Why Investors Should Combine Value and Momentum

Why Investors Should Combine Value and Momentum

March 22, 2016 $mtum, $vlue, Momentum Investing, Value Investing
Print Friendly

In the past we have discussed how to combine value and momentum strategies to improve an equity allocation.

In this piece we discuss why an investor should combine use value and momentum.*

Many investors recognize that stand-alone value and momentum strategies have historically worked. Of course, these strategies don’t work all the time and can have long streaks of terrible performance.  But long-only value and momentum investing, while interesting, don’t represent the most puzzling anomaly from financial research. The biggest puzzle is that long/short value and momentum have a negative correlation. In other words, long/short value and momentum portfolios tend to work at different times. An implication is that a long-term equity investor can manage a portfolio allocated to long-only value and momentum strategies and capture the benefits of these anomalies, but also eliminate a lot of the idiosyncratic risk associated with following either of these strategies in isolation.

Cliff Asness has a wonderful paper that highlights the need for investors to consider the portfolio benefits of a strategy, in addition to a strategy’s stand-alone investment merits. The paper is titled, “Momentum in Japan: The Exception That Proves the Rule.”

The paper examines the data from Japanese equities, which some researchers interpret as suggesting that momentum investing doesn’t work. For example, Eugene Fama recently mentioned that the Japanese data suggest that momentum might be an artifact of data-mining.

Asness takes on the challenge from a different angle. First, he highlights that long/short momentum didn’t do well from 1981-2010. A table from the paper:

momentum in japan
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

Cleary, long/short momentum wasn’t a great stand alone strategy. But modern portfolio theory highlights that the assessment of an investment strategy should not be done in isolation, but in the context of a broader portfolio (e.g., insurance isn’t bad because it has a negative expectation). Asness brings this point forward. He shows that in Japan  long/short momentum has a strong negative correlation with long/short value portfolio. The reader can visualize this in the chart below:

value and momentum in japan
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

On average, when one strategy zigs, the other zags. So from a diversification and risk-management perspective, momentum still “works.”

Of course, the research above talks about long/short portfolios and the analysis and takeaway is complex for those who aren’t card-carrying finance geeks. We’re going to simplify the situation. We will make the same point using long-only portfolios and have no requirement that the reader be a finance geek.

Let’s ask a basic question: what happens when one examine long-only value and momentum portfolios?

So others can replicate our results, we examine data from Ken French’s website. Specifically, we examine the returns to the value-weight portfolio formed on Value (B/M) and the value-weight portfolio formed on Intermediate-Term Momentum for United States stocks.

First, we examine the returns from 1927-2015 for 4 portfolios:

  1. Value: Top Decile of firms ranked on B/M from Ken French’s website. We use the value-weight portfolio returns.
  2. Momentum: Top Decile of firms ranked on Intermediate-Term momentum (past 12 months excluding last month)  from Ken French’s website. We use the value-weight portfolio returns.
  3. SP500: The total return to the S&P 500.
  4. RF: Total return to the risk-free rate (taken from Ken French’s website).

Results are gross of transaction costs and management fees. All returns are total returns and include the reinvestment of distributions (e.g., dividends).

1/1/1927 – 12/31/2015

why to combine value and momentum fig 1
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

First, one notices the outperformance over the past 88 year of value and momentum compared to the market (before fees). One sees that the long-only value and momentum portfolios are not negatively correlated (as both have market exposure), while long/short value and momentum portfolios are negatively correlated (as they have minimal market exposure due to being a long/short portfolio). However, an important finding is that value and momentum have a lower correlation (67.21%) compared to value and the SP500 (84.65%) as well as momentum and the SP500 (81.46%).

Given that these portfolios have lower correlations, we examine what happens when one invests 50% of a portfolio in Value and 50% in Momentum each month (assume a monthly rebalance back to 50%/50%).

We examine the returns from 1927-2015 for 4 portfolios:

  1. 50% Value, 50% Momentum: A strategy that invests 50% in the Value portfolio and 50% in the Momentum portfolio described below. We assume the investor rebalances the portfolio to 50%/50% every month.
  2. Value: Top Decile of firms ranked on B/M from Ken French’s website. We use the value-weight portfolio returns.
  3. Momentum: Top Decile of firms ranked on Intermediate-Term momentum (past 12 months excluding last month)  from Ken French’s website. We use the value-weight portfolio returns.
  4. SP500: The total return to the S&P 500.

