Keywords: Market, Organization, Dealing room, Trading room, Coordination, Ideal type.

In this paper, we aim at studying market-like arrangements as they are implemented inside an organization. This shall be done though the exploration of organizations operating on international financial markets: trading rooms. Compared to 'classical' industrial settings, trading rooms are both different and similar. Different because the main issue is decision-making rather than the production of material goods. Similar because they both incorporate more and more market-like practices, like subcontracting, profit centers, bonuses and internal contracting.

The first trading rooms (also known as 'dealing rooms') appeared in the US banks of the early seventies. The purpose of a trading room is to regroup a series of formerly-separated departments into a single unit placed under a single supervisor. Trading rooms gradually spread to other countries as they adopted reforms deregulating their foreign exchange and financial markets. In 1997, 4500 dealing rooms were in operation worldwide, with approximately one hundred thousand operators. Trading room organization is very standardized worldwide: all operators ('traders'), gathered in a large room ('floor'), are arranged in semi-autonomous workgroups ('desks').

In considering the environment of trading rooms - i.e. financial markets - three foremost points need to be made:

• The importance of investments made. This environment is very competitive and the investments, whether considered in terms of technology or human resources, are very important. The difference between a profit and a loss can be made through an advantage of seconds.
• The fact that a zero-error functioning is uttermost. A mistake (making a deal with a counterpart close just going bankrupt), arranged deals (a way to steal money from the bank) or processing error (a miscalculation) can all be extremely costly.
• The strong coupling of activities and decisions. Activities and decisions made by the traders cannot be considered in isolation, the behavior of all financial assets being correlated. There is a strong mutual interdependence in the profit produced by deals made within the dealing room.

Thus, despite their potential disastrous effect, dozens of deals are made every minute. The trading room of a big financial company often hosts more than a hundred operators, each of them buying and selling over a hundred million Euros daily.

The functioning of a trading room is at once very fast and very complex. Our purpose is to better understand it. More precisely, we are interested in finding whether and how the concept of 'market' is instrumental in understanding trading rooms (Q1). Going further, we want to highlight the achievements and limits of market-like arrangements (Q2). We also aim to bring a contribution to the understanding of other forms of organizations, mainly in services and the design of industrial products.

However, our problem is not only to gain access to trading rooms and to understand who does what, but also to characterize what we mean by 'market'. This concept is so rich that its meaning needs to be accurately defined. Classical economy defines mathematically what a market is, but mathematics are of little help in understanding human relations. Thus getting to define the market as a concrete organizational structure is a prerequisite. We shall deal with this from the next section on.

As for gaining access to dealing rooms, it has been accomplished through field research. Three years of participant observation (starting 1994) as a computer engineer were followed by a series of forty-five interviews in trading rooms of Paris, London and Hong Kong. Choosing field research was motivated by the difficulty of 'getting in' as well as the need to acculturate to the very specific culture of trading rooms and establish a network of relations with traders worldwide (mainly through the "Association Cambiste Internationale").

We shall initiate the definition of the market through an exploration of Weber's work, starting from the theories he developed on "the market community" and the concept of ideal type.

Of all early twentieth century sociologists, Weber is the one that had the most prominent interest for the market. After he passed his University degree in law, he took part in a study on the mechanisms of speculation and an international comparative approach of stocks markets. Weber claimed sociology must include a branch dedicated to the "sociology of the market", because this phenomenon could not be reduced to the generation of prices.

In "Wirtschaft und Gesellschaft", he asserts that social interactions on a market develop in two phases: competition and exchange.

Competition: "A market exists when there is competition (…) to secure exchange opportunities". Competition is merciless: "(the market) has consideration for things only, not for human beings nor for doing one's duty by fraternity or piety". Because of this continual competition, the boundaries of the market are blurred. They can be extended at all times in order to include more potential partners: "partners are guided in their offers by the potential action of a group of competitors of an unknown size rather than by their own action". However, if Weber agrees that if there his an antagonistic relation, the conflict remains pacific because it is motivated by the desire to precede another in the realization of an exchange. Thus conflict remains indirect and is not aimed against a single opponent. Instead of destroying an opponent, one tries to be better than he is. Weber also notes that exchange on a market is very special as it is the most rational interaction possible between human beings "the archetype of rational action" and in consequence "an abomination for any system of brotherly ethics".

