Can an ETF Collapse?

[AdamA]

http://seekingalpha.com/article/226233- ... f-collapse

This is an interesting article that asks the question, can an ETF collapse?

The summary is that there seems to be some risk due to the fact that short sellers often outnumber actually share holders by large amounts.  

Apparently this causes a lot of counter-party risk should there be a run on the ETF.  

I know we discuss this a lot, but the technical details can be educational/interesting.

Anyone have any thoughts?

[MediumTex]

I think it would depend on what kind of ETF were are talking about.  I imagine the risk of collapse of a leveraged or commodity futures ETF would be greater than something like TLT.

[AdamA]

I think it would depend on what kind of ETF were are talking about.  I imagine the risk of collapse of a leveraged or commodity futures ETF would be greater than something like TLT.

Why? 

[MediumTex]

I think it would depend on what kind of ETF were are talking about.  I imagine the risk of collapse of a leveraged or commodity futures ETF would be greater than something like TLT.

Why? 

If an ETF is based upon derivatives in the first place, it just seems like it would be much easier for something to go haywire.

Has an ETF of any kind ever collapsed?

[AdamA]

Has an ETF of any kind ever collapsed?

Not that I (after a Google search) am aware of.

They just make me nervous sometimes b/c there are now so many of them. 

[craigr]

I wouldn't buy ETFs that are really ETNs (Exchange Traded Notes) as that does have counter party risk. I suspect the main risk to the ETF is the link to the asset they are trading can drop in price just as if you owned it yourself. They can also liquidate an ETF, but the assets would be returned according to the terms laid out in the prospectus.

The benefit of the ETF is the tax management in the event of large redemptions. If you have large outflows of money from a mutual fund the remaining shareholders could be left with big capital gains taxes. With the ETF the seller is exchanging the shares with the buyer. So theoretically someone just holding onto the ETF wouldn't experience any tax impacts from large outflows.

This isn't just a theoretical issue. In the past stock market busts when people move quickly to cash in some funds they do leave the remaining holders with taxes as a result as the fund disposes of shares. In fact, I'm kind of interested to see how Cuggino at the PRPFX fund is planning on handling this when the sun starts to set on the strategy and the hot money starts leaving just as quickly as it entered.

But with any investment, the fewer pieces of paper between you and it the better. For stock index it is simply not cost effective to run your own fund and the ETFs are they way to go (or established fund). But for gold and LT bonds it is best to just own them as directly as you can handle to avoid any unforeseen risks.

[murphy_p_t]

This isn't just a theoretical issue. In the past stock market busts when people move quickly to cash in some funds they do leave the remaining holders with taxes as a result as the fund disposes of shares.

was this an issue with Vanguard in the past, more specifically their s&p 500 or total stock market funds?

[craigr]

This isn't just a theoretical issue. In the past stock market busts when people move quickly to cash in some funds they do leave the remaining holders with taxes as a result as the fund disposes of shares.

was this an issue with Vanguard in the past, more specifically their s&p 500 or total stock market funds?

I would have to research it. With the Vanguard ETF structure it is less of an issue than before (the ETFs are share classes of the fund so they can keep gains under control with smart trading between high gain fund shares and ETF sales).

But yes it can happen to any fund if there are enough redemptions or a lot of turnover in a year.

[stone]

Craigr, I thought that ETFs were traded as shares like you say but also that ETF shares were created and destroyed by authorized participants exchanging ETF shares for the basket of underlying asset. So if there is net selling of the ETF and it trades at a discount, authorised participants will give the ETF manager ETF shares and the ETF manager will give them the basket of underlying asset in exchange before annuling those ETF shares. Is it not a taxable event when the manager hands over the basket of underlying asset? The amount of assets in an ETF rise and fall just as with an open ended mutual fund.
Might some of the reason for all of the security lending by ETFs be to create slack in the system so that allocations can just be shuffled to and fro rather than being sold?
I thought that closed end funds (investment trusts) had the fund structure that keeps the stock of asset come what may and so trade at a discount or premium as a consequence. I've got the stock part in the form of investment trusts because I thought that they were the most stable and efficient format.

[stone]

The bit I pasted below really makes it look as though the regulators are asleep. Why don't short sellers have to borrow shares before selling them short? If they did that then the ETF manager would have had to get hold of the underlying equities to create those ETF shares and so the danger of more ETF shares than underlying equities would not crop up.  Am I in a muddle about this?

"owners of ETF shares often far outnumber the actual ownership of the underlying index equities by the ETF operator. One might ask how that can be possible, but the creation and redemption mechanisms inherent to ETFs mean that short sellers need not be concerned about the availability of shares outstanding when they sell an ETF short—since they can always create new shares using creation units to cover short positions in ETFs in the future. In essence, there appears to be little risk to being short an ETF since the short seller can always “create to cover”?. This has led to some ETFs having shockingly large short interest as compared to their number of shares outstanding and for every additional ETF share sold short, there is another owner of that share."

[rickb]

The bit I pasted below really makes it look as though the regulators are asleep. Why don't short sellers have to borrow shares before selling them short? If they did that then the ETF manager would have had to get hold of the underlying equities to create those ETF shares and so the danger of more ETF shares than underlying equities would not crop up.  Am I in a muddle about this?

