Billions of dollars are traded and retraded in split seconds every day. Yet men in blue jackets still run around frantically, doing I don’t know what.
It seems like these people have no other role than to be used as background film, holding their heads during a TV report on a stock plunge that day.
They do the same thing they always did, just less of it. The amount of actual pit trading on the floor of the NYSE is a fraction of what it used to be.
The men on the floor are specialists and each handles a limited number of stocks, some of which get few trades. If you want to buy or sell GE, then a message is sent to the GE specialist, who matches you with a buyer.
The system still works, so there’s no need to replace it.
perhaps the appropriate question should be, what is the point in their physically being on the floor? Why don’t they work from their own offices elsewhere, like all the other traders?
Ditto on the question. And, in your example of GE, that would seem a counter example; don’t trades get made that within a matter of seconds millions of $ are won or lost?
Nymex Trading Pits Shut Down, Marking End of an Era
Decision by parent CME to close Nymex’s open-outcry trading floor is latest step in inexorable shift toward electronic trading
I still don’t really get it.
To be fair, a lot of major stock exchanges have indeed switched to electronic trading, and NYSE is one of the holdouts who refuse to abolish open outcry entirely. There are some arguments in favour of keeping it (mostly that human specialists facilitate market-making to ensure that as many orders as possible are executed even in not-so-liquid securities), but there’s also a lot of nostalgia involved in not scrapping open outcry. “No need to replace it” is an oversimplification, since a lot of places are, in fact, seeing a need to replace open outcry with electronic trading.
I’ll miss it. I never had the slightest clue what they were doing but it was fun to watch sometimes.
The basic principle is that they are communicating verbally and via an agreed-upon code of hand signs to buy and sell securities or commodities from each other. On the floor of the NYSE, there are two fundamental types of people on the floor: Brokers and specialists.
Brokers are usually (I think self-employed brokers still exist, but are getting rarer) employed by banks or other institutions which wish to trade on the stock exchange. They can do so for different purposes: Either the institution the broker works for wants to buy or sell securities on its own behalf, or it is executing client orders. A client order is exactly what it sounds like: A client instructs his bank to buy or sell certain securities on the client’s behalf, and the bank provides that service, collecting a commission from the client for doing so. It is important to remember that the broker typically does not decide when and if to buy or sell a security; he takes orders from whomever he works for, and his job is to make sure that the order is carried out (the people who decide what to buy are situated elsewhere; they might be, for instance, portfolio managers employed by an institutional investor to run that investor’s holdings). Such orders can take many different forms; the two fundamental types are market orders, whereby the broker is instructed to conduct the transaction at the best terms available (i.e., lowest price for buying, highest price for selling), and limit orders, whereby the order comes with a price limit above which the broker is not supposed to buy, or below which he is not supposed to sell. Many different types of limits exist, however.
The other type of people on the floor are specialists. Each specialist is in charge of a particular security or category of securities; for each security there is only one specialist. The specialists have defined stalls called posts where they stand all of the time and to which they are assigned. The trading of the securities assigned to a given specialist takes place at that specialist’s post.
The shouting and gesticulating on the floor is, ultimately, communication going on between brokers and specialists about trades which people intend. Suppose a broker receives an order for a particular trade. The trade could be to buy 10,000 shares at, say, $45 each. The broker can communicate that order to a specialist, who will take it in his “book”, which used to be a physical book but is now kept electronically. It is a registry of all open order in a particular security.
The book will be sorted by buy and sell orders, and they are also sorted by the quality of the order. By quality I mean likelihood of being executed: For sell orders, the order with the lowest price is listed first (because the seller willing to sell at the lowest price will be the first to find a buyer); conversely, for buy orders the highest price will be listed first. If a specialist has matching buy and sell orders for a given security (matching meaning that the prices at which the buyer is willing to buy and the seller willing to sell match), he will execute the two against each other. The information that the order is executed is transmitted to those who placed the orders, and the information goes on to settlement, i.e., the actual delivery of the securities (which takes place outside of the stock exchange and is different from trading). The orders are then deleted from the book. As a result of this, there will typically be a bid-ask spread: The order with the highest “bid” price (that is the price the buyer is willing to pay) will be below the lowest “ask” price (the price at which a seller is willing to sell). The difference between the two is the bid-ask spread. In the case of our specialist, who is just receiving an order to buy 10,000 securities of that sort at $45 each, there could be a situation where the same specialist has an open sell order in his book for 10,000 securities of that type at $45.20. In that case, the security will be quoted at $45.00 x 10,000 / $45.20 x 10,000, meaning there’s a buy order (or sum of buy orders) over 10,000 securities at $45 and a sell order (or sum of several sell orders) for 10,000 at $45.20, and a bid-ask spread of twenty cents. This information can be communicated via hand signals but is also available electronically, e.g. on screens on the trading floor.
