Retail Fund's Big Flows Fueled By Shorts

XRT had a monster day for new money. Which is probably all short. Welcome to Bizarre Land.

The SPDR S&P 500 Retail ETF (XRT | A-46) gives me more heartburn than any other ETF. It’s a poster child for bizarre ETF trading activity. The past few days are a prime example.

The headline (like a lot of headlines these days) came in the form of flows. XRT pulled in an additional 24 percent in assets yesterday, as you can see in our daily ETF flows article.

That’s nearly $140 million in new money. That’s got to mean people are bullish on the retail sector heading into the bulk of earnings season, right? Time to pile in with the herd?

Nope. Exactly wrong. XRT has the unique distinction of consistently being the single-most-shorted ETF on the market. I wrote about its bizarre fascination for short-sellers back in 2009, and the position hasn’t really changed. XRT remains several hundred percent short:

The paradox of the “over 100 percent short” position is easily explained when you recognize that any security can be lent multiple times. I lend you my shares of XRT to short, you sell them, then the guy you sold them to lends them out again.

Only the last person in the chain actually “owns” the shares in terms of being able to redeem them with the issuer. Everyone else has “encumbered” ownership and would need to recall their shares in order to show up at the window asking for a basket of retail stocks.

There’s a great paper on this by Credit Suisse that walks through the process, but I’ll crib the best chart here:

So, why is XRT always so wildly “over-short?” Honestly, I don’t know.

The short position has fluctuated enormously since the fund’s launch, but has nearly always had far more short interest than long interest. That makes it an interesting laboratory for creation/redemption mechanics.

And that brings us to the flows data.

The first thing I always look at when I see an enormous flows day is the premium and discount chart. The easiest one to grab is on my Bloomberg terminal:

There are a few things to notice here. The first is that the two days of big creations coincided with significant down days in the fund itself. The second is that the big creations also coincided with the fund swinging from a premium to a discount. The last important detail here is that ETF fund flows are always looking a day back.

So, those big creations we see on the last day actually represent authorized participant (AP) activity happening the day prior, when XRT was trading right in line with fair value on a bad day. Then, yesterday, we saw significant selling activity which drove the ETF to a discount.

My theory—which is impossible to prove—is that the enormous creation was done in order fulfill demand for short-sellers. So let’s zoom in on the recent shares-outstanding chart:

Charts courtesy of Bloomberg

Notice, first, that short interest in XRT had been steadily climbing this year, while shares outstanding have declined. At some point that becomes untenable: People need to be able to borrow XRT to short it, and at some point, the amount they’ll pay to borrow it becomes enough that APs will simply make more shares in order to loan them out to short sellers.

In fact, XRT was on the “threshold” list for most of June, right up until the beginning of July. That means that anyone naked-short XRT (market makers) had to close out their positions by positively identifying someone to lend them the securities to deliver for settlement. Securities on the threshold list trigger a bit of a scramble as market makers call the street looking for a borrow, and eventually, someone will just make new shares.

My prediction is that we’ll see short interest pop up again in XRT, in an amount surprisingly close to the $140 million in new inflows.

The important thing to remember as a long-term investor is that this has very little, if any impact, on your experience. The premiums/discounts in XRT are miniscule compared with many other ETFs, and XRT will keep merrily doing its job—providing you with exposure to the performance of a basket of 100 or so U.S. retail stocks.

Whether you think that’s a long or a short is up to you.

At the time this article was written, the author held no positions in the security mentioned. Contacted Dave Nadig at [email protected].