The price isn’t always right

By Scot Blythe | November 14, 2012 | Last updated on November 14, 2012
9 min read

Mutual fund investors expect to beat an index, or at least not to lose money. Many have been disappointed on both counts over the past decade.

Still, a country’s economic prospects should feed into its stock prices. Some companies will be high-flyers; others disappointments. But taken together, their performance will average out and reflect growth in the economy.

Download full PDF DOWNLOAD PDF

That’s the theory behind index investing. But indexes have become more diverse, and the funds following them more complex, since the first ETFs were introduced. The result is the investor may not get exactly what she’s looking for.

That’s a result of tracking error.

Tracking error defined

Tracking error comes from many sources. Says Pauline Shum, associate professor of finance at York University’s Schulich School of Business, "Thinking about the S&P 500 index, you expect lower tracking error if you do a full physical replication, and you’ve taken into account the management fees and the trading costs."

John Youn, vice-president at ETF Solutions in Toronto, says those costs shouldn’t cause more than a 0.5% to 1.0% divergence from the index performance.

But tracking error comes in different shapes and sizes, reflecting the penetration of ETFs into somewhat illiquid markets, and the various structures ETFs are forced to adopt to get access to more alternative assets. Surprisingly, bonds are one of them, as well as the usual commodity suspects: oil and gold.

Which makes it imperative, Youn says, for retail clients and their advisors to do due diligence on the benchmark an ETF’s supposed to track, and how it plans to make the investment.

In today’s low-yield environment, bond ETFs are perhaps more salient. Most people wouldn’t associate bonds with such speculative investments as oil or gold. But the bond market, unlike equity markets, is not a transparent and liquid auction market with bid-ask prices published every millisecond. Like many derivatives markets, it is mostly over-the-counter.

Untapped markets

ON THE EQUITY INDEX SIDE, the market is saturated with products, meaning providers are searching for new opportunities. Here’s a look at current trends and upcoming products:

BMO Asset Management: In the next few years, more large firms and banks will get involved in the ETF space. For smaller firms, there’ll be more consolidation, says Kevin Gopaul, senior VP and CIO of ETFs and mutual funds. Also, advisors will start to consider how to use mutual funds and ETFs together, rather than focusing on how they differ. He predicts an influx of complex products that serve specific client needs. He adds, “The real untapped markets are new clients and prospects who have yet to enter the market, and who will drive its movement.”

First Asset: First Asset sees opportunities in the U.S., as well as in emerging markets and income-generating assets. It’s also looking at strategy-based ETFs that invest only in value or momentum stocks. Emerging-markets funds would only attract risk-on investors, says Barry Gordon, president and CEO. “More conservative investors like boomers will be looking for steady and predictable income generation.” He also expects a few companies to enter the Canadian market, driving up competition but not necessarily lowering prices. “Costs can only go so low before they don’t cover your expenses. There’ll be margin compression overall and some opportunities to offer cheaper products, but the excess costs of competing products like mutual funds are more likely to drop.”

Horizons: The company is pleased with the success of its covered call ETFs. Howard Atkinson, CEO, also says actively managed ETFs—only about three years old in Canada—have been the company’s biggest source of growth, and there’s room for more — the instruments have only $10 billion in AUM globally, with about $1.5 billion of that in Canada. A recent McKinsey & Company study predicts they’ll hit $1 trillion in assets globally in the next decade.

Invesco Powershares: “The industry has run slightly ahead of demand in terms of new product launches,” says Michael Cooke, head of distribution. “We’ll likely see fewer launches and perhaps some closures over the next two years. The ETF market has become too competitive for providers who aren’t fully invested in and committed to their ETF business. As we have seen in 2012, trading volumes may decrease slightly but inflows remain quite strong.” He says growing demand for income will provide opportunities.

iShares: iShares plans to focus on strategy-based ETFs in the coming year, as well as on funds that target equity market returns without exposing clients to high risk and volatility. Cash is not the place to be and clients want market access, says Mary Anne Wiley, managing director. The company doesn’t support leveraged funds or target narrow segments, which serve only a small number of investors, she says.

Vanguard: In September, Vanguard proposed five new funds: four stock ETFs and one bond. If approved, the offerings will double their fund lineup in Canada. Atul Tiwari, managing director, Canada, says most companies will continue to expand both fixed-income offerings, due to low interest rates, and their equity income lineups. He also predicts institutional investors and defined-contribution plans will integrate more ETFs, and says we’ll see more actively managed ETFs.

Investment dealers transact with each other before delivering a bond to the end investor, be it retail or institutional, or supply it from inventory, based on interdealer bid-ask prices.

That makes it virtually impossible for an ETF to hold all the bonds in a given index (see "Currency conundrum," this page). Adds Howard Atkinson, president of Horizons ETFs, "it comes down to investability in the underlying names. That holds for the equity side. Then fixed-income investability becomes a big issue because there is not one common market. [There have been] lots of cries for a common bond market, but it’s a dealer market so you’ve got lots of different levels of liquidity, different bid-ask spreads, depending on the dealership.

