1. How electronic trading has changed the world
Transparency for all
Transparency used to be for a very limited group. In the old world, the public never saw the limit order book. The best that traders and investors could see was the last trade, and they might be told the current bid and ask in the pit. Of course, specialists at stock exchanges held the order books for their stocks, which meant that they were the only market participants who could see them. Big floor brokers in the futures markets would see the portion of the order book created by their own customers’ orders, so again they had an exclusive view of at least the piece of the order book for the futures contracts they filled orders for.
But in the new electronic world, there is a lot more transparency and the public knows much more about resting limit orders than it did in the floor-based world. In fact, the exchange itself holds the entire limit order book,and it is technically feasible for it to show the entire book to the public. Largely due to bandwidth costs, exchanges have chosen not to reveal the entire order book, at least not for free. More data takes more space and wider pipelines so exchanges will often charge higher fees to supply deeper looks into the limit order book. In addition, the real time book has great value to traders, and exchanges are trying to get traders to pay for that value. As a concrete example, firms that wish to see the entire NYSE order book updated in real time can do so for $60,000 per year, while firms wishing to gain access to the full limit order book for the Euronext half of the trans-Atlantic exchange can do so for EUR 40,000.
Illustrating the bandwidth problem, the CME began its electronic trading venture by showing the best 5 bids and best 5 offers, along with quantities. When it acquired the CBOT in the summer of 2007, that rival exchange was showing twice the depth – the best 10 bids and offers. Due to congested pipelines, the CME had to reduce the view of the CBOT order books from 10 bids and offers to five as it migrated these products on to its electronic trading platform, GLOBEX. However, at the end of 2008, the CME required its market data vendors to switch over to the FIX FAST protocol, which compresses market data by 70% and allows more data to be distributed with the same bandwidth.This enabled it a few months later to begin converting all of its products, beginning with the CBOT products whose order book views it had previously reduced, back to the 10-best-bid-and-offer-CBOT standard.
Speed – Algo trading and co-location
There was a time when customers were delighted to have a trade executed within a minute of placing an order. Computer-based trading has made such an exercise about 900 times faster. While numbers are not generally published, insiders will tell you that it currently takes about 30 ms (milliseconds, i.e. thousandths of a second) to get an order from the box to the CME exchange matching engine. It takes about 5 ms for the matching to take place and then another 30 ms to get the confirmation back to the box, or about 65 ms total for the round turn.
Mark Gorton, cofounder with fellow engineer and Scotsman Alistair Brown, of Lime Brokerage LLC, says that at 6000 orders per second, his Chinatown-based firm trades faster than any other company on Wall Street.While it is difficult to verify who is fastest, in 2008, the firm was trading about 175 million shares per day and on peak days more than twice that. It claimed to be able to grab price quotes in 0.1 milliseconds and then do the trade in another 0.1 milliseconds. These speeds, which will soon seem like an eternity (discussions of speed are beginning to shift to microseconds – millionths of a second) are gained largely by improving network architecture. One way to do this is to minimize the number of nodes that a message passes through between the customer and matching engine, which is why customers insist on direct market access, which will be discussed below. If there are two or three intermediary nodes, this doubles or triples the time it takes to reach the matching engine.
One of the major effects of electronic trading has been the inevitable trend toward algorithmic trading, that is toward having computers replace humans in deciding on and executing trades. Since the successful exploitation of an arbitrage opportunity depends upon a trader seeing the opportunity and getting her order in before others, speed becomes very important. It is clear that lines of efficiently written code resident on a server can recognize and exploit trading opportunities much more quickly than a human
watching a screen and moving a mouse or pressing keys. While a scary thought to some, the vision of computers trading with one another is fast becoming a reality. With everyone looking for speed, algorithmic trading has spread rapidly.
According to a September 2009 survey by the Tabb Group, algorithmic trading jumped from 30% to 70% of all US equity trades between 2005 and 2009. Because algorithmic trades tend to be of smaller sizes, the 70% of trades in 2009 translates into 60% of volume.Penetration in other countries is generally lower but on the rise. One surprising result is that the firms that are doing much of this trading are often new start up proprietary trading firms rather than the traditional investment banks. For example, a New Jersey hedge fund called Tradeworx, which employs mainly twenty-something physics and computer science grads, started high frequency trading in January 2009 and by the end of the year was doing 3% of the volume in the SPDR, one of the world’s most popular ETFs.s While Tradeworx is a hedge fund, many of the high frequency firms are simply trading their own capital. Getco, for example, is a little known proprietary trading firm founded in 1999 by a couple of former floor traders from the CME and CBOE. Trading about 1.5 billion shares a day, the firm accounts for about 15% of equity trading volume in the US.
