Achieving Goal Congruence in Complex Environments: The Case of Welfare Reform.
This raises familiar questions for those who study or practice public management. Is it possible to achieve and maintain congruent goals among actors in complex, intergovernmental, and multiorganizational policy systems? If so, what conditions foster greater congruence in goals? If not, what are the consequences for the performance of public programs and for the achievement of policy objectives?
Welfare reform provides a particularly interesting case for considering these questions. In 1996, Congress passed a major legislative reform that devolved many aspects of public welfare from federal to state auspices. The radical restructuring of welfare appears to have been motivated, at least in part, by policy officials' frustration with the organizational impediments that undermined earlier reform efforts. Nathan and Gais (1999, 1), for example, suggest that the law "seeks to modify two kinds of behavior, the personal labor force and sexual behavior of poor family heads and the bureaucratic behavior of the agencies that administer welfare and related programs" (emphasis in original).
One element in the success of this effort may be the extent to which state and local welfare systems can align the interests of multiple organizations and their staffs with new policy goals. A number of scholars have described problems of goal incongruence in social welfare programs. The frequent uncoupling of the day-to-day work priorities of front-line staff from larger policy goals helps explain the failure of earlier welfare reform efforts to bring about enduring bureaucratic reform. The devolution of welfare programs in 1996 gave state public officials and agencies greater authority to address these problems at the state and local levels. This has created a natural experiment both for observing the extent to which public welfare systems achieve greater congruence between policy and operative goals and for comparing goal congruence under a variety of political and institutional conditions.
In this article, we compare aspects of the governance and management of organizations that deliver welfare reforms in three states, purposively selected to illustrate implementation contexts that vary in policy and institutional complexity. We ask whether and how local welfare organizations have been able to articulate and operationalize goals that are consistent with the substantive policy goals of state officials. We address this question by using data collected through on-site interviews conducted at local welfare sites, each of which is a network of public and private organizations charged with the delivery of welfare and related services. Across the three sites, we describe the congruence between formal policy goals, agency priorities, and operative goals from the perspectives of agency managers and frontline staff. In order to consider the role of the political and institutional context, we compare sites that differ in terms of the complexity of policy goals and the complexity of the implementing network.
BACKGROUND AND LITERATURE REVIEW
Goals in Complex Organizations
Case studies of effective and high performing public organizations have repeatedly stressed the contribution of clearly articulated missions and goals (see, e.g., Rainey 1991; Rainey and Steinbauer 19Q9; Behn 1991). Studies of effective leadership in public organizations have also stressed leaders' abilities to communicate clear organizational missions and goals (Riccucci 1996). Empirical studies suggest that goal clarity increases motivation by linking staff efforts to mission-related tasks, by insulating staff from organizational politics, and by encouraging innovation and risk taking within organizations (Rainey and Steinbauer 1999; Witt 1998; Golden 1990; Behn 1999; Bozeman and Kingsley 1998). Goal conflicts, in contrast, are seen to compromise performance by c, reating conditions under which staff seek to minimize their exposure to risk and to pursue individual, rather than organizational, interests (Thompson 1967). Goal conflicts and ambiguities may increase goal displacement within work units and, in the extreme, "organizational cheating" to manipulate performance appraisals (Bohte and Meier 2000).
Researchers who study policy implementation have drawn similar lessons about the role of goal clarity for policy achievement. Those working in the so-called top down tradition of policy implementation argue that policy-making officials who fail to specify clear goals weaken their ability to direct and monitor implementing agents (Mazmanian and Sabatier 1989; Sabatier 1986). Researchers employing bottom up and contingent models of implementation have frequently noted the extent to which interest conflicts and goal ambiguity undermine policy achievement (e.g., Pressman and Wildavsky 1984; Lipsky 1980; Brodkin 1990 and 1997).
Although public management and implementation scholars tend to agree about the importance of clear goals for public organizations, they are equally persuasive about the difficulties of formulating such goals within complex organizational systems. Achieving goal clarity is difficult, in part, because goals serve multiple purposes within complex organizations, from defining organizational missions to establishing measures of success (Scott 1992; Hall 1996). Goals also represent and advance the interests of multiple individuals and groups within organizations. Because the goals of individuals and groups within the organization correspond only partially to those of the organization as a collectivity, they are frequently a target of competition, conflict, and compromise (Perrow 1961 and 1986; Scott 1992). For public organizations, external political and regulating authorities bring additional interests to bear, and these interests are often neither unitary nor unified (Nachmias and Rosenbloom 1980).
For several reasons, the already formidable difficulties associated with specifying clear and unitary goals for complex organizations may be even greater in social welfare organizations. Public opinion data and the vacillating history of welfare reform reveal a persistent tension, in the views of both the general public and policy officials, between the policy goals of alleviating disadvantage and discouraging reliance on government assistance (Heclo 1994). As Evelyn Brodkin (1990 and 1997) has pointed out, when policies are highly conflictual, unresolved policy tensions are often passed from the political to the administrative arenas in the form of vague, even contradictory, policy directives. Given the tension between the goals of providing public assistance and deterring people from it, it is not surprising that welfare agencies have historically operated under ambiguous and even contradictory policy directives (e.g., Hasenfeld 1992; Lipsky 1980). These ambiguous and irresolute policy directives, and the high level of discretion exercised by frontt-line staff, commonly result in the uncoupling of social welfare policy goals from the day-to-day priorities of front-line staff in the agencies that deliver these policies (Lipsky 1980; Meyers and Dillon 1999; Hasenfeld 1983 and 1992).
Additional complexities arise when we consider the goals of multi-actor organizational networks. Many if not most public programs are now delivered not by a single organization, but by networks in which whole or parts of separate public and private agencies must work together to achieve policy objectives (Hindmoor 1998; Kettl 1993; O'Toole 1993; Provan and Milward 1995; Hjern and Porter 1981; O'Toole and Meier 1999). This is particularly notable in social welfare programs, in which operations have undergone a substantial shift from public to private auspices, and from an agency-based to a systems-based approach, in recent years.
The implications of these structural dimensions for goal clarity may be mixed. On the one hand, network structures are seen by some to increase the capacity of the public sector to pursue complex goals by promoting organizational specialization and increasing the resources, flexibility, and responsiveness of delivery systems (O'Toole and Meier 1999; Provan and Milward 1995). On the other hand, the increased interdependence and uncertainty inherent in network structures may exacerbate the difficulty of formulating clear goals. At the agency level, greater permeability to unpredictable and sometimes competitive environments can fuel agency-level efforts to buffer core activities and promote parochial, agency interests (Thompson 1967; Salamon 1989; Smith and Lipsky 1993). Front-line staff who cope with this permeability on a daily basis may develop collective belief systems to make sense of the resulting uncertainty by further differentiating the priorities and values of their own agency from those of other agencies in the network (Sandfort 1999). At the network level, multiple agency interests can create the same types of uncertainty that arise from competition and the divergence of interests within organizations (Considine and Lewis 1999; O'Toole and Meier 1999).
Do Organizations Need Goals?
As a result of these and other complexities, "the concept of organizational goals is among the most slippery and treacherous of all those employed by organizational analysts" (Scott 1992, 285). Given the complexity of defining, much less reconciling, multiple goals, some organizational researchers suggest abandoning the concept of goals altogether in assessing performance. The most radical of the bottom up implementation scholars sometimes make a similar argument about the futility of trying to identify unambiguous policy goals that can serve as a standard for evaluating implementation success.
