Principal, Xamax Consultancy Pty Ltd, Canberra
Visiting Fellow, Department of Computer Science, Australian National University
Version of March 1994
© Xamax Consultancy Pty Ltd, 1997
This document is at http://www.anu.edu.au/people/Roger.Clarke/EC/PaperOLTCBOT.html
I paid a short visit to the Chicago Board of Trade (CBOT), on Friday 25 March 1994. My purposes were to learn about the nature of the trading operations there, and the way in which information technology (IT) is used to support them. I acquired a considerable amount of documentation from the Public Relations Department which runs the Visitors' Area on the 5th Floor, and from the Bookshop on the Ground Floor, including a copy of the 'Commodity Trading Manual' (350 pp.).
These Notes are written to summarise my impressions. They are based on my experience a dozen years ago at the International Stock Exchange in London, supplemented by my studies of livestock exchanges in Australia and Singapore.
CBOT was established in 1848, and is the world's oldest and largest futures exchange. Its primary business is in commodity futures, with particular reference to the grains grown in the Mid-West of North America. On this base have grown futures trading in other commodities (including gold, silver, fertilizer, and most recently structural timber panels), derivatives trading (futures options), futures trading in financial securities (since 1975), and options trading in a range of commodities and securities (since 1982). It has 3,500 members and associates, and directly employs 700 staff.
Two further ventures have been entered into. The century-old MidAm Exchange has been taken over by CBOT. It offers trading in smaller lots, of agricultural commodities, precious metals, financial securities, foreign exchange and a limited range of options. Several years ago, CBOT entered into a strategic joint venture with Reuters and the Chicago Mercantile Exchange (CME) to develop the Globex international electronic exchange for futures and options trading, which operates outside Chicago trading hours. It commenced running in mid-1992, and was then installed in members' offices in Chicago, New York, London and Paris. To date, it appears not to have been an altogether successful investment. Apart from an old brochure, little information was available from the two sources approached.
Chicago is also home to the Chicago Mercantile Exchange (CME). CME's operations appear to be largely complementary to CBOT's in terms of agricultural commodities and foreign currencies futures trading; partly competitive in financial securities futures and options trading; different in terms of its strategic partnership with the Singapore International Monetary Exchange; and related, at least for the time being, in the form of strategic partnership in the Globex international electronic exchange.
CBOT trades on two floors of a building in the Chicago Central Business District, using open outcry. Open outcry seems like an anachronism in the information age, but it has survived partly because of tradition and volume, but also because it is a highly refined and specialised arrangement, which works.
The floors are segmented into different areas, 'pits' or 'rings'. Each specific futures or options contract is traded only in one pit. The pits comprise steps, and each step corresponds to a particular month of delivery for that particular contract. Traders stand on the appropriate step, indicate (by hand-signal supplemented by voice) whether they want to buy or sell, visually locate traders willing to take the other side of a bargain, exchange confirmatory hand-signals, and record the completed bargain on their cards. CBOT employees sit in 'pulpits' over each pit, recording changes in price for each security.
Each trade takes only a few seconds. The simplicity and speed are possible because the data flows required are quite limited, and the communications have been formalised into simple visual and spoken protocols which are functional despite the surrounding pandemonium. Of the thirteen items of information involved in a trade, nine are provided by the context of the trading activity, and only four (whether selling or buying, the number of lots, the number of ticks and acceptance) need to be communicated between the traders. See the Attachment.
The underlying purpose of the CBOT marketplace is to enable commodity producers to sell their produce in advance of delivering them. In this way they are able to 'hedge', i.e. ensure a minimum price which they will receive, and hence secure financing from their bank, or just have peace of mind. This 'price-setting' function is performed by providing a meeting-place for producers and the ultimate purchasers of the commodities, e.g. in the case of grain, millers.
To serve the interests of producers, it is necessary to stimulate an active market in all commodities and all months of delivery. Hence, rather than merely providing a meeting-place for producers and ultimate buyers, CBOT (like all exchanges) actively encourages the participation of gamblers, who seek to buy and sell futures purely to take a margin on the rate. The participation of speculators increases the likelihood that a sale can be made, i.e. that that a current market price exists. It also injects into the market an additional party willing to accept risk in return for an expected margin. Relatively risk-averse producers are complemented by specialists whose livelihood is made by managing risk.
In stock and share trading, traders are expected to only sell securities which they already possess, i.e. 'short-selling' is generally prohibited. In futures trading there is no such limitation, and hence speculators can enter the market as buyers or as sellers. Moreover, CBOT permits its members to trade on their own account. This is generally banned in stock and share exchanges on the grounds that there is a conflict of interest between the trader as agent for clients, and as principal. Presumably members must submit their own orders through the same channels as those of their clients, so that there is a date-and-time-stamp control over the sequence of receipt, together with rigidly-enforced FIFO handling processing of orders. In addition, unusually for exchanges, it permits some speculators to participate directly on the floor.
There is a considerable amount of IT support involved in CBOT, in such ways as the early capture of trades, the display of prices at each pit and on the wall, the publication of price and trading statistics (all of which are continual during the trading day), the examination of patterns of trading to detect possible breaches of CBOT rules, and the feeding of external information relevant to trading (e.g. bad weather, politics, strikes, transport disasters) to displays accessible to the traders). There is, however, no support for the trading operations themselves.
This is set to change. Hand-held machines have been trialled, and a design agreed upon. I understand that they are intended for implementation throughout the two floors in the third quarter of 1994. It appears that very limited details of the application are publicly available at present, viz. 200 words on the inside back-cover of the booklet 'Action in the Marketplace' (version EM42-2, 9.92.17000).
