A Tale of Two Liquidities
One thing that I've repeatedly noted about the recent stock-market crash is that it hasn't really spread to bonds, which to the casual observer looks like a reason to doubt worries about bond market liquidity. But bond market liquidity is a worry for all seasons. You can always find someone to worry about it. Bloomberg found Carsten Stendevad, who runs Denmark's $100 billion ATP pension fund:
“It’s not normal that we have a massive event in equity markets that triggers such a small response in bond markets,” he said in an interview on Tuesday.
“There may be a number of plausible explanations for this, including reduced bond market liquidity,” Stendevad said. “But the essential point is that we have to pay a lot of attention to the fact that historical trading patterns won’t necessarily hold up.”
Now in some ways this is a weird thing to say. Traditionally, what we talk about when we talk about "liquidity" is something like "a market's ability to facilitate the purchase/sale of an asset without causing a change in the asset's price." If bond prices aren't changing much, despite all the sound and fury in the equity markets, then that seems like a sign that liquidity is unusually good.
But it does make a kind of sense. There are two ways for illiquidity to manifest itself: gradually, and suddenly. The sudden form of illiquidity puts the emphasis on "without causing a change in the asset's price": If you want to sell your stock, and you sell it a second after you make that decision, but it's down 20 percent from where it was two minutes ago, you could plausibly complain about illiquidity. Especially if it's right back up again two minutes later. Much of the fear and loathing in current equity-market (and perhaps Treasury-market) structure is about this kind of illiquidity. We used to have patient intermediaries who would commit capital to smoothing prices, or so the story goes. But we've said goodbye to all that: Now we are in a brave new world where jumpy high-frequency traders race to incorporate information into prices, which move much more quickly than they used to in response to changes in demand.
The gradual form of illiquidity, on the other hand, puts the emphasis on "ability to facilitate the purchase/sale." If you want to sell your bonds, and you can't sell at a price that you're willing to accept, then you might also complain about illiquidity. But in this case the observed price of the bonds wouldn't change at all. One day you had bonds that last traded at $100. The next day you still have bonds that last traded at $100. Because you sat around waiting for Godot to come buy your bonds, but they never traded.
Or perhaps you don't sell your bonds because no one wants to buy them and also you don't want to sell them. Then you wouldn't exactly have cause to complain about illiquidity, but outside observers might nonetheless justifiably say that there's not much liquidity in your bonds. This happens a lot in bond markets: All of a company's shares of stock are alike, but each series of its bonds trades under its own Cusip, and there is a long tail of bonds that never or rarely trade. Here's a chart from Matt King at Citigroup:
If a bond almost never trades, its observed price almost never updates in reaction to new information. Of course if you own that bond, you may update your internal valuation in response to other observed prices in other markets. Some funds' bond valuations might be slouching toward lower levels to respond to the bonfire of the global equity markets this week. But without constant prints of trades, you won't get things like the 21 percent drop and almost complete recovery that General Electric's stock experienced within five minutes on Monday morning. You smooth out volatility not by having a lot of liquidity -- by being able to trade a lot without moving prices -- but by having none: If you don't trade at all, you won't move the price.
That doesn't exactly describe the bond market this week. Bonds are still trading. But volumes are down, and it's plausible that that's partly as a volatility-smoothing strategy: If you don't sell your bonds, you don't have to sell them at a loss. That looks like poor liquidity if you focus on the amount of trading (no trades!), but it looks like great liquidity if you focus on market impact (no price moves!).
If your question is whether to be or not to be worried about bond market liquidity, though, this seems like the sort of illiquidity that shouldn't bother you too much. Forced trading at fire-sale prices driven by investor redemptions has the potential to cause a death spiral. Just sitting around not trading is bad for price discovery, but good for the nerves.
But there are other shades of this liquidity divide. This morning I mentioned this Wall Street Journal article about how exchange-traded funds acted weirdly on Monday, with their prices falling far below their net asset values:
When the market sold off in the first six minutes of trading, many stocks were halted after triggering circuit breakers, including stocks that are included in popular exchange-traded funds.
Because this happened so quickly, many ETF market makers, or the broker-dealers who buy and sell those products, were unable to accurately calculate the value of the underlying holdings or properly hedge their trades. That caused them to lowball their buy offers and overprice their sell orders to ensure they didn’t take on too much risk. This sent ETF market values tumbling, too, and caused disruptions in the trading of other assets.
