Thursday, September 23, 2010

ETF Mechanics are Positively NOT Beyond Comprehension - Contrary to Herb Greenberg

Herb Greenberg did a segment on CNBC today about Bogan's "Can ETFs Collapse Piece."  I used to read Greenberg 10 years ago when he was the voice of reason in finding overvalued internet stocks, and found him to be intelligent, informed, and useful.  I guess times have changed.  In this segment, he talks about ETFs like he has absolutely no knowledge of stocks, markets, or mechanics whatsoever and ascribes to them a complexity that would be more aptly associated with particle physics.  If you watch this piece, you'll shake your head in awe that a guy like Herb Greenberg, who should know better, describes hysterically the "complicated" creation units.

Greenberg repeats the hysterical plea of "WHO WILL BE LEFT HOLDING THE BAG?" Which, of course, we already went over in my earlier piece.  Bogan, in the video, even bites my "fractional reserve lending" analogy, but puts a pejorative spin on it.  It happens to be how all stock lending works, not just ETFs.

Greenberg has an article on the subject too, but frankly, when a guy who has been writing about stocks in depth for as long as Greenberg has writes this paragraph, it makes me, simply, sad, and it's an indication that you'll be doing yourself a disservice by reading his article:

"I can’t stress the complexity of the structure. If the very nature of these “creation units” is beyond the comprehension of most investors the actual mechanics of ETFs involve an even far more complex matrix of transactions."

Ok kids - it's time to dump Herb Greenberg as your source for lessons on market mechanics.  Instead, come to Kid Dynamite's World, because, you see, these "Creation units" are such a simple concept that I could explain it to your grandmother in 5 minutes.  In fact, of course, I already did.  Here's what I wrote:

"One of the great things about ETFs is that they can be created and redeemed.   This means that "authorized participants" (read: big broker dealers) can take a basket containing the underlying stocks of the ETF, in specific weights, deliver them to the ETF trust and receive the ETF shares - that's called creating.  They can also do the opposite:  deliver the ETF itself to the trust, and receive the underlying basket of individual stocks - that's called redeeming.  "

(SARCASM ON)  Holy cow - that is incredibly complex - and nearly impossible for all but the most elite among us to understand. (/SARCASM OFF)  

You take a basket of stocks and give it to the Trust - they give you the ETF.  Or, you take the ETF and give it to the trust - they give you the basket of stocks.  WTF is complex about that?

A key realization is that the larger the short interest in an ETF is, the higher the probability that you get a short squeeze if and when people want to redeem their shares.  This short squeeze kicks off a "self healing" process that I described in the previous post, resulting in arbitrageurs creating the shares that are needed, and avoiding the feared collapse.
Amazingly, neither Bogan nor Greenberg realize that an ETF is probably more likely to "collapse" - in the sense that everyone wants to redeem and can redeem because they haven't lent their stock out -  when there is NOT massive short interest.  Imagine the simplest case where there is ZERO short interest - no shares are lent, and long holders can simply redeem their shares. (In order to redeem, you have to actually deliver your shares to the Trust, which you can't do if they are lent out, until, of course, you get them back). In the massive short interest case (Bogan's XRT example), shares have been lent (to short sellers), and thus cannot be redeemed until they are recalled from the short seller (again, I explained all of this in the other piece already) - and that borrow recall process results in short squeezes and creation of new shares, which self-heals the problem of "missing" shares (which, of course, aren't really missing at all - because only the shares that are actually outstanding can be redeemed.)

Just to help you guys sleep at night, after Greenberg rants aloud "Who will be left holding the bag? Will it be the Federal Reserve?"   No - it will not be the Federal Reserve. I answered this question already also.  In the event that an ETF gets "redeemed" out of existence (which, again, Herb and Andrew, is MORE likely to happen if there is LESS short interest),  shorts owe longs the value of the underlying basket.  Cricket.....  Cricket.  The sky doesn't fall,  the oceans don't boil, locusts don't swarm, and the Fed doesn't have to pick up the tab - it's just an exchange of cash from the shorts to the longs (and it's already collateralized when the short seller shorts the stock and posts margin).

It would be somewhat understandable for a mass media business section author to write a piece like the one Greenberg did today, but Herb - you have no excuse - you should know better.


author's note:  If you haven't read my first piece on this subject, you should do so.

also, a question for my readers:  In the first few seconds of the video clip, Faber mentions "the gold ETF" failing to track the price of gold.  Does anyone know what he's talking about?  Surely not GLD, which has tracked the price of gold tightly. (And this is not an invitation for people to leave "GLD IS A SCAM" comments - I just have no clue what Faber was talking about)


foo said...


Thanks for these posts on ETFs.

