New ETF Rule’s Hidden Costs

June 29, 2018

3. Trading disclosure (P.107): The SEC has discovered that the internet is kind of a big deal. Different exemptive reliefs have required different kinds of disclosures over the years, and this rule will standardize—and complicate—those disclosures. The big ones are:

                Premium/Discount: The rule requires market price, net asset value and the derived premium/discount to be published each day, and historical data made available. The first problem here is that there are many reasons why P/D is a dumb statistic for some ETFs (international ETFs have timing issues, bond ETFs have pricing issues, and so on). Still, many ETFs do this reporting now as a matter of course; however, implementation among small issuers is spotty. The giant monkey wrench is that the definition of market price is changing from “closing price” to “closing price, or bid/ask midpoint if closing price is bad.” While this is theoretically awesome for investors, and avoids spurious premium/discount prints, it could be a real pain to administer. The rule goes on to suggest they’re considering intraday premium/discount calculations and summary statistics, which, well … don’t even get me started.

                The bigger issue around the premium/discount requirement is that every ETF needs a system to flag the website if the premium/discount exceeds 2% for more than seven days. These kinds of triggers sound great on paper, but they’re a real headache for any web-publishing process. This will put small issuers in the position of having to remember to do this manually, and will end up with lots of room for human error.

                Bid/Ask Spread: The rule says that median bid/ask spread for the prior year be disclosed. On the surface, this sounds great. But I have personal experience trying to actually calculate this statistic from the tape, and let me tell you, it’s a nightmare. For a fast-trading security, there can quite literally be millions of quote updates a day. The only “correct” way to calculate the median spread for today (much less a year) is to calculate how long each side of the best bid/offer dwells on the book, and then time-weight the quotes. Along the way, you need a pretty robust set of heuristics to filter out bad data. This is the process used on the fund pages Trading tab (powered by FactSet). But more importantly, doing a median over a YEAR is a completely useless statistic. The median bid/ask spread changes, and it changes a lot.

Consider the B/A spread of the ETFMG Alternative Harvest ETF (MJ) over the past year:



Spreads collapsed from whole percentages to 30-40 basis points when the fund changed overnight from being a Latin American Real Estate fund to a marijuana fund (surprise!). How useful is the “median 1-year spread” here for investors? Not.

Worth The Lift?

Overall, I’m extremely pleased to see the SEC cleaning up Dodge after 25 years of frontier regulation. But as always, there’s a “be careful what you wish for” component to these things.

For small issuers, the carrot of custom baskets is probably worth the stick of the increased disclosure, but it will hardly be free to implement.

For larger issuers, I wonder if the constraints around how custom baskets can be used will chafe. Technically, the SEC is planning to automatically rescind any prior exemptive relief for funds that could be covered by the new rule when it goes live. That means the game could change quickly—both for the big guys and the newcomers.

How fast the SEC moves after the 60-day comment window expires is always anyone’s guess, but I suspect they won’t sit on their hands for long.

Dave Nadig is managing director of He can be reached at [email protected]

Find your next ETF

Reset All