In-kind redemptions give ETFs an edge in tax efficiency, but other factors are bigger when comparing them with mutual funds.
This article originally appeared in the April/May 2012 issue of MorningstarAdvisor magazine. To subscribe, please call 1-800-384-4000.
Exchange-traded funds have been promoted as being especially tax-efficient. While broadly true, some investors mistake ETFs as being uniquely tax-efficient. In order to dispel some myths, we investigate the structure and tax-efficiency record of ETFs. We find that physically replicated ETFs with in-kind creation and redemptions tend to be more tax-efficient than their open-end counterparts, but the magnitude of the benefits pale in comparison to factors like expense ratios, tracking error, index methodology, and replication methods. We also take a critical look at iShares’ claims that Vanguard’s dual share-class structure imposes additional costs on ETF-class investors.
When Are ETFs Tax-Efficient?
To understand the tax-efficiency argument, it is helpful to compare and contrast a typical ETF with a typical mutual fund. Neither mutual funds nor exchange-traded funds pay any taxes on interest payments, dividends, or capital gains. Instead they “pass through” distributions to their shareholders on a prorated basis, which are then taxed according to each shareholder’s tax situation. As long as there are no flows in or out, a mutual fund and an ETF identical in every aspect (underlying structure, index, embedded cost basis, etc.) will distribute identical capital gains.
ETFs don’t shield dividend and interest income. The ETF’s edge in tax efficiency comes from liberal use of in-kind redemptions. When an ETF experiences redemptions, it can hand over a basket of the fund’s underlying securities instead of cash. It can also pick which shares to hand over, ridding itself of tax lots with the greatest embedded capital gains. Because the trade is conducted in-kind, no capital gains are realized. While mutual funds can and occasionally do engage in in-kind transfers to meet big redemptions, they usually just liquidate portfolio holdings and distribute the proceeds to meet withdrawals. This is a practical constraint and not a legal obligation. Mutual funds deal with the public, so it would be impractical (and infuriating) to meet all redemptions with in-kind distributions. ETFs deal directly with institutional authorized participants (not the general public) when handling fund flows; they do not have to worry about generating and doling out cash.
This does not mean an ETF can exclusively conduct in-kind trades and avoid realizing any capital gains. Whenever an ETF changes its composition (as opposed to its size), it must trade securities like any other mutual fund and realize capital gains and losses. With these facts in mind, there’s no surprise that really only two simple criteria predict how likely an ETF is to distribute capital gains.
Did it change its portfolio composition frequently? This is distinct from but closely related to whether it had high turnover. An ETF can experience significant inflows and outflows, driving up its turnover, without making a single taxable trade. Passive, market-cap-weighted ETFs pass this test with flying colors.
Did it have many opportunities to conduct in-kind redemptions? Strong performance and inflows can result in capital gains accumulating faster than they can be purged. Some ETFs cannot accommodate in-kind redemptions. ETFs tracking less-liquid markets, such as high-yield bonds and some emerging markets, and ones based on swaps or futures are prominent examples.
The Capital Gains Distribution Test
Equity and fixed-income ETFs distributed capital gains less frequently and on a smaller dollar-weighted basis than open-end mutual funds of both the actively managed and passive varieties. However, distributions fall most dramatically going from active management to passive for equity funds. We ran the numbers for all categories, but for brevity’s sake we reproduced the results of five-year dollar-weighted capital gains distributions for active and passive mutual funds and for ETFs. Active funds make hefty distributions; passive mutual funds make almost none; and ETFs completely have eliminated them. (Exhibit 1)
A few patterns stick out. First, passive strategies, regardless of wrapper, are far more efficient at limiting capital gains distributions, which stems largely from their low turnover and less-fickle investor base. Second, ETFs can improve upon this efficiency gain to a smaller degree. This improvement can entirely eliminate capital gains distributions in some cases.
While the difference between active management and passive management explains most of the differences in tax efficiency between ETFs and mutual funds, we are still left with the task of explaining why capital gains differ among similar passive strategies, sometimes dramatically.
To answer that question, we looked at ETFs and index mutual funds tracking the same benchmark. The S&P 500 trackers were the most common and oldest mutual funds. Neither of the biggest S&P 500 ETFs has ever distributed capital gains. S&P 500-tracking mutual funds, on the other hand, distributed capital gains in the ballpark of 1% of NAV annually. Similar results held for other ETF-mutual fund pairs that tracked the same indexes. (Exhibit 2)
The bottom line: The biggest differences were between active and passive funds, regardless of structure. There were some differences on the margins between mutual funds and ETFs tracking the same indexes.
Morningstar Tax-Cost Ratio Analysis
The focus on capital gains distributions neglects the composition of distributions (income versus short-term gains versus long-term gains). The Morningstar Tax Cost Ratio accounts for this information, measuring how much a fund’s annualized return is reduced by the taxes investors pay on distributions. It is analogous to an expense ratio.
We compared tax-cost ratios over the trailing three-, five- and 10 years for ETFs, passive, funds and active funds. However, left unadjusted, the tax-cost ratio punishes low-cost funds. This unintuitive result occurs because expenses are taken out of distributions first, meaning the most expensive funds will be the ones least likely to distribute dividends and income. We adjusted for this bias by multiplying a fund’s expense ratio by 15% if it’s an equity fund or by 35% if it’s a bond fund, and adding the result back into the tax-cost ratio. This implicitly assumes that any dividend an equity fund pays is taxed as qualified and any income a bond fund pays is taxed at the highest income bracket. (Exhibit 3)
Nearly across the board, ETFs are more tax-efficient than both active and passive mutual funds, though their advantage is muted in fixed income. At first glance, the differences between active and passive mutual funds are modest. However, looking at the actual holdings reveals that some index fund companies are especially good at keeping a lid on tax costs: In the large-blend category, three Vanguard funds were in the five most tax-efficient passive funds; in foreign blend, two.