Abandoning goals as a criterion is problematic, however, for both public management and public administration. Public management scholars continue to define their object of interest in terms of goal achievement--as O'Toole and Meier suggest, "the set of conscious efforts to concert actors and resources to carry out established collective purposes" (1999, 510). In terms of policy delivery, even the most committed proponents of intergovernmental bargaining models would likely agree that implementation efforts should be directed toward the achievement of public purposes specified in advance by public officials.
We need therefore to devise some standard against which the organization of multiple interests within public organizations might reasonably be compared and evaluated. One such criterion is that of the congruence of the goals that motivate and direct the efforts of actors within organizational systems.
Although the operational definition of goal congruence is debated (e.g., Vancouver, Millsap, and Peters 1994), the concept has the advantage of providing an observable outcome that is independent of an evaluation of the substance of the goals themselves. In examining the implementation of policy reforms, the most straightforward definition of goal congruence may be the most useful: the extent of agreement between the official or formal policy goals of political officials and the operative goals of the organizations or networks charged with delivering that policy. Here, operative goals are the "ends sought through the actual operating policies of the organization" (Perrow 1961). It is the operative goals that "tell us what the organization actually is trying to do, regardless of what the official goals say are the aims" (p. 855). While operative goals are typically less codified than formal goals, they are nevertheless stable. As Hall (1996, 262) describes, "Operative goals are thus a derivation or a distillation from official goals.... They persist beyond the life of a particular interaction and become the standards by which the organization's actions are judged and around which decisions are made." They provide an important indicator of "policy as delivered" because, they "identify the preferences and choices made by those who control the organizational resources" and "predict the type of human service technologies that the organization is likely to adopt to achieve its goals" (Hasenfeld 1983, 87).
THE CASE OF WELFARE REFORM
The 1996 federal welfare reforms provide a particularly interesting case for studying the role of goal clarity and congruence in public programs. By devolving control over welfare programs to the state level, these reforms created a natural experiment in which states now vary in the purpose and in the structure of their welfare programs. This provides an opportunity for researchers to compare systems on the extent of, and precursors for, goal congruence in public welfare organizations.
Many political and scholarly observers have suggested that the 1996 reforms were motivated, in part, by a desire to reform or even reinvent the institutional structures through which welfare services are delivered. Despite several earlier waves of welfare reform, by the mid-1990s the welfare system was widely criticized for what Bane and Ellwood (1994) dubbed an "eligibility compliance culture" that was believed to encourage long-term welfare dependency. Studies of the largely disappointing results of earlier welfare reform efforts suggest that policy officials frequently failed to link their policy reforms to appropriate administrative and organizational reforms (e.g., Hagen and Lurie 1994; Meyers, Glaser, and MacDonald 1998; Meyers and Dillon 1999; Brodkin 1992 and 1997). One consequence was the uncoupling of welfare reform rhetoric from welfare operations. Meyers and Dillon (1999) describe the largely unsuccessful efforts of California policy officials to reform service delivery in the early 1990s:
Although elected and bureaucratic officials talked about comprehensive welfare reform, there was little evidence that workers and supervisors in local welfare offices understood the reforms in these terms. Workers understood the mission and goals of their organizations in proximate and technical terms that focused on producing payments, controlling errors, and minimizing client complaints. They understood welfare reform in similarly narrow and proximate terms that incorporated the means, in the form of rule changes and benefit reductions, while ignoring the anticipated ends in clients' work and self-sufficiency behaviors.
Alongside these pessimistic case studies, however, other observers provide examples of dramatic changes in seemingly entrenched welfare bureaucracies. Studies by Behn (1991) and Mead (1999), in particular, pointed to the role of organizational reform and leadership in effective change efforts in Massachusetts and Wisconsin. These studies underscored the importance of administrative and structural reforms that increased the congruence between policy objectives and agency operations. As a long time observer of welfare reform, Lawrence Mead, suggests, "In the end, welfare cannot be reformed by rhetoric. It must be reinvented" (1999, 615).
The 1996 Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) substantially increased the opportunities for state and local welfare systems to experiment with new administrative and organizational designs to address these problems. The most dramatic change under the PRWORA was the elimination of federal entitlements for cash assistance through Aid to Families with Dependent Children (AFDC), for child/day care, and for job placement and training services under the Job Opportunity and Basic Skills (JOBS) program. Federal funding for these services is now provided through two block grants, the Temporary Assistance to Needy Families (TANF) and Child Care and Development Fund (CCDF). The elimination of federal entitlements has important implications for the governance of welfare. Under TANF and the Child Care and Development Block Grant (CCDBG), states are no longer required to provide assistance as an entitlement or to adhere to federal eligibility rules. This gives state officials substantial new authority to design and deliver their own cash aid, child care, and employment services.(1)
The 1996 welfare reforms thus gave states new opportunities to reinvent their welfare systems. One of the outstanding questions is whether, with new local autonomy, state welfare systems will be able to improve the extent of congruence between policy and front-line, operative goals. Reform-minded rhetoric can now be heard at virtually all levels of government. However, simply broadcasting new goals for welfare cannot be expected to induce change. As Behn (1999, 72) reminds us, "Establishing goals does not automatically produce an individual innovation, let alone create an entire innovative organization. Merely articulating goals is not enough." For welfare organizations to succeed in reinventing themselves, many would argue, these goals must be understood and adopted by the individuals, agencies, and networks that will need to cooperate in their realization. In other words, multiple actors will need to achieve congruence in the purposes or goals that drive their efforts.
Two state-varying factors are likely to influence the extent of goal congruence within state welfare systems. One is the clarity and complexity of policy goals. As research on organizational goals suggests, ambiguous or poorly defined goals can exacerbate the uncoupling of policy and operative goals, as well as competition between intraorganizational interests. The PRWORA defines the purposes of the TANF program broadly: providing assistance and reducing dependency on that assistance.(2) This gives states considerable latitude, within the federal rules regarding work and time limits, to set specific state goals and priorities that resolve the historical tension between providing and deterring public assistance.
The second factor likely to affect goal congruence is the institutional complexity of the delivery system. As we have noted, the shift of service delivery from single agencies to multiagency networks may have important, and still poorly understood, implications for achieving goal congruence. In the case of welfare delivery, devolution appears to have moved the boundaries of program authority both downward and outward, fueling the trend toward more complex, network-based service delivery systems. The most obvious shift of authority has been from the federal to the state level, and many states have used their new authority to engage in further second order devolution that shifts responsibility downward, to county or local governments, and outward, to new public and private entities. Of the nineteen state TANF programs studied by Nathan and Gals (1999), for example, ten are now relying almost entirely on employment departments or commissions to deliver the employment and training portions of the TANF program; another four have established a system of joint implementation between the welfare and employment agencies.
RESEARCH QUESTIONS, DATA, AND METHODS
In this study we use comparative case studies to consider whether and under what implementing conditions local welfare systems achieve goal congruence by linking operative goals (at the agency and network levels) to welfare reform policy goals. By using data collected through semistructured interviews with welfare officials, managers, and front-line workers in local programs in three states (Georgia, Michigan, and Texas), we compare formal policy goals, local agency priorities, and operative goals as they are perceived by organizational members and as they are translated into concrete operations and performance measures. We address the following questions:
* What is the extent of congruence between formal policy goals and operative goals at the network and agency levels in each of the three sites?
* Does the extent of congruence vary with the complexity of the organizational network structure?
* Does the extent of congruence vary with the complexity of the formal policy goals?
Although three case studies do not provide sufficient variation to test hypotheses about the relationship between the implementing context and the extent of goal congruence, prior literature suggests the direction of the association we would expect to observe. Given evidence that goal conflicts are greater in highly decentralized and network based organizational systems, we expect that goal congruence will be greater in sites that have less complexity in organizational structure. Given evidence that goals are often uncoupled in organizations that are directed to pursue multiple, competing ends, we expect that goal congruence will be greater in sites with less complexity in formal goals.