Everything about markets is faster, and traders are continually looking for ways to trade faster. More efficient computer code, greater bandwidth and closer proximity to the exchanges’ matching engines are all pursued. The extreme case of getting closer is co-location, which involves the placement of the servers containing the black-box software in a room either close to the exchange’s matching engine or connected by extremely fast cable to that matching engine. High frequency trading and co-location go hand in hand. Again, co-location is simply is the electronic version of a practice engaged in by floor traders for decades. Futures trading floors had a species of trader known as the scalper or local. They made their living by buying and selling all day long, always ready to buy a little lower and sell a little higher than the last price. In essence, they were market makers, supplying liquidity and trying to profit from the bid-ask spread. They found that the closer they could stand to the big order fillers, the more likely they would get to trade opposite large orders coming in. Being right next to the big order filler also allows the local to build a relationship (floor traders chat during slow periods) that can lead to more favorable treatment. In fact, physical co-location on the trading floor was so important that some traders would arrive long before the opening to grab their little piece of real estate in the pit, and at times, disagreements over where one stood resulted in arguments and even fistfights. In the electronic world, you don’t have to arrive early or fight to co-locate; you simply have to pay for the service of locating your server in a place near the exchange’s matching engine.
The price of a trade
The shift to screens has dramatically reduced the cost of trading in three ways. First, it has ushered in huge increases in volumes in many assets, resulting in smaller bid/ask spreads. Second, it has resulted in substantially lower trading fees, and third, it has slashed the price of market data, especially for small traders. We will deal with the second and third of these sources of cheaper trading.
Brokerage firms have dropped their commissions drastically as the electronic revolution has pushed down costs and increased competition. Back in the early 1980s, it was not unusual to pay a 1.25% commission to trade stock at a full service brokerage firm in the US. For 300 shares of a $50 stock, the commission ran $187.50. Today, the cost of that same transaction would be between $1.50 and $10 – from 95% to 99% lower.The bid-ask spreads have narrowed considerably as well. In the 1990s it was not unusual for a NASDAQ stock to have a bid-ask spread of 25 cents. Today that same stock, following new order handling rules imposed by the SEC along with decimalization, would easily have a one cent bid-ask spread – a 96% decline.
Exchange fees did not have as much room to drop, since most exchanges were not-for-profit entities. The biggest change is that fees have gotten more complex. A combination of electronic trading, clearing, and a new-found concern over profitability that comes from being a publically traded company has resulted in a plethora of prices. Fees on new products being promoted can be reduced, while products in which the exchange has a secure dominance or incurs greater costs can be increased. Let’s take the CME as a case in point. Before the CME went public and had any serious degree of electronic trading, it had a single price for all futures customers – 70 cents per contract.It didn’t matter whether the customer traded one or 10,000 contracts, or whether the product was pork belly or Eurodollar futures, the price was 70 cents per contract.
While 70 cents was the fee received by the exchange, the customer also had to pay the floor broker $1 to execute the trade on the trading floor (an expense that disappeared with screens). So the total cost of dealing with the exchange was the sum of these two fees or $1.70 to either buy or sell. Today, the exchange fee depends upon the product traded and there are 13 different fees ranging from 16 cents for micro foreign exchange contracts (1/10th regular size) to $2.30 for full size equity futures, like the S&P 500. So equity futures cost 30% more, interest rates 30% less and commodity and foreign exchange futures within 10% of the $1.70 cost a decade back. The CME has to pay McGraw Hill and others undisclosed payments for the exclusive right to list the S&P 500 and other stock indexes, so it is logical to see equity index fees higher. The smaller fees for mini and micro products are proportional to the smaller product sizes. Weather products are innovative creations that have generally not found significant traction and need lower fees to make them more attractive. Based on February 2010 volumes, the mean CME fee is about $1.23, or 27% lower than the $1.70 fee charged before it went public and became substantially electronic. This is a much smaller drop than that seen in brokerage fees. The difference is that the brokerage industry is fiercely competitive, whereas the futures exchange world is characterized generally by exchanges with exclusive products and thus much less competition.
