Analysis of John Hancock Multifactor ETFs
August 14, 2017
In September 2015, John Hancock Investments launched six strategic (smart) beta John Hancock Multifactor ETFs, with underlying indexes designed by Dimensional Fund Advisors LP (DFA). By now, the product suite has grown to a total of twelve ETFs, three “core” and nine “sector” ones.
Traditionally, DFA mutual funds were available only through advisors operating within the company’s program. With John Hancock Multifactor ETFs, retail investors can access DFA strategies without paying an advisory fee, which is typically 1% of assets under management (AUM). However, since DFA offers a large selection of mutual funds, it is not clear which of them can be replaced by the ETFs.
Let’s start with the John Hancock Multifactor Large Cap ETF (JHML). To identify the best candidates for substitution, we will use the correlation of rolling 52-week returns (conventionally, we would use rolling 36-month returns, but John Hancock ETFs have insufficient history). Although high correlations do not imply product identity, there are a good starting point for further analysis. Here are the correlations of DFA core and large-cap funds with JHML:
Of the candidate funds, the DFA US Large Cap Equity Portfolio (DUSQX) and DFA US Large Company Portfolio (DFUSX) had the highest correlation with JHML. Let’s see what total returns and traditional statistics looked like for the candidate funds and the ETF:
Indeed, the performance of DUSQX and DFUSX was similar to that of JHML, although the volatility of the ETF was slightly lower than that of the funds.
Next, let’s take a look at the John Hancock Multifactor Mid Cap ETF (JHMM). DFA does not offer an explicitly-named mid-cap fund, so we will try the core and small-cap funds. Here are their correlations with JHMM:
Based on this criterion, the DFA US Core Equity 1 Portfolio (I) (DFEOX) and DFA US Core Equity 2 Portfolio (I) (DFQTX) were the best candidates for substitution.
The DFEOX tracked JMHH most closely, although at a lower annualized return and a slightly higher standard deviation.
Finally, let’s analyze the John Hancock Multifactor Developed International ETF (JHMD). This ETF was launched in mid-December 2016 and, as of this writing, does not have 52 weeks of history. Therefore, to determine its correlations with DFA International funds we will use a rolling 26-week period:
The DFA International Core Equity Portfolio (I) (DFIEX) and DFA International Large Cap Growth Portfolio (DILRX) had the highest correlations with the ETF. The ETF most closely tracked the former fund:
Although John Hancock Multifactor ETFs have a relatively short history, we have identified specific DFA mutual funds that these ETFs can effectively substitute. However, it should be noted that ETFs trade at market prices and not at net asset values (NAVs) as mutual funds do. Therefore, ETF premiums/discounts and spreads may negatively affect investors’ returns. Nevertheless, these ETFs are worth a consideration by those investors who like DFA’s multifactor strategies but do not want to pay recurring advisory fees to gain access to DFA mutual funds.
To learn more about the performance of John Hancock Multifactor sector ETFs, please register on our website.
January 17, 2017
Entering an Exclusive Dimension
A recent piece in Barron’s proposes an investment into seven actively-managed mutual funds. This recommendation is motivated by the following observation:
A long, humiliating period for professional stockpickers might be giving way to something different. Stocks that have moved in near unison in recent years are beginning to chart more distinct paths. Data points that haven’t mattered in a decade, like the relationship between prices and fundamental measures of value, are starting to have more sway on returns. The divide between cheap stocks and expensive ones remains exceptionally wide, which could mean last year’s shift in favor of value investing is just the beginning.
Supposedly, were on the verge of entering the “stockpicker’s market,” as shown in this chart:
The myth that low correlations between stock returns lead to active manager’s outperformance has long been debunked. Similarly, a high active share is cited as one of the reasons actively-managed funds will outperform their passive peers. Please refer to our earlier post for a discussion of this topic.
So, this post will instead focus on the long-term performance of the funds featured in the article:
The following charts with related statistics show the cumulative RealAlpha™ for each fund that has at least ten years of history through 2016 (to learn more about this and other patent-based performance measures Alpholio™ uses, please consult our FAQ). In all analyses, the number of ETFs in the reference portfolio was limited to no more than seven. The ETF membership and weights in each reference portfolio were constant throughout the entire evaluation period.