Our analysis is limited in several important respects. The case study data used for this analysis do not provide performance data that would allow us to address the larger question, Does goal congruence matter for agency and network performance? This study examines conditions that may foster or impede goal congruence; in future research we hope to extend our inquiry by adding measures of agency and network level performance. Given the small number of cases, we are not able in this article to draw definitive conclusions about goal congruence. These case studies are best understood as concrete illustrations of how formal and operative goals are coupled (or uncoupled) in varying policy and organizational contexts.
Case study data were collected as part of a larger study conducted through the Rockefeller Institute of Government, which examines the implementation of welfare reform in four states (Michigan, Texas, Georgia, and New York). These four states were selected to maximize variation in policy, administrative structures, and political culture. Within each state, data have been collected at three sites in order to provide variety in community context (including urban, suburban, and rural sites), in management practices, and in client characteristics.
For the analysis in this article, three of the twelve local sites were selected for comparison. To consider under what conditions operative goals are most congruent with policy goals, we selected cases to illustrate two contextual dimensions that the literature suggests will be particularly salient: the extent of complexity in policy goals,, defined as the clarity and stability of policy directives, and the complexity of the institutional system, defined here in terms of the number and integration of implementing organizations. In order to examine the link between these contextual variables and goal congruence, we selected cases that represent three conditions (illustrated in exhibit 1): low complexity in policy goals and low complexity in institutional arrangements (case 1); low complexity in policy goals and greater institutional complexity (case 2), high complexity in policy goals and high levels of institutional complexity (case 3).(3)
Exhibit 1 Contextual Factors that Influence Goal Congruence Institutional Complexity Policy Goal Complexity Low consistent and stable policy goals CASE 1 Low Unambiguous goals single agency or Simple or highly integrated tightly coupled system institutional systems High CASE 2 multiple agencies in Unambiguous goals loosely coupled system Complex or fragmented institutional systems Institutional Complexity Policy Goal Complexity High inconsistent and/or shift-in policy goals Low single agency or tightly coupled system High CASE 3 multiple agencies in Ambiguous goals loosely coupled system Complex or fragmented institutional systems
Although a fourth condition--high levels of complexity in policy goals and low levels of institutional complexity (hypothetical case 4)--is possible, we found no clear illustrations of this combination in our twelve case studies. This may be more than an accident of sampling. Given the dynamic relationship between institutional structure and policy formulation, it may be less common to observe cases in which policy goals are inconsistent or unstable in systems that are dominated by a single agency or a tightly coupled organizational system.
It should be noted that we classify sites using dichotomous categories for heuristic purposes. In reality, variation is observed along continuua of complexity--from more to less hierarchical systems, and from more to less complexity in formal goals. This complexity is a central feature of public systems that deliver public policies that are themselves complex, often ambiguous, and even contradictory. We return to a discussion of these issues in our conclusions.
In this analysis, we are not interested in the content of the goals per se because states vary in the specifics of their reforms. Nor are we concerned with whether individual staff personally agree with the direction of policy reforms. Rather, we are interested in goal congruence on the organizational level, defined here as the extent to which formal policy goals, as articulated by political officials, correspond at the agency level and network levels with staff understanding of policy goals; staff understanding of the actual priorities of the organizations; and the operative goals or standards against which from-line workers are judged on performance.
For each of the sites, we begin by describing the context in terms of the complexity of formal (or espoused) welfare reform goals and the complexity of the institutional system. We next examine the congruence between the formal goals espoused by policy officials, the understanding of these policy goals by local managers and front-line workers, and the actual priorities and operative goals that dominate agency operations. Given the shift of responsibilities from hierarchical agencies to multiagency networks, we consider whether formal and operative goals are closely coupled on both agency and network levels. We conclude by considering some of the implications for policy achievement and for public management.
A site in this study is the network of organizations that are jointly responsible for three related forms of assistance under the TANF block grant: cash assistance, employment services, and the authorization of and payment for child care. These networks vary in number of organizations that are responsible for the delivery of this package of assistance and in the public/private divisions among these organizations.
We classify local sites in terms of institutional complexity by considering two dimensions. The first, following O'Toole and Meier, is "[t]he extent to which public programs are located fully within a single (hierarchical) agency or spread across parts of two or more organizations--within a single government, located across governments ... or encompassing links between public agencies and business or not-for-profit organizations" (1999, 508). The second dimension is the degree of integration among these organizations, including especially the formal management structures for distributing resources and monitoring performance. Sites are considered less complex if fewer organizations must coordinate their efforts or if multiple agencies are involved but are tightly integrated through stable administrative structures that more closely approximate traditional hierarchical structures.
We classify sites in terms of goal complexity by considering the extent to which policy officials have communicated--through legislation, administrative directives, and other forms of official communication--internally consistent and stable goals for welfare reform. Formal goals are defined as the objectives of welfare reform espoused by public officials in the relevant political jurisdiction(s)--the state, and as appropriate, the county.(4) A variety of sources are used to ascertain the formal goals, beginning with state statutes and other policy documents and drawing on earlier state reports produced by the Rockefeller Institute State Capacity Study project field researchers (see Nathan and Gais 1999).
In classifying formal policy goals as more complex or less complex, we focus on the consistency and stability of the directives from policy officials rather than on the substantive consistency of state policy goals. Given the dual emphasis of the federal TANF law--to provide "assistance to needy families so that children may be cared for in their own homes ..." and to end "the dependence of needy parents on government benefits ..."--it could be argued that the welfare reforms are inherently contradictory (and thus complex). As states have responded to the law, they have taken various approaches to resolving the tension between these emphases. From the standpoint of implementing organizations, however, it is the consistency of the state approach that most directly determines the complexity of the policy environment.
To compare goal congruence across the sites, we define congruence as the correspondence between the formal goals stated above and three site-level goal measures. The first is the local understanding of formal goals, measured as the response (at different levels of the organization) to questions such as: What is the main message that you have received from policy officials about the goals for welfare reform in this state? What do you think is most important to policy makers? Our second site-level measure, agency priorities, we define as the day-to-day priorities that organization members believe drive the work of the agency. To measure agency priorities, we asked staff at all levels of the organization: How do you think policy officials define the success of welfare reform for your office? Do certain measures of success "drive" you more than others? We asked agency managers and supervisors to tell us what priorities they believed drove their staff, what aspects of staff work always got done, and what aspects were skipped "when they have to cut corners." We asked parallel questions of front-line staff about their own work.
Our third site-level measure, operative goals, we define as the standards against which staff members judge, and are judged, with regard to performance. We obtained these measures by reviewing detailed information about agency management information systems, evaluation, performance appraisal, and related procedures. To ascertain the actual role of these tools, we also asked staff at all levels of the organization about the implementation and importance of evaluation procedures. We asked agency managers and supervisors to list the specific criteria by which they evaluate staff, how they obtain performance information and monitor performance, and what they consider a "good job." With front-line staff, we asked parallel questions about the methods and criteria for both formal and informal performance evaluation. Although performance measures do not provide a complete accounting of' operative goals, they do provide information about what staff are told to accomplish--what the agency defines as the concrete ends of their work.
We will now describe the goal and institutional context of each of the three sites (summarized in exhibit 2). We will compare the sites with regard to local understanding of formal policy goals, agency priorities, and operative goals (summarized in exhibit 3). Finally, we will consider the level of goal congruence in each site (summarized in exhibit 4).