According to contemporary strategic IS theory, IT can be applied to support any of the key sources of strategic advantage: cost leadership, focus, differentiation or alliance.
There would appear to be considerable scope for cost-savings in CBOT's present operations. They are extremely labour-intensive, involving not only the traders, but also order-takers at the edges of the floor, runners between the order-takers and the traders, data-capture operators to capture both sides of the deal from traders' cards, and price-recorders in the pulpits beside each pit. The two sides of the deal require matching, in batch-mode overnight or perhaps progressively during the trading day. With the existing arrangements, it would be reasonable to expect substantial numbers of mis-matches, which cost time, effort and nervous energy to fix.
Capture on hand-held devices at the point in place and time that the trade is made would at least remove the need for data-capture. To dispense with the runners would require additional features, in particular connectionless (i.e. RF) transmission between the hand-helds and the system. [The only written information available states that "order fill information will be entered into the Broker Workstation", which implies that the runner brings the hand-held (or a card from the hand-held) back to the order-taker. However I thought I understood the Public Relations Manager to say that runners would be eliminated by the new system].
A completely on-line trade recording system would involve electronic transmission of orders from the order-taker to the trader [which the booklet implies will be the case], and of deals back to the order-taker. It would seem more effective to transmit to the CBOT system, and have CBOT make the data available (immediately) to the order-taker. This is all the more likely to be the approach adopted because CBOT (or, more accurately, a separate corporation BOTCC) is used to contractually separate buyer from seller, i.e. all contracts for sale and purchase are legally effected with BOTCC rather than with the other trader. The reason for this is to reduce the number of legal relationships that exist from the potential 1,400 x 1,400 if every member dealt with every other member to a mere 1,400 between each member and BOTCC.
Using such a scheme, matching could be performed much more quickly than with the existing, manual system. Conceivably it could be undertaken in something close to real-time, because when the slower of the two traders transmits the data, it should be able to find an immediate match. The frequency of errors may not be reduced all that significantly, but they would be identifiable much earlier, and should as a result be fixed at the end of the trading session when memories are fresh, rather than the following morning.
There are problems to be overcome, however. The hand-held must make the data on new orders readily available to the trader, and displays on hand-helds are usually quite limited in size and readability. There is the question of fall-back in the event of a hand-held being dropped and broken, the (8-hour) battery failing, or the hand-held device, the broker workstation, the communications linkages or the main processor malfunctioning. This would presumably be addressed by way of a pool of spare devices and workstations, a fault-tolerant processor configuration (i.e. Tandem or Stratus), and considerable investment in quality assurance for the application software and the system integration process. It is vital that all parties feel confident in a new technology-supported scheme, because there is so much at stake.
Beyond cost reduction, the use of IT to contribute to other aspects of CBOT and member strategy requires careful consideration. From the viewpoint of the traders, automation of trading would be a collaborative arrangement, although individual traders may be able to leverage on it through additional investment in their own, internal systems, and hence offer their clients improved or additional services.
From the viewpoint of CBOT, it is not clear that its 9am-1:15pm services are subject to any direct competition, and hence assault on market-share or defence against a competitor's assault do not seem to be relevant. The afternoon session is designed to coincide with the opening hours in Japan, and IT may be able to make some contributions to an alliance in that marketplace. It may also enable expansion to other time-zones and hence markets, in particular in Europe; or into other forms of securities, such as foreign exchange, and stocks and shares.
A more revolutionary approach, consistent with the information age, would be to abandon floor-based trading, as stock exchanges such as NASDAQ, London and Australia have largely done. Traders could sit in their own offices, orders could be displayed on their screens, and trades could be effected over the network. The existing, very economical trading dialect could be readily replicated in screen- and roller-ball-based form.
One advantage of this approach would be the transfer of the real estate costs from CBOT to its members. More importantly, the articifical restriction to set floor-trading hours could be overcome, enabling CBOT to market its services to the world, in all time-zones. There may be a conflict here, between the aims of CBOT as a corporation, and those of its shareholders/members.
It appears, however, that the intended vehicle for more revolutionary application of IT to commodities trading, the strategic alliance with CME and Reuters, may not have been as successful as was anticipated. A longitudinal case study of Globex would be very informative.
The following paragraphs list the data-items involved in effecting a trade, and how they are economically communicated by visual cues alone and/or voice:
For example, a deal may be initiated by a buyer extending the hands above the head, palms inwards, with two fingers extended. On meeting the eyes of another trader who is indicating 'sell two', the buyer may move the arm to the horizontal, with one finger extended, indicating one tick above the market price. If the other trader indicates assent (using any mutually acceptable signal, such as a nod), the deal has been done, and both traders write the details on their cards.
In practice, most trading appears to be done using the visual cues alone. Nominally, however, the signs are meant to supplement rather than replace voice. There is an agreed dialect which, for the above example, would involve the buyer calling price then quantity. If the trade is initiated by the seller, the call is quantity, then price.
Go to Roger's Home Page.
Go to the contents-page for this segment.
Send an email to Roger
Created: 14 October 1995
Last Amended: 21 November 1997
|These community service pages are a joint offering of the Australian National University (which provides the infrastructure), and Roger Clarke (who provides the content).|
| The Australian National University|
Visiting Fellow, Faculty of
Engineering and Information Technology,
Information Sciences Building Room 211
|Xamax Consultancy Pty
Ltd, ACN: 002 360 456|
78 Sidaway St
Chapman ACT 2611 AUSTRALIA
Tel: +61 6 288 6916 Fax: +61 6 288 1472