"Most bond index ETFs don’t replicate because the tracking error would be huge. They tend to sample and there is no tracking error there."

That’s very unlike large-cap equity indexes, where it’s generally easy to trade into the underlying stocks. Even there, however, not all stocks are liquid. In the MSCI All-Country World Index, a portfolio manager may also choose to sample the index rather than own every stock.

Youn gives the following example: Sometimes portfolio managers will own Australia and not New Zealand because Australia is more liquid than New Zealand, and that may add to the tracking error. Beyond that, some ETFs use rules-based indexes, rather than market-cap weighting. With an ETF that uses equal weighting or fundamental weighting.

"You’ve got lots of rebalancing costs because it’s rebalanced every quarter," says Atkinson. "[So,] you can have significant costs and tracking error on those mandates."

Banks and ETF providers

Then there’s synthetic replication, in which the ETF provider enters into a swap contract with a bank. For a fee, the bank agrees to provide the ETF with the total index return: capital gains plus dividends.

This shifts responsibility for tracking error to the bank. At the same time, there is a tax advantage: dividend income is transformed into capital gains, which are taxed at a lower rate.

The bank also posts collateral. In Europe, where such structures are now common, regulators have raised concern about the composition and quality of the collateral basket. For example, equity ETFs may have as collateral lots of corporate debt that the bank happens to have in inventory.

"The collateral basket can hold something completely different than the ETF, so the two values can really fluctuate," explains Shum. "When you have a very volatile market, the values of the two are not always equal. Anytime you have volatility in the market, things diverge from each other and counterparty risk would increase."

You expect lower tracking error if you do a full physical replication, and you’ve taken into account the management fees and the trading costs."

That was true during the financial crisis. However, with the collateral basket,"it’s only at certain points in time that we see that value. The net asset value of the fund we do see at a high frequency, but not the value of the collateral basket, and that’s something regulators take issue with: lack of transparency," says Shum.

Still, Atkinson notes, with swap-based ETFs, "they’re trading off the very low probability of a counterparty default; and it’s a defined counterparty default risk that can’t exceed 10% of the ETF’s net asset value. That risk only exists if you’ve got a mark-to-market gain in the ETF." If the ETF is in a loss position, it owes the counterparty money.

Another synthetic structure can be found in inverse, leveraged and commodity ETFs. These use futures contracts or other derivatives that roll over frequently (daily in the case of both inverse and leveraged ETFs; quarterly in the case of commodity ETFs). That leads to frequent rebalancing, which has caused investor confusion because futures markets will diverge from spot prices.

That’s why futures-based ETFs now say they track the futures index rather than commodity index itself. With contango, says Shum, where the price of the next futures contract is higher than the spot price, "every time they roll over, there’s a roll yield and it’s often negative. So every time they roll over a short-term futures contract, they are losing value. That adds to the deviation from the actual index."

Divergence stems from the nature of the commodity, adds Atkinson. "Oil, for example: it’s extremely impractical, if not impossible, to store it to get the physical exposure — unless you’re an oil company, with large storage tanks. So you’re forced to use the futures market, which of course introduces basis risk from the spot price." It requires a benchmark that’s differentfrom the spot price. "For our commodity contracts, we use an index that has a methodology that says over these four or eight days we roll [this proportion] from the near-month to the second-month — so it’s all rules-based and therefore that index will generate a return that doesn’t necessarily mimic spot. But it is an index that we can then come as close as possible to tracking."

Global issues

Currency conundrum

HERE’S AN EXAMPLE of tracking error in action. Take a global bond manager who wants exposure to Taiwanese bonds. They’re hard to buy, so the manager resorts to forward contracts on Taiwanese currency.

There is a close link between currency movements and bond prices: investors seeking bonds from a particular country will have to buy its currency.

But it’s not an exact fit, and hence it becomes another source of tracking error.

One issue concerns Europeans more than North Americans: the mechanics of security lending. Funds can earn fees by lending the securities they own to short sellers. What concerns regulators is how the revenue is shared, and, just like with swap-based ETFs, the quality of the collateral.

"The revenue that an ETF brings in by lending securities offsets some of the costs" and lowers tracking error, explains Atkinson. "You do have counterparty risk, such as a short squeeze on a stock in the market, if the lender can’t get the securities back from the borrower when it needs them. While the collateral offsets that risk, it may not be sufficient under the circumstances."

Atkinson wants better definitions of counterparty risk, and rules on revenue-sharing between the fund and the manager. "It’s fair that the manager is compensated for costs and time for securities lending. But there should be some parameters for what percentage the managers can take for themselves."

He also suggests an explicit limit on how much a fund can lend out. "That would go a long way to providing clarity and uniformity in the market," he says, "and help investors understand just what risks they’re taking in the ETF because of the securities lending."

The advisor should know what the benchmark is, what the manager’s philosophy is, and how the fund has performed against its benchmark in the past. Says Youn, "People should be aware of tracking error because you essentially want to know what the ETF does and what it’s supposed to do."

Scot Blythe is a Toronto-based financial writer.

Back to ETFs In Depth mainpage >>

Scot Blythe