Exchange: the next phase is different: the deal is made and the transaction executed. The deal is a compromise by which counterparts barter things or trade goods against money in reciprocal compensation. Although they are the central purpose of the market, exchanges are reduced to a minimum, as well as their social side. Once completed, they are forgotten "market shows a simultaneity and succession of rational sociations each of these being so ephemeral that it dies with the exchange it was initiated for". After a deal is made, all sides are free of mutual engagement.

Weber's idea of the market we just described is not an image of the market that can be observed in the 'real world', it is an ideal type.

Weber claims no concept can represent the whole of reality, and suggest to resort to the ideal type as a conceptual tool to understand complex social phenomena. Weber himself invented many ideal types. In this way, he designed the "Bureaucracy" as a mean to study the impact of rationality on organizations. Another ideal type, the "spirit of capitalism" was correlated with the rise of Protestantism.

How to construct an ideal type? Weber compares the building of an ideal type to the work of a caricaturist. The caricaturist, guided by his instinct, produces a drawing which highlights a series of features of the person he represents. The scientist designing an ideal type does just the same. However, features of an ideal type must be mutually articulated and form what Weber calls "an homogenous chart of thought". From the point of view of science, what is at stake is to put together a logical construct, even if the building of the construct itself demands some intuition. Thus, the ideal type cannot be created outside a subjective point of view, but at the same time, it goes beyond subjectivity because it is rational.

When he tries to further define the status of the ideal type, Weber chooses to tell what it is not.

• The ideal type is not reality. The scientist must not try to 'give a name' to the things he sees but only to articulate some of their characteristics.
• An ideal type is not the product of an investigation. It cannot be exhaustive since some features of reality are rejected. There is no method to select these features. The ideal type is "not defined by features common to individuals nor by average characteristics". The ideal type is a utopia.

Where then lies the interest of ideals types? In their heuristic value. The ideal type is not an hypothesis, but something to help formulating hypotheses. It does not describe reality, but only to guide its exploration and put in light emerging phenomena.

Using the concept of ideal type as it has been created by Max Weber leads us to propose an ideal type of the market, defined by three groups of features:

1. the importance of explicit and quantitative exchanges
2. the omnipresence of concurrence
3. the emphasis on autonomy and accountability.

Exchange of goods or services is the reason why a market exists. The division of work and specialization it allows is essential to the development of society. This is described by Adam Smith as a domination of 'exchange value' over 'use value' and leads to a weakening of traditional social relations. We have shown the primary importance of this to Weber, but most sociologists of this period support this idea. Marx describe how bourgeoisie "eradicated the complex ties binding feudal man to his natural superiors, (…) to leave no other bond between man and man than cold interest (…) (Bourgeoisie) drowned the sacred thrills of religious ecstasy, knightly enthusiasm and cheap sentimentality into the freezing waters of selfish calculation". Thus market stimulates coordination and exchange, but at the expense of social integration. Parties aspire to a reduction of their mutual dependence, not to a lasting association. All sociologists of modernity refer to this "great transformation" of social relations.

This explicitness of the "give and take" is supported by the expression of quantities. Market relations are bound to the expression of amounts expressing the reciprocal flows exchanged: prices. However, quantification is not restricted to the expression of prices. Other amounts, like profits, also play a major role.

Use of prices is heralded by Hayek who sees them as means to share and broadcast information. Although human communication is imperfect, prices can convey information accurately. Although human knowledge is incomplete, prices can summarize, coordinate and optimize decision.

The presence of exchanges and prices is not enough to characterize a market. These exchanges must take place in a situation of concurrence. As we remarked before, social interactions on a market develop in two phases: competition and exchange. Then the cycle starts again with a phase of competition.

This is why understanding the market cannot be reduced to analyzing a series of bilateral relations, but must encompass the existence of other bidders and buyers. All parties need not be gathered on a single location. As Weber notes, the "physical gathering in a given place, whether the market square, the fair or the stock exchange" is not necessary. The boundaries of the system that need to be taken into account are greatly widened, and one of the sources of uncertainty comes from the fact that new competitors can 'step in' at all times.

Rules applicable to all parties must be elaborated and enforced. These rules must at once be minimal to give parties a maximum of autonomy and be sufficient to avoid cheating. These rules themselves must be global (apply to every actor of the market), minimal (thus less costly to enforce) and non-negotiable (they cannot change all the time). As Coase remarks, setting up an arrangement for the elaboration and enforcement of this rules amounts to a cost.

