This blog post is available as a PDF here.

Summary­­

  • As an investment strategy, momentum focuses solely on prior returns.  
  • Being valuation agnostic, however, does not mean that a momentum strategy does not have first-order valuation effects on portfolio construction.
  • Using historical US sector data, we find that both cross-sectional and time-series momentum strategies may serve as good diversifiers to the potential risks of large structural repricing events in environments of high absolute valuations.

As an investment strategy, momentum should be agnostic of valuations.  Caring only about prior returns, whether total return changes comes from yield, earnings growth, or valuation changes, matters little.

Being agnostic, however, does not mean that the strategy does not have first-order effects in the dimension of valuations. 

Consider that, generally, momentum will buy appreciating assets (either relative or absolute, depending on whether we are talking about relative momentum or trend-following) and avoid depreciating assets.  We would expect that at least some of that appreciation or depreciation would come from multiple expansion or contraction.

To quote a recent Morgan Stanley Research Foundation piece,

“We think this helps momentum and trend avoid buying too early in stressed markets, allowing the cheap to get cheaper and the rich to get richer. We think this is why momentum and trend-based portfolios can be good diversifiers, doing better when stocks are down than up.”

With absolute market valuations at historically high levels, can the first-order effects of momentum in the value dimension help momentum diversify re-pricing risk?

 

Valuation Data and Assumptions

To explore this idea, we used sector and industry group data (30 unique sectors) from the Kenneth French data library.

Without the availability of traditional value measures, we must use an easily available proxy. In this analysis, we use dividend yield.  We defend this choice for multiple reasons:

  1. The most common measure of value is book-to-market (B/M), which assumes that the fair valuation – or at least on average a reasonable valuation -  of a company is its book value.  Another such model of valuation is the dividend discount model.  If we assume a constant growth rate of dividends (g) and constant cost of capital for the company (r), then book value should be proportional to total dividends (D), or, equivalently, book-to-market proportional to dividend yield
    bm-proportional-to-d
  2. Similarly, if you assume a constant long-term payout ratio (p), dividends per share (D) are proportional to earnings per share (E), which makes yield inversely proportional to price-to-earnings, another popular valuation ratio.
    e-proportional-to-d

Under these assumptions, yield is proportional to commonly quoted valuation metrics: book-to-market and earnings yield.  We argue, therefore, that by normalizing yield relative to historical levels, we can approximate relative valuation changes over time.

Each month, we compute a valuation z-score by evaluating current yield versus yields over the prior decade.  This z-score is then turned into a percentile.

 

Relative Momentum

To explore the first-order effects of relative momentum on valuation exposure, each month we sorted the 30 sectors by their relative momentum (using standard 12-1 month total returns) and then evaluated the change in valuation percentile over the next 1-month period.

The long-term averages are reported below.

 average-percentile-change-by-mom-rank

Source: Kenneth French Data Library.  Calculations by Newfound Research.

 

What we can see is that those sectors exhibiting low relative momentum (1-5) generally saw their valuation percentile increase over the next month by 0.5-1.0%.  On the other hand, those sectors exhibiting high relative momentum (25-30) saw their valuation percentile decrease over the next month by an average of 0.5-1.0%.

It is important to remember here that valuation percentiles increase means that the security is getting cheaper while decreasing percentiles indicate a security that is getting more expensive.  

If relative momentum were truly value agnostic we would expect all of these values to be near zero.

The take away is that relative momentum provides exposure to securities that are getting more expensive while avoiding those getting cheaper.

Of course, long-run averages going back eighty years do not help set short-term expectations.  Below, we’ve plotted the rolling 3-year spread between the average change of high momentum sectors (ranked 20-30) and low momentum sectors (ranked 1-10).

 rolling-3-year-spread

Source: Kenneth French Data Library.  Calculations by Newfound Research.

What we see is a volatile, yet fairly consistent negative spread, indicating that relative momentum will put you in sectors getting more expensive and avoid sectors getting cheaper.

Trend Following

While theoretically relying on similar behavioral biases, trend-following invests in those securities exhibiting positive prior returns and avoids those exhibiting negative prior returns (hence its alternative names of time-series momentum and absolute momentum). 

In this study, we use a simple metric to identify if a sector is exhibiting a positive trend or not: whether it is above its 10-month moving average.  If the sector is above its 10-month moving average, the next month is identified as being a period of positive trend.

Because there may exist periods where all sectors are exhibiting positive trends or all sectors exhibiting negative trends, computing a rolling spread change, as we did with relative momentum, is less informative. 

What we can look at, however, is the average results by sector and decade.

