The first fully electronic derivatives exchange is one that almost no one remembers today. No one remembers INTEX because it opened for business over a quarter-century ago and because it was a total failure. No one keeps records on failed ventures, especially old ones. However, it is worth recounting the story of INTEX not only because it was first but also because its failure spotlights an important aspect of the shift from floor to screen—namely, that strong forces have shackled the U.S. exchanges to traditional floor-trading while the rest of the world was speeding ahead in the electronic trading race.

INTEX was born of a personal frustration with the trading floors of the 1970s and 1980s. A futures broker named Eugene Grummer had spent a career at Merrill Lynch and was tired of being abused and poorly treated by the floors of the various exchanges. The pit of a trading floor in the 1980s was not very transparent compared to the modern screen-based market. Say a customer wanted to buy July corn futures at the CBOT. She would know two things: She'd know the price of the last transaction, and if she were speaking directly to the floor, which only larger, more active customers could do, she'd know the last bid and offer. Traders were fond of saying that the bid or offer on a floor was good only as long as the breath was warm—in other words, about a second. So, usually the last bid and offer was not a price at which anyone could necessarily trade. Often the price at which an order was executed would seem to be a little worse than the most recent bid or offer.

And then there were the delays. In those days, when orders came in more rapidly than floor brokers could really handle, a “fast market” was declared by the pit committee and many of the obligations of executing brokers were temporarily waived. It could sometimes take two hours before a customer would know whether his order was executed and at what price. In an extreme case during the bull market in precious metals in 1979 and 1980, the main metals exchange, COMEX (which was later absorbed into NYMEX), was choking on paper. In some cases, customers did not get confirmations that their orders were filled for several days. There were clearly problems with floor-trading.

So Grummer decided that a transparent electronic marketplace would solve many of these problems he had faced for so long. He first needed money, so he raised venture capital from a Texas oilman named Wallace Sparkman. He needed an experienced person to run the company, so he hired as president Junius (Jay) Peake, who had been a partner at the securities firm Shields and Company. Grummer himself became chairman. INTEX was born in 1981, though it would be several years before the first trade took place.

Grummer and Peak wanted to start the exchange within the United States. That's where they resided and where most of their potential customers lived. So they went to see the federal derivatives regulator, the Commodity Futures Trading Commission (CFTC). The result was not good. They were told informally that an electronic system would likely take years to gain regulatory approval. It wasn't that the CFTC didn't like the idea of electronic trading; on the contrary, they loved it.

The CFTC staff knew that screen-based trading would allow them to know precisely when each trade was entered, received, and executed. They would be better able to see whether, for example, a broker was trading ahead of his customer. ³ It was very difficult to pin down the precise time that a pit trade occurred, since execution consists of a shout and a reply and the only record of the time of execution was the time the floor broker or trader wrote down on an order or trading card. Often the floor guys would execute a number of trades and then, when they had a second to do so, write them down, guessing a little at the precise time each trade took place. In a very busy market, considerable time could elapse between actual execution and the time at which the trade was written down. All this makes it very difficult for either the exchange or the regulator to know whether a broker traded ahead of his customer in a floor-based system. According to someone involved with INTEX at the time, the reason that the CFTC informally advised Grummer and Peak to go offshore was because the big floor-based exchanges in Chicago and New York would do everything they could to slow down the approval and the process would probably take years instead of months.

So, to get up and running more quickly, the INTEX leaders decided to establish the new exchange in Bermuda. Why Bermuda? INTEX had chosen the London Commodity Clearing House (LCCH) to clear the Exchange's trades, and LCCH already had a presence in Bermuda, so it would be relatively quick to get permission to both establish the Exchange and import the state-of-the-art computer equipment into the country. What was state of the art at the beginning of the PC revolution was Tandem computer equipment placed in Bermuda at the INTEX offices and then hardwired via transatlantic cable to DEC 10 computers at member desks in the United States.

