his paper is concerned with the role played by the size of an interest group in the production of collective political pressure, and with its success in the redistribution of political income. Most of the regulation literature sustains the hypothesis that the optimal size of an interest group is somewhat small. This support can be traced back as early as 1908 when The Process of Government, seminal book written by Arthur Bentley, was first published. Arthur Bentley characterizes the government as a process, in which interest groups are the protagonists. This process is the activity of the groups in their relation with one another. Groups are in constant activity, pressing one another, cooperating, competing, forming offensive and defensive alliances, splitting apart, and disappearing, while new groups are being formed. Strong groups dominate, and delineate the existing state of society; state that under this framework has to be appraised as an equilibrium, given that it is the end result of the pressure exerted by a multiplicity of interest groups. Bentley argues that usually, in the political game, the successful groups are not majorities but minorities, "First of all, the number of men who belong to the group attracts attention. Number alone may secure dominance. Such is the case in the ordinary American election, assuming corruption and intimidation to be present in such small proportions that they do not affect the result. But numbers notoriously do not decide elections in the former slaves states of the South. There is a concentration of interest on political lines which often, and indeed one may say usually, enables a minority to rule a majority." (Arthur Bentley, 1908, p. 215) Since the early seventies some economists (i.e., George Stigler, 1971; Sam Peltzman, 1976; Gary Becker, 1983) have followed Bentley's insights, beginning the challenging work of building a general theory able to provide conditions that favor the existence of successful interest groups. All of them agree with Bentley on the advantages of a small group in the political competition. This advantage is founded, for example, on the free rider problem faced by large groups (Mancur Olson, 1965). The following quotation illustrates Olson's argument: "It has often been taken for granted that if everyone in a group of individuals or firms had some interest in common, then there would be a tendency for the group to seek to further this interest... If we ponder the logic of the familiar assumption described in the preceding paragraph, we can see that it is fundamentally and indisputably faulty... If the consumer or worker contributes a few days and a few dollars to organize a boycott or a union, he or she will have sacrificed time and money. What will the sacrifice obtain? The individual will at best succeed in advancing the cause to a small (often imperceptible) degree. In any case he will get only a minute share of the gain from his action. The very fact that the objective or interest is common to or shared by the group entails that the gain from any sacrifice an individual makes to serve this common purpose is shared with everyone in the group... Since any gain goes to everyone in the group, those who contribute nothing to the effort will get just as much as those who made a contribution. It pays to - let George do it -, but George has little or no incentive to do anything in the group interest either, so... there will be little, if any, group action. The paradox, then, is that... large groups, at least if they are composed of rational individuals, will not act in their group interest." (Mancur Olson, 1982, pp. 17-18) While most of the authors acknowledge that small groups have advantages in the political game, not all the scholars agree. For example, Kaveh Mirani (1984), argues that the 3 optimal size of a group may be somewhat large if the type of political pressure exerted by that group is violent. Mirani claims that violent political pressure has a distinctive characteristic: the fact that there exists a nonzero probability of being apprehended and punished. Since this probability decreases with an increase in the size of the group, he argues that for violent political pressure the optimal size of a group will be larger than for nonviolent political pressure. The following quotation summarizes Mirani's hypothesis: "The theoretical model suggests that the existence of substantial economies in size which arise out of the reduced risk of apprehension and punishment when the size of the group increases implies that the optimum size of a revolutionary group is relatively large. This is in contrast with the conclusion reached by the existing theories of collective action and interest group competition which associates political activity with small group size." (Kaveh Mirani, 1984, p. 95) This paper will postulate the hypothesis that even when it is considered the possibility to exert violent political pressure the conclusion reached by the existing theories of collective action that associates political activity with small group size may still remain valid. The organization of the paper is the following: section 2 describes the small group argument and introduces Mirani's hypothesis; section 3 proposes a simple model from where the usual small group argument may be derived even when it is considered the possibility to exert violent political pressure, and provides some empirical evidence in order to illustrate its plausibility; section 4 concludes. II. M