The price of a price
While trading fees have fallen tremendously following the advent of electronic trading, the price of real time quotes has fallen much faster and helped to fuel the growth of online trading by retail clients. As can be seen in table 2, the monthly cost to a retail trader of receiving unlimited quotes on bid/ask and last price on NYSE stocks fell from $168.50 in 1983 to $1 today. To be fair, in 1984, a retail trader had to pay the same amount as a professional, $168.50, and virtually no one did. Most retail traders had full service or discount brokerage accounts and conveyed their orders over the telephone after their brokers told them the bid, ask and last price available on the brokers’ screens. The non-professional category was added in 1984 and resulted in a 92% reduction in monthly costs. Over the next four years, as a result of unrelenting pressure from online brokers trying to build their businesses, the monthly fee for real time bid/ask and last prices dropped another 95% for non-professional traders and helped to attract a huge increase in clients to the online brokers.
The price battle was tough and fascinating. Despite the fact that exchanges had relied on sales of market data for a significant share of their revenue, the online brokers argued that charging for prices was sort of like Wal-Mart charging for a price list as you walked into the store. But the issue was even more interesting in that it raised the question of who really owned these prices. It was clear that the exchange was the entity capturing the market data resulting from transactions on its floor or electronic platform, but it was often the customers of the brokerage firms who were parties to the transactions, so shouldn’t they be considered part owners of the prices, since it was their actions that created the prices? As a practical matter, when quote vendors and brokers entered into contracts with exchanges to receive the market data, they had to sign a document saying that the exchange owns the data. But this didn’t fully settle the issue. Ownership of prices has been hotly contested in the US, but at present, the courts are saying that exchange settlement prices are not the intellectual property of the exchange and others are free to make use of them as they wish.
New order types
Many observers would argue that electronic trading has resulted in an explosion of order types. To others, there was nothing quite like the human broker, who could do virtually anything a computer could do and much more besides. The truth, as it often does, lies somewhere in between. One long time floor broker told me of a customer who would tell him to sell eggs when pork bellies went above a certain price. This was not a normal CME order type, and it required monitoring two markets, but he was young and hungry at the time and would bend over backwards to please customers.
There is little doubt that the number of order types in the electronic world is much larger than it was during the floor years, but some new orders are not actually all that new. Throughout the development of electronic trading platforms, many old timers quipped that the human floor broker could never be replaced by a computer. A case in point is when a customer would give a floor broker a large order and tell the broker to work the order gradually so that it could get executed with the least amount of market impact. Such orders were called “Disregard Tape,” “Not Held” and “Take Your Time.” This was considered the kind of value added that only an experienced human broker could supply.
In today’s electronic world, that same order is known as an iceberg order, which essentially allows the trader to display only a small portion of his bid or offer while hiding below the surface the much larger total size of the order. For example, if a trader wishes to buy 50,000 shares of stock, but does not want signal this size to the market and allow other traders to lie in wait for him, he may enter an iceberg order to show only 1,000 shares at any given time, hiding the remainder for gradual release to the market. Only after the visible portion gets hit by market orders does more of the order become visible. Such order types can be provided by the exchange, the broker, software licensed from a third party (called Independent Software Vendors or ISVs) or by proprietary software created by the trader. There is an irony in the iceberg order and the other new hidden order types. While the limit order book has become more transparent, and traders have a much better sense of the depth of the displayed market, the hidden order types actually make the limit order book less transparent.
In the old world of floor trading, the number of order types varied by exchange but generally there may have been a dozen at most – “market,” “limit,” “stop,” “market on open,” “market on close,” “not held,” “good till cancelled,” “all or none,” etc. The number of order types is today truly larger. One large, retail brokerage firm that handles both securities and derivatives offers 50 different order types and algorithms to its customers.Some of these 50 are the traditional orders available in floor trading, but many are new and possible only in an automated trading environment. One of these is the VWAP, or volume weighted average price, which allows the customer’s order to be gradually executed throughout the day to capture something close to the average of all the day’s trades weighted by volume. Such a price can be easily programmed, but really couldn’t have been executed before the advent of electronic trading.
