to extract what in effect are rents from newer ones is, however, limited by the dependence of these new members. The less their dependence, the less extensive the obligations they will incur.
In summary, then, the greater the dependence of members, the greater the extensiveness of group obligations. The extensiveness of a group’s obligations alone, however, has no necessary implications for group solidarity. What also matters is the probability that members will comply with these obligations. It is to this problem that I now turn.
The Probability of Compliance
Rational egoists may desire the benefits derived from group membership, but they hope to receive these unconditionally. If members value the joint good, they are willing to commit themselves (or to be obliged) to help produce it; yet they will still have an incentive to free ride. Even though all the members of the protective association place a high value on security, still they would prefer to receive it without honoring their full obligations (say, by understating their assets or ‒ better yet ‒ by refraining from making any contribution at all). This is precisely the difficulty encountered with the provision of public goods: since they can be consumed by anyone, then rational egoists will not help to produce them. Even though members may place a high value on some joint good, free riding can be curtailed only if there is some means of assuring compliance with corporate obligations. A group’s ability to do this is a function of its control capacity. While extensive obligations arise only in groups that provide immanent joint goods, control is an issue in all groups – even in those producing marketable commodities.3
The relationship between control and compliance is intricate for two reasons. In the first place, the group must have sufficient resources at its disposal to effectively reward or punish its members contingent on their level of contribution or performance. This ability to provide what are essentially selective incentives can be called the group’s sanctioning capacity.
By virtue of the fact that members are more or less dependent (by definition), all groups have at least one potential sanction ‒ namely, exclusion from the group. Exclusion is the ultimate sanction in that it denies individuals access to the jointly produced good that they value. Yet some groups use this sanction more readily than others. For example, intentional communities are more likely to expel deviant members than are families. Though the effectiveness of the threat of expulsion varies with the member’s dependence on the group, the group’s willingness to employ it as a sanction is analytically distinct from the dependence of its members.
In any case, many groups employ additional sanctions that fall short of expulsion to motivate compliance with corporate obligations. Although these sanctions are collectively produced, they are quite different from the benefits that lead people to join the group initially. Like exclusion, many of these sanctions are negative and therefore cannot count as benefits at all: if members do not live up to their obligations, they will suffer the consequences. In further contrast to the good that motivates membership, the provision of sanctions need not be regular or guaranteed but can be intermittent and provisional. A union’s strike pay, for example, can be an incentive for picketing, but it only comes into play during a strike.
In order to be effective, these sanctions must be distributed to members selectively. Whether these sanctions are material or nonmaterial, their supply is never unlimited. Thus, to attain maximum compliance, groups must not only devise means of producing or procuring stores of adequate sanctions, but they must also convince all members that they will receive the particular sanction that is appropriate to their past behavior. If compliant members are consistently punished while noncompliant ones are consistently rewarded, then the overall level of compliance will be at its nadir. And if there is too long a delay between behavior and subsequent sanctioning, the efficacy of a sanction declines.
The second reason for the intricacy of the relationship between control and compliance is that the group must be able to detect whether individuals comply with their obligations or not. This is its monitoring capacity. Monitoring is problematic because individual behavior is often difficult to observe, much less to measure. Some acts ‒ those conducted in utter privacy ‒ are intrinsically harder to monitor than others ‒ those carried out in the full view of other members. When a group tries to attain attitudinal as against visible behavioral compliance, its monitoring task is all the more demanding.
True, not all members have an interest in concealing their behavior. Deviants alone have this incentive, but the compliant can usually be relied upon to publicize their virtue. Yet this does not mean that groups composed of the relatively virtuous can do without monitoring. Monitoring is required not only to ferret out the noncompliant but also to check on the allegedly compliant, for claims of virtuous behavior can never be taken at face value. In the absence of monitoring, deviants or shirkers are also likely to describe their past behavior as virtuous. Hence, all self–reports of compliance must be sifted to separate the wheat from the chaff, and this, in turn, requires monitoring.
Altogether, then, noncompliance with obligations (and with rules of any sort) can have at least two separate roots: it can be due to inadequate sanctioning or to impaired monitoring. Since each of these activities is costly, the total costs of control constitute a severe constraint on any group’s ability to attain compliance.
What determines a group’s control capacity? Many considerations come into play [… ], and for illustrative purposes I shall only mention two of them here.
The first determinant is the measurability of the individual’s contribution. Whenever an individual’s contribution to the production of a joint good cannot be reliably indicated by an output ‒ as is the case, for example, in teamwork ‒ control is problematic. Since acts carried out in privacy are more difficult to monitor than public acts, another factor is the group’s ability to limit the privacy of its members. It is in the interest of members to extend their privacy, just as it is in the interest of the collective to limit it.
A group’s survival depends upon the adoption of effective techniques to control its members. Yet insofar as control enables group members to produce joint goods, it must be considered a second–order collective good (Laver 1981: 62–71). As such, the provision of control is itself subject to the free–rider dilemma. While each member may gain from the overall solidarity of the group (because solidarity is an enabling condition for the supply of the joint good), free riding remains each rational agent’s best strategy. Members will not voluntarily assume the burden of control without sufficient compensation. It follows that all long–lived groups must include some individuals ‒ sometimes called agents ‒ who are compensated for providing control and are motivated to do it on this account. Without such agents groups cannot secure routine compliance. But agents come in varying sizes and shapes. In informal groups everyone is simultaneously an agent and a member; in more complex structures agents and members are differentiated and perform mutually exclusive roles. In some groups (American academic departments) members rotate into and out of the agency role; in others (capitalist firms) access to this role is more restricted. Different institutional arrangements ‒ particularly those that affect the distribution of the joint good ‒ determine the relations between agents and members in all groups.
Whereas dependence characterizes all voluntary groups, it is insufficient to solve the free–rider problem. Without control, group solidarity is, at best, a chimera. Large groups with relatively great control capacity are fundamentally different from those lacking this capacity. They are likely to have clear and consistent corporate goals, for these are necessary to precisely define the members’ obligations. Control promotes the stability and exclusivity of groups.
To recapitulate, solidarity varies with the extensiveness of corporate obligation together with the probability that members fully comply with these obligations. The theory suggests three conclusions. (1) Since groups that produce goods for the marketplace can compensate their members with wages, solidarity will be confined
