VTI vs SPY

In the ETF portfolios models I work with, I usually use Vanguard's Total Stock Market (VTI) instead of the more well known S&P 500 SPDRs ETF (SPY). Since I usually have a few questions about the purpose of this substition and the significance of the differences, I thought I'd address them here (note: all statistics quoted here are as of October 20, 2006).

As both are index based equities, it's important to pay attention to the index and its characteristics as that drives much of the behavior of the ETF. SPY is based on the S&P 500 Index, holding all of the S&P 500 stocks. For those that are curious, SPY was constructed as as SPDR (an undivided ownership interest) when ETFs were first emerging and has a somewhat different operational model than the more prevelant index-based passive mutual model that most newer ETFs (including VTI) use.

VTI is based on the MSCI US Broad Market Index, which typically spans about 1,300 of the largest stocks traded on NYSE, AMEX and OTC markets. VTI is a Vanguard ETF product.

Comparing the indexes, VTI is much broader than SPY and this accounts for the primary difference. Looking at SPY's top 10 holdings, compared to VTI's top 10, you'll note that SPY a higher concentration of individual holdings in the top rankings. For instance, SPY contains 2.99% of General Electric (GE), compared to VTI's 2.35% GE. This is accounted for by VTI's much broader coverage. Comparatively, SPY is a narrower, deeper pool compared to VTI's wide, shallow pool. Likewise, VTI's holding turnover is nearly double what SPY's is given its breadth. If you're paranoid about the S&P's weighting model and the impact of individual security problems (like investors deciding Google is a dot-com 2.0), SPY's higher concentration provides greater risk.

So what does that mean for an investor in the two equities? Both have comperable yields (1.74 vs 1.72 SPY to VTI), composite P/E ratios (14 vs 15) and basic risk ratios, with 5-year standard deviations hanging around 12.70 for SPY and 12.87 for VTI. In many respects, they perform in a similar manner according to most risk measurements, but historically, VTI's breadth gives a little boost in returns, as illustrated in this five-year comparison chart.

And don't forget to consider the expense ratios. SPY comes in at a modest 0.1%, while VTI is a low 0.07% (common for Vanguard ETF products). While 0.03% is insignificant to many, that's still extra money in your pocket.

Downsides to VTI include underperformance when the largest of largecaps are outperforming the market comparatively (as has occurred this year), causing VTI to have a bit of penalty for its breadth. Likewise, VTI lacks the shear trading volume, market cap and liquidity that SPY has. SPY's average daily volume is around 50-60 million shares a day, compared to VTI's meager 50-150 thousand range.

That said, VTI usually represents a solid proxy for SPY with very close index coverage and risk, with a slightly better returns on slightly lower index expenses. Combined with its greater diversification which I believe should be considered in any index containing Google and you've got a good candidate for broadbased large-cap index coverage.

Disclaimer: I own both SPY and VTI in my personal portfolios.