Exhibit 2 Local Sites: Context Institutional Policy Goal Complexity Complexity Low consistent and stable policy goals Low CASE 1: Summer County, Georgia single agency [index] Governor and public welfare agency or tightly articulated consistent emphasis on employ- coupled ment and welfare deterrence system [index] Public welfare bureaucracy controlled and tightly integrated all aspects of welfare-related services [index] Local public welfare agency contracted-in multiple support services High CASE 2: Duncan County, Michigan multiple [index] Dominant political coalition championed agency in consistent policy emphasis on employment loosely combined with welfare support coupled system [index] Formal goals of welfare reform estab- lished at the state level [index] Implementation divided between public welfare and employment agencies, and between public and private contractors Institutional Policy Goal Complexity Complexity High inconsistent and/or shifting policy goals Low single agency or tightly coupled system High CASE 3: Bradford County, Texas multiple [index] Formal policy goals shifted between agency in employment, quality control (error reduc- loosely tion), and welfare deterrence coupled system [index] State policy officials pursued aggressive decentralization, downsizing, and privati- zation of welfare system [index] Implementation divided among public welfare and quasi-public workforce boards, with extensive contracting-out of employ- ment services Note: The names of the counties have been changed to protect the anonymity of our respondents. Exhibit 3 Local Sites: Formal, Informal, and Operative Goals CASE 1: Summer, Georgia Welfare Employment Local understanding of Keep clients from Employment formal policy goals "crossing the line" to welfare receipt Local understanding of Deterrence and Employment actual policy goals employment Local understanding of Deter welfare use Get clients agency priorities and get clients into jobs into jobs Outcomes monitored Deterrence rates; Employment and evaluated for TANF participation referrals; employ- front-line workers rates ment placements CASE 2: Duncan, Michigan Welfare Employment Local understanding of Work First and Work First formal policy goals zero cases without earnings Local understanding of Employment Employment actual policy goals Local understanding of Efficient case Get clients agency priorities processing into jobs Outcomes monitored Service timeliness; TANF participation and evaluated for client complaints rates; employment front-line workers placements CASE 3: Bradford, Texas Welfare Employment Local understanding of "Get them a job, Employment formal policy goals get them a better job, get them a career." Local understanding of Caseload reduction Caseload actual policy goals and quality control reduction Local understanding of Reduce error rates Meet federal par- agency priorities and detect fraud ticipation rates for TANF clients Outcomes monitored Accuracy in claims Min-fed TANF and evaluated for processing; service participation front-line workers timeliness rates Note: The names of the counties have been changed to protect the anonymity of our respondents. Exhibit 4 Local Sites: Goal Congruence Institutional Policy Goal Complexity Complexity Low consistent and stable policy goals Low CASE 1: Summer County, Georgia single agency or High levels of policy and operative tightly coupled system goal congruence Tightly coupled goals at the network and agency levels High CASE 2: Duncan County, Michigan multiple agencies in Network level policy and operative loosely coupled system goal congruence Partial uncoupling of goals at agency level Institutional Complexity High inconsistent and/or shifting policy goals Low single agency or tightly coupled system High CASE 3: Bradford County, Texas multiple agencies in Uncoupling of policy and operative goals loosely coupled system at network and agency levels Agency priorities dominated by expediency and risk avoidance Note: The names of the counties have been changed to protect the anonymity of our respondents.
CASE 1. We classify our first case, Summer County, Georgia, as a site in which policy goal complexity is low, because policy officials at both the state and county levels have communicated a clear and consistent direction for welfare reforms. We also classify institutional complexity as low because, while several agencies are involved in some aspect of welfare delivery, their roles are dictated and tightly coordinated by the public welfare bureaucracy.
Policy Goal Context. Welfare reform in Summer County moved steadily and consistently toward the goals of employment and welfare deterrence over a period of several years. The state of Georgia launched employment programs that were precursors to its TANF program as early as 1988 and expanded them to all 159 local welfare systems by 1993. In 1995, the emphasis shifted from training to WorkFirst, under the leadership of the governor and the commissioner of the Department of Family and Children's Services (DCFS). The WorkFirst initiative was billed as a "change in philosophy" that shifted "the goal of welfare reform from income maintenance to employment.' This emphasis was carried over into the state's TANF plan, which requires employment or employment preparation from all adult recipients with a child over the age of one, imposed tough "two strikes and you're out" sanctions for noncompliance, and established a forty-eight-month time limit for assistance.
While state officials have the ultimate authority to determine welfare policy in Georgia, county governments have considerable influence over policy details. Local advisory boards for the DFCS hire the agency director at the county level and approve local program initiatives; with the encouragement of the governor and the DFCS commissioner, the policy-making authority of local boards and agency directors has grown. "The mantra," according to a local manager, "was proceed until apprehended." At our local site, the director of the Summer County DFCS has taken full advantage of this devolution to formulate her own goals for welfare reform, including vigorous efforts to increase employment in current welfare cases and to divert new applicants from cash assistance. "Diversion is our initiative," she explains. "I don't think other counties even have diversion. The definition was ours. The idea of keeping them from `crossing the line' [into welfare receipt], because once they cross the line, you've got to push them back."
Institutional Context. We classify Summer County as low in institutional complexity because the local delivery system has tightly coupled and hierarchical structure under the control of the public welfare agency. Although welfare-related services are from a number of providers, policy and programmatic authority are highly centralized under DFCS at both the state and the county level.
At the state level, DFCS controls policy making, administration, and funding for cash assistance, food stamps, Medicaid, child care, and welfare-related training programs. Under the TANF block grant, some employment-related services have been contracted to the Georgia Department of Labor. DFCS has retained significant responsibility for employment services, however. In Summer County, the public DFCS agency occupies the central position in the welfare delivery system, managing all welfare, Medicaid, child welfare, and child support services. Although job placement services are contracted to the local Department of Labor, which co-locates job placement specialists at the DFCS office, other services are contracted-in to DFCS from other county and private service providers (such as the Community Action Agency and the department of health). The DFCS agency is thus able to directly provide to clients a range of services that go well beyond cash assistance. On-site or nearby services include, for example, job counseling, employment skills training, child day care, community work placement, a teen parent program, and drug treatment services. DFCS staff act as case managers in the coordination and allocation of services in what is described by local managers as a tightly linked system of providers. As the DFCS director describes it: "It's really like we all work for the same agency."
CASE 2. We classified our second site, Duncan County, Michigan, as a case in which policy goal complexity is low, because state policy makers have formulated and imposed clear goals for welfare reform on local agencies. We consider institutional complexity high because public officials have distributed responsibility for achieving those goals across a network of independent public and private organizations that operate with only limited integration at the local level.
Policy Goal Context. The dominant political coalition in Michigan has been unusually clear and consistent in emphasizing employment as the goal of state-level welfare reform. The governor launched employment-supporting reforms as early as 1992 under federal AFDC waivers and considerably strengthened the employment focus of these reforms in 1994 with the Project Zero Demonstration Project (for zero welfare cases without earnings). On the eve of the 1996 federal reforms, the state legislature passed a welfare reform bill based almost entirely on rules that had been drafted and implemented between 1992 and 1995 by the governor and the state welfare commissioner. The goals of the Michigan welfare reform law reflect the governor's goals of combining employment expectations with welfare support, reading in part: "Efficient, fair, cost-effective administration of the family independence program" and "[p]rovision of family independence assistance to families willing to work toward eventual self-sufficiency." In keeping with this emphasis, the Michigan law provides generous earnings and does not impose a time limit on aid as long as recipients are engaged in employment activities. However, it does impose immediate and ongoing work-engagement obligations on all recipients. The formal policy goals of the state are passed directly to local agencies through policy directives, program structure, and contracting provisions.