First, the average monthly valuation percentile change for periods identified as positive trends. 

 

1930s

1940s

1950s

1960s

1970s

1980s

1990s

2000s

Food

-2.12%

-0.84%

-0.94%

-0.03%

-0.91%

-0.96%

-0.39%

-0.15%

Beer

-2.47%

-0.59%

-1.19%

-0.68%

-0.54%

-0.85%

-0.14%

0.29%

Smoke

-1.94%

-0.25%

-1.96%

0.25%

-0.65%

-1.09%

-0.62%

-0.90%

Games

1.48%

-0.59%

-0.75%

-0.22%

-0.82%

-0.88%

-0.26%

1.07%

Books

-2.36%

-0.80%

-0.90%

0.09%

-0.94%

-1.10%

-0.46%

0.92%

Hshld

0.21%

-0.53%

-0.79%

0.01%

-0.58%

-0.95%

-0.13%

0.56%

Clths

0.57%

-0.41%

-1.24%

-0.20%

-0.58%

-0.98%

0.05%

-0.20%

Hlth

-1.03%

-0.26%

-1.35%

-0.02%

-0.32%

-0.88%

-0.59%

0.61%

Chems

0.13%

-0.44%

-0.63%

-0.53%

-1.51%

-0.50%

-0.90%

-0.67%

Txtls

-2.23%

-0.78%

-1.30%

-0.93%

-0.30%

-0.95%

-0.04%

-1.53%

Cnstr

-1.80%

-0.18%

-1.02%

-0.85%

-0.16%

-0.27%

-0.06%

-1.33%

Steel

-0.95%

-0.03%

-1.02%

-0.44%

-0.25%

-1.25%

-0.33%

-2.24%

FabPr

-2.87%

0.06%

-1.00%

-0.57%

-1.41%

-2.46%

-0.43%

0.24%

ElcEq

-1.98%

-0.43%

-0.78%

-0.79%

-0.54%

-0.93%

-0.72%

-0.19%

Autos

-1.03%

-0.30%

-0.90%

0.04%

-1.39%

-0.37%

-0.38%

-1.73%

Carry

-3.56%

0.33%

-0.90%

-0.10%

-1.22%

-1.70%

-0.31%

-0.31%

Mines

-1.42%

-1.24%

-0.28%

-0.45%

-0.57%

-1.44%

-1.46%

0.07%

Coal

1.76%

0.04%

-0.86%

-0.60%

0.70%

-1.16%

0.28%

0.38%

Oil 

-1.96%

-0.39%

-0.87%

-0.80%

-1.24%

-1.09%

-0.77%

-0.08%

Util

-1.42%

-1.05%

-0.54%

0.50%

-0.26%

-1.04%

-0.24%

-0.14%

Telcm

-1.31%

-0.58%

-0.71%

0.34%

-0.16%

-0.97%

-0.19%

-0.26%

Servs

-4.43%

-0.31%

-0.60%

-0.18%

-0.06%

-0.65%

-0.55%

0.48%

BusEq

-3.31%

-0.53%

-0.66%

0.12%

-1.07%

-1.25%

-0.27%

0.76%

Paper

-2.25%

-0.56%

-0.96%

-0.47%

-1.51%

-0.95%

-0.23%

-0.83%

Trans

-1.72%

-0.39%

-1.41%

-0.46%

-1.11%

-0.56%

-0.44%

0.78%

Whlsl

-3.03%

-1.28%

-0.61%

-0.25%

-0.05%

-1.13%

-0.37%

-0.52%

Rtail

-1.17%

-0.81%

-1.06%

-0.34%

-0.86%

-0.91%

-0.25%

0.92%

Meals

-7.89%

-0.16%

-0.87%

0.00%

-0.26%

0.11%

-0.22%

0.79%

Fin 

-0.51%

-0.89%

-0.70%

-0.57%

-0.93%

-1.71%

-0.34%

-0.70%

Other

1.25%

-0.22%

-0.72%

-0.27%

-0.51%

-0.79%

-0.92%

-0.29%

 

We see, by in large, that regardless of sector or decade, periods of positive trends were periods of valuation multiple expansion. 

Not surprisingly, we see the opposite for periods of negative trends.

