Global Equity Funds: Get More, Pay Less
When the whole isn’t greater than the sum of its parts, is it really different?
When the whole isn’t greater than the sum of its parts, is it really different?
When the whole isn’t greater than the sum of its parts, is it really different?
I’ve been trying to figure out the cheapest way to access the global equity markets. I’m looking for straight-up vanilla exposure—comprehensive, cap-weighted—the global opportunity set.
My answer should have made waves, because I recommended mutual funds. Two of them.
Two Vanguard mutual funds—VTSAX and VTIAX, which access the exact same portfolio as the ETFs Vanguard Total Stock Market (VTI | A-100) and Vanguard Total International Stock (VXUS | A-99)—provide the most comprehensive, cheapest exposure; better, it turns out than Vanguard Total World Stock (VT | B-100).
Surprisingly, I, the director of research at ETF.com, ended up favoring the two mutual funds above—that’s right, the mutual funds, and not the ETFs—for global equity exposure. It was about avoiding trading costs and steering clear of intraday trading at premium or discounts to net asset value.
The world’s response to my ETF-lover’s heresy was a giant yawn. No snarky tweets, no fire-her-now notes to ETF.com’s top brass.
Instead, I got this query from a reader:
Does VTI + VXUS really give you VT? The total holdings don't seem to add up.
The reader’s math was incontrovertible.
VTI holds 3,742 stocks
VXUS holds 5,667
VT holds 6,363
Clearly, 3,742 + 5,667 = 9409, and not 6,363
Share counts per Vanguard, as of Aug. 31, 2014
What gives? What happened to the missing 3,046 securities? Is there something big missing from VT? Is there huge overlap in the two portfolios?
Not quite. It’s about optimization in VT, and microcaps and initial public offerings in VTI.
Optimization
VT’s underlying index, the FTSE Global All Cap, had 7,459 constituents as of Sept. 30, 2014. VT ought to hold about 1,100 stocks more than it does. For efficiency’s sake, VT’s portfolio manager has optimized its portfolio. She’s bought most, but not all of the 7,459 index constituents, doing her best to replicate all of the FTSE Global All Cap’s country, sector, size and valuation ratios, while avoiding the least liquid, hardest-to-trade securities.
According to Joel Dickson, Vanguard’s head of investment research and development, VT’s missing 1,096 stocks make up about 1 percent of VT’s overall market cap.
That explains about one third of the missing securities. What about the other 1,950?
The “missing” securities have been hiding in plain sight. They’re the long tail of U.S. micro caps, and shiny new IPOs. It’s all made plain in index providers’ rule books.
Let me explain.
My House, My Rules
Vanguard’s international funds follow FTSE indexes, but their flagship U.S. size and style funds track CRSP indexes. The U.S. component of VT’s underlying FTSE index—the FTSE US All Cap—is not the same as VTI’s CRSP U.S. Total Market Cap Index. FTSE and CRSP each have their own rules.
The first big difference is in depth of coverage. FTSE’s All Cap series aims to cover 98 percent of each region’s market cap. FTSE covers large, mid- and small-cap stocks. CRSP ups the ante with micro caps. Indeed, CRSP claims “close to 100 percent” coverage of the U.S. equity market.
There’s also variation in IPO seasoning rules. Indexers require new stocks—IPOs—to prove themselves before they qualify for inclusion. Typical requirements include minimum market caps and daily trading volumes.
Once an IPO is seasoned and passes its liquidity and market-cap tests, it can be included at the next index review date. FTSE and CRSP have different seasoning periods and also different review frequencies.
CRSP’s 20-day seasoning period and quarterly reviews means that stocks enter VTI far faster than they would if they had to wait for FTSE’s three-month seasoning requirement and semiannual reviews (though FTSE can fast-track the process for large-caps).
Rules Drive Stock Selection
These rule differences cause real differences in fund portfolios. As of Aug. 31, 2014, a whopping 1,801 U.S.-incorporated and -listed companies are in VTI but not in VT.
VTI’s micro caps include my hometown-based biotech firm Xoma and watchmaker Movado; VT doesn’t have these.
Then there are the IPOs. As of Aug. 31, 2014, CRSP-tracking VTI included Ally Financial; the revived GMAC, which IPO’d in April; and CBS Outdoor Americas, the billboard firm, which listed in March. FTSE-tracking VT will probably see these firms added soon, but meanwhile, CRSP has the jump.
Adding It All Up
If Vanguard were to switch VTI to FTSE USA All Cap, then VTI would wind up with 1,985 constituents (I interpolated this, because FTSE doesn't offer a USA All cap index fact sheet on its website. I took the number of U.S. constituents from the FTSE North America All Cap index fact sheet).
FTSE’s Global All Cap Ex-US held 5,564 names as of Sept. 30, 2014.
5,564 + 1,985 = 7,459, so FTSE Global Ex-US All Cap + FTSE US All Cap = FTSE Global All Cap. In an all-FTSE world, the VTI + VXUS = VT equation would hold. But that’s not the world we currently live in.
Do We Care?
FTSE US All Cap Index and CRSP US Total Market Index’s returns have been nearly identical since the CRSP series launched in April 2011.
I compared the two return series in a regression (using total returns in both cases). I found a goodness of fit between the two of 0.9998 and a CRSP versus FTSE beta of 1.0045 (CRSP is ever so slightly more volatile than FTSE). Statistically, these indexes are all but identical.
In practice, there’s been a difference in overall returns, but it’s not worth getting excited about—at least not yet.
The FTSE USA All Cap index outperformed CRSP’s US Total Market by 14 basis points a year between April 1, 2011 and Sept. 30, 2014. But this difference was not statistically significant. In other words, there was a healthy chance that this outperformance was random.
Joel Dickson called the difference “basically meaningless.” I agree.
Perhaps, over time, if statistical significance improves as the data set expands, Joel and I will change our minds. For now, the big pile of microcaps and IPOs doesn’t move the returns needle.
That's why the big picture equation of VTI + VXUS = VT still holds, even in a FTSE/CRSP combination.
It’s not often that you get more for less. Yet that’s exactly what happens if you break VT into its components VTI and VXUS—you get more stocks, and more complete coverage, for less cost. I’ll take that deal any day.
At the time this article was written, the author held no positions in the securities mentioned. Contact Elisabeth Kashner, CFA, at [email protected].