Those who wanted to join INTEX had to pay about $36,000—$20,000 for the membership and another $16,000 for the computer. INTEX had 600 memberships available, and by the Fall 1984 launch date, some 285 of the 600 had been sold.

The launch date had been announced and missed a number of times during the period of INTEX's gestation. One participant recalls at least four official launch dates. The problem was the old one of “the perfect being the enemy of the good.” The exchange tried to incorporate all the suggestions given by potential participants. During the planning process there had been discussions of listing both gold and U.S. Treasury bonds (T-bonds) on day one, but the Exchange finally decided to put only one egg in the opening basket, and that egg was gold. The idea was to later add T-bonds, freight rate futures, and possibly silver to the contracts available for trading.

So finally, on October 25, 1984, INTEX successfully opened its electronic doors. Opening day was disappointing. The world's first electronic futures exchange saw only 142 gold contracts traded in an abbreviated four-hour session. ⁴ And things never got much better. INTEX was a failure.

Why did INTEX fail? It was partly the choice of contract. The conventional wisdom is that new contracts have the greatest chance of success in a bull market. They generate more of a buzz and more speculative interest. By the time INTEX launched gold in October 1984, the metal was in anything but a bull market. Gold, with a few reversals, had been in a long bear market since its $850 peak in January 1980 and was just under $340 that Thursday morning in October when INTEX opened its doors. The fact that it was headed down below $290 by March didn't help things.

Even more important, there were strong forces at work to ensure that INTEX would fail, no matter what product it chose. All the floor-based exchanges in Chicago and New York saw INTEX as a threat to their way of life. The several thousand individual members of the CME, CBOT, and COMEX knew how to make a living in a floor-based world. They did not want their exchanges to go electronic, and their elected leaders who ran the exchanges did what they could to maintain the status quo. So, when the big exchanges were given an opportunity to participate in INTEX, they all declined. But they went even further: They made it clear to the brokerage firms that they'd better not support INTEX either. If they did, things could get nasty. We don't know whether any threats were carried out, but we do know that firms were threatened with the loss of prime booth space on the trading floor (that is, near the pit) if they threw any business to INTEX.

So, what happened to INTEX? Four and a half years after the abortive launch of the INTEX exchange, the parent, INTEX Holdings, joined with Telerate, a supplier of financial data, to create an electronic trading system for the London International Financial Futures Exchange (LIFFE). ⁵ This screen-based system was called APT (for Automated Pit Trading) and became an after-hours electronic system, similar to GLOBEX at the CME, Project A at the CBOT, and Access at NYMEX.

So the first attempt at screen-traded derivatives failed. We now turn to the second attempt, which started in a rather surprising place and which, unlike INTEX, was at least a modest success.

It's clear why the first automated exchange was set up in Bermuda: because it would have taken forever to get approval in the United States. It was really a U.S. exchange run by U.S. citizens largely directed to U.S. customers but set up in a place where regulatory approval was possible. But why was the second one set up in New Zealand? The entire country of New Zealand has a population about half the size of Chicago. It has never been considered to be on the frontier of technology. But little New Zealand was absolutely a pioneer in the development of screen-based trading. It's almost like no one ever told them they couldn't do it.

It all started with seven wool traders who were spread around four wool-trading centers: Auckland, Wellington, Christchurch, and Napier. They had traded with each other over the telephone and had traded wool futures on the London Wool Terminal Market, a futures market for Australian and New Zealand wool. They decided it was time to set up a local futures market in New Zealand, a market that would reflect local New Zealand prices. But there was a problem: Because they were scattered around the country, they could not all be present for trading on a physical floor in any one of the four wool centers. So they decided that the exchange needed to be computerized.

First, they needed to join forces with some deeper pockets, so they enlisted the support of 10 financial institutions. Together the wool traders and financial institutions would own the exchange. Second, they established a relationship with the International Commodity Clearing House (ICCH), a subsidiary of the London Commodity Clearing House. ICCH became the new exchange's clearinghouse, but just as important, it helped them develop the new automated trading system, cleverly called ATS for, yes, automated trading system.

