Traders question value of stock-market circuit breakers
wsj.com
wsj.com
So what they actually do is not "suspend trading" as is often stated, but suspend order matching. Orders can be placed in the book and are matched at the end of the suspension via an auction process (all the buys and sells that have price limits such that they can be matched off) are matched at the end of the suspension (usually at a single auction price print) and then continuous trading resumes.
This is essentially the exchange pausing "exchange time" so everyone can position and then resuming it again, and is exactly the same matching process that happens at the beginning and end of each trading day (and iirc after lunch in Japan) to help the market price reflect the consensus of opinion of the information that came out overnight while markets where closed.
There is a reasonable line of thinking that continuous trading is somewhat overrated and some research advocates for matching to be done entirely through auctions. So say for instance you have 1 auction per minute throughout the day.
This feels like something that could be very intuitively explained with visuals. Does anyone know of such a visual explanation of this?
The core concept is that there are a bunch of folks that more or less promise to either post bids (promises to buy) or offers (promises to sell) continuously, and for that promise they (sometimes) get special treatment as 'market makers.'
In an orderly market, there are both bids and offers for every instrument.
A market maker (or really any market participant) desire to participate in the market is a function of a) the rest of their portfolio (whether they are net 'long' or net 'short') and b) their expectations of when and at what price they can unwind the trade if they get 'hit' on their bid (aka someone sells to them at their posted price) or 'lifted' on their offer (aka someone buys from them).
When markets are moving 5-6% an hour, it becomes almost impossible to have any certainty on a) what the rest of your portfolio really looks like and b) when and at what price you might be able to unwind a particular trade.
This results in market makers (and HFT, and other market participants) basically exiting the market.
When people say the 'bid-offer spread is wide' this is usually what they are referring to. All the usual players basically take a step back and say 'too rich for my blood' and then you have wide markets, and 'risk premiums' for assets all over the place.
Circuit breakers are an attempt to slow time down to give people a chance to understand their portfolios better, and hence return to the market. Wide bid-offers represent highly profitable market opportunities, if the market is orderly.
When it's not, you get stuff like an ETF on US government bonds trading as a steep discount to the 'intrinsic value' of the underlying bond portfolio.
Aka chaos.
https://ycharts.com/companies/TLT/discount_or_premium_to_nav
[1] http://people.stern.nyu.edu/jhasbrou/TeachingMaterials/STPPm...
So let's say you are looking to buy a stock. A limit order puts your bid price on the market and someone can come along and sell it to you at that price, but only after all the higher bids and all the bids at the same price that were entered before yours were executed.
The other option is to try to pay the asking price of someone who is trying to sell the stock. Let's say you see an ask at $80.00 and you are willing to pay that price so you put in a market order to buy. But someone could see your order and "front-run" you by buying up all the $80.00 asks and relisting at a higher price so you get stuck with it at that new higher price.
The safer option is to try to match that $80.00 ask with a limit order of your own at the same price, but there is a risk that it might not go through if someone beats you to it.
A vol break is like everyone is still able to swipe on tinder but the dating is going to happen in one big bang (pardon the pun) when the market resumes.
Investopedia and Wikipedia can explain the different order types.
In my market, the “auction” was just a period where the answer to the question “at what price do we clear the most volume currently sitting in the order book?” was being calculated by the computer.
edit:
“Isn’t that just the price where the best bid matches the best ask?”
No, because matching stops for a little while before an auction. Also, you can place some orders that will only participate in the auction.
So your order book might have a whole lot of bids at different prices encroaching into the offer side (or vice versa). Unlike during continuous trading where I don’t think bids can cross into offers (because they’d be matched and removed).
The job of our auction computer was to figure out which of those price levels would remove the most volume from the book.
I think you'd line up all the buy orders by descending price, and line up all the sell orders by ascending price. Then there are three possibilities: the buy orders are exhausted, the sell orders are exhausted, or there are enough of each that their prices cross an equilibrium point.
In the latter case, you set the price at the crossing point, and clear all trades that you can at that price.
If the buys are exhausted, you set the price at the sell which clear it. And if the sells are exhausted, you set the price at the buy that will clear it.
It would, under such a system, be advantageous to always have some buys or sells put in that you don't expect to clear, but would produce profitable trades in case of sudden market moves in your favor.
In the past, that function would have been filled by the "market maker" trader.
"During the batch interval (e.g. 1 second), traders submit bids and asks as price-quantity pairs. At the conculsion of each batch interval, the exchange 'batches' all of the recieved orders and computes the market-level supply and demand curves. If supply and demand intersect, then all transactions occur at the same market-clearing price."
The demand curves can be degenerate, which means we have to address two edge cases: "Bids and asks of exactly the market-clearing price may get rationed (pro-rata). If there is a range of market-clearing prices, choose the midpoint"
[1] https://www.cftc.gov/sites/default/files/idc/groups/public/@... [2] http://people.stern.nyu.edu/jhasbrou/TeachingMaterials/STPPm...
If you hold inventory, you might be able to sell one unit at $x, but the buyer willing to buy the next 99 units may only be willing to offer a lower price.
