ETFs Defy Traditional Classifications


Due to their unique structure and operations, exchange-traded funds don't fit neatly into traditional fund classifications. Data providers and industry observers still haven't decided for sure how to pigeonhole ETFs.

In particular, several methods can be used for calculating ETF returns, and each approach can yield significantly different results.

James Novakoff, president of registered investment advisor firm Levitt Novakoff, recently penned a report that showed Chicago fund tracker Morningstar uses last trade data to calculate ETF returns for its widely used Principia Pro database. Morningstar confirmed that it uses last trade data for Principia Pro, and that it lists ETF returns based on the fund's net asset value (NAV) as well as market returns (using last trade) on its website.

Using last trade data for ETFs can lead to potential reporting inaccuracies for generic returns, especially for thinly traded ETFs. To take an example, assume an ETF's last trade takes place at 10 am on the last day of the month or quarter when performance is recorded. The 10 am price is potentially "stale," especially for volatile ETFs, because ETF bids and offers are constantly updated by market makers and specialists throughout the trading day and tend to bracket the funds' intraday portfolio prices.

When determining the returns for closed-end funds, it is acceptable to use the last trade.

"With a closed-end fund, you can get significant performance movement without a move in the actual NAV," said Lee Kranefuss, CEO of individual investor business at Barclays Global Investors. "Closed-end funds trade at more significant premiums and discounts, so market returns are more appropriate."

However, ETFs have an arbitrage mechanism that generally keeps the price of the fund in line with the NAV of the underlying index portfolio (for recent data on ETF bid/ask spreads and premiums and discounts, click here). Therefore, using last trade prices to calculate generic returns may create a distorted picture with ETFs.

"The difference between ETFs and individual stock pricing is value transparency," says ETF researcher Brad Zigler. "You don't really 'know' the minute-by-minute book value of IBM during the trading day. You do know this with ETFs."

The National Association of Securities Dealers (NASD) has required ETF managers such as Barclays Global Investors and State Street Global Advisors to list both market and NAV returns on their websites. To calculate market returns, fund managers use the mean of the bid/offer spread at 4:00 pm. NAV returns are calculated simply by using the net asset value of the underlying portfolio at the end of the day.

The fact that ETFs trade throughout the day like stocks is the main source of the classification problem, says Kranefuss. Mutual funds can only be bought and sold at the end of the day, and the price is the NAV of the fund at the close. Using closing NAV to calculate mutual fund returns has been an accepted practice for decades; how ETF returns should be standardized hasn't been settled yet.

Although Kranefuss noted that the new ETF returns regulation from the NASD shows there isn't uniform industry acceptance for any one method yet, he believes using NAV returns is acceptable for calculating generic ETF returns.

"The passive ETF is actually a simple model - a fund that tracks an index very closely throughout the day," said Kranefuss. "Problems arise when you try to think about it in a traditional open-end or closed-end way. Using NAV is an easier metric for looking at long-term performance. The issue of using NAV or the mean of the bid/offer spread at 4:00 pm really comes down to a few basis points if you're looking at monthly or yearly returns."

The bottom line is that ETF investors and analysts are looking at ETF returns in different lights.

"If you're an ETF investor and you sell during the day, then what you care about is market return," said Kranefuss. "That's your gain or loss. Market return is what you can take home and eat."

Novakoff concurred by saying returns of individual client portfolios should be calculated using actual ETF purchase and sale data.