Hacker News Reader

Comment by mrep | Hacker News Reader

How would they lose liquidity? Authorized participants [0] are always in the market for ETFs. If an ETF share price is crashing out of line with the index it tracks, they will step in and buy shares, swap them with the ETF issuer for the shares of the underlying stock in the index, and sell those shares for an arbitrage profit.

Even if one of the underlying stocks becomes illiquid, a big enough price divergence on all of the other liquid stocks would make it profitable to eat the loss or hold the illiquid ones (risky, but remember, there are many authorized participants competing with each other so if there is some way to make an easy arbitrage profit, they will find a way). You'd basically need the entire market to become illiquid.

[0]: https://www.investopedia.com/terms/a/authorizedparticipant.a...

View on HN

Burry's point wasn't that the ETFs would lose liquidity, but rather than the lightly-traded names in the index would.

If the index instruments have a lot more liquidity than the underlying names, then that means they won't be able to gracefully absorb the liquidity shocks from an unwind in the index.

Here's some back of the envelope calculations. Between SPY, IVV, and VOO alone, there's $500 billion in S&P 500 index ETFs. Consider if an unwind event leads to 20% of index assets being redeemed in a single day. That's $100 billion from the above ETFs alone.

Now consider a typical thinly traded single-name stock like Chubb Limited (symbol CB). CB makes up 0.3% of the S&P 500 by weight. So in the hypothetical scenario above, the APs would have to collectively sell $300 million worth of CB in a single day. Chubb's entire ADV is only $238 million.

Trying to sell more than 100% of a stock's ADV in a single day is guaranteed to produce huge market impact. The current liquidity providers in CB almost certainly cannot absorb that amount of trading volume all in one direction. In that scenario, Chubb's stock might fall by 20%, for something that had nothing to do with the company itself.

I think the overall point is that a lot of single-name stocks nowadays don't really have much of an individual market. Names like TSLA, FB or TEVA definitely have a robust market with a lot of traders still focused on company specifics. But a lot of the more boring, lower volatility, mid-cap stocks (like CB) mostly just trade along with the index nowadays. If there are technicals related to index capital flows, stocks like that are going to get taken for a ride.

[1] https://www.etf.com/channels/sp-500-etfs [2] https://www.slickcharts.com/sp500 [3] https://finance.yahoo.com/quote/CB?p=CB

This is a really good explanation, but you would think market forces would kick in. If a company drops 20% (or even 3%!) from an event that doesn't effect the business itself, you're going to get smart money buying. I'd have a hard time believing some niche hedge fund somewhere wouldn't make a killing off this by providing liquidity.

This is under the assumption that there will be capital available to flow, if there isn't because of a truly bad scenario, yes it's a downward spiral, but it will take quite an event to get us to that point.

The "smart money" would start buying, but I imagine the concern is that as passive instruments become the majority of the market, there wouldn't be a deep enough pool of assets held by "smart money" to provide offsetting liquidity in the way you describe.

But wouldn't this reach some sort of equilibrium point?

IE if a smart person with a lot of money think there is no "smart money" left, wouldn't (s)he just start their own smart money hedge fund to provide / do this?

If nobody is left to do thing X AND thing X is basically guaranteed profit, isn't it natural for people to step in and do thing X?

But we already have that. Gold. Or CHF. The problem is that it takes roughly 10 years to outshine the stock market.

I know next to nothing about investing, but I thought the article said that passive instruments are a very small portion of the market...and also that in house indexing has always been a thing.

But doesn't it seem like that, as total ownership of the market shifts towards passives, higher returns would be driven by funds running active strategies such as these, thus making them more attractive investments? It seems like this sort of thing chases an equilibrium.

> I'd have a hard time believing some niche hedge fund somewhere wouldn't make a killing off this by providing liquidity.

Typically the opportunists here would be stat-arb hedge funds. They bread and butter of their strategy is to isolate the non-explainable ("idiosyncratic") movement in single name stocks, then bet on that factor mean-reverting.

E.g. if Microsoft goes down 0.5%, but the market's up 1%, and the tech stocks are up 1.2%, and various other factors that move Microsoft don't explain it being down. Then they'll bet that the movement is driven by random noise, and buy Microsoft, betting that it will re-converge with where it's expected to be. They'll also use various techniques to isolate whether idiosyncratic movements are likely driven by company specific factors or market noise. Like NLP on a newsfeed to look to see if there are any breaking stories, or tracking recent analyst revisions on the stock.