The notion of a "market price" for a stock is misleading in the same way. Price represents consensus, but it is only the most shallow consensus, as many participants would not transact that price but would instead offer a worse price. Liquidity is the notion of how much the price for large volumes differs from the price for smaller ones.
Liquidity is arguably the more important concept, and so artificially batching transactions into less frequent time slots reduces liquidity and thus increases risk.
Circuit breakers, like many finance regulations, are intended to address the emotion-driven aspect of prices as perceived by humans, but should not be necessary as trading becomes more algorithmic and less superstitious.
Circuit breakers are, in effect, a clove of garlic worn around the neck of animals who are better suited to cheering on football clubs than they are to executing abstract strategies.
I was under the impression that algorithms were created by humans.
I'd rather there be circuit breakers than risk the markets losing 50% in one day, which seems possible in a panic selloff.
Exactly this. 20 minutes is enough time for people to make phone calls, check code pushes, check that the books look like what they expect.
Yes, this is what is referred to as a "gap down" - it produces what is almost a discontinuity in the graph of the price. This is also why the opening price for an asset is often different than its closing price from the previous day.
What theory?
Intraday pauses are meant to the market time to digest new information. A trading halt at the Monday open requires a different justification.
The counterargument to these halts, by the way, is that they draw attention to the price drop. That may prompting more selling than would have otherwise occurred. (For example, as people rush to get their orders in before the next halt.)
For example, you could agree to buy or sell at the average price over the next 30 days.
Sell 3 shares in the morning and hey you're down 7%, sell at mid-day and you might be only down 4%.
"Sell at the average over 30 days" changes a lot based on implementation. So now again, you're talking about a derivative product with specific parameters...which in turn, will actually be bid against: if I, another trader, know you're going to reliably put shares up for sale at a certain time each day, then my inclination is to start anticipating sales - and oh look you're now functionally a futures market.
[1]: "Why active bond investors can beat the index when active equity investors can’t" The Economist, https://www.economist.com/finance-and-economics/2020/02/27/w...
Crashing equities inhibit capital formation. That happens in both the public and private markets. That, in turn, puts pressure on credit markets.
The firms that survive will cut back on spending. The ones that are too leveraged will go under. Both have tangible effects.
In this case, there is the uncertainty around demand destruction. Markets falling more than expected communicate a belief that more demand will be destroyed in the coming weeks than was expected. That, in turn will prompt corporate planning to reduce outlays. Not only is cash from financing expected to drop, but cash from operations to do so more than expected.
Nobody is lost or vanishing due to Covid-19, it's just a change in their respiratory rate.
In the pre-automated era I can see circuit breakers working in a constructive way for the markets. These days, with trading being mostly automated and algorithmic, to me it feels more like a bureaucratic tool to give government institutions some time to play the markets.
>The mechanism, which some complain does little good, was triggered twice last week during a coronavirus-fueled selloff,
Well that headline went out of date real fast.
For those that don't know, the breaker triggered again at 9:30:01 (time according to CBOE) this morning.
For example: https://www.who.int/emergencies/diseases/novel-coronavirus-2...
The amount of emails we received from customers complaining they lost a lot of money while they were asleep was part of the reason I left, and now also part of the reason I think whichever traders WSJ interviewed (can't see article bc it's paywalled) are wrong.
edit: I think maybe you were referring to ForEx currencies only.. well, there's a difference between a broker and an exchange and I'm constantly surprised traders don't know this.
PS: I worked at Euronext too (many years ago)
If price discovery and liquidity where solved problems, then of course a crypto-like model of 24*7 no holds trading would be a nice to have but the truth is very different here..
And as the entire stock market is based on trusth, the long-term and even mid-term results are horrendous and totally not worth it.
1) Forbid any trading of stock futures. To me, reading about stock futures reaching circuit breakers would invoke panic and the need to sell as soon as the market opens. Let the market open at 9:30 EST, to whatever price the market feels is appropriate at that time. Do not give them a clue as to what it can be. Let the market decide for itself.
2) Forbid short selling. I feel most of the selling is done by long-term holders, but short selling does excacerbate the situation. I'm in favor of forbidding short selling forever, let alone in a crisis, as it's an artificial technicality of the market. You can't short sell houses, or your possessions, right? So why can we short sell stocks (if you want to hedge, go buy put options)? I sure as heck didn't give anyone the right to borrow my stocks and sell it. It's mine - I own it.
2) If short selling actually did exacerbate the situation, wouldn't that create a buying opportunity for other market participants? In reality, short selling does not systematically force markets to price things incorrectly.
2) I think short selling helps misprice stocks in both directions. It creates artificial price increases simply b/c of shorts covering as well. Let stocks go up because more people want to buy it than they want to sell. And let stocks go down because more people want to sell than they want to buy. That's what a market is for other assets - whether it's ebay, amazon, groceries, etc. Don't let other technicalities affect the price.
1) No, it is not just because people see the price pre-market. It is because it beat earnings. Futures are not an arbitrary marker price.
2) There is so much arbing going on that I don't agree with you.
In both of the cases you mentioned, futures and short selling, if they are prohibited, what would happen is the resulting price would only reflect the opinion of people who were positive and the market would not reflect all the available opinion, and therefore would be less value in determining the clearing price of capital.