TAA vs buy and hold in overvalued markets

US stocks are expensive. There seems to be article after article on the expensiveness of US stocks these days. Plus, bonds in general are really really expensive. Both US stocks and all bonds are in the top (90th+ percentile) tiers of expensiveness relative to history. As Cliff Asness of AQR points out, the problem is that they are both at these expensive levels at the same time, which hasn’t happened ever. That means that a portfolio of US stocks and bonds (50/50, 60/40, take your pick) has a very low expected return going forward. Maybe the lowest ever. OK. Now what? The standard advice to all of this is generally to stick to a buy and hold investment approach, save more money, and to lower your expected portfolio return. That is perfectly fine advice. The problem is most investors can’t stick to it when markets start to down trend and that there are better approaches.

Ben Carlson, from A Wealth of Common Sense, wrote a great piece yesterday on asset allocation. My favorite line was

Behavior trumps all because even the greatest portfolio or strategy in the world does you no good if you can’t get yourself or your investors to stick with it.

In my opinion this is the biggest problem with the incessant drumbeat of buy and hold investing. Despite the mountains of evidence that most investors can’t stick to a buy and hold approach it is still the most recommended approach to any asset valuation concerns. As Ben goes on to point out in this piece there is a better approach for most investors – basically some type of simple rules based TAA approach. I want to highlight one particular aspect of a TAA approach here in this post. I want to show how the simplest of TAA approaches has compared to buy and hold when the stock market has been overvalued.

I’m going to use the Shiller P/E for the SP500, or CAPE, as the indicator of value for the US stock market. There are several other popular ones. Take your pick. They all showing something similar. I’m sure you’ve seen some version of the chart below going around for years.

It’s a horrible short term indicator of performance but not so bad over longer periods of time. It’s average is about 17 going back to 1929. We’re at about 29 right now. Anything over 20 is usually taken as a sign of high valuation. The market has spent 36% of all years with a CAPE above 20. Let’s take all years starting from 1929 when the CAPE was over 20 and look at subsequent 1 year, 3 year, and 5 year compounded returns and the range of those returns. Then I’ll do the same thing for maybe the simplest yet one of the best TAA strategies, Antonacci’s GEM strategy. The results are below.

Note: I must say I was quite unfair to the GEM strategy. Since published results only go back to 1973 I created my own simple GEM strategy using SP500 returns going back to 1929, using 12 month momentum of the SP500 vs cash at 0%. Basically a simple absolute momentum strategy. Real life results would be better.

The first thing to note is that yes, during periods of high valuation, CAPE > 20, stocks return less than normal, 5% vs the long term historical 10% or so. However, the range of those returns is crazy wide. From a high valuation you could have been unlucky enough to earn -12% a year for 5 years or lucky enough to earn almost 24% for 5 years. That’s why CAPE is a horrible indicator of short term performance. But on average your returns are lower than other periods with better valuations. Now look at the GEM strategy. About double the average return and a smaller range. Also, importantly it is only the low end of the range is cut off, not the upper end. If the market continues to higher valuations you get to participate in the upsides and if the market starts a turn to lower valuations you have a limited down side. All with a clear set of simple rules.

In summary, a TAA strategy is a powerful alternative to buy and hold, especially in periods of high market valuation. While returns for TAA in periods of high valuation are also lower than their long term averages they are significantly better than buy and hold and maybe more importantly with much lower drawdowns. Just like with any strategy that is different it also comes with its own set of challenges but the odds are that it will have better outcomes going forward.

Full Disclaimer - Nothing on this site should ever be considered advice, research or the invitation to buy or sell securities. These are my personal opinions only.

12 thoughts on “TAA vs buy and hold in overvalued markets

  1. Very timely reminder in the face of all the overvaluation talk. I was surprised to see 5yr returns for the SP500 so high when CAPE >20. Most of the articles in the popular press seem to imply returns would be negative over this period when valuations are high. Good to see the actual data!

    1. 1994 to 1998. CAPE went from 20.5 to 32.3. And then it kept going up for two more years to 43. Crazy times!


  2. Nice write-up Paul! With my TAA strategy implemented, I’ll be ready when valuations return to historical norms (unlike in 2000 when I was B&H). The one thing I still struggle with is the tax bill from using TAA strategies in my non-retirement accounts. Any suggestions?

    1. Best suggestion I have is to use the lowest turnover TAA strategy you can. Like GEM.


      1. Would really love to see a post on tax strategy for TAA and quant portfolios. I always assumed the excess returns of AGG3 or AGG6 made them better than GEM even in taxable accounts, considering that most of the short-term trades in the AGG strategies are losses and the big winners tend to be held for more than a year. Would definitely be interested to see whatever info you have on this.

        Would also be interested to hear about any backtests you have done on modifications to GEM (using total U.S. stock market instead of SP500, TAA bond instead of BND, etc.).

          1. Any resources you recommend for this then? Any tools that can backtest strategies with taxes taken into account?

  3. How come the GEM strategy was posed as a comparison? What about the IVY (GTAA) Portfolios? Thanks for all your information.

    1. Because it is the only one that I could construct a data series back to the great depression.


    1. Yes, I have. The guys at AllocateSmartly have done that as well. I leave that up to them.


Comments are closed.