Today I want to talk about momentum. I’m always struck by the hand wringing I see from investors as prices reach new highs. This seems particularly evident over the last year during a great run for stocks. This is odd when you look at the historical evidence for momentum. Momentum is the most powerful factor in investing by a significant margin. Higher prices tend to lead to higher prices. Yet momentum is probably the hardest factor for most investors to internalize and implement in their portfolios. It is counter intuitive, especially if you have a heavy value bias as many do. In this post I want to highlight a very counterintuitive yet very effective strategy to emphasize just how powerful momentum can be and why momentum should be a part of your portfolio.
First, let’s remind ourselves of the evidence for momentum. Below is a table I’ve shown before. It is the AQR factor model (an enhancement to the famous Fama-French model).
Momentum (UMD in the table) is by far the most powerful of the factors. Almost twice as strong as the next strongest factor. Antonacci in his research on asset class momentum has some of the best and easily understandable evidence for it. Basically, you should consider applying momentum to your investing in some form or the other to increase your chances for market beating returns and lower risk. Now, let’s look at a very simple way to apply momentum to equities.
One of the most powerful expressions of momentum that I’ve seen is the Buy At the High strategy. Jake at Econompicdata describes the strategy at the bottom of this post. In this strategy you invest in stocks only if the previous month closed at a an all time high, otherwise invest in US treasuries. Sounds kind of crazy but look at the results.
Not only are the returns higher than buy and hold but more importantly drawdowns are reduced by a huge amount. High prices tend to lead to higher prices, until they don’t. This strategy would rank up there with the top TAA strategies that invest globally, don’t trade often, and beat buy and hold which I discussed in Choosing your first TAA strategy. It would also take you a few minutes a month to implement.
I ran this strategy through my database to get a better feel for the results. Since 1970, for US stocks (SPY), this strategy would have only had 5 down years with the worst year being a return of -6.7%. For global developed stocks, e.g VEA or EFA, this strategy would have had only 4 down years with the worst year being a return of -10%. In addition to the good returns and low drawdowns think about the emotional impact of these types of results. It definitely has a high ‘sleep at night’ factor. Yet, most investors would never consider this strategy. Something to think about.
In short, momentum is a powerful factor in investing but it is counterintuitive and thus hard for most investors to implement in their portfolios. A very simple TAA strategy, of owning equities only when they are making new all time highs, illustrates momentum’s effectiveness very clearly.
10 Comments
Alan · November 7, 2017 at 5:54 pm
Paul,
I started to get interested in momentum several month ago after coming across an article that showed how applying momentum to Merriman’s Ultimate Buy-and-Hold Portfolio increased it’s return from 10.0% annualized to 16.6% annualized from 1973 to 2015. Researching momentum I came across your site and have been reading all the articles for the last several month. Great work Paul!
One issue I have is that I am more of a Buy-and-Buy investor, maybe dollar cost averaging investor. I have been investing at least 20% of my steadily rising income since mid-eighties and, although my investment strategies changed over the years, I have yet to make a withdrawal. Most charts, like the one above, show growth of an initial investment over time and are very dependent on the particular time period chosen. I would be very interested to see how various TAA strategies do when invested in on a continuous basis in real terms. Perhaps you can address this in a future article.
Nick · November 7, 2017 at 8:20 pm
Great article – and you are right on about how counter intuitive it is. As the old saying goes, “Buy Low, Sell High.” But, this actually suggests the opposite.
paul.novell@gmail.com · November 8, 2017 at 5:36 am
Yep. exactly.
Tim · November 8, 2017 at 7:48 am
Good post as always Paul.
As a value investor I also struggled buying companies with a share price that has already moved up, not even to mention companies hitting a 52 week high.
But I have gotten over it – mostly.
What helped a lot was this article I researched – if you do the work it does a lot to convince you of the validity of an indicator.
https://www.quant-investing.com/blogs/general/2015/01/13/10-myths-about-momentum-investing-squashed
This may be an even better momentum indicator – have you taken a look at it yet?
https://www.quant-investing.com/blogs/general/2017/05/15/how-to-find-stocks-on-the-move-with-a-better-momentum-indicator—exponential-regression
paul.novell@gmail.com · November 8, 2017 at 1:55 pm
Thanks Tim. I’ll take a look at your links. I’ve gotten over it mostly as well even though implementing momentum on individual stocks is even harder but even more rewarding. I remember when I started with a pure momentum quant portfolio and one of the first buys I had to make was NVDA. Price was 50% above it’s 200 day SMA. I thought it was nuts and really had to fight myself hard to make the trade. That was almost 2 years ago and it’s still part of the same portfolio…Crazy thing is that is not such an uncommon thing.
Paul
Tim · November 9, 2017 at 7:22 am
I still have the same but a LOT less than before.
I also use a strict trailing stop loss (also really a momentum selling strategy) – that way I have had over 200% gains on companies I already thought was overvalued.
Dave W · November 8, 2017 at 11:47 am
Another value investor coming to terms with this, but the data talks….
Talking of data – have you run this for other indices ? You show MSCI but say you ran it on SPY…
For an earlier starting point than 1969 ? (Malkiel starts in 1940 for most of his data sets)
For a different time since peak (eg last quarter ?)
For a more local peak (eg 12 mo high in the last month ?)
Also, what Treasury did you use ? 30 year ?
paul.novell@gmail.com · November 8, 2017 at 1:48 pm
Yep. It’s tough overcoming those deep value biases.
I’ve run it myself for SPY and EFA indices. I’ve also run it on a shorter time frame for QQQ and it works just as well. The graph in the post is from the blogger I referenced and he ran it for the MSCI world index.
The earliest data for bonds I have is 1948 so I have only run the SPY set back to 1948. The foreign and word indices don’t go back that far. Monthly seems to work best. I tired weekly and it doesn’t work nearly as well. Too much noise.
I always use intermediate treasuries, i.e 10 yr. Results are worse for shorter term bonds. Haven’t tried longer ones.
Paul
Bernduffy · November 15, 2017 at 8:19 am
Hey Paul
Great thoughts as always. Btw not finding EP newsletter sign in. Is it possibly because I’m on iOS phone?
paul.novell@gmail.com · November 15, 2017 at 8:38 am
Shouldn’t make a difference. This is the link to login.
Comments are closed.