And Now, A Hot Take Rant On ETFs And Liquidity…

At this point, there is no question in my mind that the ETF “revolution” is out of control.

“There’s an ETF for that” is to investing what “there’s an app for that” is to daily life. No matter what it is you’re looking to do – from simply tracking a benchmark to venturing into corners of the credit market you have no business being in to replicating hedge fund strategies – you can be sure there’s an ETF out there that will let you do it. In most cases, the cost will be enticingly low.

Is this a good idea? No. Obviously it is not. In fact, when you see things like, for instance, ETFs that allow retail investors to sell protection on a basket of European corporate bonds, you’re left to wonder how some of these “innovations” are even legal.

Further (and I’ve almost given up on this debate, because everyone’s penchant for not letting common sense get in the way of blindly accepting the assurances of people with a vested interest in the products is maddening), it is unquestionably the case that high yield and emerging market debt ETFs are structurally flawed. Denying that is to accept the philosophically impossible notion that a thing can be more liquid than the basket of things it represents. No matter how ingenious the transformation mechanism, that isn’t possible. In a pinch, those ETFs will be as liquid as the underlying bonds and no more. Period. Citing a well-behaved NAV basis during times of market turmoil doesn’t change the fact that the vehicles are flawed anymore than it makes sense to suggest that because my dry-rotted tires didn’t blow out on the highway, the dry rot isn’t a problem. If my tires have advanced dry rot and they don’t break apart at high speed, the only thing that proves is that those are some resilient tires. That’s great, but it doesn’t in any way, shape or form address the underlying problem.

Equity ETFs and liquidity are a different discussion, for a variety of reasons. My gripe with equity ETFs primarily revolves around the preponderance of absurd offerings (e.g., gimmicky niche vehicles based on what Jesus would buy), the myriad risks associated with factor crowding (there’s a ton on that in a post I wrote for Dealbreaker last year) and, more broadly, the extent to which the whole kit and caboodle is absurd on its face.

What do I mean by that latter point? I mean the following (and this is excerpted from an October post documenting how investors rushed to ETFs to access liquidity during the selloff):

Unsurprisingly, 25% of ETF volumes this month are concentrated in SPY. That, Goldman notes, is more than 160bps above the 6-month average.

“This points to investors leaning on S&P500 ETFs (as they have S&P500 futures) to seek liquidity in a fast moving market”, the bank writes, stating the obvious.

As ever, folks will point to that as a positive development to the extent it means ETFs are serving as a handy liquidity conduit during drawdowns. But for the umpteenth time, I’m still not sure that makes sense from an intuitive perspective.

Yes, the “passive” characterization is meant to describe the nature of the vehicles, but implicit in the notion of an index-tracking product is the idea that people who buy it are passive investors. To deny that is to be deliberately obtuse.

In the past, I’ve been accused of conflating one “passive” with another “passive” when it comes to ETFs, but let me ask you this: Isn’t the whole idea of indexing to facilitate long-term investing with a mind towards passively tracking broad-based measures of U.S. corporates (i.e., benchmark indices)? Isn’t that the point? Is it not at least a little bit disingenuous to assert that the real purpose of index ETFs is to serve as liquidity conduits for people to tap in the event something goes wrong? Isn’t there at least some sense in which index ETFs are now to stocks what CDX and iTraxx are to single-name CDS?

I’m not in a particularly forgiving mood today when it comes to hypothetical pushback, so let me just go ahead and answer my own questions as posed in that latter excerpted paragraph: “Yes”, to all of them.

Well, in a new “ETF spotlight” note dated Thursday, Goldman touches on some of the topics that are near (and not-so-dear) to my heart and there are some passages and visuals worth highlighting in the context of equity ETFs, selloffs and liquidity. A lot of this is familiar, but it’s fresh and in light of recent market turmoil (both in Q4 and in May), highly germane.

A couple of pages in, Goldman reiterates that ETF volumes as a percentage of the total tape have remained generally constant over time, but tend to spike during selloffs. “Case in point, ETF volumes have recently averaged 29% of the total tape on the back of market volatility, an increase from 24% last month but still well below the levels seen in 4Q 2018”, the bank writes. Do note the scatterplot on the right, which shows that as volatility increases, ETF dominance of the tape does too.

(Goldman)

Goldman’s exposition is lengthy, but scanning it for sections that are relevant to this discussion finds me on page 6, where the bank says ETFs are still “almost exclusively a retail product”. After briefly rehashing why active managers might use ETPs, Goldman writes that “one area that seems relatively less explored is the potential [for ETFs] to be a source of liquidity, particularly in times of stress when moving around single stock positions is more challenging.”

That prompts a quick discussion of liquidity more generally. Obviously, market depth has been disconcertingly impaired during times of stress recently and that’s due in no small part to the liquidity-volatility-flows feedback loop. Goldman notes that for single stock pickers, the liquidity environment has become extremely challenging when markets are falling and volatility is rising. The relationship between volatility and market depth is well known – as volatility rises, liquidity disappears. Describing the chart on the left in the set of visuals shown below, JPMorgan’s Marko Kolanovic earlier this year noted that the relationship “is very strong and nonlinear e.g., market depth declines exponentially with the VIX.”