Results are gross of transaction costs and management fees. All returns are total returns and include the reinvestment of distributions (e.g., dividends).

1/1/1927 – 12/31/2015

why to combine value and momentum fig 2
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

As expected, the overall 50%/50% performs well, as an investor in value and momentum strategies receives some free diversification benefits of investing in both factors. The combination portfolio is nice because sometimes Value works, while other times Momentum works. We try to show this visually through two graphs below.

To assess the ability of combination value and momentum portfolios to smooth the path to long-term performance, we examine the spread between 5-year compound annual growth rates for a specific strategy relative to the passive benchmark (S&P 500).

First Half (1/1/1927 – 12/31/1973)

why to combine value and momentum fig 3
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

Second Half (1/1/1974 – 12/31/2015)

why to combine value and momentum fig 4
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

The evidence suggests that a value and momentum system, which combines both pure value and pure momentum into a single portfolio, may prevent a value-only investor or a momentum-only investor from suffering through extended, long-term stretches of poor performance. Of course, not all pain can be erased, and investors must always be aware that they should expect to endure sustained stretches of volatility and relative underperformance, even with a globally diversified value and momentum equity portfolio. Nothing in life is easy, especially when it comes to investing!

Another way to view the volatility from year to year for Value and Momentum is to examine the spread between 1-year compound annual growth rates for a specific strategy relative to the passive benchmark (S&P 500). The chart below (from the 2nd half) shows that from year to year, long-only value and momentum can quickly flip from being a big winner to a big loser, however the combination portfolio (on average) is a bit more stable year to year.

Second Half (1/1/1974 – 12/31/2015)

why to combine value and momentum fig 5
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

As many remember, momentum did well in the internet bubble, and poorly thereafter. Similarly, Value did well in 2014, but got crushed in 2015. Bottomline: combining value and momentum helps manage long stretches of underperformance.

Despite the clear benefits of maintaining roughly static exposures to both value and momentum, readers always ask us if there is a way to “time” the factors. We think this is a risky proposition at the outset, because we give up a “bird in the hand” in the form of a well-established diversification benefit, to take on a “potential bird in the bush” benefit of capturing some extra returns by timing across value and momentum. We present the data below on the most common concept people ask us to test.

A simple idea is the following — keep a track of the total return to Value and Momentum over the past 12 months, and whichever factor has done better, allocate more to that factor. In other words, we can deploy a basic relative-strength rule across the two portfolios.

To test this idea, we examine the returns from 1928-2015 for 4 portfolios (Note we lose 1927 as we need to compute the 12-month return during this year):

  1. 50% Value, 50% Momentum: A strategy that invests 50% in the Value portfolio and 50% in the Momentum portfolio described above. We assume the investor rebalances the portfolio to 50%/50% every month.
  2. Relative Strength (25%, 75%): Based on the past 12 months total return to Value and Momentum, every month this portfolio invests 75% in the strategy (Value or Momentum) with the higher past 12 months return, and 25% in the strategy (Value or Momentum) with the lower past 12 months return. We assume the investor rebalances the portfolio every month.
  3. Relative Strength (0%, 100%): Based on the past 12 months total return to Value and Momentum, every month this portfolio invests 100% in the strategy (Value or Momentum) with the higher past 12 months return, and 0% in the strategy (Value or Momentum) with the lower past 12 months return. We assume the investor rebalances the portfolio every month.
  4. SP500: The total return to the S&P 500.

Results are gross of transaction costs and management fees. All returns are total returns and include the reinvestment of distributions (e.g., dividends).

1/1/1927 – 12/31/2015

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request. These results were updated 5/10/2016.

The 50/50 portfolio, which always captures the diversification benefits of value and momentum, can be improved at the margin by using relative momentum across the two portfolios. However, there are (1) small costs of switching back and forth between the factors and (2) possible tax implications of making a 100% switch from value to momentum (or vice versa).

Big picture — a static value and momentum portfolio has known diversification benefits, as one factor tends to work well when the other factor is performing poorly. If one attempts to “time” factors, relative strength is a decent idea. However, when it comes to value and momentum, we recommend that investors don’t get cutesy: 1) invest in concentrated value; 2) invest in concentrated momentum; 3) exploit portfolio benefits (50/50 portfolio); 4) have long horizon. We provide the tools and the products to facilitate.