Suffice to say that stat-arb funds in this day and age are very very good at this. In a liquidity unwind event, the signals stat-arb traders use would absolutely be lighting up. The biggest question would be whether stat-arb funds in aggregate have enough capital to counter the absolutely humongous flows that a potential index fund unwinding would release.

Another complicating wrinkle is that most stat-arb desks are no longer independent hedge funds, but units within larger multi-strategy funds (like Two Sigma or Millennium). There's good and bad. The good is that if there's all of a sudden massive opportunity the multi-strat fund can quickly reallocate more capital to the stat-arb desk.

The bad part is that stat-arb desks may be unwound due to arbitrary contagion in other parts of the market. If the multi-strat fund sees a big loss in another unit, it may pull capital from the stat-arb desk to meet margin calls or redemptions. For example this happened in August 2007 [0]. The subprime mortgage market blew up, then all of a sudden a bunch of esoteric stocks started behaving crazy for the next weeks. And that was largely because multistrat funds were pulling capital to meet margin calls on their mortgage portfolio.

[0] http://web.mit.edu/Alo/www/Papers/august07.pdf

Doesn't the "circuit breaker" process on the stock exchange exist for this scenario? To give people time to line up buyers when there is a sudden unexpected repricing?

yes, possibly. but i think that may feed back into some weirdness for the index funds

"i'm supposed to sell 500m of this small stock but i only got to sell 300m before it was frozen out...now my basket of holdings is slightly off from the actual index...hopefully i can rectify it tomorrow..."

I hate when people use acronyms without first defining them, they are literally impossible to google. What is ADV?

Average Daily Volume

Yes, if indexes sell off any many times their average daily volume that means components will as well. That's fine. If something happens where people are selling the S&P 500 down 20% in a single trading session (the full day circuit breaker limit) there should be chaos amongst the components because shit has hit the fan. Savvy players will scoop up deals, most people will do nothing and life will move on.

This is a great explanation. Can you explain why this isn't a problem for actively managed funds, though? Do they just avoid stocks like CB?

No, I think in such a scenario active investors would wind up getting just as whalloped as passive investors. By definition active investors in aggregate have the same average exposure as cap-weighted indexers to any given stock.

Arguably, after the crash they might be able to take advantage of the opportunity. If they concentrate into the stocks with the biggest price displacements, then as those stocks return to normal, they may make up some or all of their initial losses. Passive indexers can't do this, because they don't have the mandate to deviate from their pre-defined allocations.

But I think the broader issue is that excessive amounts of indexing present potential systematic risks for everybody. Burry's hypothesis is more relevant for policymakers than it is for investors.

As an individual, long-term buy-and-hold investor, low-cost index funds are without question the best option for investing. The problem is behavior that's rational on an individual level, may produce irrational results at the collective level.

I'm not sure I stated my question clearly enough. I'm not asking whether active funds would fare better than index funds in a downturn, though it's interesting to hear.

I'm asking for a comparison of two scenariors. In the index fund scenario, we're in the world where index funds represent a large and growing chunk of the market. In the active funds scenario, index funds aren't a thing. Why is it that in the active funds scenario, what you're describing in the previous post doesn't happen? Why doesn't a 20% marketwide downturn cause chaos in these smaller stocks? Is it because there would be higher volumes if those stocks were held by active funds? Or because the active funds wouldn't hold stocks like that?

Basically, I'm asking for a comparison of the scenario you described with what happens in a no-index world, because I can't quite think through the difference myself.

The difference is hypothetical, because if enough people exit the active fund, even if the fund manager thinks it's a folly, eventually they have to sell whatever makes up the fund, and eventually those sells will push prices down. That's why crashes are fast and hard.

Yes in theory the manager can try to alter the composition of the fund by selling the stocks that are still going strong. But why would they? That just exposes them to known hazards even more.

So, probably the active funds will be the first ones to drop the small stocks first in a crash.

That's an interesting point. While it's definitely true that any major unwind of that size will disrupt the market, there's some reasons to believe an ecosystem dominated by active investors would do better.

First as mentioned before, active investors have discretion. There's strong reason to believe that as the sell-off's happening that they'd move into the most dislocated stocks. That acts as kind of a negative feedback loop. It wouldn't stop a market-wide selloff, but would keep things more balanced between single-name stocks vis-a-vis the rigid rules governing passive index managers.