The main activity of dealing rooms operators is to strike deals. All these deals are formalized by oral contracts. Most of the time they consist in agreeing on the amount exchanged and a price. Traders often calculate continuously what profit or loss they make. Operators from the same dealing rooms make deals with one another but they can often choose to strike the deal with a counterpart not belonging to the same firm. These transactions have all the characteristics of market exchanges. Sophisticated information and communication technologies put all parties in a situation of continual concurrence with the rest of the world.

Traders or desks managers are accountable for the profit or loss made: when there is a profit they receive their share of it. Bonuses that can make for a very important part of the total salary. But when losses are made, they are at a risk to lose their job.

Thus, competition can be very hard, even between traders from the same company. These rivalries are not shunned by the trading room manager who does not always consider them unhealthy.

Traders are given a lot of freedom and delegation. The decisions they take often involve millions of Euros. They are also enabled to buy and sell different types of financial products in order to fund or hedge their position. They can even try to anticipate price variations and speculate. With this growing autonomy, each operator and each desk becomes a profit center in itself: "we have a budget for everything like a small firm: material, telecommunications, capital" a quota for overheads like the administrative processing of transactions and computer maintenance is added. Therefore traders often hold the decision over computer engineers when buying computer software, material or feeds.

Before trading rooms appeared, banks featured a series of departments, each one specialized in the trading of one category of financial products. Considering the ideal type of the market, the grouping of all these departments under a single supervisor does not only bring economies of scale. It also allows for:

• Having all desks abide by the same (global) rule,
• A better communication of prices, and consequently a stronger concurrence

This rule are also minimal in that they set objectives and limits to individuals or to teams but do not specify anything in terms of choices to make. Theses decisions are left for traders to take … and retain responsibility for their consequences.

Comparing the ideal type of the market with the situation inside dealing rooms can bring up a better understanding of their functioning, although some features or situations do not match with our ideal type. They mainly fall in two categories:

1. As we have shown, the market can only exist within a frame that can produce rules but also control and enforce their application. Our ideal type cannot help in determining how this is done.
2. Situations of much richer communication than through quantities are frequent. Situations of amicable arrangement or acts that could seem gratuitous occur. Giving things or services without an immediate counterpart is not in the logic of the market.

We shall revert to these categories later.

Two categories of achievements and two categories of limits can be singled out:

Thus, coordination through the action of the famous 'invisible hand' of the market is implemented. In this way, all traders of the dealing room are coordinated, not to embrace a given strategy, but to make profits. The dealing room itself is very versatile and can adapt quickly to crises and changes in its environment.

Trading on a financial market is very risky and some theories claim it is impossible to forecast future prices. The setting up of a market seems to solve this problem: although trading rooms sometime account for big losses, they are on average a source of profit. Of course, a trading room do not earn money through speculation only, but also by acting as an intermediary for the bank's customers or through other means.

For young and talented individuals, working in a trading room is an attractive job. They can have more responsibilities and get better paid than in most other positions. However, working in a market-like organization is very demanding. The workload is sometimes enormous, stress stands high and the dismissal can come very fast.

Employee turnover is very high and can have harmful side effects:

• The relationship between the employee and employer is restricted to the very short term, each trying to get the most out of the other. This is a source of lasting uneasy feelings.
• It disrupts the advent of team dynamics and prevents the organization from working smoothly. When you work on putting in place a specific software program and it turns out the trader you made it for changes job, the investment is lost.

High bonuses can also lead to certain problems, as profits can sometimes be obtained at the expense of their colleagues.

• People sometimes express discomfort and uncertainty about their identity "do I work for myself, for the company, for a product line?".
• High individual bonuses make people work hard, but sometimes become very individualistic too. "People with high bonuses work hard and are aggressive, but are bad when teamwork is required".
• People only care for profit in the short term, without consideration for the future. This prominent source of risk will be mentioned again later.

Finally, as market permanently keep people 'on their toes', it tends to 'burn them out': "a trader is like a high-level sportsman, he can get excellent results for a time, but he wears off quickly". Most traders cannot stay in this job past forty

In principle, a market as a whole is not exposed to risk. Exchanges allow each individual to take some risks, and he has to manage it by himself. If he fails and goes bankrupt others individuals are not affected since the parties are independent. They simply go on trading with other counterparts.