Institutional Context. The complexity of the institutional structure in Michigan arises from the division of responsibility between public sector entities and between public agencies and private contractors. The 1996 Michigan welfare reform law divided oversight of welfare-related activities between two state-level entities with distinctly different administrative histories and structures. Authority for cash assistance, child care, and various employment supports such as transportation assistance remains with the Family Independence Agency (FIA).(5) The FIA delivers services through a centrally managed system of county level offices, in which local FIA offices exercise relatively little discretion about program design or delivery. The staff of the FIA are responsible for eligibility and benefit determinations for TANF, food stamps, Medicaid, and child care.
On the employment side, the welfare reform law moved all employment services out of the welfare agency and into the Michigan Jobs Commission (MJC). Since its establishment as a cabinet-level agency by executive order in 1993, the MJC has taken over an increasing number of employment programs in the state.(6) In sharp contrast to the welfare agency, the MJC operates through a decentralized system of twenty-six local Michigan Works! contractors and is largely free of detailed procedural oversight at the state level.
At the local level, employment services are often further divided between public agencies and private contractors. In our Duncan County study site, the public Employment Services agency operates as the Michigan Works! agency to oversee delivery of all welfare-related employment programs (Work First services). Employment Services operates with considerable local autonomy in the design and delivery of employment programs, but it is prohibited from providing services directly to TANF clients. Instead, Employment Services uses a competitive bidding process to select local contractors to provide Work First services. The successful bidders in Duncan County were three well-established nonprofit organizations.
The split of operations between the FIA, for welfare, and the Employment Services Agency and its three Work First contractors, for employment services, creates a mixed delivery system in Duncan County. Managers and staff describe the working relationship between the local FIA agency and the private Work First contractors as cooperative but complex. Coordination has increased demands on supervisors and staff. According to one FIA supervisor, "Maintaining relations with three Work First agencies is very time-consuming. Compared to the past, its more intensive and structured: what can you do, what can we do, how can we make it work." Communication burdens are increased by new and partially incompatible management information systems. The result, as one manager described it, is "a fair amount of back and forth and slippage" in caseload tracking. Coordination difficulties are increased by differences in agency-level performance standards set by the state-level FIA and the MJC. In Duncan County, for example, the FIA operates under the state directive to see that all welfare recipients will be immediately and continuously engaged in employment activities. Employment Services and its local Work First contractors, on the other hand, have not been provided with sufficient resources by the MJC to serve all TANF clients who are mandated to enroll in their services. The Work First contractors prioritize their intake of new clients, which complicates the work of the FIA workers who are responsible for imposing sanctions for any clients who are not compliant with mandatory Work First activities.
CASE 3. In our third site, the Texas county of Bradford, complexity is high on both dimensions of our typology. We classify policy goal complexity as high because the local welfare system has been given inconsistent and shifting policy directives from state policy officials. We consider institutional complexity high because services are delivered through a decentralized network of public and private entities.
Policy Goal Context. Welfare services in Texas have been shaped by several waves of reform that have sent sometimes consistent, oftentimes contradictory, messages about policy goals. Particularly influential changes were initiated by Texas House Bill 1863, passed in 1995. The law was notable for using the federal waiver process to impose stringent time limits on welfare receipt several years in advance of the federal reforms. Of equal importance, the law advanced an aggressive campaign by the governor to decentralize, downsize, and privatize Texas' public bureaucracies. HB1863 moved employment services out of the public welfare; system and into the quasi-private Texas Workforce Commission (TWC).(7)
While the priority to decentralize and privatize welfare operations has been clear, the substantive goals of welfare reform have shifted repeatedly. For example, policy goals have changed with the turnover of state welfare commissioners. Under one commissioner, all responsibility for TANF employment was relinquished to the TWC; his successor launched Texas Works, a program that requires front-end employment search at the welfare agencies. Policy goals also have changed in response to federal actions. Soon after the state launched Texas Works, for example, local offices were told to shift their focus away from front-end employment assistance in order to comply with directives from federal officials to reduce the error rate in the food stamp program.
Institutional Context. The drive to decentralize and privatize welfare services, combined with vacillating policy directives, creates considerable complexity and uncertainty in the institutional context for local welfare delivery. Formal responsibility for welfare service delivery is divided at the state level between the public welfare agency, the Department of Human Services (DHS), and the TWC. The DHS operates the TANF, food stamps, and Medicaid programs through regionally managed public welfare offices; local offices exercise little discretion in either the goals or the organization of these services. The TWC, in contrast, manages all employment services, child care, and employment assistance by distributing funds through a highly decentralized system of quasi-public Local Workforce Development Boards (LWDB). Each LWDB, composed of employers, agency representatives, and services providers, is then free to select local contractors for service delivery, including public, nonprofit, or for-profit entities. In sharp contrast to the DHS welfare system, LWDBs are expected both to set local priorities for services and to design locally responsive systems for achieving them.
The rapidly changing governing structure has resulted in considerable turmoil in local sites such as Bradford. Front-line staff in the DHS have been retitled Texas works advisors. There have been few changes in their actual job responsibilities, however, and they continue to refer to themselves informally as eligibility workers for the TANF cash assistance, food stamps, and Medicaid programs. All employment-related assistance--including child care and transportation--is now managed through the LWDB and its contractors. In Bradford, this contract was awarded to Jobs Plus (a pseudonym), a large, for-profit corporation. Jobs Plus also delivers employment assessment, services, and placement to TANF clients, along with referrals to outside community services such as adult education and training.
Although they are managed through different state systems, there is considerable overlap and blurring of boundaries between the facilities and staff of the Bradford DHS and the Jobs Plus contractor. Jobs Plus is co-located at the Bradford DHS office. The majority of the front-line workers at Jobs Plus are former DHS staff who had been laid off when employment services (under the JOBS program) were shifted out of DHS.(8) Since Jobs Plus had limited experience with job services, it adopted the procedures, forms, and other program details from the earlier DHS program. Day-to-day interdependence between the agencies is also high, and it is complicated by differences in the management priorities of the two systems. The Bradford DHS is held accountable by the state DHS, for example, for the rate at which TANF recipients participate in the employment services provided by Jobs Plus; for Jobs Plus, meeting performance standards set by the LWDB (under the direction of the TWC) for client participation depends, in part, on the rate at which DHS sanctions its recipients for noncompliance with these activities. This interdependence creates conflicts when agency-level priorities differ. For example, while DHS staff are told to divert new applicants away from the welfare system, those at Jobs Plus hope to see cases open so that they can fulfill contractual obligations and draw down contracted funds by providing employment and support services.
Local Understanding of Policy Goals, Agency Priorities, and Operative Goals
The cases in this study illustrate considerable variation in the goals and the organization of the welfare services devolved to the states in 1996. Variation is equally great in the local understanding of formal policy goals and in the translation of policy goals into agency priorities and operative goals in welfare and employment agencies.
CASE 1. In Summer County, Georgia, managers and staff understand the overriding concern of policy officials to increase employment and reduce welfare use. The local slogan of stopping new applicants from going over the line to receive welfare is closely linked to both the day-to-day priorities of the agencies and the performance standards for staff.
Local Understanding of Formal (Policy) Goals. The Summer County DFCS director credits the state commissioner with giving consistent directions about policy goals. "We've always thought we knew what our customers needed," she observes. "[The commissioner] gave us the go ahead. I'd call and he'd say, `Don't let policy get in the way.' The message was to move people toward self-sufficiency." She has used this discretion, her control over staffing and other aspects of program design, and substantial local funding to develop her own approaches to reducing welfare reliance--including aggressive efforts to divert new applicants to aid.