Active equity investing really isn’t that complicated, but it sure isn’t easy.

Endnotes

*We aren’t the only researchers to suggest this coarse of action. For example, we recommend readers explore the paper, “Value and Momentum Everywhere,” which is outstanding.

Appendix — Net of Fee returns

We examine the returns from 1927-2015 for 4 portfolios, taking fees off the portfolios for trading costs:

  1. 50% Value, 50% Momentum: A strategy that invests 50% in the Value portfolio and 50% in the Momentum portfolio described below. We assume the investor rebalances the portfolio to 50%/50% every month. This returns stream is net of the fees described below for value and momentum.
  2. Value: Top Decile of firms ranked on B/M from Ken French’s website. We use the value-weight portfolio returns and deduct 0.25% annually for the annual rebalance.
  3. Momentum: Top Decile of firms ranked on Intermediate-Term momentum (past 12 months excluding last month) from Ken French’s website. We use the value-weight portfolio returns and deduct 3.00% annually for the twelve monthly rebalances (0.25% per rebalance).
  4. SP500: The total return to the S&P 500, this return stream is gross of any fees.

Results are net of transaction costs and management fees described above. All returns are total returns and include the reinvestment of distributions (e.g., dividends).

1/1/1927 – 12/31/2015 (Net of Fees)

why to combine value and momentum fig 7
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

The results (net of fees) are similar to the gross of fee table above, albeit with lower returns an investor would receive after accounting for transaction costs.

Below we split the study into first half and second half using the same fee assumptions.

First Half (1/1/1927 – 12/31/1973) Net of Fees

why to combine value and momentum fig 8
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

 

Second Half (1/1/1974 – 12/31/2015) Net of Fees

why to combine value and momentum fig 9
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

 

***

Please remember that past performance is not an indicator of future results. Please read our full disclaimer. The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. This material has been provided to you solely for information and educational purposes and does not constitute an offer or solicitation of an offer or any advice or recommendation to purchase any securities or other financial instruments and may not be construed as such. The factual information set forth herein has been obtained or derived from sources believed by the author and Alpha Architect to be reliable but it is not necessarily all-inclusive and is not guaranteed as to its accuracy and is not to be regarded as a representation or warranty, express or implied, as to the information’s accuracy or completeness, nor should the attached information serve as the basis of any investment decision. No part of this material may be reproduced in any form, or referred to in any other publication, without express written permission from Alpha Architect.

***

If you liked this post, don't forget to subscribe to our blog.




About the Author

Jack Vogel, Ph.D.

Jack Vogel, Ph.D., conducts research in empirical asset pricing and behavioral finance, and is a co-author of DIY FINANCIAL ADVISOR: A Simple Solution to Build and Protect Your Wealth. His dissertation investigates how behavioral biases affect the value anomaly. His academic background includes experience as an instructor and research assistant at Drexel University in both the Finance and Mathematics departments, as well as a Finance instructor at Villanova University. Dr. Vogel is currently a Managing Member of Alpha Architect, LLC, an SEC-Registered Investment Advisor, where he heads the research department and serves as the Chief Financial Officer. He has a PhD in Finance and a MS in Mathematics from Drexel University, and graduated summa cum laude with a BS in Mathematics and Education from The University of Scranton.


  • Thomas Musselman

    If you time on your list of “to dos” it would be interesting to see the above compared to combining the value and momentum factors (but not all the other possible factors you use in your more complicated investment strategy) instead of in two halves of the total tested portfolio, i.e. the momentum part of the value subuniverse, and the value part of the momentum subuniverse. Instead of 50% in each.

  • Jack Vogel, PhD

    I believe I answered this question in the following post:

    http://blog.alphaarchitect.com/2015/03/26/the-best-way-to-combine-value-and-momentum-investing-strategies/#gs.2nVa1Fs

    I show the top momentum decile split by value, and the top value decile split by momentum — splitting Value by momentum works, while splitting momentum by value does very little.

    Is this what you were suggesting?