I'm remember reading this article a while ago:

In the event of a rush of redemptions where the creation basket (and I assume redemption basket?) doesn't exactly match the underlying holdings of the ETF, wouldn't there be some interesting effects on the instruments that are held in different allocations between the creation basket and the ETFs' holdings? Or am I missing something?

Kid Dynamite said...

foo - that's a good question. I haven't had experience with ETFs that have creation units that are different from the portfolio of underlying stocks owned.

In the end, the ETF never gets fully redeemed anyway - there are measures in the prospectus that halt redemptions and unwind if the fund gets too low. But what would happen is that the redemption basket would reflect the trust's actual holdings of course, so there's really nothing odd/crazy that would happen there. Instead of an ETF that tracked the portfolio, you'd end up with the actual portfolio (or the cash value of it, in the worst case)

That article is saying that the basket is optimized, making it easier to create and thus easier to trade (tighter spread, more liquid)

I'd be interested to see how the creation unit tracks the portfolio weights over time - it's possible that little things like secondary share issuances can temporarily skew the basket weights, and that the author caught it at an unusual time, and it's also possible that the creation unit overweights some stocks one day, and undeweights them the next.

This doesn't happen with the big domestic ETFs though - as far as I know.

Anonymous said...

Herb started in on the for-profit stocks August 4th and put the bottom in. He is out of touch and washed up.

sign me,

Off his (david) rocker

Kid Dynamite said...

Hey -= question for my readers - did anyone notice the first 10 seconds of the video clip - Faber says something about the "gold ETF failing to track the price of gold"

WTF is he talking about? GLD tracks the price of gold very well... ??? anyone have any ideas on what Faber was talking about? (and please, no "GLD IS A SCAM" idiots... just point out mispricing, if it's there recently and I missed it.

Unknown said...

I sent Herb the FT Alphaville article you sent with Steve's comments, not sure if he ever got a chance to read 'em, but you should reach out to Herb, far as I know he's a pretty good guy and open to learning more.

Anonymous said...

Re Fabers comment, I'm not an etf expert but I know Nat Gas etf's have had tracking error issues due to contango in the forward market, maybe something similar is going on in some gold etf's to a lesser degree? Just a guess.

IF said...

It would be useful if you could tell us how to spot a swap based ETF and other funny things that are a little bit more complicated than what you described. I am not afraid of the simple ones, it is the ones that hold a lot of Japanese stock plus swaps that start scaring me.

About gold: maybe ask Jesse, he seems to track this.

Kid Dynamite said...

Anal_yst: i looked for an email address on Herb's CNBC page, but gave up quickly.

IF: great question, and it relates to Anon's comment on the Nat Gas etf.

first rule: 1) levered ETFs: they employ derivatives/swaps to achieve daily leveraged returns (Which, we all know, do not translate into multi-period leveraged returns. 2) ETFs containing futures/options, like the Nat Gas ETF (UNG), oil (USO) etc, or even VIX (VXX)

it's always good to read the prospectus. But it's good of you to point out that there are certainly lots of ETFS that don't have easy create/redeem mechanisms - they are a different animal.

I'm over-generalizing "ETF" to mean the big State Street "Spiders" - those are you XLF, XRT, XLP, XLV, XLY, etc... then the other big basket ones: QQQQ, MDY, DIA

MacroTrader said...

The GLD etf was not tracking the Nymex price of gold as there was a low print as the last trade the previous day. Nymex price was showing up approx. $17 (or 1.3%) while the GLD and the other gold bullion etfs showing marginal positive or negative performance based on the open. The actual price of spot gold was only slightly positive for the day as the large move was recorded in the previous day, unlike the Nymex which was showing a larger gain as they didn't price in the entire move.

I believe this is the cause of the discrepancy between the GLD price and spot. Looking at a longer term chart the ETF bullion prices (GLD in specific) track bullion almost exactly (minus the MER of 0.40%).

The concerning part about the situation IMHO was the fact that I called the World Trust Gold Services (the sponsor of GLD) and they were unable to explain why there was such a large tracking error on Wednesday.

Kid Dynamite said...

thanks Macro - I don't understand what you found concerning though - you explained what caused the discrepancy.

MacroTrader said...

KD, I would think that plan sponsor would understand the factors that are moving their etf. Especially when it is a widely traded issue and there is a vast amount misunderstanding about ETF mechanics as of late (as detailed in your recent posts).

Kid Dynamite said...

Macro - it's just a time-sync issue - you already explained it! it doesn't require further explanation...

it's like how the S&P Futures trade until 4:15, while the cash closes at 4pm... if MSFT release earnings at 4:01, and futures trade up and close 15 pts rich, then the next day, the index is up 15 pts, but the futures are unchanged...

I'm pretty shocked that you called the WGC and they couldn't explain that to you - you must have gotten some monkey on the phone.

wcw said...