(JPMorgan)

He continued, noting that “the higher the VIX, the more liquidity is driven by the VIX, and recently up to ~80% of liquidity variations were explained by the VIX.”

Goldman underscores this on Thursday. “Not surprisingly, periods of low liquidity often line up with periods of increased volatility… suggesting that in periods where investors often need liquidity the most, it is often the most difficult to find”, the bank writes, before observing that “these periods also tend to line up with the largest increases in ETF volumes both in absolute terms and as a percent of the overall tape.” Imagine that, right?

(Goldman)

You want details? Fine. Here is Goldman digging a little deeper:

Not only do they often account for a large volume of the tape, they often trade multiple times the amount the average underlying stock does. In addition, the average bid-ask spread tends to be a fraction of the underlying stocks. These trends tend to get even further exacerbated in times of stress. For example, we find that the average ETF trades roughly 60% more dollar flow than its average weighted underlying stock counterparts and often trade twice the dollar volume of their stocks in periods of market volatility. We’d also note that the average ETF bid-ask spread trades at roughly 40% tighter versus its comparable average weighted spread of stock constituents of the past year, which becomes even more pronounced in periods of high volatility such as 4Q18.

The problem with this, conceptually, is that the less something gets used, the further into disrepair it falls. I’ve used the same argument when discussing credit ETFs and the underlying bonds. The evidence seems to suggest that the dependence on ETFs during times of turmoil is contributing to lower liquidity in the underlying names.

Again, this is at least conceptually similar to what’s going on in credit markets. Earlier this year, I cited some commentary from BofA on this, although almost every bank has written about it voluminously over the past five or so years. BofA updates the charts you see below fairly often, but I’m just going to use the same set I referenced in January.

The bottom line is that ETFs and CDX IG are shouldering a larger and larger share of the “burden” (if you will), a reflection of folks going where the liquidity is. “Investors have significantly increased the use of high grade ETFs and the CDX IG index”, BofAML wrote earlier this year, on the way to breaking things down. Trading volumes in CDX IG rose 58% in 2018 and trading in high grade ETFs soared to record levels relative to bond trading volumes. Here are the visuals:

(BofAML)

I should note that BofA is not in the doomsayer camp on credit ETFs and/or credit market liquidity, although they do acknowledge the trends and risks both for IG and HY.

The point in bringing the credit discussion up is that the same simple logic applies to equities in the context of ETFs. The more concentrated the action is away from the single names, the less liquid that market will be both in absolute terms and especially relative to the ETFs.

Critically, that trend encourages still more trading in ETFs and synthetics, in a self-fulfilling prophecy. If you want proof of just how self-fulfilling this really is, read the following hypothetical proposed by Goldman for how an active manager might go about navigating an illiquid market for shares of a small-cap regional bank:

As an example, suppose you were looking to exit a position in PBCT in December 2018 which is also the most highly correlated stock to the KRE (an ETF comprised of Regional Banks). The stock bid-ask spread averaged 6.6bps vs. the ETF of 2bps and generally volumes in the KRE were >10x the level of PBCT. Given the high correlation between PBCT and the KRE (92%), one could argue the KRE could be used as a more cost-effective way to synthetically get out of the PBCT position (e.g., sell the equivalent exposure of the PBCT holding, then manage out of the PBCT position over a period of days while at the same time closing the KRE position) without taking on significant tracking error.

That may be great advice in terms of how someone might seek to mitigate challenging circumstances in an illiquid environment for single stocks, and in that respect, the ETF (in this case KRE) is doing the active manager a great service.

But panning out to a 30,000-foot level, the ETFs themselves appear to be at least partially responsible for putting everybody in a scenario where active managers have to engage in the type of thought experiment described by Goldman. If that’s the case, it’s a problem.

Goldman goes on to note the obvious, which is that “not all managers may be able to [employ the kind of strategy outlined above] within their mandate”. And that’s not because their mandates were haphazardly conceived. Rather, it’s because this state of affairs is so regrettably ridiculous that nobody thought to plan for it ahead of time.

Finally, when explaining why active managers have “shied away” from ETFs, Goldman says the predominate concern is “the perception that their own clients are paying them to be stock pickers and are not interested in paying the ETF management fee on top of the fee they are already paying to the active manager.”

That’s probably true in most cases. But don’t tell it to all of the people who, in a post-crisis world where understanding how to make money in two-way markets isn’t a prerequisite for opening up shop and calling yourself an “asset manager”, have become rich and pseudo-famous by accepting “client” fees to do nothing more than sit around and buy ETFs for people. Some of that crowd are regulars on the CNBC circuit.

Now go think on all of that.

<end rant>

---