The policy goals for employment and welfare diversion are widely echoed by supervisors and front-line staff in both DFCS and the Department of Labor. The goal, according to one DFCS worker, is "to do what we can to get these people to work." Because jobs are specialized in DFCS, the specifics of how staff address this goal vary. But the language they use to describe the intent of welfare reforms is consistent. Keeping new applicants from crossing the line is repeated in policy documents and even emblazoned on staff T-shirts. Staff out-stationed by the Department of Labor describe similar goals for welfare reform. "We are trying to get them a job. So it starts as soon as they come in. They hear it from DFCS. They hear it from us. If they hear the same from all of us, we are all on the same sheet of music." Although they are not responsible for benefit determinations, the Department of Labor staff echo the goals of welfare deterrence. "I send them out on jobs right away," explained one worker. "Our job is to avoid that [welfare] check ever getting there. To get them a job."
Agency Priorities and Operative Goals. Agency priorities and operative goals in Summer County are also directed toward employment and welfare deterrence. Under the direction of the Summer County DFCS director, the agency has developed specialized staff positions that are tightly coordinated to provide front-end diversion and on-going monitoring of employment activities.
The efforts of multiple, specialized DFCS and Department of Labor staff are focused on employment and deterrence, in part, by the direct and persistent attention of the DFCS director to outcomes. Starting in 1997, she began to personally track the status of all cases. She also keeps close track of deterrence rates. "We divert close to 70 percent of people from even going on welfare," she reports. "There are maybe 250 that have come on the system and stayed on since welfare reform began in January 1997, out of thousands of applicants." She is equally attentive to participation and employment rates among open TANF cases.
The director's ability to monitor outcomes is aided by a specialized management information system (MIS). Frustrated by their inability to extract the information from the state MIS, Summer County DFCS developed a local, PC-based system that maintains detailed records on all current and former welfare recipients in the county. From-line staff are responsible for seeing that data are entered to track client status and activities. Managers can use the system to generate reports on client activities and costs--on the client, worker, or administrative unit levels.
Although performance standards and evaluation criteria vary according to job specialization, they reflect the priority given to welfare diversion and employment. Because most of their contact is with new applicants, eligibility workers in DFCS say that it would be difficult to know if their pending applicants or clients go to work. They are keenly aware, however, of the rate at which applicants are diverted from the system. New applicants who find work during the initial job search period, or who simply fail to follow through on the job search or other eligibility tests, are treated as procedural denials. They now account, by one worker's estimate, for one-half of new applications.
Ongoing case managers at the agency, in contrast, are expected to track the status of all employable cases and to report every month on the work participation status of each of their clients. For those clients placed in mandatory work experience slots, the case manager is expected to call weekly to monitor participation. Supervisors report that they expect workers to document their plans and their follow-up in case records, which serve as the basis for performance evaluation. The MIS also provides monthly reports about caseload status. According to supervisors, "All our workers know we want to meet federal participation requirements." As for other priorities, such as quality control (error reduction), one worker observed, "The emphasis is so much on work, we don't think about it as much anymore."
The Department of Labor job placement specialists who are out-stationed at DFCS describe themselves as responding to two groups of stakeholders. One is the DFCS, which sets numeric job placement goals in its contract. The second is the local employer. As one described it, "I work with employers too. I know that they [clients] have to be ready.... If they aren't dressed right or if they don't feel good about themselves they [employers] will know." Between the two groups of stakeholders, the influence of DFCS goals for job placements is clearly most salient. Supervisors describe their standards for the performance of front-line workers: "The accuracy of client information. The match between client needs and what is provided.... The realism of the job choices." Front-line staff report that their dominant concern is getting jobs for clients. They know that their agency has job placement targets. Although these goals are not translated into numeric targets for individual staff, they create performance pressure. "As long as we meet them, we're OK," is how one worker described it. "But we are kind of, sort of, under the microscope."
CASE 2. In Duncan County, Michigan, managers and staff in both the welfare and employment agencies are clear in their understanding of policy officials' goals, and they express little conflict over the dual emphasis on employment and welfare assistance. The welfare and employment agencies differ, however, in how closely this goal is linked to agency priorities and operative goals.
Local Understanding of Formal (Policy) Goals. Whether or not they fully agree with the changes, staff at the Michigan site widely credit the governor for consistency in approach: "The way [they] approached it ... the message was constant. `While we are going to require work, we are going to do these things.' The message has been you have to go to work first--but they did back it up with a tremendous increase in day care and an increase in resource limits."
At the welfare office (the FIA), managers are clear that the goal is to get every able-bodied person employed. Getting clients off welfare is not the formal policy goal, according to these managers. If anything, managers expect their workers to make every effort to open applicants' welfare cases in order to get them support services that may increase employability. As one manager summarized it, workers "wouldn't dissuade them from getting on welfare. Because our role is to provide support." From-line staff in the Duncan County FIA are equally clear that the formal goal of reform is to increase employment. In the past, they commented, it was possible to put someone into a work support or training program. But now, "the focus of Work First is to get them a job." Staff are also clear that it is not their job to keep applicants out of the welfare system. As one described it: "People have the right to apply. [We] wouldn't dissuade, because we have supportive services."
The staff of Employment Services and its Work First contractors are equally clear about policy-makers' goals. The same goal is echoed by the staff of the Employment Services and contractor agencies: "to see that the participants are placed in the best job and have them stay there." They do not see their jobs in terms of welfare reduction, per se, but they believe that the two are inextricably connected. The goal, as described by one job specialist, is "to get people jobs so that they won't need welfare; so there won't be any more welfare."
Agency Priorities and Operative Goals. While local managers and staff in Duncan County concur about the formal goals of policy officials, their understanding of the informal goals or priorities for their own agencies differs substantially between the welfare and employment agencies.
In the public welfare agency, although they are cognizant that the formal goal of welfare reform is rapid employment, there is less agreement about whether this is the priority for the work of their agency. The FIA manager is emphatic that employment should be the goal for his staff: "The only string on the TANF funds is employment. There is no standard of promptness. No error rate. Staff say we can't control that. But I say, then why are we here? If we have no role in employment, why are we there?" Yet managers also describe the overriding priority for their workers as "getting a [welfare] case open." If their workers have to cut corners, the work that "always gets done" is the computer paperwork to start welfare benefits. What workers let slide, according to supervisors and front-line staff, is the counseling, the home calls, and other direct client contact designed to remove barriers to employment. Nor is extensive employment counseling described as appropriate for new applicants who come in: "They have a need. They need food. They need money." From-line workers in the FIA are even less ambiguous on this point than are their managers. They describe their job as getting benefits and emergency assistance to clients; getting clients to work, in their view, is the job of the local Work First agency. As one described it: "That is their [Work First] problem. That's their issue. When a customer is referred to Work First, determining that they participate twenty-five hours a week is their responsibility."
These priorities are clear in the operative goals promoted by the FIA information systems, performance standards, and evaluations. Under welfare reforms, the job of from-line workers was enlarged to provide support services and to encourage employment in quarterly home visits. Despite this job expansion, FIA workers are not evaluated on measures relating to employment services or outcomes. The management information system does not provide systematic information about their clients' work status to the FIA workers or their supervisors. Employment information, when it is received, is used to adjust client benefits rather than to evaluate staff performance. The FIA performance evaluations are dominated by traditional standards for case processing. In supervisors' terms, this means meeting the standard of promptness and quality control. Asked whether there are consequences for staff who don't meet the required number of home visits or provide employment services, supervisors say, "No. There should be, but no."