In fairness, as noted in re IWM, ETF trusts can see and have seen substantially all their outstanding shares show up as redemptions. Even this is not in any way a problem (unless you actually go to zero), so long as the underlying retail assets are not involved. There is, of course, one thing that will kill an ETF: lack of retail interest. The result is a vicious cycle of lower volumes, higher spreads and lower sponsor margins. Hence, the ETF death watch blog (does that still exist)?

In re basket mismatch, there is *for APs* no such thing as basis risk with ETFs. The sponsor publishes a basket daily, and the basket (there is only one) is what APs create and redeem. The possible difficulty here is creating and redeeming when basket components are not easily amassed, which is why there are cash-in-lieu provisions in AP contracts. There _could_ potentially be a problem *for the ETF sponsor* if said sponsor mismanages its own portfolio, so it is worth tracking published ETF holdings against the baskets. I have not seen this myself, but I suppose there are enough incompetents in any industry that this must have happened. And naturally at a certain point you could then have counterparty risk with the sponsor, but this is much more limited than in the case of ETNs.

Hammer Player a.k.a Hoyazo said...

Are you ever going to do a post on the inherent problems with these leveraged ETFs and how they lose their value over choppy times even if the value of the underlying basket of assets is at the same place it was earlier? That remains one of the most amazing open "loopholes" in our system to this day that the SEC, Congress, etc. seem to be basically ignoring, even though it is ( think) a factually true phenomenon. Love to hear your thoughts on this someday.

Kid Dynamite said...

Hoyazo - I don't see any reason to do a post on it, because the Leveraged ETFs do exactly what they say they will do - leveraged DAILY returns. The long term effects of that are simple math, and many have pointed out the obviousness of the fact that if an index goes from 100 --> 120 --> 100 on consecutive days, the leveraged ETFs do not finish where they started. I don't know what to add - it's math, and it's in the prospectus. (obv, 100 --> 120 is a 20% increase. 120 --> 100 is a 16% decrease)

Congress shouldn't be involved at all, except that American investors are by and large completely idiotic and can't grasp this basic math. Maybe there should be more "warnings" - to protect people from their own ignorance?

so - yes, I'll probably do a post, but not until i read something on CNBC about how "SKF is down 85% over the same period that the underlying index it's supposed to track doubly short is down 60%"... and I'll have to explain why the author is an idiot, why people should understand what they are investing in, and how compounding works...

EconomicDisconnect said...

I think they should have those little captcha math thimgies when you try to place an order for any stock or fund. If you cannot solve that you should not be buying anything. Should be in place for houses too, think of all we could have avoided!

I left a comment over at Prag Cap, he still has not answered your question on 10 trillion or all the treasuries.

EconomicDisconnect said...

Great work all day Kid, I thought you were tending mulch and stuff? I really respect PragCap but I can also not get on board.

kh said...

Re; basket mismatch.

Published baskets never reflect underlying fund holdings precisely. Share quantities are rounded, assets are excluded for various reasons, cash components are often not known at the time of creation (eg; dividend entitlements of unknown value).

The largest funds make this explicit. Eg; EEM publishes a creation basket, a redemption basket and a pure calculation basket daily. APs deliver the creation basket and receive a perfect slice of the fund; they are indeed exposed to basis risk as a result.

Foo's point is valid and interesting.

Jon Beyer said...

Perhaps Faber was referring to PHYS, which is a gold closed-end fund that has been trading at a crazy premium to NAV. (mutters) crazy goldbugs....

Anonymous said...

The most common way people make that mistake in GLD is by comparing the GLD closing price with the closing price of the Gold futures contract. Since they close 2 1/2 hours apart it creates noise in an otherwise perfect relationship.

Anonymous said...


Thanks for the posts on ETFs. Very insightful and informative logic in the midst of so many who don't know what they're talking about.

One question about your discussion of the Fed not needing to step in: what about a scenario where interest rates were spiking and the market value of the collateral used for these transactions (e.g. T-bills) went below the value of the collateralized positions? If an institution became unstable and couldn't meet margin calls, then it would be the Fed who would come in to bail them out. No?

Kid Dynamite said...

Anon - the process you just described applies to all short sales - not ETF specific, and it's why we have margin calls, of course.


has nothing to do with ETFs... the same applies to all collateral situations...

I'm not sure if the Fed is the one on the hook in that case - in the real world, what's SUPPOSED to happen is that the institution's bond holders restructure debt and accept new equity - they cover the losses. This is exactly what should have been done in our financial crisis, but instead the taxpayer (Treasury) and the Fed stepped in, horrifically.

Anonymous said...

You're right, I suppose there isn't much in a scenario of collateral problems that would be unique to ETFs (vs. any other type of security).

I completely agree with you on the bailouts, bondholders should have been dinged.