  • pcavatore

    Great post…both on the combination and timing topics.
    About timing factors I generally agree with you that’s quite a risky proposition.
    What do you think about other common timing approaches like rule-based (ie. Markov regime switching model) or macroeconomic scenario-based (ie. come up with a set of relevant indicators to assess the current macro environment and over/under-weight factors that did well/bad over the same previous environments) ones?

  • Sergey Vedernikov

    Did you try “relative weakness” rule – i.e. invest more heavily in the factor with weaker performance (a kind of contrarian strategy) over 1-5 years.

  • Eduardo Gonzatti

    Hello Jack, thanks for the excellent post.

    Could you further post, or even edit this same here, about the results in the LS portfolio for us geeky “equities market neutral” ones?

    Also, would you care to comment about this topic on a “quantocracy.com colleague” of yours? https://blog.thinknewfound.com/2016/03/beware-bad-multi-factor-products

    Maybe this could lead us further into a systematic approach where one could time in different frames each strategy, don’t you think?

    Best Regards

  • Thomas Musselman

    Exactly; thanks for the memories (if I may quote Bob Hope)

  • Jack Vogel, PhD

    In general, timing factors is difficult. We did test timing value and momentum using EBIT spreads between value stocks and growth stocks. The post can be found here:

    http://blog.alphaarchitect.com/2015/03/04/do-valuation-spreads-matter-for-market-timing/#gs.MMPAy_g

    The hypothesis is that if there are large spreads in valuation across the cheapest stocks and the most expensive stocks, this may indicate a value premium (“stock picker’s market”), while a small valuation spread may indicate a momentum premium (“a trending market”).

    End result — We find little evidence that valuation-spreads can time value and momentum

  • Jack Vogel, PhD

    In our tests above, we assume a monthly rebalance into 50% value and 50% momentum — so this test will, by construction, invest more into the factor with poor past performance and sell some the factor with better performance. As far as moving the weight in each strategy beyond 50%, we have not tested this.

  • Arjun K

    I think the 75/25 contrarian split instead of a momentum split could be interesting to look at.

    The Research Affiliates team seem to think contrarian is the way to go (http://www.researchaffiliates.com/Our%20Ideas/Insights/Fundamentals/Pages/442_How_Can_Smart_Beta_Go_Horribly_Wrong.aspx) based on the last line of the article:

    “And we will objectively test the comparative efficacy of performance chasing in the factor zoo (i.e., favoring the factor tilts and smart beta strategies with the best recent performance) against investing in strategies with abnormally poor recent performance. The results are eye-opening to say the least.”

  • Jack Vogel, PhD

    I will possibly write another post on long/short value and momentum portfolios later — high level results are the following (1) value L/S and Momentum L/S are negatively correlated (2) combining the 2 portfolios is again a good idea.

    I agree with the post you mention. In the post here (http://blog.alphaarchitect.com/2015/03/26/the-best-way-to-combine-value-and-momentum-investing-strategies/#gs.7nzv9=I) we show that sorting cheap stocks on momentum is a good strategy. We prefer to sort cheap stocks on quality as this causes a lower correlation to a standalone momentum portfolio — however, both strategies are backed by evidence.

    I wish we had a great system to time value and momentum (that also does not cause a ton of trading). Here is a study we did on the topic:

    http://blog.alphaarchitect.com/2015/03/04/do-valuation-spreads-matter-for-market-timing/#gs.MMPAy_g

  • Sergey Vedernikov

    Maybe it worth to try – if results for “relative strength” are worse than 50%/50% – it is hint that the opposite rule (“relative weakness”) may work better. At times it may increase risk a bit (as usual with reversion strategies), but compensate this with better reward over the long-term.

  • Jack Vogel, PhD

    We do a similar test here:

    http://blog.alphaarchitect.com/2015/03/04/do-valuation-spreads-matter-for-market-timing/#gs.MMPAy_g

    We allow the weights to move from 20% to 80% based on valuation spreads (this is a contrarian signal). Results show some improvement, but maybe not worth the brain damage compared to 50/50 (especially if in a taxable account, as it causes more trading).

    In the tests in the post above, we assume a monthly rebalance into 50% value and 50% momentum — so this test will, by construction, invest more into the factor with poor past performance and sell some the factor with better performance.

  • So I take it the an underlying portfolio mix of combining the QVAL and QMOM ETFs would reasonably replicate these results ?