Second, by definition active funds are more differentiated from one another. Passive indexing produces a mono-culture with analogous ecological risks to what we see in nature. The typical active fund only holds about 50 positions at any given time. So, on the whole while active investors in aggregate hold 0.3% of their portfolios in Chubb, at an individual level most funds hold zero. And some minority may hold 1%, 5%, 10% or more of their assets in Chubb. So if one fund fails, that's less likely to spread contagion to every other fund in the universe.

In that type of unwind scenario, some funds will do pretty decently, and some funds will do horribly. But the point is there will be a dispersion of results. That makes the market as a whole more robust. Panicking investors are more likely to re-allocate their capital from the bad funds to the good funds, rather than pull all their out in a flight to quality.

Wouldn't a random selection process for basket assets and/or weights do just as well as active fund management to stave off this systemic risk?

When I actively managed money for mutual funds, we had limits on both total % ownership and how many days of volume we could own. We were small cap growth, so sometimes we could be stopped from owning as much of a company as we wanted because it was too illquid. Passive doesn't care - it buys what it needs to replicate the returns of the underlying. Also, it's well known that if a hedge fund (or mutual fund) is in trouble, they sell more of what they CAN rather than what they want. That's one reason why we always knew who else owned the stocks we did. Index funds have no leeway here. They need to sell - whether there is liquidity available or not. My gut is, there will be much less liquidity available during the next crisis than you'd expect. A lot of "non passive" investment is quant driven, which ends up becoming very homogenous in nature across different firms. Many also have momentum factors in play, where they buy short term positive momentum and sell the reverse. Once again,vthey don't have a lot of leeway for humans to say "this market is different".

>You'd basically need the entire market to become illiquid.

Yes. It has happened before.

OK, but in that case is there a distinction between index funds and actively managed funds? Is this a risk that index funds are uniquely exposed to?

Also, another thing to keep in mind is that this only affects people who are trying to sell at the bottom. Buy and hold investors care little for liquidity issues during a crash.

> in that case is there a distinction between index funds and actively managed funds?

Yes. Active managers can choose what to sell based on prevailing market conditions. Index funds must sell across the board. That could involve getting hosed on names in a short-term squeeze.

> this only affects people who are trying to sell at the bottom

There are lots of index funds. For a broad-market fund, you're probably right--a patient investor can ride out the bloodshed. For leveraged or specialized funds, on the other hand, a rout could permanently impair the portfolio.

Equity market collapses, furthermore, have a habit of transmitting into the real economy. A sustained downturn could impair funding conditions, which in turn could affect the fundamental characteristics of a portfolio.

There are some escape clauses in Vanguard's index funds:

The fund may temporarily depart from its normal investment policies and strategies when doing so is believed to be in the fund’s best interest. ... Vanguard funds can postpone payment of redemption proceeds for up to seven calendar days.

And a lot of index fund investors are buy-and-hold so it's unclear if a recession would even cause a liquidity / redemption crisis.

> And a lot of index fund investors are buy-and-hold

But index funds themselves aren't. They have to sell stock when units are destroyed and vice versa. In addition to other factors, this can result in weird tax effects as well as tracking error to the index.

Well,

1) To track the actual index, index funds must continually rebalance their portfolio. In a liquidity pause they may not be able to do this, thereby becoming a non-index fund. Actively managed funds have the portfolio they have -- unless they're defrauding the public somehow. An index fund that becomes a non-index fund would fall in this latter category.

2) The index (not the funds) is assumed to reflect all the information that can be used to make some money by arbitrage. This process is referred as "price discovery". But in a liquidity pause, price discovery grinds to a halt. Actively managed funds have their own idea of what are the fundamental prices beyond what the public leaderboard says; their price discovery is not beholden to the existence of a liquidity market. Actually -- if the market goes for years with very low liquidity, it becomes more likely that people who, say, are shorting Herbalife for fundamental reasons, have more knowledge than the index. In this way the index is like an AI that can become starved for data.

> basically need the entire market to become illiquid.

Which came dramatically close to happening in 2008, see, e.g. [0].

[0] http://pages.stern.nyu.edu/~sternfin/pschnabl/kacperczyk_sch...

> Even if one of the underlying stocks becomes illiquid

Burry’s money quote in the original Bloomberg article was on limited liquidity for a largish number of stocks - over a 1,000 stocks in Russell 2,000 weren’t traded heavily (by his benchmark).

Reply on news.ycombinator.com