"owners of ETF shares often far outnumber the actual ownership of the underlying index equities by the ETF operator. One might ask how that can be possible, but the creation and redemption mechanisms inherent to ETFs mean that short sellers need not be concerned about the availability of shares outstanding when they sell an ETF short—since they can always create new shares using creation units to cover short positions in ETFs in the future. In essence, there appears to be little risk to being short an ETF since the short seller can always “create to cover”?. This has led to some ETFs having shockingly large short interest as compared to their number of shares outstanding and for every additional ETF share sold short, there is another owner of that share."

A trade involving a short on one side has a few days to clear (like any other stock sale).  The shares sold "short" are purchased by the buyer like any other purchase, and these shares must be delivered to settle the trade.  Generally, shares sold short are supposed to be borrowed before the short sale.  The lender is typically a fund (which makes a little extra money by lending out the securities it owns) or a brokerage house which lends securities from its customers' margin accounts (also to make a little extra money - for the brokerage, not for the customer).  After the short sale settles, there are two owners for any shares involved - the original owner from whom the shares were borrowed and the new owner who bought the shares sold short.  The short seller is on the hook to eventually return the borrowed shares (by buying the same number of shares from somebody else at some point in the future).

In the scenario you're suggesting, an AP sells shares short and then instead of borrowing the shares to settle the trade creates shares by buying the underlying assets.  This indeed may be how APs act, but this is not "short" selling - it's just selling.  When such a trade settles there are no borrowed shares, and no shares with multiple owners.  Short sales always involve multiple ownership.  If a stock reports that X% of its shares are sold short, I believe it is reporting that there are in reality 100% + X% of its shares currently "owned" (i.e. the X% sold short were borrowed and resold to another owner). 

[Ad Orientem]

Can an ETF collapse?  Of course it can.  Just because it hasn't happened yet doesn't mean it can't.  ETF's haven't even been around that long.  That said some perspective is in order.  Every form of investment carries risk.  Barring a catastrophic event ETF's as a class are probably very low risk, which is not quite the same as risk free.  Some are probably better than others and I would shy away from those that are either new or run by small time fund families.  In the event of a catastrophic event the effects would likely be felt much farther than than the collapse of an ETF.  That's why gold at the least should be held directly if possible.

I had an unrelated conversation a couple of weeks ago with a friend who insisted that something could not happen.  Which brought to mind the fact that exactly 100 years ago in Belfast Ireland, builders were putting the finishing touches on the "Engineering Wonder of the Age."  She was the longest, tallest, most luxurious passenger ship in the word.  And she was universally acclaimed as "unsinkable."

Think the unthinkable, and you will rarely be surprised.

[stone]

rick S, so in the example used, there are a bunch of people who own the S&P retail ETF who bought that ETF from short sellers. There are also a bunch of different people who own the underlying shares directly. Because of the short selling, the price of the underlying shares was not pushed up by the popularity of the etf. When the shorts think that it is time to exit, then they might have to buy the underlying shares so as to create new ETF shares to cover. Is it the case that if someone bought a large proportion of just one component company of the underlying index, then they could induce a short squeeze?

[rickb]

rick S, so in the example used, there are a bunch of people who own the S&P retail ETF who bought that ETF from short sellers. There are also a bunch of different people who own the underlying shares directly. Because of the short selling, the price of the underlying shares was not pushed up by the popularity of the etf. When the shorts think that it is time to exit, then they might have to buy the underlying shares so as to create new ETF shares to cover. Is it the case that if someone bought a large proportion of just one component company of the underlying index, then they could induce a short squeeze?

When the shorts of the ETF cover, they have to return shares of the ETF.  If someone bought a large proportion of one component company of the underlying index, this would make the NAV of the ETF shares increase, which (indirectly) tends to make the price of the ETF increase.  So, yes, this (also indirectly) squeezes the ETF's shorts regardless of whether the shorts are APs (who can create shares to cover a short position by buying the underlying assets) or not (non-APs have to cover with ETF shares bought on the open market).

I'm not sure where you're going with this.  My point was that because APs who are short can cover by creating shares doesn't mean that while they're short there aren't shares with multiple owners.  Fundamentally, it doesn't matter whether the short is an AP or not.  In either case, while the short position is open (after it settles) every shorted share has two owners and to cover the short position shares have to be returned to whoever they were borrowed from.  As long as the ETF tracks the NAV of the underlying assets, buying shares of the ETF on the open market should cost about the same as buying enough of the underlying assets to create the shares.

Maybe you're talking about the interval between the sale and the settling of the sale.  During this interval (a few days), to settle the trade a regular (non-AP) seller has to surrender shares they already owned (a "regular" sale) or shares they borrow from some existing share owner (a "short" sale).  An AP has a third option - to buy underlying assets and create shares to settle the sale.  I suppose an AP using this option is technically making a "naked" short sale (since they don't already own the shares being sold), but I wouldn't call this "selling short" (since there's no borrowing involved) - it's closer to a regular sale.  I think the defining characteristic of a short sale is that it is settled with borrowed shares.