Unfortunately, there are limits to this approach of risk-management:

• As traders get their share of profits made in the short term, they tend to take excessive risks to get a quick return for themselves.
• In the case of dealing rooms, bonuses work when there is profit to share, but not when there is a loss to take: it doesn't matter for the trader whether the loss is big or very big. In other words, bonuses allow the sharing of profits, but not of losses.
• The way bonuses are calculated is sometimes complicated or dependent on a good administrative processing of deals. Failure of this control system is frequent because the ideal type of the market demands a maximum of freedom and consequently hinders control.

Most disasters that occurred lately can be charged on this inability of the market to manage risks. The bankruptcy of the Barings Banks is a prominent example of this. It is spectacular to see how a single man can lead to the bankruptcy of a whole firm.

Nowadays, market-like structures are more and more adopted by firms. The first organizations to be concerned are those whose main task is to manage information. In this way, trading rooms can be paralleled with control towers in air traffic control, integrated engineering floors in the automotive industry or telephone help line desks in insurance companies. More: market arrangements are implemented in situations were they were thought to be impracticable. In manufacturing, internal contracting is becoming widespread and new factories are implemented in which independent firms manage different sections of a single assembly line.

In this paper, we try to characterize what a market can be by building an ideal type and testing his relevance in understanding the functioning of dealing rooms. We have shown how market helps setting up a form of organization which both versatile and able to face a very complex and turbulent environment. We have highlighted some weaknesses: mainly the inability to manage important risks and the stress it puts people under.

A study of the empirical organization also shows that, although the ideal type of the market is relevant, the organization of trading rooms cannot be understood without resorting to other concepts. As a suggestion for further study, we suggest adding two other ideal types, which would be called "hierarchy" and "network". Testing their relevance and mutual interaction could be done on the same field: trading rooms.

ADLER P. ; ADLER P.A. 84, "The social dynamics of financial markets", JAI Press

BACHELET R. 97, "Organisation et gestion des risques en salle des marchés financiers : appareil, marché, réseau", doctorat en sciences de gestion, Université Paris-IX Dauphine, décembre 97.

BACHELET R. 97, "Organisation et gestion des risques en salle des marchés", actes du Congrès de l'AFC et de l'IAAER, Paris, octobre 97.

BACHELET R. 98, " Les salles des marchés financiers : décision et organisation", Humanisme et Entreprise, n° 228, avril 1998.

BAKER W.E., 84, "The social structure of a national securities market", American Journal of Sociology, vol 89, n°4.

CAILLé a. 83, "Comment on écrit l'histoire du marché" in CNDP 93, "L'institutionnalisation du marché", DEES n°92 juin pp. 11-34 from La revue du MAUSS

COASE R. 1937, "The nature of the firm", trad. française - Revue française d'économie, vol II, n° 1, 88

CORIAT B. ; WEINSTEIN O. 95, "Les nouvelles théories de l'entreprise", Livre de poche

DUMONT L. 77, "Homo æqualis", Gallimard in CNDP 93, "L'institutionnalisation du marché", DEES n°92 juin

Duval Guillaume 98, "L'entreprise efficace à l'heure de Swatch et de Mac Donald", coll Alternatives économiques, ed. Syros.

FAMA E. 1965, "The behavior of stock market Prices", Journal of Business, 38(1), jan 65, pp. 34-105

FRYDMAN R. 89, "Marché (économie de)", Encyclopédia Universalis, Corpus Vol 14, pp. 494-499

GRANOVETTER M. 94, "Les institutions économiques comme constructions sociales : un cadre d'analyse", in Orléan A., "Analyse économique des conventions", PUF

HAYEK (Von) F.A. 45, "The use of knowledge in society", American Economic Review, (1945), 35, pp. 519-530

MENARD C. 90, "l'économie des organisations", éd. La découverte, 2^ème éd 1993

POLANYI K. 44, "La grande transformation", Gallimard 83

POWELL W.W. 90, "Neither market nor hierarchy : network forms of organization", Research in organizational behavior vol 12, pp. 295-336

SIMON Y. ed. 97, "Encyclopédie des marchés financiers", Economica.

SMITH A. 1776, "An inquiry into the nature and causes of the wealth of nations", Cannan éd. 1976

SWEDBERG R 94, "Markets as social structures", in SMELSER N. ; SWEDBERG R., "The handbook of economic sociology", Princeton University Press

WEBER, M 1922, "Essai sur la théorie de la science", presses Pocket 92

WEBER M. 1922, "Économie et société" trad. Julien Freund et al., Tome 2, 1995 éd. Pocket, chap. VI, pp. 410-416.