Therefore, we must qualify the notion that passive is necessarily more tax-efficient than active. Some passive funds, possibly because of fickle investors, still incur large tax costs. The results suggest that the best passive mutual funds are competitive with ETFs.
Does Vanguard’s Dual-Share-Class Structure
Contribute to Tax Inefficiency?
The rhetoric between some ETF giants has become more heated recently. We think much of this is due to a change in asset leadership in a few key categories. In early 2011, the immensely profitable iShares MSCI Emerging Markets Index EEM shed nearly $10 billion in assets over a matter of months, losing its throne as the biggest emerging-markets fund to Vanguard’s much cheaper fund Vanguard MSCI Emerging Markets ETF VWO. By the end of 2011, another of its category-leading ETFs, iShares Barclays Aggregate Bond AGG, lost king-of-the-hill status to Vanguard contender Vanguard Total Bond Market ETF BND. BlackRock has fought back by claiming Vanguard’s dual-share-class structure may impose additional costs on investors, particularly in tax efficiency.
The data seem to bear this out. For the three biggest pairs of ETFs for which the firms track the same indexes, the one- and three-year tax-cost ratios are consistently lower for the iShares offerings. (Exhibit 4)
The gap can be significant. IShares’ EEM beats Vanguard’s VWO by 23 basis points on the one-year measure. The irony is the ETF is a share class of Vanguard Tax-Managed International. Across the board, iShares ETFs lead Vanguard ETFs in tax efficiency. However, iShares’ desired narrative of Vanguard’s ETF investors paying the price for other investors’ behavior doesn’t hold up to scrutiny, at least in this case. Vanguard seems to be a victim of its success.
To see why, consider how an ETF gains its tax efficiency. Much of it comes from the in-kind redemption process, which allows the ETF to dole out the most appreciated tax lots without triggering a taxable event. However, strong inflows reduce an ETF’s opportunities to rid itself of appreciated tax lots. This is likely what’s happening with Vanguard’s ETFs. Looking at estimated net flows for the year as a percentage of starting assets, in every single year, for every ETF, Vanguard’s asset growth beat iShares’ by huge margins. Vanguard ETFs’ growth advantage over iShares ETFs’ averaged 92 percentage points over the past four years.
It’s also worth looking at Vanguard’s mutual fund flows. Over the past five years, the volatility of flows was modest. Even during the depths of the financial crisis, Vanguard shareholders didn’t engage in destructive liquidation that could have resulted in capital gains for ETF share-class owners. We think Vanguard’s emphasis on asset allocation and passive investing has resulted in a mutual fund clientele less prone to portfolio churning, mitigating concerns that ETF shareholders will be left with the bag when it comes to taxes.
Granted, there was at least one example of Vanguard ETF shareholders paying for other share-class investors’ behavior. In November 2008, Vanguard Extended Duration Treasury had three share classes: EDV with $50 million in assets, VEDTX (an institutional share class) with $253 million, and VEDIX (which virtually had no assets). The fund’s year-to-date return through November was 22%. Through December, it was 55%. In December, the fund had a $108 million outflow, or 32% of the fund, all of which came from the mutual fund, not the ETF. This resulted in an $11 capital gains distribution, 11% of the net asset value at the start of the year.
It seems the biggest tax risk to Vanguard ETF investors is when the ETF share class is tiny compared with the mutual fund class, allowing a handful of investors redeeming shares to trigger capital gains. With the biggest Vanguard ETFs, this should not be much of a risk.
Tax Efficiency Is Only Part of Efficiency
We find it hard to credit concerns that Vanguard’s dual-share-class structure poses a significant risk to ETF investors. Granted, a couple of Vanguard’s ETFs have been less tax-friendly than iShares’ in 2011, but the funds’ massive inflows explain most of the difference. Looking forward, we expect the potential for spillover taxes to fall along with the share of total assets held by Vanguard’s mutual fund share classes.
That said, the risk of spillover taxes will always be there. But if Vanguard mutual fund holders continue to exhibit tendencies to redeem funds less often than other fund families, then the dual-class structure has several benefits over both stand-alone mutual funds or newly launched ETFs.
Dual-class-structure ETF investors have thus far been compensated by the built-in liquidity and improved ability to fully replicate the index a pre-existing asset base provides. These factors have attributed to their ETFs demonstrating strong overall performance from the day of their launch. Success begets success in passive, market-cap-weighted index funds. Investors have shown skill in selecting good index funds and ETFs, and scale generally begets lower fees and more efficient tracking in index funds. This may also help explain why, since the SPDR’s SPY inception in January 1993, the Vanguard 500 open-end mutual fund has superior pretax and aftertax performance despite being a stodgy old mutual fund.
Perhaps the most efficient aspect that ETFs have brought to the indexing landscape isn’t related to taxes at all. It is their unfettered distribution across platforms: No longer do index investors have to settle for subpar index funds that charge egregious fees.
Now that the competition is available to all investors, no matter your brokerage, perhaps it is time to eliminate 150 or more of the “platform filling” S&P 500 index mutual funds with poor tracking, high fees, and more frequent capital gains distributions in favor of the more efficient mutual funds and ETFs that are delivering on indexing promises.