The wonders and worries of direct market access
In the floor world, if a trader wanted direct accessBut as speed became more important, some customers, such as hedge funds and proprietary trading shops (prop shops), demanded direct links to the exchange’s matching engine, without being slowed down by passing through the firm’s risk management system. Given the amount of business such traders brought to the firm, it was difficult to refuse them, and firms would allow these customers to trade directly at the exchange in the name of the brokerage firm, without any risk management filter. Of course, the brokerage still had to guarantee the trades and make good on them if the customer failed to do so. In these circumstances, the best they could do was to carefully examine the customer’s risk management system prior to allowing direct access, and to monitor all orders submitted in real time. So instead of looking at the order before it hit the exchange’s limit order book, as in the past, the brokerage firm got out of the way and saw the order a fraction of a second after it had hit the exchange. This small delay was caused by the order going to the exchange on an express route (which was quite expensive) and coming to the firm via back roads (i.e. the free internet). If the firm saw something terribly wrong, it had the right to go in and cancel the order if it had not already been executed. But this kind of best practice followed by the big firms was not necessarily followed by all their smaller brethren. The tradeoff between giving customers the fastest possible access to the exchange’s matching engine and protecting the firm from losses resulting from algorithms gone wild is one that will continue to require thoughtful attention by exchanges, firms and regulators. to the matching engine, which in those days was the trading pit or specialist’s post, there was only one way to do it: purchase a membership and stand in the pit or around the specialist’s post. In other words, only exchange members had direct market access. The closest that customers got to direct access in those days was to call down to their brokerage firm’s booth on the trading floor and have their order carried into (or later as markets sped up, hand signaled into) the trading pit. This allowed the clerk on the floor to notice if an order was ridiculously large or otherwise unreasonable and take appropriate action. In the new screen world, many electronic exchanges initially preserved this common sense risk management by allowing only exchange members to have access to trading terminals. Customers continued to phone their orders into their brokers, who would in turn enter them into the electronic system, after checking trading or position limits that had been established for each account. Even as customers were allowed to enter orders via their own computers (or PDAs and cell phones), it was a simple matter to block orders that exceeded pre-established trading limits.
A number of people worry about the risk of this direct market access. It is a fact that algorithms can be buggy. A colleague of mine lost $258,000 in 11 seconds as a result of a poorly written program. It could have been much worse had he not been staring at the screen at the time. Another, better known firm lost $4 million in a very short time due to a bad program. Another firm actually went bust in 16 seconds after an employee inadvertently turned on a program.And there are many such examples like this, which are not reported, because of the need for secrecy and the fact that the firms are privately owned. But the fear is not that firms lose money. None of us should care if a firm that was taking speculative positions with its own capital does damage to itself. That’s the nature of the business. The fear is that bad code could result in a program going crazy and selling huge quantities of stock, triggering another crash.
While some argue that direct access customers are smart enough to design well written programs and the brokers could stop something gone awry quickly enough, errors do happen.NYSE Euronext fined a Credit Suisse Group subsidiary for slowing down trading by jamming the system with hundreds of thousands of cancel and replace requests when the original orders had never been sent. The U.S. SEC is sufficiently concerned that in January 2010 it proposed a ban on naked access.
Out trades out, fat fingers in
In the old world of trading floors, traders would shout out bids or offers and shout out an acceptance of a bid or offer.They would then record the details of these trades on special trading cards, if trading for their own account or on the customer order if they were acting as a broker. They would write down quantity and price and commodity involved in the trade, along with the time or time bracket in which the trade took place. They would also jot down the opposing trader’s floor name and the number of his clearing firm, both of which appeared on every floor trader’s badge. This information then would be keypunched into the exchange’s clearing system. If any one of these trade descriptors didn’t match, the system rejected them as an out trade, a trade that did not match and did not stand until the two parties resolved the differences. There were many, many out trades during the floor years, especially in very active markets. Sometimes a trader would simply make a mistake in writing down a trade description. Sometimes there would be a miscommunication. One trader says she’s selling 60 contracts and the other trader hears 16. Sometimes there would be a keypunch error by the exchange employee.
Back in the 1980s, there was a period when as many as 25% of all trades done in certain pits at the CME had errors in them and had to be resolved by the human intervention of special out trade clerks after the trading session or the next morning before the market opened. Each broker had such a clerk to represent him in negotiating a settlement to such an out trade. In some cases, it was clear that one of the two traders was at fault and he would eat the error. In other cases it was unclear and the two parties would simply split the cost of the error. Out trades were just considered one of the costs of trading and regulators didn’t much like them because they gave the traders the opportunity to change certain dimensions of the trade after the fact. Some exchanges would impose penalties on clearing firms if the percent of their trades that resulted in out trades exceeded a certain level.
