Social structure and reputation: the NASDAQ case study
Centre d'Economie de l'Université Paris Nord, Université Paris 13
Correspondence: Samira Guennif, Centre d'Economie de l'Université Paris Nord, Université Paris 13, 99 avenue Jean-Baptiste Clément, 93430 Villetaneuse, France. E-mail: guennif{at}seg.univ-paris13.fr Valérie Revest, E-mail: revest{at}seg.univ-paris13.fr
In 1996, two investigations conducted by the Securities and Exchange Commission and the American Department of Justice reported non-competitive practices among market makers on the NASDAQ. These reports also mentioned the influence of the NASDAQ social structure on market makers' behaviours. Most market makers adopted social norms in order to increase significantly their income at the expense of the customers. This paper aims to explain the rise and long-term effects of non-competitive practices through the integration of a concrete view of embeddedness. We propose the use of game theory tools to achieve this goal. A re-reading of Kreps's model of reputation sheds light on its structural dimension and illustrates the way social structure governs individual behaviours.
Key Words: NASDAQ • non-competitive behaviours • embeddedness • social structure • game theory • reputation • trust • JEL classification: A12, A14, B41, L14, G14, G15
1 Since 1985, several authors have proposed a view of embeddedness and described the way social structures influence individuals' behaviours in ways neglected by standard economy theory (Zuckin and Dimaggio, 1990; Uzzi, 1996; Gulati and al., 2000).
2 In a dealer's market, the market makers or dealers offer bid and ask prices continuously. The alternative organization is an order-driven market where the prices are determined by the confrontation between the buyers' orders and the sellers' orders.
3 Although the NASD is a self-regulation organization, it is still under the control of the SEC.
4 The Department of Justice conducted numerous telephone and in-person interviews of current and former NASDAQ stock traders and investors and listened to approximately 4500 hours of audiotapes of telephone calls between stock traders employed by the defendants and other market makers. It also analysed data for all market makers quote changes on NASDAQ during a 20-month period between December 1993 and July 1995.
5 Recently, US equity markets have moved to decimal quotations.
6 Similarly, Breton and Wintrobe (1982) claim that trust results from belief formulations in repeated transactions. The perspective of future gain motivates ex post good behaviour and legitimates ex ante trust between players (Meidinger and al., 1999).
7 This is a critical point since actors are unable to draft contingent contracts; they face difficulties in measuring individual performance, in observing compliance (Alchian and Demsetz, 1972; Hart, 1996) and in predicting. Finally, bounded rationality impedes the drafting of complete contracts (Simon, 1978).
8 In other words, to trust B, A2 must have information about B's reputation (his past behaviour), and A1‘s reputation (his capacity to give good information about his partners’ behaviour).
9 If the transaction ends up with problems between A1 and B, A1 can declare that he has been cheated by an opportunist. In the same circumstances, B can declare he has fulfilled his obligations, that A1 was unlucky. Therefore, the question of the credit A2 gives to A1 arises. A1 may declare that he has been cheated to hide his undue requirements. Then B may refuse to commit his reputation and may renounce trade with such individuals (Lorenz, 1988).
10 Given their pertinence to standard economic theory and its under-socialized view of economic behaviour, Granovetter argues that the reputation models proposed by game theorists are useless for capturing the essence and effective influence of embeddedness (1985, p. 490).
11 In the same vein, Kandori describes ‘community enforcement’: ‘dishonest behaviour against one partner causes sanction by other members in the society’ (1992: 63). This ‘community enforcement’ also prevails for peers who do not sanction the defector (64). Structural enforcement is also defined as ‘community governance’ or ‘network governance’ (Bowles and Gintis, 2002; Jones and al., 1997).
12 Analysing industrial districts, Saglio reveals how individuals observe a set of social rules that contains economic rules (1991, p. 531). Studying industrial vertical disintegration in Japan in the 1960s and 1970s, Dore notes that subcontracting relations involve both ‘benevolence and self-interest’ (1983, p. 170). Hirschman insists on ‘sweet trade’. According to him, trade is communication based on closed contacts among people who make promises, trust each other, claim and complain (1984, p. 69). Williamson admits that social norms such as ostracism sustain transactions. They limit self-interest, ‘individual aggressiveness’ (1975, pp. 106–7).
13 The balance between egoist and non-egoist motivations is also presented as a particular trade-off between competitive and co-operative behaviours within industrial districts, networks or communities (Piore and Sabel, 1984; Bernstein, 1992; Gulati and al., 1999).
14 Some sociological studies on financial markets shed light on the role of the network and of local norms (Baker, 1984; Hassoun, 2000; Mackenzie and Milo, 2003).
15 A market maker admitted that in the absence of the price convention and in certain instances, he would have used odd-eighths quote increments (US Securities and Exchange Commission, 1996).
16 Public limit orders are posted by the investors in an order-driven auction market and specify a price and a quantity.
17 This kind of procedure is facilitated, ceteris paribus, by the specific NASDAQ organization of exchanges (Revest, 2001). For example, unlike the NYSE, the NASDAQ and other dealer markets do not have time priority rules requiring market orders to be executed with the entity who first posted the current best quote (Grossman et al., 1997).