So, on January 20, 1985, the New Zealand Futures Exchange opened its computer screens for trading in three financial contracts: prime commercial paper, 5-year government bonds, and U.S. dollar/New Zealand dollar. The 10 financial institutions wanted to lead off with the financial products. Wool was added a few months later. Volumes were small; 200 contracts was a big trading day. However, due to the 1990 Gulf War, volumes picked up and the system slowed down. So, the system was modified to handle more volume and, more important, to accommodate the individual share options added in 1990. The exchange also added two more words to its name, becoming the New Zealand Futures and Options Exchange (NZFOE). The name no longer exists, since the 17 owners sold the exchange to the Sydney Futures Exchange (SFE) in 1991, which was merged with the Australian Stock Exchange in 2006. The SFE, incidentally, had developed its own electronic system, called SYCOM, which was created not as the primary trade matching system but only as a venue for trading once the pits had closed for the day.

So there are two answers to the question of why New Zealand was the number-two fully automated derivatives exchange in the world. First is that screens solved the logistical problem of traders spread around the country, and second, the exchange was brand new, so there was no painful transition from a floor.

The first electronic exchange, INTEX, failed. The second electronic exchange, NZFOE, survived. But we have to get to number three before we begin to see really spectacular success.

The Tokyo Grain Exchange (TGE) was an early innovator, moving its floor system onto screens on April 1, 1988. Though this was almost three years after OM and just a few months before SOFFEX, the TGE has the distinction of being the first floor-based exchange to shift to computer screens. All the exchanges that preceded the TGE into the world of screens—INTEX, New Zealand, OM, and SOFFEX—were new exchanges that were born electronic. The TGE chose to convert from the old method.

Why was the TGE the first exchange to convert from floor to screen? For one thing, it was easier to convert the traditional Japanese trading system to screen than it was to convert the traditional Western system. Here's the difference: In the Western continuous double auction, traders can bid and offer, buy and sell, at any time during a four- to seven-hour trading period, and there might be thousands of transactions at hundreds of different prices during the day, with real-time price feeds sending every bid, offer, and transaction out to a waiting public. In the Japanese single-price auction (known as Itayose or session trading), which has been widely used in futures on physical commodities in Japan, there are a small number of trading sessions during the day, in which all interested buyers and sellers are matched at a single price. For example, the Tokyo Grain Exchange has six daily sessions for trading corn futures—9:00, 10:00, and 11:00 a.m. and 1:00, 2:00, and 3:00 p.m. At each of these sessions, the exchange posts a provisional price that it feels represents a balance of potential buy and sell orders. At that price, all members post the number of contracts they and their customers are willing to buy and sell. If there are, for example, more contracts offered for sale than bid on, the provisional price posted is too high and is lowered, and participants again indicate how many contracts they are willing to buy and sell at the new lower price. The process is continued until the number of contracts offered equals the number of contracts bid on. All parties are matched at that price, and the price and quantity transacted for the session are disseminated to the marketplace.

So, beginning on that crisp spring day in Tokyo in 1988, computer terminals were placed directly on the trading floor so that instead of standing and indicating buy and sell orders to exchange representatives via hand signals, the members did the same thing via terminals. As time went on, brokers were allowed to have the terminals in their offices.

Though most Japanese commodities were traded in these single price auctions, in April 1991 the Tokyo Commodities Exchange (TOCOM) converted its precious metals—gold, silver, and platinum—to an electronic version of the continuous double-auction trading system used in the West. This was a more difficult conversion, but technology had advanced by three years since TGE's conversion. Gradually all the Japanese derivatives and securities exchanges converted to an electronic model. The process was completed on September 3, 2007, with the conversion of the final holdout, the Central Commodity Exchange of Japan.