In this situation, there are several possibilities. If nothing else changes, no trade takes place because buyers and sellers don’t match. But it is possible that now a buyer who posted the $45 bid ups his bid and changes it to $45.20. Then the orders match, the transaction is executed, and a price of $45.20 is posted as the “last price”, i.e., the last price at which a transaction occurred. Complete price data on the stock exchange always includes all three components: The current highest bid price in the book, the current lowest ask price in the book, and the last price at which trading took place. Conversely, it is also possible that the seller who submitted the $45.20 sell order lowers his ask price to $45, and that becomes the last price. It’s also possible that buyer and seller meet somewhere in between, say at $45.10. Lastly, it is possible that the specialist himself jumps in: Specialists are allowed to execute orders against their own portfolio, and they have an incentive for doing so because it generates transactions and keeps the bid-ask spread low. Specialists are constantly evaluated against the bid-ask spread in the books they manage and try to keep it as low as possible. In that scenario, the specialist could jump in by submitting, on his own behalf, a buy order for $45.10, thereby reducing the bid-ask spread from twenty cents to ten cents. This is called market-making and is considered desirable, because it makes the market more “liquid”, which means that more trading takes place and it is easier for orders to be executed. Stock exchanges are competing against each other for a high degree of liquidity, since investors prefer to do their business on liquid markets.
Brokers are also allowed to trade directly with each other without recourse to a specialist. In that case, they are obliged to report the price at which they traded to the stock exchange (for this purposes there are papers lips available all over the trading floor; the brokers fill in the slip with the details of the trade, scan it at one of the electronic readers also available all over the floor, and the data is disseminated).
So in essence, the shouting and gesticulating is a coded form of communication whereby people look for other people who are willing to buy securities from them, or sell securities to them, and the associated price and quantity negotiations. It may look like chaos at first glance, but there’s actually a sophisticated code behind it that mkes it possible for everyone involved to unambiguously understand each other’s signals.
All this may be true, but that communication is happening among the relatively small number of other people who happen to be in that location. Couldn’t the exact same thing be done with each person sitting in their own office, communicating with far more people? Wouldn’t the larger population enable them to get better prices for a given stock, by finding the small number of people - or the one person - who wants what I have for the price I’m asking?
Sure, but you have to find a practically feasible way to do so. Modern telecommunications and IT allow you to - institutions that wish to trade enter their orders in a big electronic order book, and an IT system executes matching orders against each other. A lot of this is happening, and a regulatory term used for such a kind of set-up is multilateral trading facility in the EU and alternative trading system in the United States.
But in the olden days, the open outcry system was pretty much the closest you could get, in a practically feasible way, towards giving as many people as possible access as fast as possible to trading. Banks or other institutions who wish to trade would be physically represented by a broker acting on their behalf; and all the brokers meet in in a physical venue with direct face-to-face contact to all other brokers representing all other parties interested in trading.
Yes - and it is. The vast majority of securities trading is done electronically now.
The only reason the trading pits at NYSE continue to operate is that there’s a contingent of old-school traders who believe it gives them an edge to see the “live” market in human form. And maybe they’re right. But I think it’s just largely down to tradition. The smaller commodities and stock exchanges went all-electronic years ago (and NASDAQ was always electronic.) IMHO it’s just a matter of time before NYSE does as well.
The NYSE (not MYMEX) still has a small amount of open-outcry trading, I think.
Thanks for the detailed explanation, but Leo Bloom’s question (and mine) is: Why can’t all you describe be done automatically via computer? Prices are discovered in milliseconds, far faster than even the fastest human trader can wiggle his fingers!