The stocks that were halted exhibited the second kind of illiquidity, in its strongest possible form: You couldn't sell them at any price, at least for a while. Their collapses were nasty, brutish and short, and were halted within a few minutes to give them time to recover. The ETFs were not halted, and they exhibited the first kind of illiquidity: You could sell them, but the price would be shockingly far below where it was a few minutes ago. (Of course, then in many cases the ETFs were halted too.) This looks a little weird. But the first kind of liquidity is, if not necessarily better than the second, at least more liquid. Aaron Brown of AQR Capital Management said in an e-mail:
The fact that equity ETFs traded below their net asset values due to trade halts in individual stocks is not a defect of ETFs, it's an advantage that increases total liquidity. It means you can continue to bet on stock prices in real time when the exchanges are trying to slow you down. You might argue on policy grounds that you don't want that -- after all, the whole notion of a trading halt is that reducing liquidity is good -- but as an investor you're always better off having the option to trade.
As Brown acknowledges, it is not a truth universally acknowledged that more liquidity is always good: Many criticisms of modern market structure explicitly argue that it promotes too much liquidity, too much trading and too much jumpiness. Stock-exchange circuit breakers intentionally reduce liquidity (by preventing trading) in an effort to contain price moves, because as a policy matter we prefer smooth price moves to constant trading. The ETFs that fell further after their component stocks were halted added liquidity -- you could trade more -- but exacerbated observed price impact, which arguably made liquidity look worse. And because most of them recovered quickly without much new information, those price moves look irrational in hindsight, which is perhaps evidence that the liquidity they provided was in some sense "excessive."
One thing that you might very tentatively say about these two kinds of illiquidity is that the first kind -- huge price impacts -- is better suited to markets where price discovery matters a lot, while the second kind -- no trading at all -- is better suited to markets full of systemically important levered investors. Stock prices move in real time, stocks are rarely halted, and everyone knows that their price moves regularly overshoot what is justified by the fundamentals. But stock market crashes, on their own, tend not to be that damaging, in part because everyone knows that stock-price swings can be so dramatic. And stock market levels are among the most widely cited pieces of economic news, not because stock prices are so fundamental but because they're so current. On the other hand, "safer" assets tend to be levered more, and to come with higher expectations of safety, so they cause more damage when they crash. Many of them also tend to trade less. And if you can prevent a crash by just not trading them, that's might be the best of outcomes, even if it might also be the worst kind of illiquidity.
Disclosure: Some ideas in this post originally appeared in a different form on Ello.
That's BlackRock's definition of "market liquidity," and seems reasonable, though BlackRock also says that "the term 'liquidity' has become a catch-all phrase for several different concepts." Similarly, BlackRock's Richie Prager has said that "market liquidity in this context is the ability to buy and sell an underlying security at a price without unduly disrupting the market." Matt King at Citigroup (quoting Kevin Warsh) defines it as "the ability to transact quickly without exerting a material effect on prices." The New York Fed says that "liquidity typically refers to the cost of quickly converting an asset into cash (or vice versa) and is measured in a variety of ways."
As one wag put it on Twitter: "The bond markets are so illiquid prices aren't moving despite huge shifts in demand!"
There are skeptics.
A related worry described in "Flash Boys" is that there is, as it were, liquidity, liquidity everywhere, but not a drop to drink: Stocks are quoted at tight bid/ask spreads, but as soon as you try to buy shares the quote vanishes. This is much of what IEX's "magic shoebox" is intended to prevent.
Consider the view of Krishna Memani of OppenheimerFunds that "Bond buyers, who currently have no incentive to buy, will eventually step up when the market declines and owning debt becomes more lucrative again." That's a view that apparent illiquidity -- a lack of bond trading -- is really just a lack of interest in trading bonds.
BlackRock complains about this a lot, and has urged issuers to issue fewer standard benchmark bonds to improve liquidity. But as long as investors keep saying, "Please, sir, I want some more" when confronted with new fragmented issues, that idea seems unlikely to get too much traction.
Call it patience, or lack of liquidity, or just paralyzing fear. Whatever it is, bond traders certainly seem immobilized amid the turmoil. Fixed-income trading volume remains about 60 percent of what would be expected in an accelerating decline, Jim Vogel, an interest-rate strategist at FTN Financial, wrote in a note Monday.
Are there 50 of them? Ugh I don't know that one may be a little too much even for me.
Unless of course stock market investors have tons of margin debt.
The classic bond market liquidity worry is not quite about leverage in the traditional sense. But it is about risk-averse investors fleeing bond funds with generous liquidity terms, which from a certain angle looks like borrowing short (from fund investors) to lend long (in bonds), even though the fund investors technically hold equity.
The interesting case that seems to be in both categories is Treasury bonds, which trade like stocks (only more liquidly) but which get levered like bonds (only more so). This works, as far as I can tell, because on-the-run Treasuries exhibit neither kind of illiquidity: They trade a lot, and their prices move smoothly. Treasuries are just great. The rare exceptions scare people.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
To contact the author on this story:
Matt Levine at email@example.com
To contact the editor on this story:
Zara Kessler at firstname.lastname@example.org