In Employment Services and its contractor agencies, in contrast, agency priorities and operative goals relate directly to client employment at both the agency and the staff level. Managers at Employment Services identify two distinct groups of stakeholders to which they are accountable. One is the local business community. In their work with employers, managers describe as a priority making sure contractors are finding "qualified individuals for employers." The second group is the public oversight agency (the Michigan Jobs Commission), which pushes the agency toward concerns about placing clients in employment activities and reducing costs per client. Achievement of Michigan Job Commission's goals for client placements is rewarded with incentive funds; failure to meet the goals is penalized both by a loss of funds and by loss of status as a manager. Employment Services, in turn, passes employment-related performance measures on to their contractors: "Enrollment levels, placement levels.... All the numbers the state requires are looked at." The private Work First contractors use these performance targets to set priorities for their staff. Although they are not translated into individual targets, staff at the Work First contractors know that their agencies have employment placement targets and believe that failure to meet employment goals could result in penalties.
CASE 3. In Bradford, Texas, managers and staff in both the welfare and employment agencies express considerable confusion about the formal goals of welfare reform in Texas. The priorities and operative goals of agencies in both systems reflect proximate work pressures, rewards and penalties, which are inconsistently related to formal policy goals.
Local Understanding of Formal Policy Goals. Staff and managers in Bradford are aware that employment is the espoused goal of welfare reform in Texas. They are widely dismissive of this goal, however, describing policy officials' real goal as saving money by deterring welfare use, detecting fraud, and closing cases. On the welfare side, when managers in the Bradford DHS are asked about policy officials' goals, they are quick to use the language of the Texas welfare commissioner: "You get them a job. You get them a better job. Then you get them a career." As the managers describe it, this is the vision of the welfare commissioner, and therefore it is the vision of the agency. They are acutely aware that policy directives from the state have been neither clear nor stable in the past. Given DHS managers' confusion, it is not surprising that they believe that formal policy goals are not widely understood or embraced by their own staff. "They are so used to rolling with the punches," is how one manager described it. "Policies change all the time. They see it as just another process." Many managers and staff also dismiss formal policy directives as euphemisms for officials' political objectives. As one supervisor described it, the governor's goal was "Getting people off welfare. It's no secret, its in the papers. But that is politically incorrect so they say it is about getting people jobs." Skeptical of policy officials' motives and reliability, DHS staff! consider new policy directives to be just another demand on their time and attention, and they freely communicate this to their clients. One supervisor justified this in frank terms: "Clients complain about the new forms. You know what I have [workers] tell them? Well, the policy changed and you have to do it now. And you know what? When you come back in six months is going to be different again."
Managers and staff in Bradford employment agencies express nearly as much confusion about the formal policy goals of welfare reform. For the staff of the LWDB and its contractor Jobs Plus, confusion is exacerbated by what are perceived to be contradictory demands from external stakeholders. Staff describe the goal of state officials in terms of meeting the federal participation rate for current welfare recipients. In contrast, it is widely believed that the members of the LWDB, dominated by employers, "don't give a whit about participation.... The board perceives their primary customers to be employers. They want us to prepare people to be good employees." The expectations create competing and irreconcilable demands on staff. As one described it: "Meeting participation rates doesn't really work in concert with getting people off welfare and into jobs. There is a conflict because participation means that people are doing some activity other than working. If I got everyone a job, I wouldn't meet participation."
Agency Priorities and Operative Goals. Given widespread confusion and cynicism about the formal goals of welfare reform, it is not surprising that staff understanding of the priorities and operative goals at the local level reflect, instead, the proximate operational demands of each agency.
In DHS, staff have been told that the new policy goal of the Texas Works initiative is to foster employment and redirect applicants away from cash assistance. In terms of performance monitoring, however, DHS performance measures are specific in their emphasis on case processing and quality control. As one local manager describes, he is held accountable for "Managing service delivery. Timeliness. Ninety-six percent accuracy in casework."
These same priorities are passed on to front-line staff in DHS. Attention to employment-related outcomes is limited because supervisors lack the tools to encourage, observe, monitor, or reward staff efforts related to employment. As one supervisor observed, "There's been a big decline in caseloads, but we don't know where they've gone." With no systematic data on employment services or outcome, supervisors report that they are aware of staff efforts to promote employment only "when they tell us about it" and staff performance on this dimension is neither monitored nor rewarded. Workers are carefully monitored, on the other hand, on the speed and the accuracy of their eligibility determinations. Quality control (error reduction) is given particular attention; it has been made easier by new database linkages with other state agencies. If an error is caught by the supervisor, the worker is allowed to correct it. But if it is caught by the outside quality control auditors, consequences are severe: "They have to correct it and go to a counseling meeting at regional.... [They] get raked over the coals, can't get superior rating on evaluation which means they cannot get merit pay."
For managers and staff at the Jobs Plus contractor, the formal policy emphasis on employment-related outcomes is clear. These outcomes are operationalized narrowly, however, in terms of federal participation targets for current welfare cases. Frontline staff are given numeric targets for the participation rates of their TANF caseloads, based on the minimum standards in the federal welfare law--dubbed the min-fed by local staff. Staff performance on this standard is monitored monthly through the MIS, and individual staff face penalties if they fall below the target but receive cash bonuses if they raise participation rates. Clients' actual employment outcomes, in contrast, are not systematically tracked, monitored, or rewarded. In fact, workers are actually penalized on performance ratings if they get clients into employment because: these clients no longer count toward their monthly min-fed targets. For staff, the agency priority is clear, but the conflict between operative and formal goals is confusing. As one Jobs Plus worker reflected: "With all the changes to make this a temporary program, the philosophy of putting people to work is in conflict with what they continue to harp on, collect data on, base funding and staffing on. You have to wonder, are we really tracking anything that is relevant. If the goal is getting people to work, why aren't we measuring how many people are put to work? Why aren't we rewarded for getting people to work?"
The three states in this study had similar opportunities under the 1996 federal welfare reforms to reinvent the goals and institutions of welfare services. They responded to these opportunities in ways that reflected the particular political and institutional context of the state. Although we have dichotomized two dimensions of this response--policy goals and organizational arrangements--this is only a heuristic device for describing the actual complexity of states' responses. Given a federal mandate to pursue the dual goals of providing assistance and discouraging dependency on assistance, these states communicated policy goals and structured organizational systems in ways that may be more accurately understood as representing a continuum of responses--from those that were reductionist and simplifying to those that were more complicated in both substance and organization.(9)
Although our two-way classification greatly simplifies this reality, it is useful for analytic purposes. Considered within this typology, the variation in goal congruence across the three sites is seen to correspond in the anticipated fashion to the complexity of the implementing context. Goal congruence appears to be more problematic either when policy goals are complex or when organizational systems are complex; when both are complex, substantial uncoupling of formal and operational goals is observed (see exhibit 4).
CASE 1. In Summer County, Georgia, state and county officials have resolved the potential tension between providing welfare services and reducing welfare dependence by defining their welfare services as employment support. Services in support of the policy goals of employment and welfare diversion are provided through a tightly linked interagency system that is dominated at both the state and the county level by the public welfare bureaucracy. Consistent goals and low levels of institutional fragmentation contribute to high levels of congruence between formal policy goals, perceived agency priorities, and the operative goals or ends of agency operations. From-line staff and managers in both welfare and employment services are aware that the formal policy goal of reform is to get people to work and, in the local terms, to keep them from crossing the line into welfare receipt. Agency operations are consistent with these goals. Agency caseloads, deterrence, and work participation rates are closely monitored by the director of the public welfare agency. Operations support the same goals through job specialization, a supportive management information system, and performance monitoring that emphasizes welfare deterrence and employment placements. Although staff are not given individual performance targets, they are aware that their performance on these dimensions is watched--so much so that, in the words of one, they feel that they are "under the microscope." Other areas of performance, such as quality control to reduce error rates, are explicitly understood by staff to have lower priority.
