within broader asset classes instead of across all of them: http://imgur.com/a/oGz0W

    It should be noted that the debt part had the best performance, so this should be a concern. But taking away treasuries isn’t equivalent to taking away all debt. And things aren’t too terrible for the other classes either. Stupidly enough I don’t know which ETFs/funds I included in every class but you can assume it’s the usual suspects.

    • 5 MarketSci

      Hello Qusma – all of the alternatives you mention exhibit a much higher positive correlation to equities than UST, especially during market crises. The worst of the bunch, HY bonds, are no better than equities in today’s market in terms of diversifying risk. A lot of sovereign debt suffers from the same one-two punch of low yields today and a promise of rising yields in the future. And all of the options mentioned are far different in terms of the ability to diversify risk than the UST-centric world presented in most TAA backtests.

      I’m not arguing TAA is going totally off the rails (otherwise I would stop trading it altogether), but each of these assets mentioned provides unique challenges/opportunities for TAA models, and that discussion (at least to my knowledge) isn’t happpening.


      • You are completely right, of course. Corporate debt in particular does not have the countercyclical qualities of treasuries. In terms of TAA maintaining high returns in all environments, I think the core issue here is countercyclicality and not treasuries in particular. What’s important is having assets that we can expect, with high probability, will have negative or at least very low correlation to equities during the next bear market.

        If yields rise high enough before the next recession, the problem solves itself. If they don’t then TAA models may be in a bit of a pickle. The same is probably the case if we’re in for another “lost decade”.

        So let’s say we’re in an environment where we could reasonably expect zero returns from both US equities and treasuries…You are right, TAA probably won’t do as well as it has in the past, but how will it compare to other LT approaches? It’s not like 60/40 is going to be great in that sort of environment. In lieu of good alternatives, I’m still willing to trust relative/absolute strength.

  2. 7 david varadi

    if the macro-forecasters are correct, there will be a lot of up and down cycles for both the equity markets and the bond markets. if one is able to forecast momentum (or returns) accurately it won’t really matter as much, and arguably there will be more swings to take advantage of–especially with a shorter time horizon (ie depends on the nature of the TAA model). however in practical terms, assuming many inputs are not that forecastable,and use traditional momentum there will probably be a lower return to taa strategies since a large portion of the return is due to efficient beta capture.

    however to expand the range of possible outcomes- it is important to note that portfolio rebalancing tends to do especially well in such environments and can produce a positive return even if both time series have a zero return or are negative. of course the taa approach taken dictates the ability to take advantage of this effect. furthermore it is also worth noting that Japan was in the same spot as the US is currently in the early 90’s, and bonds were still the best performer over the decade despite uniform opinion that rates would rise. a tactical approach did much better in Japan (albeit with lower returns/risk-adjusted returns than the prior history during good times) than buy and hold or most investment schemes. I think that it is safe to say that TAA returns in an absolute sense will probably be low relative to their long-term track record- especially if everyone is chasing the same strategy. However, the jury is out as to how it will do versus a pure passive or strategic approach. Versus buy and hold–I think that TAA is likely to outperform under the conditions that both equity and bond markets struggle. I feel that a 50/50 or any other strategic mix will do well as a function of the added entropy, perhaps that explains the recent alpha of such strategies. They do a good job of taking advantage of rebalancing alpha. Ultimately for those that wish to forgo trading options or other non-linear approaches, TAA probably remains a good alternative to buy and hold, and is probably worthwhile combining with a strategic approach in some proportion.

    • 8 MarketSci

      Hello David – I agree with the majority of what you wrote “if all goes according to plan” (the future being unknowable and all that).

      I think that you hit on an important point I’ve been trying to make which is that expectations need to be tempered. I see a lot of big numbers thrown about based on a few decades of an environment very conducive to TAA, but very little tempering of expectations based on the challenges we face today.

      I diverge in stressing the fact that there is real limited upside remaining in UST, a key component of just about every model I see. That’s not my prediction, it’s a mathematical certainty.

      If in fact, as in your Japan example, UST is the best performer in the coming decade, that means the equity market will have been a very ugly place to invest. And if that’s the case, then it just reconfirms the need to temper expectations (because it’s a little overly-optimistic me thinks to believe market timing is going to wipe away all that ugly, leaving the same smooth equity curve these models would have produced historically).


  3. 9 david varadi

    i agree—the better question is–what is/are the alternative(s) to getting robust and reasonably high returns given that we cannot know the future? especially for investors that do not wish to trade very actively or use leverage.

    • 10 MarketSci

      There’s the gazillion dollar question.

      For investors who don’t trade actively, I think TAA (or one of its variations) is the best choice

      I’m not poo-poo’ing the concept of TAA – I still manage a big chunk of my LT wealth with it for all of the reasons we’ve talked about.

      I’m poo-poo’ing our collective lack of appreciation for the fact that the next decade will necessarily be a different environment than the previous few on which all of these backtests are based on.

      For investors who do trade actively, there are always other options.


      • 11 finsovet

        maybe this topic is not much covered by quants (unfortunately I have nothing to offer here), but definitely it was by value investors. If you are interested, J.Montier, J.Grantham, bill B.Gross and R.Arnott — all have been making a point how useless UST are these days, and offering lots of alternatives from cash to real assets.
        They all rightly point out the rates have been long under pressure to lure investors to long term or yield-rich assets. Money authorities the world over are trying to create ‘wealth effect’ to induce spending/growth on one hand and help banks’ balance sheets on the other. Where I disagree with the above mentioned gentlemen (who rendering long treasuries as almost worthless) is that I think long term, de-leveraging and this race to the bottom will continue for years to come. This is bullish for stocks AND bonds! Stocks are playing catch up with the bonds at the moment, but as the bond yields rise, this will mean revert in due course.
        I think TAA is the best vehicle to navigate these waves. I am using momentum/value/mean-reversion effects in my TAA.
        Bottom line, treasuries will come back, I would better stay the course, having in mind that abandoning a sound strategy is probably the most and worst investors’ trap.

  4. 12 MB

    I haven’t researched this so I don’t have any empirical evidence to back the claim up, but Ray Dalio said something interesting once ( well, likely more than once and unfortunately I don’t have a source for his quote). He said that while markets are often correlated, spreads are not. So maybe the evolution of TAA is moving from the base assets classes to spreads among or between them?

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