facebook twitter instagram linkedin google youtube vimeo tumblr yelp rss email podcast phone blog search brokercheck brokercheck
%POST_TITLE% Thumbnail

MTUM For Momentum

Posted by Kevin Means, CFA on Jul 9, 2019 8:45:00 AM


A recent article showed that among 7 major “smart beta” factors, momentum was the one with the highest risk-adjusted return in recent years.  The ETF (exchange-traded fund) used to represent the momentum factor was iShares Edge MSCI USA Momentum Factor ETF (MTUM).  Because of its low costs and tax efficiency, it is a very attractive way to invest in the momentum factor over the long-term.  Currently, MTUM has characteristics, especially very strong upward revisions of EPS forecasts of its constituent stocks, that make now a particularly good time to initiate a position.        


The Case for Momentum Investing

When I was studying Economics as an undergrad and Finance as an MBA student in the 70s and 80s, the overwhelming consensus among academics was that markets were efficient.  Most adherents to the Efficient Market Hypothesis (EMH) believed in the “semi-strong form” of market efficiency—that current market prices reflect all publicly available information (past prices, fundamentals, news, etc.).  Nearly everyone held at least the less-restrictive “weak form” of market efficiency—that past returns could not help forecast future returns.  Respectable opinion was that any use of technical analysis, trend-following, or use of price momentum was essentially the financial equivalent of reading entrails.

When Jegadeesh and Titman published their paper, “Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency” in the Journal of Finance in 1993, it was a direct challenge to the Efficient Market Hypothesis, even in its weakest form.  They found compelling evidence that past stock returns were helpful in predicting future stock returns—winners tended to keep winning and losers tended to keep losing.  Most academic studies ever since have pointed back to this paper as the opening salvo in the debate regarding the causes of superior risk-adjusted returns from momentum investing.

Most published studies have focused on momentum among individual stocks.  Some studies have found that momentum is mostly a sector or industry effect.  Other studies have found impressive returns from momentum effects using “style” portfolios of large vs. small and growth vs. value portfolios.  Still other studies find that momentum works well when applied to international stock market indexes, government bond market indexes, commodities, and currencies:

“The existence of momentum is a well-established empirical fact. The return premium is evident in 212 years (yes, this is not a typo, two hundred and twelve years of data from 1801 to 2012) of U.S. equity data, dating back to the Victorian age in U.K equity data, in over 20 years of out-of-sample evidence from its original discovery, in 40 other countries, and in more than a dozen other asset classes. Some of this evidence predates academic research in financial economics, suggesting that the momentum premium has been a part of markets since their very existence, well before researchers studied them as a science.” (“Fact, Fiction and Momentum Investing,” Asness, et. al., Journal of Portfolio Management, Fall 2014.)

The evidence for a momentum anomaly is so strong that even those high priests of market efficiency, Fama and French, in a 2008 Journal of Finance article called it "an anomaly that is above suspicion…the premier market anomaly."

MTUM:  The Best Momentum ETF

One of the biggest challenges for those who wish to pursue momentum investing is the difficulty of implementing it effectively.  For one thing, by its very nature, momentum investing requires a high degree of turnover.  This means that transaction costs will tend to loom large and eat into the excess returns of the strategy. 

Also, with high turnover comes high realization of short-term capital gains, which are taxed at ordinary income rates.  This is a major drag on after-tax returns for taxable accounts.

Fortunately, there are several momentum ETFs available which greatly ameliorate these problems.  Because of the structure of ETFs, they are usually able to avoid distributing capital gains to shareholders, eliminating the tax problem from a high turnover strategy.  Also, the sponsoring institutions are usually very good at controlling transaction costs, in part because as passively managed funds, they must perform as closely as possible to their underlying indexes, which have no transaction costs.

