Highlighting the preference for efficiency and liquidity with Morningstar data
The ETF industry knows how to exploit the first-mover advantage. Being the first to launch an ETF in a new category draws the interest of investors, which creates liquidity, which, in turn, draws the interest traders thus improving assets under management and liquidity further in a self-reinforcing process. But being the first in a category and even being the most liquid are not the only factors in determining ETF success.
SPDR S&P MidCap 400 MDY has a five-year head start and nearly 5 times the trading volume of iShares S&P MidCap Index IJH. Both funds track the same index, and, while traders have traditionally used MDY, investors are increasingly using IJH as their preferred choice for exposure to the S&P MidCap 400 Index. In the past year through the end of January, IJH gained about $100 million in assets while MDY lost nearly $1.7 billion. To understand why requires some knowledge of ETF history and mechanics.
UITs: First Generation ETF Technology
When MDY launched in 1995, it followed a unit investment trust structure, the same as SPDR S&P 500 SPY. Unit investment trusts face restrictions against holding anything not in the index, reinvesting dividends and lending shares. The last year in which an ETF came to market using the unit investment trust structure was 2002. Five years after MDY, iShares launched an ETF following the same index but organized as a regulated investment company. This is the same legal framework followed by most mutual funds and gives IJH the freedom to lend shares and to reinvest dividends. Both of these help the ETF provider better track an index. Revenue generated through share lending is shared with investors in the fund, and, by reinvesting dividends, the drag created by excess cash is reduced.
Enter Morningstar Data Points
That flexibility afforded by the fund structure has allowed IJH to return performance closer to its index. Over the past year through the end of February, IJH has returned 2.39% and MDY returned 2.24% while the index returned 2.55%. In a frictionless world, we would expect both ETFs to underperform their benchmark by the expense ratio. But IJH delivered performance better than one would expect given its 0.21% expense ratio and MDY delivered worse performance given its 0.25% expense ratio.
In the real world, index fund performance does not necessarily equal index return less the expense ratio due to various factors, most notably trading costs. The realized cost of replicating an index can be seen in the Estimated Holding Cost data point found at the bottom of the Performance tab on the Morningstar quote page for an ETF. On average, we would expect the Estimated Holding Cost to be slightly larger than the expense ratio and to reflect the efficiency with which an ETF tracks its index. For IJH, the Estimated Holding Cost has been around 0.16% over the past year compared with 0.31% for MDY.
Liquidity Matters When It's Time to Trade
In addition to the holding cost, another factor to consider when evaluating ETFs is the trading costs. Trading costs include not only commissions, but the impact of trading itself (for example, a large purchase order temporarily inflating prices, which reverses after the buy order is complete). With nearly 5 times the average daily dollar volume, MDY still has a slight advantage when it comes time to trade, but that advantage is smaller than one would expect given the vast difference in volume and is meaningful only for extremely large trades. Between the two, long-term investors should focus on the estimated holding cost whereas short=term or large dollar traders might focus more on the market impact costs.
Looking at their market impact cost estimates, a hypothetical $100,000 trade in IJH might move the market at most 0.03%, 3 times larger than MDY, but still in the top sixth percentile in terms of all the most liquid ETFs. If we divide the assets under management by the average daily dollar volume, we can get a rough estimate of the average holding period. For MDY, that number is 17 days compared with about 90 days for IJH. So the average holding period of IJH is about 5 times longer than it is for MDY. Still, MDY’s slight advantage in market impact cost is not enough make up for IJH’s 0.15% advantage in its estimated holding cost, and IJH will probably continue to gather assets faster than will MDY.
While fund structure helps explain why IJH has been more popular than MDY, it does not explain why SPDR S&P 500 is still more popular than iShares S&P 500 Index IVV in terms of assets and flows. SPY has an estimated holding cost of 0.16% compared with IVV’s 0.08%, so the UIT structure still shows a cost disadvantage. However, liquidity may play a larger role than holding cost in the fund selection process. IVV has an average holding period of about 53 days compared with just 3.4 days for SPY, resulting in a typical holding period in SPY that is 15 times longer for IVV. Another way to think of it is that for every dollar that trades in IVV, that same dollar is traded 15 times in SPY. So, despite the Estimated Holding Cost advantage in IVV, traders and hedgers still prefer to use SPY, which has about a 1 basis point market impact cost advantage over IVV.