 

1930s

1940s

1950s

1960s

1970s

1980s

1990s

2000s

Food

0.35%

2.24%

0.33%

1.35%

2.94%

0.68%

1.31%

2.67%

Beer

-0.24%

1.85%

1.23%

2.44%

1.53%

0.32%

0.58%

1.47%

Smoke

1.09%

1.75%

3.47%

1.66%

2.06%

1.18%

3.10%

2.29%

Games

0.18%

1.92%

-0.85%

0.67%

2.20%

-0.06%

0.77%

-0.17%

Books

1.37%

2.06%

0.41%

0.54%

3.04%

0.78%

0.97%

0.67%

Hshld

-1.19%

1.99%

-0.53%

0.52%

2.17%

0.04%

0.63%

1.14%

Clths

0.01%

2.40%

0.62%

0.55%

2.40%

0.57%

-0.13%

2.19%

Hlth

-0.13%

1.63%

2.28%

0.40%

2.32%

0.73%

1.43%

1.00%

Chems

0.65%

2.15%

-1.15%

2.56%

2.48%

0.92%

0.94%

3.40%

Txtls

-0.13%

2.74%

0.21%

2.58%

1.45%

0.45%

1.18%

0.93%

Cnstr

0.68%

1.51%

0.58%

1.73%

2.12%

-1.75%

0.14%

2.55%

Steel

1.88%

1.26%

0.55%

1.90%

0.54%

1.42%

-0.01%

4.80%

FabPr

2.39%

1.54%

0.53%

1.44%

4.35%

3.56%

-0.15%

1.74%

ElcEq

1.98%

1.41%

-0.06%

1.93%

2.10%

1.07%

2.16%

2.36%

Autos

1.16%

1.66%

0.12%

2.02%

1.51%

0.49%

-0.34%

1.28%

Carry

3.02%

0.50%

0.77%

0.99%

3.16%

2.12%

0.63%

2.41%

Mines

5.22%

1.37%

-0.83%

1.52%

2.50%

0.48%

1.81%

-0.34%

Coal

2.25%

0.81%

-0.24%

1.08%

0.83%

-0.10%

-0.29%

0.31%

Oil 

1.82%

2.62%

0.04%

2.86%

3.30%

2.97%

0.47%

2.79%

Util

3.40%

1.11%

1.17%

1.47%

0.29%

0.90%

0.92%

2.34%

Telcm

2.08%

2.03%

0.62%

1.54%

0.44%

0.32%

0.46%

1.77%

Servs

-0.17%

1.79%

-1.94%

0.53%

1.67%

0.33%

1.16%

0.43%

BusEq

2.13%

2.14%

1.02%

1.29%

2.26%

0.32%

0.55%

0.68%

Paper

0.68%

2.53%

0.72%

1.38%

3.48%

0.47%

0.47%

3.04%

Trans

3.33%

1.51%

1.67%

1.09%

2.58%

0.14%

0.33%

0.76%

Whlsl

0.00%

1.98%

0.06%

0.82%

1.66%

0.72%

0.63%

2.64%

Rtail

-0.49%

2.39%

1.28%

1.27%

2.58%

0.27%

0.30%

0.54%

Meals

0.32%

1.63%

0.46%

0.06%

1.99%

-2.58%

0.23%

0.63%

Fin 

0.71%

1.56%

0.93%

1.80%

3.49%

2.11%

1.53%

1.14%

Other

-0.35%

1.02%

-0.42%

0.98%

1.56%

1.88%

0.49%

1.78%

  

As in the relative momentum case, buying positively trending assets and avoiding negatively trending ones has historically helped avoid falling knives and value traps.

 

Conclusion

While most investors explore the addition of an investment strategy for its return merits, we believe that an additional reason to consider a momentum strategy is for its potential diversification and hedging benefits.  To quote Morgan Stanley,

“Expensive valuations in traditional assets can result in big structural repricings, which are inherently difficult to time. We think that momentum-based portfolios have the ability to capture such a structural repricing. The price to be paid in terms of sub-par returns in our base case scenario (low returns, range-bound markets) while we wait for such a repricing may be worth it as the effective cost of such a hedging/diversification tool.”

 

Client Talking Points

  • Portfolio construction is a practice whereby we try to make the whole greater than the sum of the parts.    
  • In today’s environment of expensive valuations, momentum strategies may help diversify and hedge away some of the risk of big structural repricings.

Corey is co-founder and Chief Investment Officer of Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Corey is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients.

Prior to offering asset management services, Newfound licensed research from the quantitative investment models developed by Corey. At peak, this research helped steer the tactical allocation decisions for upwards of $10bn.

Corey is a frequent speaker on industry panels and contributes to ETF.com, ETF Trends, and Forbes.com’s Great Speculations blog. He was named a 2014 ETF All Star by ETF.com.

Corey holds a Master of Science in Computational Finance from Carnegie Mellon University and a Bachelor of Science in Computer Science, cum laude, from Cornell University.

You can connect with Corey on LinkedIn or Twitter.

Or schedule a time to connect.