Here is a quick comparison of a few of the largest momentum ETFs:

I prefer iShares Edge MSCI USA Momentum Factor ETF (MTUM) because of its extremely low expense ratio and bid-ask spread. The ETF selects from the MSCI USA Index, which is comprised of large and mid cap U.S. stocks.  A raw momentum score is calculated for each stock based on a 50/50 average of the stock’s return for the 6 months and 12 months prior to the most recent month, in keeping with the original Jagadeesh and Titman study.  This raw momentum score is divided by the trailing 3-year weekly volatility of the stock to yield a risk-adjusted momentum score.   The securities in the fund (currently 125) are weighted by the product of their market caps and their risk-adjusted momentum scores.  The market cap component has the effect of putting more weight on larger, more liquid names with lower trading costs.  The risk-adjusted momentum score component will result in sector weights that may depart dramatically from the cap-weighted index.  I consider this an advantage, since research indicates that momentum helps pick sectors as well as stocks.

MTUM is very tax efficient.  It has a distribution rate of 1.30%.  It distributes only its dividend income.  It has never distributed a capital gain.

MTUM:  An Alpha Machine

Very few funds of any kind, whether actively or passively managed, actually produce a meaningful level of alpha, or risk-adjusted excess return.  Nearly all of the return for nearly all funds is simply beta, or risk-based return from exposure to risk factors.  And on average, funds generate negative alpha because of their expense ratios and operating costs, including transaction costs. 

To see if MTUM is has historically provided any alpha, we must first disentangle its risk-based return from its total return.  This exercise starts by measuring the sensitivity of MTUM’s returns to four risk factors that capture much of the risk common to most ETFs:

  • Stock market risk (MKT), as measured by the S&P 500 Index
  • Bond market risk (LTB), as measured by the 10 Year Treasury Benchmark Index
  • Currency risk (DLR), as measured by the U.S. Dollar Index
  • Commodity risk (OIL), as measured by the West Texas Intermediate Crude Oil Index

I use exponentially-weighted 36-month rolling multiple regressions to measure these risk factor sensitivities (often called betas) simultaneously.  MTUM started trading on April 16, 2013, but the underlying index goes back to December 31, 1981.  I use the index returns to calculate risk factor sensitivities in order to take advantage of this longer history, since I know that as an index fund, MTUM’s returns will be very close to its underlying benchmark index (before the impact of the 0.15% fund expense ratio).  This enables me to backtest the strategy embedded in the index as well as measure the history of its risk factor exposures.

 

  

The graph above shows that MTUM’s equity market beta (labeled MKT in blue) is its only consistently significant risk factor exposure, as expected for an equity fund.  The historical equity market sensitivity has generally been between 70% and 130% (or a beta of .7 to 1.3) but in the post-2008 era it has been fairly close to 1.0.  In recent years, its LTB sensitivity has increased, perhaps because aggressive Fed easing has been driving the stock market more than usual.    

 

 

The graph above disaggregates the cumulative return of MTUM since December 31, 1999.  Returns are based on the underlying index before fund inception (after subtracting the .15% expense ratio), and on live fund returns after inception.  Much of the fund’s return is explained by its equity market sensitivity (blue), as expected for an equity ETF.   (To calculate the return from MKT sensitivity, I multiply the ETF’s previous month-end MKT sensitivity times the monthly price return of the S&P 500.  I use the same methodology for the other three risk factors.)  The residual return (black) is the total return minus the return from the four risk factor sensitivities.  This is the fund’s alpha, or risk-adjusted excess return. 

As highlighted in a previous article, it is not unusual for factor-based ETFs such as MTUM to show impressive alpha based on the backtest of the underlying index before inception.  Sometimes, fund sponsors cherry-pick the amount of history to show for the index, or even tinker with the rules governing the index to boost its historical return.  MTUM’s sponsor is BlackRock, a firm known for its index construction and its careful quantitative implementation.  The fact that they offer index returns back to 1981 (I chose to show only the period since 1999) indicates that they did not cherry pick the index history.

The fact that the fund’s residual return (alpha) has been nearly as high since its inception as it was prior to inception indicates that there really is some sort of market anomaly being captured by the index construction rules:

            Time Period                             Starting           Ending             Avg. Alpha

            Before inception                     12/31/99         3/31/13                3.33%

            After inception                        4/30/13           6/30/19                2.53%     

Remember, most funds have no discernible alpha—their return is entirely explained by risk factor sensitivities.  Both the power of and the persistence of the risk-adjusted excess return for MTUM are extremely impressive. 

