Time for the part 3 of the series on using economic indicators to time the market. In this post I’ll add a simple extension to the models analyzed in Part 2. If you haven’t read the first two posts you won’t understand this one.
I’m just extending the model to include foreign stocks, foreign developed and emerging markets. This is much more reflective of real diversified portfolios – even with the heavy home bias amongst US investors. Lets see what that does to the results.
For foreign developed stocks I’m using EFA ETF because it has the longest history. Similarly, for emerging markets I’m using the EEM ETF. With the addition of foreign stocks to the mix we’ll do two things; replace the risk on period with an equal allocation to SPY, EFA, and EEM vs just investing in SPY. We can also use asset class momentum, just like in the AGG portfolios, and invest in the top performing ETF among the three. For the safe asset, we’ll use IEF since it performs the best among the safe assets.
The results over the analyzed time period are below. Just as in the last post these ETF have not been available for a long time thus limiting how far we can look back.
Two important things to note. In this time period, adding foreign equites reduces returns over a system that just uses US stocks. This is one of those cases where backtests can be quite misleading. Going forward, over the long term, would you expect these results to be the same? That global diversification would lead to lower results. And more importantly, would you be willing to invest based on that belief? I wouldn’t. I would choose the more diversified portfolio. Also, notice that as you would expect asset class momentum further improves results over a simple equally weighted diversified model.
Good stuff. We now have two globally diversified tactical asset allocation models based on the unemployment rate as the master risk on/off switch, with the traditional 200 day SMA (or 12 month absolute momentum, whichever you prefer) as the secondary risk off/on indicator. And for those that are following the UI indicator in real time, the latest unemployment rate was reported on June 3 at 4.7% down from 5.0% in the previous month. Thus it remains below it’s 12 month moving average.
11 Comments
Tim · June 8, 2016 at 8:55 pm
Hi Paul.
Really interesting series of articles with some good food for thought. Appreciated.
There is a decent website with good interactive charts where indicators such as the above can be visualized (and where all sorts of indicators, including moving averages, can be applied). It is TradingView. Worth a look perhaps, if you haven’t already, for a serious bundle of free data.
paul.novell@gmail.com · June 9, 2016 at 8:23 am
Thanks Tim. Was unaware of that site.
Paul
Max · June 9, 2016 at 7:56 am
Hi Paul,
First just wanted to say great site, lots of good information for any investor, especially beginners.
Question about your use of Asset Class Momentum, for the last two portfolio you listed are you staying out of the top performing ETF (Based on 12 month momentum) Until it crosses it’s 200 MA or are you using it instead of the 12-Month abs. momentum for determining the top performer?
paul.novell@gmail.com · June 9, 2016 at 8:27 am
Max,
I’m using the same momentum measure as in the AGG portfolios, which is the average of the 1,3,6,12 month total returns. When the UI indicator is OFF, everything is invested in the risk asset (SPY, EFA, or EEM), whichever has the highest momentum. When the UI indicator is ON then the 200 day SMA is used to determine whether to invest in the risk asset (SPY, EFA, or EEM) depending on which one has the most momentum, or the safe asset (IEF).
Paul
Don Thompson · June 9, 2016 at 10:52 am
Paul, more great information and ways for us to take advantage. Thanks for your continuing efforts.
In your charts you sometimes include draw downs and some other times worst year. I like both, and would like to see both. For example, if a portfolio has a significant draw down during the year, and recovers much of that by year end, we might see a worst year of 5%, but a DD of 25%. If I only see a worst year of 5% in the charts, and then see my portfolio dropping way below that during the year, I doubt that I would stick with that portfolio. But, if I had seen both statistics in the charts, DD and worst year, I wouldn’t be surprised by a drop below the worst year statistic. Any thoughts?
paul.novell@gmail.com · June 9, 2016 at 11:31 am
Thanks Don. I always try and use max drawdown when the data is available. It is the best and most appropriate statistic. If I only have calendar year returns, as in the case of the long term return data I post, then I can’t post drawdowns, I can only talk about worst year. But you can use rules of thumbs to estimate drawdowns for asset classes; US stocks: -50%, Foreign developed stocks: -60%, Foreign emerging stocks: -70%, Us govt bonds: -10%, US corporate bonds, -20%.
Don Thompson · June 9, 2016 at 11:38 am
Thanks, Paul. I’ll build that into my portfolio info.
kevin tomera · June 9, 2016 at 5:17 pm
First, your work and insight is greatly apprecated.
The unemployment 12 month moving average crossover with monthly unemployment makes more sense as a prognostic indicator. than the 200 day EMA (which depends on the herd) My concern is that these are unusual times with unemployment now ending at 6 months rather than 18 months which explains why with very poor new job creation the unemployment rate fell . Only 35,000 new jobs but 645,000 people added to the 94.7 million no longer looking for work. (this does not include 1.8 million+ college graduates this june. It might be better to look at employment or labor force participation ?
paul.novell@gmail.com · June 10, 2016 at 9:57 am
Hey Kevin, I would recommend reading the source post, from Philosophical Economics linked to in Part 1, for this work I did. The UI indicator has a long track record of pretty good results, back to the Great Depression, despite changes in measurement techniques. Philo Econ also looks at a bunch of other economic indicators and the unemployment rate performs the best.
Paul
Bernie Duffy · June 12, 2016 at 7:11 am
Thanks for this, Paul. With the yield curve no longer an accurate recession indicator (due to CB manipulation of interest rates and money supply) this seems like the next best thing. One of my go to guys, David Rosenberg, just last week tilted bearish mostly based on unemployment figures. Would love to hear your thoughts on recession strategies. Seems inevitable at this point.
paul.novell@gmail.com · June 12, 2016 at 10:01 am
Hi Bernie, I have no particular recession strategies. Markets drop the most during recessions, but they can drop quite a bit outside of recessions as well. So, its best to have strategies that protect to the downside in any environment. Then, besides intellectual curiosity, there’s no reason to overly focus on recessions over plain and simple risk management.
Paul
Comments are closed.