Screen-based trading eliminated the traditional out trade along with the hordes of out-trade clerks needed to resolve these mismatched trades. Electronic trades match instantly and all relevant data to the trade is gathered automatically, so there is no room for misunderstanding. This does not mean that human traders did not make mistakes in entering orders, but the misunderstandings due to not properly hearing a counterparty or sloppily writing down the wrong name for the counterparty or clearing firm was essentially eliminated.
With the out trade being dead, where do we channel human error? Into fat finger errors. There are not as many of them, but they can be quite large. A fat finger error refers to a situation where a trader hits the wrong button – a buy instead of a sell, a wrong price or a wrong quantity. When the head of an exchange has to resign over the embarrassment created by a fat finger error, it is clear that the error is a shocking one. And that’s just what happened in 2005, when an employee of Mizuho Securities entered an order backwards. He sold 610,000 shares of J-Com at one yen per share, instead of selling one share at ¥610,000, which is what he intended.The firm only had 14,500 shares of J-Com in its inventory. But of course, traders took advantage of the cheap shares and took the other side of the order. The Nikkei 225 closed down 2% for the day. In 2001, a Lehman Brothers trader accidentally sold $300 million instead of $3 million worth of a UK company and sent the FTSE 100 down 120 points and took $50 billion off the value of the FTSE. And then there was the London Salomon Smith Barney trader who accidentally sat on his keyboard submitting orders to sell 10,607 contracts, sending the value of the Notional bond down 149 basis points.
Early on when their customer bases were growing rapidly, there were also a number of cases where broker-dealers created shoddy systems that resulted in big headaches for customers. For example, a customer at one of the largest online brokerage firms complained back in 2000 that while his account showed that he had turned his original $12,000 investment into $2.3 million, in fact he had actually been losing money.Another had a $71,000 position created in a stock he didn’t want and suffered a $53,000 loss that the broker insisted he pay because he didn’t report the error quickly enough. A third customer claimed that he entered an order to sell his 5,000 shares of a company and the online system executed the order twice, leaving him short 5,000 shares. When he called to report the error to the broker, he claimed that he was kept on hold 7 hours (the broker said it was only 3 hours) and he ultimately lost $23,000 due to the error.
Bigger and fewer
One very important effect that the shift to screens has had on the world of exchanges is that the incentive to be big is much greater for electronic exchanges. Economies of scale in the cyber world are much greater than they were in the floor world, and exchanges are racing each other to the bottom of the new electronic cost curve. In the floor-based world, adding new products or significantly increasing the volume of trade in existing products required more physical space on a trading floor and often more floor traders to handle the brokerage and market making. In the new screen-based world, new products simply require a little more space on a server. So new products and volume growth in existing products brings in more revenue without adding appreciably to costs, thus boosting profits. And profits have become very important as most exchanges have shifted from a not-for-profit member-owned structure to a for-profit stockholder-owned structure. Given that volume growth pushes up revenues much more rapidly than costs, all exchanges want to be larger and they have been doing their best to become so.
The result is fewer and larger exchanges. Euronext, the pan-European stock exchange was built from the merger of four national stock exchanges between 2000 and 2002. But Euronext went from pan-European to trans-Atlantic by merging in 2007 with the New York Stock Exchange (NYSE), which itself had merged with Archipelago, a former Chicago-based Electronic Communication Network (ECN) that had become an exchange by merging with the Pacific Stock Exchange.An extreme case of this trend can be seen in the United States. In 2004, the CME accounted for 50% of all US futures volume. By 2008, the CME’s market share had almost doubled to 97%. The reason, of course, is that the CME had acquired the second and third largest US futures exchanges, the CBOT in 2007 and NYMEX in 2008, becoming the CME Group. How did the CME emerge on top, especially given the fact that it had been the number two derivatives exchange in the US for most of the 20th century, always lagging behind the CBOT? The CME was long known as being feistier, more innovative and more aggressive than its cross-town rival. But a large part of its emerging on top had to do with its consistent leadership and long-term vision regarding electronic trading. It was able to balance the inevitable tension between the fears of its members and the need to move boldly into this new screen-based world. While the CBOT was experimenting with night session floor trading and jumping from one electronic trading platform to another, the CME marched forward with a focus on a single electronic platform – GLOBEX. And while the CME was building volume on GLOBEX, NYMEX was so married to the idea of floor trading, as late as 2005 it was still trying to expand by setting up floor-based exchanges in Dublin, London and Singapore. The CME was also the first US derivatives exchange to become a for-profit, publicly traded company, which gave it more structural and financial flexibility than its member-owned brethren. Even though the CME was slow to move into electronic trading compared to European and Asian exchanges, it did so much more effectively than its domestic rivals. NYSE subsequently bought the American Stock Exchange, a very old exchange that was the US innovator of exchange-traded funds (ETFs). A second trans-Atlantic exchange, NASDAQ OMX, was created in 2008 via NASDAQ’s complicated merger with OMX, itself a Stockholm-based company that was born of the mergers of OM with stock exchanges in Stockholm, Copenhagen, Helsinki, Tallinn, Riga, Vilnius and Iceland. A third trans-Atlantic marketplace was created by a European exchange buying an American one, namely Eurex’s purchase of the International Securities Exchange (ISE), the world’s first fully electronic derivatives exchange. This move, incidentally, restored Eurex’s position as the largest derivatives exchange globally after losing it to the CME Group for a brief period.