  • Thomas

    “To date, we have tested many “timing” ideas and we have found NONE that provide enough benefit to justify risking the known diversification benefits of allowing value and momentum to work as a system.”

    What about applying a version of double momentum ? I.e., instead of timing Quality vs Momentum, timing Domestic Quality/Momentum vs International Quality/Momentum ?

    Anyway, great post !

  • Jack Vogel, PhD

    The above post just examines US stocks — that is a good idea for a future blog post (US Value and Momentum, as well as International Value and Momentum).

  • We can’t discuss ETFs in this forum.

    To capture the expected results above, one would need to combine concentrated global value and momentum exposures, but in order to benefit as an investor the exposure would need to be affordable, tax-efficient, and insulated from “factor drift.” And even after those problems were solved, you’d need value and momentum to perform, out of sample.

  • the issue with a contrarian timing mechanism (or really any timing mechanism) is giving up the “bird in the hand” correlation benefits. That is the challenge — and the hurdle is quite high given the diversification benefits created by combining concentrated value and momentum exposures.

  • Adam Kearny

    I remember reading somewhere that value and momentum have historically tended to alternate the lead in 5-8 year cycles. What about a back-test where you simply flip out of a factor (either value or momentum) once it has outperformed the other over the trailing five or six-year period (hence, is likely nearing the end of its stretch of outperformance), rather than every 12-months? In other words, these two factors themselves seem to have 5-8 yr “momentum”–it doesn’t alternate a whole lot more frequently than that other than short-term and small amounts–if you use that long-cycle periodicity, you should capture the bulk of the timing pretty well. Also more tax-efficient than flip-flopping every 12 months.
    In recent decades, it would be like–get out of momentum in 1999, get out of value in 2005 or 2006. Get out of momentum in 2012, etc. Not perfect, but captured most of the style shifts, I would think.
    Alternatively, I’ve noticed in recent decades that each factor/style seems to have a final “blow-off” year where it absolutely crushes the other factor, and these extreme divergence years typically mark the end of the 5-8yr run. Momentum in 1999. Value in 2006. Momentum in 2015? Perhaps combine the 5-8 yr mean reversion cycle with a blow-off period indicator would work even better? Bottom line is that these seem to be long cycles.

  • Adam Kearny

    A related question: has anyone done any research on factors that identifying the worst performing stocks (e.g. some combination of high valuation, poor momentum, and poor quality)? Given the 90% max drawdowns (basically, an extinction event) in all of these long-only strategies, it certainly would be valuable to identify a systematic short-hedge that doesn’t cost too much in terms of surrendered beta or short-term big draw-down risks.

  • Tim Chen

    Can you explain why 50/50 value and momentum is considered a strong combo when the returns are lower than 100% momentum, and standard deviation and drawdown higher? Thanks!

  • Tim, that is a great question and something we should have made more clear.

    In general, if someone had the ability to withstand long and extremely painful relative performance periods, going “all-in” on momentum will likely maximize your expected return. However, you probably won’t maximize your risk-adjusted return. The biggest issue with momentum is the frictional costs. The strategy results presented above are monthly rebalanced versus annually rebalanced on the value portfolios. We keep everything on a gross basis so the reader can determine their own assumptions on frictional costs. But let’s say momentum costs you 300bps a year — which is probably too high for the last 10yrs, but way too low in the early part of the sample. All the sudden momentum is more in line with value and the diversification benefits of combining value and momentum become more clear.

    I’ll talk to Jack and see if he can post the net results as well.

  • Steve

    Tim from memory I think the net returns are in the excellent, DIY Financial Advisor book, if you’ve got it.

  • Tim,

    Jack updated the post to include the net data. The generic value portfolio — and the results in general — are heavily influenced by the inclusion/exclusion of the Great Crash. To the extent one believes this is a plausible outcome in the future and/or the data from this period is believable, one should consider these results carefully. If one includes the more recent periods when data is more reliable, the results are more clear.

  • Tim Chen

    Ha, still a great reminder to crack open the book, which has been sitting around on my Kindle for far too long.

  • Tim Chen

    I’ve seen a bunch of comparisons, but this is the first where value trounces momentum. It’s also probably the most relevant. Very interesting – thanks Wes/Jack.