I think part of the answer is that some securities (e.g. some options) are thinly traded enough that the bid-ask spread set by algorithms and end customers will be very wide, and better prices can be obtained when two specialist/brokers are in the same room. (When they hear a new bid they needn’t worry about a sudden news event — it’s just George responding to their request.) But these are, almost by definition, low volume opportunities, making this sort of trading increasingly obsolescent.
That is pretty much covered by friedo. It’s not like the thinking behind your question is wrong - there is a massive development going on towards purely electronic trading for precisely the reasons you describe. That development has also affected major stock exchanges, not just small ones - Frankfurt, for instances, has completely abolished open outcry, and does all its trading through its electronic system called Xetra. The trading floor still exists, but nowadays in merely boasts people staring at computer screens rather than shouting at each other, and the main reasons for having even that room is to give TV journalists a backdrop for their stories. A lot of other markets have gone the same route.
Yes, that is one of the reasons given by those who advocate that open outcry should be kept. The argument is that market-making is facilitated if a human specialist sees an opportunity to jump into a wide big-ask spread. But as you say, the argument does not really hold for liquid securities with high trading volumes, and even for fringe products I’m sure that artificial intelligence could be created that does a similarly good job as a human specialist. I’m with friedo here - a massive amount of trading is already done electronically, and the holdouts who stick to open outcry are doing so mostly for nostalgia. Sooner or later open outcry will be completely a thing of the past, it’s just that we haven’t fully completed the transition yet.
Just wanted to say thanks to Schnitte for taking the time to write one of the best descriptions of this baffling process that I have ever read.
Schnitte did a great job explaining the role of floor traders but as others have noted, this doesn’t explain why computers can’t do the job better. In fact, generally, they can. This article says that as of 2014, less than 15% of trades were done on the floor of the NYSE. End of an era: The NYSE floor isn't even good for PR photos anymore I think it’s less than 10% now, and perhaps more like 6%. This article quotes one guy saying it’s “pretty darn close to zero.” This is the last photo we’ll ever run of the NYSE trading floor - MarketWatch
The NYSE is, I believe, the last U.S. stock market to have floor brokers. They claim that the “human judgment” of floor brokers contributes to better liquidity (ease and speed of trading) and lower spreads (lower cost of trading). NYSE: Market Model Of course, this could just be marketing speak. The NYSE needs to distinguish themselves from other markets and as the only one left with a trading floor, it’s easy for them to use that as their hook. But even the company that operates the NYSE can’t claim with a straight face that floor traders are necessary – they run several other markets that are purely electronic.
One argument is that floor traders are necessary when the market goes haywire. End of an era: The NYSE floor isn't even good for PR photos anymore* That’s plausible. We know that algorithmic trading and “fat finger” electronic executions sometimes contribute to episodic market volatility but I haven’t seen a lot of evidence that floor traders help those markets stabilize more quickly. Still, if some people believe it and they feel more comfortable trading on the NYSE than other venues because people are there as a backstop, those traders still have value in attracting that volume to the NYSE. That’s the real purpose of floor traders.
I’ve been a professional trader for 10 years. When I began trading, the floor was already losing relevance. Now, the only time I ever execute trades through a floor broker is during the opening and closing auctions. If I’m looking to find liquidity in an illiquid security, I’ll try to send my order to the floor early enough for the broker to find offsetting interest. See here (pdf) for a description of how that works.
Here’s a site that explains the hand signals. The floor wasn’t the most politically correct place.
Interesting bit of technical knowledge.
I must say this zombie definitely is better now than when it was fresh meat.
What happens if the prices are beyond being matched? Like, say, suppose that the seller is willing to accept any price of at least $40, and the buyer is willing to pay any price up to $45. Obviously in this case there’s going to be a deal, and somebody is going to be happy that they got a bargain, but what price is the trade made at? Midway in between? The price of whichever party put their order in first?
Interesting article. I can see why you and other traders might want closing trades *after *the imbalance is announced. But why do people trade at the “close” to create such imbalances in the first place?
Trades at the opening I can understand. Any trade placed when the market is closed defaults to a “trade at the open”, right? But why do traders specify “at the closing”, especially since, as that article implies, trades at the close are likely to get a bad price if you’re on the wrong side of the imbalance which by definition most such traders will be?