2. Assessment-Are we almost done and was it a good thing?
While it seems that we are in the very late stages of the transition from floors to screens—it has been going on for over three decades now—we may still have a while to wait for the time when all exchange trading is electronic. There are serious pockets of resistance. While tremendous strides have been made, there are still many transactions in equities, bonds and derivatives that still take place on the floor, over the telephone, or increasingly, via instant messaging systems. For example, very large equity transactions are likely to continue to be arranged away from the electronic exchange’s matching engine. Such transactions are best done by brokers who know the players and can match orders with minimum market impact. Options have been a particular sticking point. With many options trades consisting of more complicated combinations of contracts, many traders prefer to negotiate these on the floor of exchanges than on screens. And many option transactions that appear to be conducted on electronic platforms are actually pre-arranged in the “call-around market.” And while OTC trading has increasingly shifted to screens, the customized nature of OTC trading makes at least a certain portion of those trades likely to be negotiated over the phone for some time.
What’s good about what has occurred? While individual traders may prefer to keep their own orders and trades hidden from view, they seem to want to know as much as possible about the orders and trades of others, so we’d have to say that the transparency brought by electronic trading is a good thing. Likewise, the falling prices of both trades and market data have both delighted traders and brought in many more market participants, some of them doing huge volumes. No one will miss the “out trades,” probably the most visible manifestation of the inherent inefficiency of the old system. Even the out-trade clerks who cleaned up the messes made by the brokers have surely found something more interesting to do.
What about the speed? Most of us love speed. Who would want to ever go back to dial up internet? Who doesn’t thrill at watching new speed records set in Olympic events or on racetracks? And we all watch in slack jaw amazement at the incredible increase in speeds at which trades can be done. Speed has become the new obsession in the trading industry. And as much as some of us might like to slow our lives down, when it comes to markets, we simply don’t have that choice. Markets have always been about competition and competition, to varying extents, has always been about speed. It’s just that we’ve learned how to continually make these speeds blazingly fast.
What are the downsides to electronic trading? While the transition to screen-based trading has brought many benefits to the world, it is a clear example of Joseph Schumpeter’s process of creative destruction. To create a new world, an old one had to be destroyed. Those who earned their livelihood on the old trading floors—the runners, the price reporters, and the traders themselves—saw their jobs disappear, sometimes overnight. The new jobs created in this transition required very different skill sets: runners could not easily become programmers, and floor traders could not easily become screen traders. The new proprietary trading shops generally preferred to hire bright fresh college graduates, rather than attempt to retrain middle-aged floor traders. They had acquired too many bad habits that were too difficult to unlearn. The kids raised on video games had the speed, the motor skills, and the open minds to continually beat the older floor traders to the punch. A few of these floor traders were sufficiently visionary to set up prop shops and hire the new kids to staff them. There is no doubt that the transition has been painful. In early 2010 a documentary film called Flooredpremiered in Chicago to sold-out audiences, made up largely of CME, CBOT and CBOE floor traders. The film, which documents the disappearance of both floors and floor traders, poignantly captures the feelings of confusion and loss experienced by once successful traders who find their skills increasingly irrelevant. One trader rails against the trend as he angrily says “I hate email. I hate computers…the computer is the worst and most evil thing of trading that I’ve ever seen….It’s more devious…trust me when I say these people are cheating on the computer…The people who are making money are the programmers.”