Here is a chart with statistics for the AllianzGI NFJ Dividend Value Fund (PNEAX; Class A shares):
The fund cumulatively returned over 20.5% less than its reference ETF portfolio of lower volatility.
Here is a chart with statistics for the DFA US Large Cap Value Portfolio (DFLVX; Class I shares):
The fund cumulatively returned about 8.5% more than its reference ETF portfolio of lower volatility.
Here is a chart for the Dodge & Cox Stock Fund (DODGX):
While the fund produced a 14% higher cumulative return than its reference ETF portfolio, by early 2016 it also lost virtually all of its cumulative RealAlpha™ generated since 2007.
The following chart is for the Sound Shore Fund (SSHFX):
On a cumulative return basis, the fund underperformed its reference ETF portfolio by over 7.7%; most of that loss occurred over the past two years.
This chart is for the T. Rowe Price Equity Income Fund (PRFDX):
The fund’s cumulative return was over 23.3% lower than that of its reference ETF portfolio of a slightly higher volatility.
The final chart is for the Vanguard U.S. Value Fund (VUVLX; Investor Class shares):
The fund cumulatively returned about 9.1% more than its reference ETF portfolio of a slightly lower volatility. However, as recently as at the end of October 2016, the cumulative RealAlpha™ was only 4.4%.
In conclusion, only three out of the six funds analyzed above added some value when compared to their respective reference ETF portfolios. The rest of the funds underperformed, and in some cases quite significantly. It remains to be seen whether a combination of the expected low stock correlations in the market and a high active share of these funds leads to their significant outperformance in 2017.
To learn more about these and other mutual funds, incl. the composition of their reference ETF portfolios, please register on our website.
To learn more about the these and other mutual funds, including the composition of their reference ETF portfolios, please register on our website.
January 7, 2014
A cover story in Barron’s provides lots of interesting details about the history and operations of Dimensional Fund Advisors (DFA). Founded in 1981, DFA has recently reached $332 billion in assets under management (AUM).
About 78% of these AUM are in stocks, and about 85% in low-cost mutual funds with an average expense ratio of 0.39%. The funds have a small-cap and value tilt, based on the Fama-French three-factor model. Lately, the firm started to augment its funds with a profitability factor.
The article states that
More than 75% of its funds have beaten their category benchmarks over the past 15 years, and 80% over five years, according to Morningstar — remarkable for what some investors wrongly dismiss as index investing.
To substantiate this, the article compares two similar funds from DFA and Vanguard:
For example, take the Vanguard Small Cap Value index fund (VISVX), which is based on the S&P 600 Small Cap Value index and is the counterpart to Dimensional’s DFA US Small Cap Value (DFSVX). The DFA fund has a much smaller tilt — its average market value is $1.1 billion, versus Vanguard’s $2.7 billion — and on all measures is much more value-oriented. So the Dimensional fund better captures the market-beating advantage of small and value stocks. In fact, a lot better: The DFA fund returned 42% in 2013, beating 88% of its peers in Morningstar’s small-cap value category, versus the Vanguard fund’s 36% return, which beat just 53%. Over 15 years, which includes periods that were less favorable to small and/or value stocks, DFA’s fund returned an average of 12% a year, beating 80% of peers. The Vanguard fund returned 10% on average, beating just 37% of peers. The Dimensional fund costs twice as much as Vanguard’s — 0.52% versus 0.24% — but the significant outperformance more than makes up for that difference.
That only tells a part of the story. According to Morningstar data, DFSVX had a lower Sharpe Ratio than VISVX in the 3-year (0.96 vs. 1.01) and 10-year (0.47 vs. 0.48) periods through 2013. This is also reflected in the generally higher volatility and upside and downside capture ratios for the DFA fund. As a result, the DFA fund produced lower returns than the Vanguard fund did in the down years of 2007, 2008 and 2011.