Chinese futures markets were introduced in the early 1990s, first at the Zhengzhou Commodity Exchange (an overnight train trip west of Beijing), then followed quickly by exchanges in Shenzhen, Suzhou, Dalian, Shanghai, and many other places; by 1993 there were over 40 exchanges in China. But the amazing thing about this rapid proliferation of exchanges is that in a country of cheap labor, where labor-intensive floor exchanges would, at least superficially, seem to have made more sense, virtually all these exchanges were electronic.

So, how did this happen? Initially, in the first couple years of the decade, the exchanges traded cash and forward products and most activity was on a trading floor. At some exchanges, traders wandered around the floor and struck deals one by one. At others, the exchanges would actually hold auctions at which traders would bid. During this period, Zhengzhou sent its president and chief economist to visit exchanges in Europe, Asia, and North America. They came to the conclusion that open outcry was too open and public in that it didn't allow for sufficient anonymity for traders and that the screen system was more orderly. They also figured it was easier to find people who could program than to find people who understood how to make markets on a trading floor. They came back to China and developed a simple electronic trading system for standardized futures, and on May 23, 1993, the Zhengzhou Commodity Exchange, the first true futures exchange in China, traded its first futures contract on a screen. The rest of the exchanges quickly followed suit.

The exception was the Shenzhen Metals Exchange, which created a London Metals Exchange-style trading floor. But it quickly found itself bogged down in errors, misunderstandings, disputes, and defaults, and it too switched to a screen-based system. It concluded that floor-based trading was actually more expensive once you factored in all the problems. For both Shenzhen and all the other exchanges, they were started by and owned by the government in some form, often a municipal government. Even though the initial startup costs were greater for electronic trading, government entities didn't mind and were willing to front the money for these projects. Electronic trading, as even the regulators of the U.S. exchanges knew, was easier to monitor and regulate and would help prevent the cheating of customers.

The Chinese exchanges were electronic, with a twist. All the terminals were located on a trading floor, arranged in circles or in a big U or in two big blocks facing one another, depending on the shape of the trading floor. In a sense it was a hybrid system. Like the exchanges in Chicago, the traders came down to the floor to trade every day. But they sat at terminals on the floor rather than standing in pits waving their hands and yelling out what they wanted to do. In fact, they are still referred to as pit traders. Some, called self-trading members, trade their own accounts, but the majority, called agent-trading members, trade for clients. Order entry and matching are electronic, but entry is handled by these agent-trading members who receive orders over the phone or by fax or email. With all traders sitting in one room, they could get up and wander over to the desks of other traders to discuss a trade, a technical issue, or the latest joke. The arrangement preserved some of the social aspects of the floor. It also had everyone in one room, where they could be more easily monitored by exchange authorities. But the biggest driver for this setup was technical: in the early 1990s in China, it was much easier to connect a trade-matching system to a bunch of computers in one room than to computers spread throughout a building, a city, or a country. The Internet had not really come to China yet.

Both the exchange and the government liked having all the traders in a single room. Since futures—and for that matter, markets generally—were new to China, the exchanges wanted something visual that they could show visitors. This was especially true for potential customers such as grain merchants and others involved in the cash commodity markets. And, since government support was crucial for these new creations, a floor made it much easier to explain to government officials what was going on. Trading floors have always been more interesting than PowerPoint slides. Of course, the government liked having all the traders in one place so it could keep an eye on them. Finally, for purposes of general promotion, there is nothing like a trading floor for the media to take photos of and for the general public to visit, just as they would the Forbidden City, the Great Wall, or the Shaolin Temple.

That was in the beginning. Today upstairs brokers have direct access to the exchange and no longer need to go through the guys on the floor. The floor terminals at China's three existing exchanges have practically all been idled. About 70% of the orders are entered through terminals off the floor, though it is still used as a convenient, though less active, visual representation of the trading process for visitors from all over.