MTUM:  Now is the Time

The residual return for MTUM has some definite variability.  There have been periods when it has been higher and some when it has been lower or even negative (when the black residual return line above has been downward sloping).  Is there any way to forecast whether the current period is likely to be above or below average?

In fact, we spend a lot of time on that very question at Sapient Investments.  We try to identify sensible, fundamentally-based factors that help give us an edge regarding the likely residual return of the ETFs and CEFs (closed-end funds) in which we invest.  For the most part, the factors that we use are various ways of quantifying value, momentum, or quality.  We get the data from FactSet, which provides a bottom-up aggregation of factor exposures for ETFs based upon the fact that the constituents of ETFs are public information available daily. 

We put factor-based ETFs like MTUM into a research universe of other fairly similar ETFs.  The “alternative equity ETFs” universe includes ETFs focused on factors such as size, value, growth, volatility, momentum, quality, buybacks, insider activity, dividend yield, IPOs, and hedge fund holdings replication.  Most limit their holdings to U.S. stocks, but some include global or international holdings.  Currently, the alternative equity ETF universe has 92 constituents. 

I would like to highlight just one of the factors that our research has found particularly helpful in forecasting the alphas of factor-based ETFs:  earnings estimate revisions.  The particular factor is a “diffusion” index that measures % up - % down.  FactSet provides both the number of estimates that have been revised up or down within a particular time period, as well as the number of companies that experienced net upward or downward revisions, for both the current fiscal year and the next fiscal year.  We have constructed a factor that combines all of these elements:  “EPS Estimate %Up - % Down.” 

Our research indicates that within the alternative ETF universe the EPS Estimate %Up - %Down factor is the single most powerful ETF selection factor in our arsenal.  The graph below illustrates one of our research tests for that factor.  We begin our test at the end of 2006 when the number of ETFs with data for this factor reached a critical mass of 27.  (Currently, of the 92 ETFs in the universe, 76 have data for this factor.)  At the end of each month, we form a “Top 10” portfolio of the 10 ETFs with the highest EPS Estimate %Up - %Down.  Each has a 10% weight.  We calculate the residual (risk-adjusted) return of this portfolio during the month and then re-select and rebalance the “Top 10” portfolio at the end of the next month, and so on.  Similarly, at the end of each month we also form a “Bottom 10” portfolio of the five ETFs with the smallest, or most negative, EPS Estimate %Up - %Down, re-selecting and rebalancing monthly.

 

 

In the graph above, the blue line is the cumulative log of residual return of an equal-weighted portfolio of the 10 ETFs with the largest EPS Estimate %Up - %Down, rebalanced monthly.  Since 2006, the average log of residual return (net of risk effects such as equity market and interest rate betas) has been 1.7% per year on average.  The orange line is the same thing but investing in the 10 with the smallest (or most negative) EPS Estimate %Up - %Down.  That return has been -4.1%.  The green line is a long-short implementation of the strategy.  Its return has been 5.8% per year.  

 

 

As of June 30, 2019, MTUM had the highest EPS Estimate %Up - %Down of any ETF in our alternative equity ETFs universe.  Based upon its factor loading and our model of the expected return to the factor, we were forecasting that MTUM would have a .54% residual return during July.   (Our models only forecast one month ahead.)  If sustained, this would result in an annualized alpha of 6.48% (.54% X 12 = 6.48%). 

The graph above depicts the test for only one factor.  At Sapient Investments we use several other factors to forecast the residual returns of ETFs.  EPS Estimate %Up - %Down is essentially a momentum factor.  (Thus, it is not surprising that MTUM would have a very high loading on this factor.)  We also use various value-related and quality-related factors.  While MTUM is not highly loaded on all of these, the incremental expected residual returns from these other factors add incrementally to its overall residual return forecast. 

Conclusion

  • Momentum is one of the strongest and most persistent factor anomalies
  • It is well-supported in the academic literature
  • MTUM is a very low-cost and tax-efficient way to exploit the momentum effect
  • Very strong estimate revisions for its constituent stocks indicate that MTUM is likely to achieve above-average alpha in the near future