• ETFs

New Multifactor ETFs from iShares (Part II)

Last week we looked at the iShares Edge MSCI Multifactor Canada Index ETF (XFC), one of a trio of new ETFs based on multifactor indexes from MSCI. Recall that we found a few surprises in our analysis: a large part of the index’s performance could not be explained by the size, value, momentum and quality factors. Now let’s see how the US and international multifactor indexes hold up to the same analysis.

iShares Edge MSCI Multifactor USA Index ETF (XFS)

My first issue with the iShares Edge MSCI Multifactor USA Index ETF (XFS) is that its underlying index doesn’t use a broad-market parent index (like the MSCI USA IMI Index). Instead, its universe is the MSCI USA Index, which excludes small cap companies. A multifactor ETF that tilts towards small companies is going to have a hard time doing so when their parent index contains only large and mid cap companies. iShares argues that mid cap companies have also historically outperformed large cap companies in the US, so we’ll leave it at that.

The management fees on this product are 0.45%, making it 6-7 times more expensive than the broad market iShares Core S&P U.S. Total Market Index ETF (XUU). XUU also tracks an index that includes 3,817 stocks, while XFS tracks an index of only 135 stocks.

The tracking error of the MSCI USA Diversified Multiple-Factor Index has been lower than most of the other single factor indices, so this should also help investors with their behavioural biases.

Monthly Tracking Error to Index: January 1999 to December 2015

TrackingErrorXFS

Sources: MSCI, Dimensional Returns 2.0

 

A 5-factor regression analysis of the multiple-factor index reveals an unexplained alpha of 0.47%, though this is nowhere near as troubling as it was in our Canadian example. Although the index did not have a heavy size or momentum tilt, it did include a healthy dose of value (+0.14) and quality (+0.27), as we should expect.

5-Factor Regression Results: January 1999 to December 2015

RegressionXFS

*Data shaded in light grey have a t-stat greater than 2, which indicates a statistically significant result
Sources: MSCI, AQR Data Library

 

While the regression analysis showed few surprises, I do find it odd that none of the individual factor indices outperformed the multiple-factor index over the measurement period. As you can see in the table below, the individual factor indexes had annualized returns no higher than 7.41%. Yet somehow the multi-factor index returned 7.88%. This could be due to a rebalancing bonus of some kind, but the amount just seems high. (Imagine a portfolio of ETFs that earned a higher return than every one of its individual holdings.) Also, the largest tilt in the multiple-factor index was to the quality factor, and yet this was the worst performing individual factor index of the group, at 5.42%.

Performance Summary: January 1999 to December 2015

PerformanceXFS

Sources: MSCI, Dimensional Returns 2.0

 

iShares Edge MSCI Multifactor EAFE Index ETF (XFI)

Similar to XFS, the iShares Edge MSCI Multifactor EAFE Index ETF (XFI) sorts its stocks from an index that excludes small cap companies.

The management fees on XFI are also 0.45%, which is roughly double the cost of the broad-market iShares Core MSCI EAFE IMI Index ETF (XEF). XEF also tracks an index of 3,164 stocks, while XFI includes just 209 companies.

The tracking error of the MSCI EAFE Diversified Multiple-Factor Index relative to the broad-market index is the lowest of the three multiple-factor indices, at 1.15%. It is also lower than any of the individual MSCI factor indices listed below. This should lead to fewer cases of regret among investors who have become accustomed to plain-vanilla indexing.

Monthly Tracking Error to Index: January 1999 to December 2015

TrackingErrorXFI

Sources: MSCI, Dimensional Returns 2.0

 

Performance Summary: January 1999 to December 2015

PerformanceXFI

Sources: MSCI, Dimensional Returns 2.0

 

The 5-factor regression showed a significant tilt towards all factors except momentum. Nothing else in the results really jumped out as being out of place:

5-Factor Regression Results: January 1999 to December 2015

RegressionXFI

*Data shaded in light grey have a t-stat greater than 2, which indicates a statistically significant result
Sources: MSCI, AQR Data Library

 

If you’ve made it this far, you’ve probably already made the decision to stick with your low-cost, plain-vanilla indexing strategy. Factor investing is enticing, but as we’ve seen with these multifactor indexes, it can be difficult to put into practice. If you use a factor-based ETF that claims to capture the premiums from value or momentum stocks, for example, it’s hard to know whether you’re getting what you’ve been promised. Do you really want to be running 5-factor regressions as part of your due diligence process?

These strategies also have more turnover than a passive strategy (MSCI estimates their global multiple-factor portfolio turnover at 40.1%, relative to 3.1% for their global passive index strategy). In taxable accounts, this increased turnover could result in higher capital gains distributions and lower after-tax returns. I still recommend choosing a strategy that you are comfortable with and that is easy to understand.

By | 2016-12-20T03:24:13+00:00 August 22nd, 2016|Categories: ETFs|Tags: , |2 Comments

2 Comments

  1. Zach August 28, 2016 at 1:41 pm - Reply

    I know that past performance does not guarantee future returns but is your argument against multiple-factor ETFs that because of the ‘black box’ nature of these active ETFs it is difficult to understand what the performance drivers of the higher returns from 1999-2015 were, which in turn makes harder to predict what influences the market will on return in the future?

    • Justin August 28, 2016 at 7:17 pm - Reply

      @Zach: The main reasons I would recommend broad-market ETFs for DIY investors (rather than multifactor ETFs) are:

      1. Simplicity
      2. Lower cost
      3. Less likely to cause bad investor behaviour
      4. More tax-efficient

      The biggest argument for using multifactor ETFs is:

      1. Possible outperformance relative to broad-market ETFs (you could say this about any active strategy), before-taxes/before-fees/before bad investor behavior

Leave A Comment