But the process of creative destruction has expanded beyond the exchanges to the institutions broking the trades. As mentioned above, small start-up firms are doing a tremendous portion of all trading both in equities and in derivatives. With direct access to the exchange, these firms have little need for the traditional broker/dealer or futures commission merchant. They only need these firms to clear and guarantee their trades, and do not need the traditional services like research and execution.
While the destruction of the old order naturally generates nostalgia and sadness amongst those who grew up in it, as well as concern over individuals caught in the process, it is an inevitable part of technological change and does not require any regulatory intervention.
Regulators are happy about some things and worried about others. Electronic trading has solved one of the biggest regulatory headaches of previous decades, and that was the lack of a decent audit trail for trades. In floor trading, while outside orders were timed stamped when they came in, floor traders did not record the precise time they executed each of their trades. At one point the futures exchanges improved things by requiring all brokers and traders to record the 30-minute time bracket (and later the 15-minute time bracket) in which their trade took place. In addition, both the exchanges and the CFTC would use this information to attempt to reconstruct the trading day, so there would be a reasonable idea of when a trader made his own trade and entered a trade for a customer. But without a precise time for each trade, neither the exchanges nor regulatory personnel were able to tell with certainty whether a broker had traded ahead of his customer, a major violation of both exchange rules and government regulations in most countries. This allowed front running and a host of other possible abuses to go undetected. In the screen-based world, we know precisely when each transaction took place, making it much easier to detect violations.
At the moment, a few things seem to bother regulators, especially amidst the populist sentiment of the electorate following the recent financial crisis. High frequency trading generally and flash trading in particular, along with co-location, strike some regulators as unfair, giving some traders an advantage over others and especially over the general public. And a number of regulators feel that market integrity and even financial system integrity could be compromised by a scenario of a well financed but buggy algorithm gone wild when combined with direct market access. It is essential that these scenarios and risks are thoroughly explored and well understood by regulators contemplating various kinds of bans.
The technological changes taking place outside of the exchange servers will also continue to amaze, confuse, and startle us. Now that the technology genie has been let out of the lamp, he’s not going back in, and markets will continue to evolve and develop. There will be abuses, false starts, unexpected twists and turns, and things that keep traders, exchange officials and regulators up at night. There will be so much that we simply cannot yet imagine, and it will be an exciting adventure.
Michael Gorham is Industry Professor and Director of the IIT Stuart Center for Financial Markets at the Illinois Institute of Technology. He is also Adjunct Distinguished International Professor of Finance at EGADE, the graduate business school of Monterrey Tec, at the Santa Fe Campus in Mexico City. In addition, he currently serves on the board of directors for the CBOE Futures Exchange in Chicago and, until July 2008, served on the board of the National Commodity and Derivatives Exchange in Mumbai, India. He serves on the business conduct committee of the Chicago Mercantile Exchange, the editorial boards of the GARP Risk Review and of Futures Industry magazine. He is regional director of the Global Association of Risk Professionals for Chicago.
He is co-author of two books; India’s Financial Markets: An Insiders Guide (July 2008) and Electronic Exchanges: The Global Transformation from Pits to Bits (May 2009), both published by Elsevier.
From 2002 to 2004, Mr. Gorham served as the first director of the Commodity Futures Trading Commission’s new Division of Market Oversight, a division of 100 economists, lawyers, futures trading specialists and others dedicated to the oversight of the nation’s 12 futures exchanges. Earlier, Mr. Gorham was an economist at the Federal Reserve Bank of San Francisco and vp of international market development at the Chicago Mercantile Exchange.
He has been involved in consulting projects to create a stock index futures market in India, establish a commodities market in the United Arab Emirates, evaluate the feasibility of Parmesan cheese futures in Italy, and to modernize financial markets in Egypt. He also served as Managing Editor of the Journal of Global Financial Markets.
He has written for newspapers, journals and magazines in Argentina, China, Japan, Mexico and the U.S. and has given talks on derivatives in 14 countries. After university, he was a Peace Corps volunteer working on an agricultural modernization project in Malawi, Africa. He holds a BA in English literature from the University of Notre Dame, an MS in food and resource economics from the University of Florida and a Ph.D. in agricultural economics from the University of Wisconsin.