The Ugly Truth About ETFs Is Revealed (In A Footnote)

Heisenberg is going to lose his fucking mind if he here’s one more otherwise smart person try to assert that high yield ETFs are more liquid than the underlying assets.

As I’ve said on more occasions than I care to count, that proposition is a philosophical impossibility.

It makes no fucking sense.

An ETF cannot be more liquid than what it represents. It can pretend to be more liquid, and it can trade that way when everyone is stuck in the proverbial Matrix, but ultimately, corporate bond ETFs represent ownership of an underlying asset and given that, the idea that they can be more liquid than those assets is patently ridiculous.

Some people seem to get this. Howard Marks gets it. Carl Icahn gets it. And as far as I can tell, some folks at Barclays get it.

Let me just be clear: the only reason this charade works is because the flows are diversifiable and that allows people to swap portfolio products as opposed to transacting in the underlying. But without the dealer middleman (who, you’ll recall, basically disappeared following the crisis), there is no cushion if the flows become unidirectional.


If (or more likely “when”), everyone is selling, someone, somewhere is going to have to trade the underlying bonds here.

And when that day of reckoning finally comes, everyone is going to discover that the very thing folks like Vanguard and BlackRock keep mistaking for the “solution” is actually the problem. Look at this chart:


Here’s how Vanguard spins that:

For every €1 in trading volume, only 1 cent resulted in primary market trading. Put another way, 99% of the trading volume resulted in no portfolio management impact and no trading in underlying securities.

If you listen to the people pushing this fiction, that’s supposed to be a good thing. And maybe it’s ok for equities. Hell, maybe it’s even ok for IG. But it goddamn sure isn’t ok for HY.

Because what it means is that the proliferation of ETFs is leading to less and less trading in the underlying and that can only mean that the market for the underlying is getting less liquid literally by the session.

Compounding this problem is the substitution effect you’re seeing in CDX versus cash. If everyone is using CDX indices to allocate/place directional bets (as opposed to hedging), then that, by extension, means they aren’t expressing those directional views in the cash market. Again, the less that market is used, the more illiquid it becomes. And the more illiquid it becomes, the more likely it is that when we need it, it will have fallen into disrepair. That’s a recipe for a fucking firesale.

Just look at this chart:


If you plot that against HY spreads, you’ll see that to the extent anyone is willing to inventory this shit, that willingness evaporates almost completely in a pinch. That of course isn’t at all surprising, but the point is that it only serves to underscore the points made above.

Finally, look at this chart:


Ok, here’s how Goldman interprets that:

The mechanics of ETFs has made continued efficiency gains. These efficiency gains essentially allow the ETF price gap vs. NAV to close relatively quickly (Exhibit 2) even as fund flow volatility moves higher. As we discussed on previous occasions, the rising volatility of ETF flows reflects a higher velocity in the ETF create/redeem mechanism, the process that allows ETF managers and authorized participants to minimize the ETF’s tracking error. This higher velocity, a byproduct of more aggressive liquidity provision and improving technology, shortens mispricing periods and thus pushes the volatility of the NAV basis lower.

If I’ve said this once, I’ve said it a hundred times: that completely misses the point. You can’t solve the inherent liquidity mismatch (ETF shares trade like stocks but represent illiquid bonds) with “velocity.”

See the thing is, you cannot explain away an absurd model by citing “volume.” I’ve variously illustrated this point by referring readers back to a classic 1988 SNL skit about a fictional bank called “CitiWide Change Bank.” The bank’s business model was simple: you bring them one denomination and they’ll give you any kind of change you want. The punchline comes from one of the bank’s executives who says this:

People ask us all the time: how do you make money doing this? The answer is simple: volume.

That’s the same ridiculous argument everyone is making about HY ETFs. Namely that the underlying problem (no liquidity for the bonds that the ETF represents) can be solved by more trading in the ETF (volume).

So there’s your rant.

And here’s SocGen trying to make Heisenberg have a nervous fucking breakdown…

Via SocGen

Fixed income ETFs have grown considerably in size in recent years. In Europe, total assets have tripled since 2010, reaching $160bn, i.e. a 25% share of the European ETF market. In the US, fixed income ETFs amount to $450bn (3x since 2010) and 17% of the market. When subjected to closer scrutiny, it is apparent that this growth has been driven mainly by a strong demand for less-liquid strategies, including investment-grade (IG) corporate bonds, high-yield (HY) corporate bonds and emerging market (EM) bonds in local or hard currencies.

Appetite for these strategies is rooted in the search for yield amid a steady decline in global interest rates, at least until recently. When fixed income investors have the analytical skills internally to carry out bond picking, they tend to favor cash bonds. However, this may not be the case with non-standard strategies, such as HY corporate or EM bonds. In these instances, ETFs are viewed as a convenient tool to gain market exposure that would have been more complicated to access otherwise. Actually, ETFs can be easily traded on exchange or via a liquidity provider/market maker during trading hours and under normal market conditions.

Beyond the ease of trading ETFs throughout the day, ETF wrappers offer a decisive advantage in terms of liquidity (and distribution) over non-listed traditional mutual funds. By definition, ETFs combine the liquidity features of listed instruments and open-ended mutual funds, and hence enjoy several layers of liquidity under normal market conditions:

  • As other listed instruments (stocks, future contracts, etc), ETFs may be traded through the order book (“on-exchange liquidity”) or off the order book via a market maker (“off-exchange liquidity”), like in the case of block trades involving stocks.
  • As other open-ended mutual funds, ETFs can create and redeem units or shares at the end of day. The only difference with non-listed funds is that the primary market is only open to eligible market makers – referred to as “authorized participants” in the case of ETFs – and to all investors in the case of non-listed mutual funds.

Put differently, ETFs enjoy intrinsic secondary market liquidity – be it on or off exchange – that is incremental to the primary or underlying market liquidity.

For this reason, ETF liquidity is superior to that of the underlying market.


Thankfully, SocGen acknowledges as much with the following caveat:

[This assumes] the ETF secondary market remains balanced between buyers and sellers. this higher liquidity profile of ETF assumes secondary market liquidity does not disappear, which may be the case in theory during a sell-off or a one-way market. In these instances, the client (sale) order execution would require tapping underlying market liquidity. ETF liquidity would then just be the same as that of the underlying assets. If these underlying exposures are not liquid, ETFs would be not liquid either,


The problem is, that “trivial” point is buried at the bottom of a giant blue box with the following header:


See why that’s deceptive?

The title should be “Special Focus: Why ETFs Aren’t Really Liquid” and then, if you want, you can talk about the phantom liquidity that exists at the bottom. But not the other way around.



4 thoughts on “The Ugly Truth About ETFs Is Revealed (In A Footnote)

  1. Yes. And then yes again.

    But so long as the music is playing and central banks keep the liquidity flowing into the market, nobody (almost nobody) cares. What will be the timing or event that really makes a difference?

  2. At least on the retail side, aka my side, I imagine almost nobody could tell you the majority of the stocks or bonds held by a given ETF that they have shares in, and only a slightly larger proportion could tell you how that ETF tracks them.

    I suppose ETFs that hold government bonds might be an exception. Still, if history has shown me anything, it’s that the best investment strategy is not to know what companies you’re invested in and to never try to time the market by selling ahead of a crash.

Speak On It