Compounding this ambivalence is the further problem that by their nature, collective goods contain important elements of ambiguity. Is the output that is valued the availability of a particular service, is it the altered behavior implied by its usage, or is it the consequences of the usage? The access working mothers have to quality day-care under the auspices of government, for example, may increase both the number of children enrolled in approved day-care arrangements and the number of mothers who are working. Potential supporters of the program may want to' 'purchase" the high quality, the increased usage by additional children, or the improvement in the employment or educational status of women (perhaps under the guise of making welfare mothers earn their keep). In another situation, enforcement of a regulation proscribing certain behavior may lead to punishment of violators (say, drug dealers), deterrence of destructive behavior (drug consumption), and the alleviation of problems (crime) associated with the proscribed behavior. Again, actual and potential supporters may value these outputs so differently that they will oppose one another over programmatic features.
Faced with such contradictions, social welfare executives must induce voluntary political approval for social programs by working out arrangements that attract or compel active political support for and belief in the value of certain kinds of output (for example, by child development specialists, welfare reform advocates, and potential employers of mothers of children). They must continually recognize and sort through the various ambiguities involved in any program and try to communicate an intelligible sense of purpose that somehow transcends or neutralizes destructive conflicts among a program's various constituencies.
In securing support, social welfare executives can assume an entrepreneurial role.14 They can attempt, for instance, to recruit a subset of group members by offering them selective inducements; an example is designing programs in accord with the concepts of output of organized service providers such as physicians, vocational rehabilitation counselors, mental retardation or child development specialists, psychologists, or police. They can propose new program initiatives which combine the interests of several constituencies. More generally, according to Mancur Olson, "a leader or entrepreneur, who is generally trusted (or feared), or who can guess who is bluffing in the bargaining, or who can simply save bargaining time, can sometimes work out an arrangement that is better for all concerned than any outcome that could emerge without entrepreneurial leadership or organization."
Alternatively, social welfare executives can seek to mobilize larger, more dispersed constituencies, such as interest groups representing clients or civic groups representing taxpayers. They might, for example, issue regulations that are broadly popular with influential parties: prohibiting physicians from charging Medicaid clients other than standard fees, or keeping single mothers from using unapproved day-care arrangements. Or they can offer the promise of rewards, such as the governor's support for coveted legislation or budgetary allocations.
However, by their nature, redistributive and regulatory policies such as these are relatively coercive. What frustrates effective entrepreneurship is the awareness or suspicion by affected groups that entrepreneurial executives either cater to special interests to gain their support or coerce the citizenry: the public, whose earnings are being "confiscated" for the benefit of the poor; beneficiaries, whose behavior is being manipulated as the condition of survival; and service providers, who are being told how to practice or what prices to charge.
Moreover, the would-be entrepreneur might well be regarded as superfluous to goal achievement by self-interested actors. As Russell Hardin observes, "Groups can often control their own effects on policy more specifically when they actively lobby, campaign, or go to court than when they more nearly passively influence the choices of entrepreneurs." The prevalence of negative attitudes toward their departments and their efforts, and the difficulties of overcoming them, explain much of the pressure and vulnerability associated with the social welfare executive's job.
The actual provision of social welfare services is also hampered by a number of obstacles. Even the identity of those to be served is often at issue: Does an official seek change in clients or in service providers, in parents or in their children, in job seekers or in employers? The greater the change being sought, the greater the pressure on program administrators is to demonstrate actual effectiveness (and the greater is their vulnerability to evidence of ineffectiveness).18 Human-service organizations thus tend either to prefer clients for whom success is more likely, avoiding the most troublesome cases, or to emphasize their processing responsibilities, which are less problematic, at the expense of seeking change in the functioning of the difficult or recalcitrant.
Further, in making a benefit or service available, social welfare executives necessarily alter the incentives facing those who are eligible and those administrative units that deal with eligible people. The result may be behavior quite different from what is desired. Certifying employable people as disadvantaged and thus eligible for a wage subsidy, for example, actually reduces their employment prospects, since employers shy away from applicants who bear the disadvantaged label despite the incentive of subsidy. Making available publicly assisted daycare services may induce some people to abandon perfectly adequate private day-care arrangements in order to take advantage of subsidized care. Offering reimbursements for the cost of service to people who fall into certain categories of need encourages service providers to place people in these categories in order to increase the level of reimbursements even if such labeling is of questionable validity. These "moral hazards" produce unintended and onerous consequences, usually in the form of higher-than-expected costs and caseloads.
The more fundamental problem, however, is that social welfare executives must achieve their service delivery goals by securing the willing cooperation of a dispersed network of self-interested agents-employees and contractors who, though they act on behalf and at the behest of the executive, may or may not share fully the executive's preferences and who are in positions deliberately or otherwise to frustrate the achievement of executive purposes. Management has the least influence over the bureaucracy's production of services when direct service jobs involve considerable discretion, when choices must be made among multiple objectives or work tasks, and when policy involves changes in practice in an established bureaucracy-three circumstances that are prevalent in social programs.22 Disagreements between the principal and his or her agents may be sharp and, more troublesome, may remain covert or unexpressed.
The penalties or inducements at the executive's disposal may be insufficient to elicit the desired behavior from these agents. Financial incentives may be too weak to overcome existing reward systems. The agency's ability to enforce sanctions may be poor. Interest groups may know that continued resistance will be supported in the legislature. Injunctions and restraining orders may be sought by the aggrieved.
Having to rely on self-interested agents thus introduces substantial uncertainty into the prospect of achieving service production goals. Social welfare executives must attend to such tasks as structuring the relationships with clients or agents (for example, determining the form of contracts with proprietary or not-for-profit service providers), monitoring agent performance, and enforcing compliance, all of which are costly in terms of executive time and attention and have a low likelihood of being entirely successful. Indeed, the participants in social program provision may have little motivation to cooperate in ensuring effective and efficient program performance, and there may be little the social welfare executive can do about it.