The article says that a deliberately paced trading as well as market making in the 14,000 stocks DFA owns both add to its outperformance. However, DFA faces an ongoing criticism: since its funds are sold exclusively through 1,900 rigorously screened and trained financial advisors, they are not easily accessible to individual investors, especially those with a small amount of investable assets, not willing to pay advisory fees or already having an unaffiliated advisor. This is what creates an “exclusive dimension” of DFA, which Alpholio™ can help investors enter. Following up on one of the previous posts, let’s analyze DFSVX in more detail.
The following chart shows the relative performance of the fund vs. its reference portfolio of ETFs:
An investor who committed to the fund in early 2005 would have gained only a modest amount of cumulative RealAlpha™ by late 2013. This was mostly caused by the fund’s underperformance in the three years mentioned above. In addition, at about 22.7% the annualized volatility of the fund was 2% higher than that of its reference portfolio in the overall analysis period.
The next chart illustrates ETF weights in the reference portfolio in the same period:
The fund could effectively be emulated by a collection of just four related ETFs: iShares Russell 2000 Value (IWN; average weight of 34.9%), iShares S&P Small-Cap 600 Value (IJS; 30.1%), iShares Morningstar Small-Cap (JKJ; 18.5%), and iShares Morningstar Small-Cap Value (JKL; 13.7%). (The remaining two ETFs accounted for only 2.8% of the reference portfolio on average.)
The weighted expense ratio of these four ETFs is currently only 0.33% compared to the fund’s 0.52%. In addition, while an investor trading these ETFs might incur some commission, spread and premium/discount costs, he/she would not have to pay a recurring advisory fee of about 1% (or be forced to switch advisors) to gain benefits similar to those offered by DFA funds. Over time, dedicated factor ETFs will likely make such fund substitution even easier. Thus, entering an exclusive dimension of factor investing is no longer as hard as it has been.
To get a unique perspective on the DFA and other funds, please register on our website.
July 30, 2013
A recent article from InvestmentNews describes the popularity of Dimensional Fund Advisors (DFA) mutual funds with financial advisers:
For the third time in four years, Dimensional Fund Advisors tops the list of mutual fund companies best positioned to increase its share of assets with financial advisers… The Austin, Texas-based mutual fund company, best known for its factor-based investing philosophy and high barriers to entry (for the mutual fund world, at least), scored 25% higher than bond megashop Pacific Investment Management Co. LLC, which came in second. The Vanguard Group Inc., the world’s largest mutual fund company, finished third.
Why do advisers prefer DFA funds? Three reasons come to mind: exclusive access, smart beta, and superior performance. The first reason is the same as the “high barriers to entry” mentioned in the above quote — the DFA funds cannot be purchased directly by individual investors, but only through qualified advisers. This means advisers can position themselves between inexpensive index-like vehicles and investors’ assets, thus improving the justification for an advisory fee.
The second reason emphasizes characteristics that make DFA funds supposedly different from regular index funds. For example, DFA Core Equity funds skew towards small-cap and value stocks:
Increased exposure to small and value companies may be achieved by decreasing the allocation of the portfolio’s assets in large growth companies relative to their weight in the US universe. Securities are considered value stocks primarily because a company’s shares have a high book value in relation to their market value (BtM).
However, this tilt is well known through Fama-French research and can be achieved through other means, including specialized factor exchange-traded funds (ETFs).
The third reason requires more scrutiny. While it is true that most DFA funds earn above-average ratings in their respective categories according to Morningstar, how does the performance of these funds look like from the Alpholio™ perspective? Let’s analyze the first DFA fund on the US Core Equity list, the US Core Equity 1 Portfolio (DFEOX). Here is the cumulative RealAlpha™ chart for the fund:
The chart clearly shows that on a truly risk-adjusted basis, the fund did not generate any alpha in the past seven years. This is further supported by performance statistics:
Similar results are observed for all of the DFA US Core Equity funds since their inception:
These data indicate that an individual investor would realize better risk-adjusted returns from substitute portfolios of ETFs than from these DFA funds, while in the process gaining intra-day liquidity and full control of investments. Alpholio™ provides composition of such substitute portfolios on an ongoing basis. Thus, the three reasons for preference of DFA funds, in particular the exclusive access, no longer hold.