CASE 2. Michigan lawmakers resolved the potential tension between the goals of providing assistance and reducing dependence by embracing both. The policy message--that welfare reform is about getting every able-bodied recipient to work and about providing appropriate welfare assistance--appears to be widely understood by program managers and their staff in Duncan County. One reason that the system is able to promote dual and potentially conflicting goals is that functions have been split into separate spheres for welfare and employment agencies. This structure helps to reconcile potential goal conflicts and to direct the efforts of the network, as a whole, toward multiple goals. This structure may have the consequence, however, of partially uncoupling policy and operative goals at the agency level.
In the Duncan County Employment Services and its private Work First contractors, the formal policy goal of employment is tightly coupled with operative goals through agency and staff-level performance targets. In the public welfare agency, in contrast, congruence between policy and operative goals is much weaker. From-line staff have been given new titles and new job responsibilities; they have been told that they are expected to help achieve the new employment goals of welfare reform. The day-to-day priorities of their jobs have not changed, nor have the performance standards against which they are evaluated. This creates a considerable gap between their understanding of the intent of welfare reform and the daily demands of their work. Staff reconcile this demand by doing, as one manager describes it, "what they are comfortable with--which is the old [eligibility] stuff. And what we measure."
CASE 3. In Bradford County, Texas, the welfare system is characterized !both by contradictory and unstable policy goals and by high levels of institutional complexity. Policy officials have communicated multiple and changing policy goals. Policy officials also have communicated expectations that staff find difficult to reconcile--for example, to spend more time providing supportive, up-front employment support to new applicants and to spend more time scrutinizing applicants' documents to detect fraud. Political officials at the state level have initiated a major restructuring to decentralize and privatize the welfare system, dividing responsibility for welfare services among the public welfare agency, the quasi-public Local Workforce Development Board, and private contractors. These systems differ both in their performance standards and in their management structures. One result of contradictory policy directives and institutional fragmentation appears to be widespread uncoupling of policy and operative goals.
Whether they are considered on the network or the agency levels, the operative goals of welfare agencies in Bradford appear to be shaped by expediency and to be largely disconnected from espoused policy goals. This can be traced, in part, to the skepticism of managers and staff who believe they have received multiple, conflicting signals from political officials who have ulterior, but unspoken, political objectives. Managers and staff protect themselves by focusing on those aspects of work for which the concrete penalties for poor performance are most severe. In the welfare agencies, staff direct their efforts toward accurate case processing by collecting extensive documentation and verification from welfare applicants. To the extent that this discourages applicants, the increased vigilance may contribute to caseload reduction. It does little, however, to support the official policy goal of increasing employment, because it creates an adversarial encounter with clients and deflects applicants from the system before they can access employment services. At the Local Workforce Development Board and its private employment contractor, staff direct their attention to meeting min-fed (participation rates) by placing TANF recipients in job preparation activities and assuring that those who do not comply are given sanctions. But this conflicts with the larger goal of getting clients into employment, and staff express frustration that they are neither measured nor rewarded for performance on the formal policy goal of getting people to work.
CONCLUSIONS AND DISCUSSION
These cases demonstrate varying levels and forms of goal congruence in the delivery of welfare services in the wake of welfare reform. They illustrate the conclusions of a substantial literature on organizational goals by suggesting that goal congruence is more easily achieved when the formal goals of policy officials are consistently communicated and when the organizational context is less fragmented. In this, they suggest that clarifying policy directives and rationalizing organizational systems could reduce the often-noted uncoupling of operative goals in welfare and other human service agencies.
These case studies cannot answer the larger question of whether goal congruence matters for agency performance and goal achievement. We can only begin to address this question when we consider issues that are raised by prior research on the role of goals in policy implementation and public management.
Goal Congruence and Goal Achievement
Implementation scholars suggest that congruence between formal policy goals and agency-level operative goals may be a prerequisite for substantive goal achievement. To the extent that formal policy goals represent the actual interests of elected officials, this would suggest that congruence between policy and operative goals provides evidence that public resources are being directed toward appropriate public purposes.
On this dimension, the lessons to be drawn from these case studies are rendered ambiguous by the nature of the reforms themselves. The federal welfare reform law communicated two directives: to provide assistance to those who need it, and to discourage dependency on that assistance. For the purposes of this analysis, we have defined goal complexity in terms of the consistency of the message communicated by policy officials. But as some observers have noted, goal clarity may be difficult to achieve given the dual focus of the federal law. As states have expanded their roles in welfare policy development, their interpretation and emphasis on these goals has varied. The question of the importance of goal congruence for goal attainment depends, in part, on how states have defined their own policy objectives.
If the goal of welfare reform is simply to reduce caseloads, the congruence of policy and operative agency goals may not matter. Caseload reductions may be achieved under a variety of conditions--when the formal policy goal of increasing employment and deterring welfare use is linked to similar operative goals, as observed in Summer County, or when formal policy goals for employment are uncoupled from operative goals for vigilant fraud detection, as observed in Bradford, Texas. In each of the states in this study, TANF caseloads declined 40 percent or more between 1995 and 1998, above the national average of 34 percent.
Officials in these three states did not espouse caseload reduction as their goal, however. Michigan law states the goal of welfare as "[p]rovision of family independence assistance to families willing to work toward eventual self-sufficiency." In Texas, the welfare commissioner describes the goal of reform as helping clients: "Get them a job. Get them a better job. Get them a career." In Georgia, the goal is to provide up-front and ongoing assistance that will prevent applicants from going over the line to receive welfare.
If these goal statements accurately reflect the intentions of policy officials, the extent to which implementing agencies and their staff see a link between their work and policy goals may matter a great deal. Front-line welfare staff in Summer County, Georgia, who believe that their job is to get clients jobs may be better equipped to provide supportive assistance than, for example, front-line welfare workers in Michigan who believe that linking clients to employment is "not our responsibility" or employment counselors in Bradford, Texas, who believe that what counts is meeting min-fed participation numbers for active welfare cases
Wholesale uncoupling of policy and operative goals raises the most obvious concerns in this regard. The partial uncoupling of goals at the agency level raises more subtle and perhaps interesting issues about the delivery of public programs through multi-agency networks. Both Texas and Michigan divide TANF program responsibility between the public welfare bureaucracy and private contractors who report to local job commissions or boards. This helps reconcile potentially conflicting charges to impose strong and potentially deterring employment expectations, on the one hand, and maintain a program of cash assistance, on the other hand. The uncoupling of policy and operative goals at the agency level may matter less in these cases, if multiple goals are pursued by the network as a whole. If applicants and clients encounter each agency in the local service provision network of Duncan County, Michigan, or Bradford, Texas, this institutional arrangement may in fact capitalize on agency specialization to provide more-varied forms of assistance. Problems arise, however, if applicants and clients do not encounter each agency--an outcome that becomes more likely as greater efforts are made to deter, redirect, or divert new applicants from welfare. For these individuals, who encounter just one node of the local network, the displacement of policy goals at the agency level may have important implications for access to public services.
Goal Congruence and Public Management
Public management researchers concern themselves with goal congruence for other reasons. In the search for models of effective management and agency leadership, scholars generally agree about the importance of clear agency missions and goals. Their emphasis is supported by empirical research that suggests that agencies and their staff are more innovative, more satisfied, and more productive when the demands and rewards of their work are consistent with their understanding of the larger mission of the organization.
These cases provide some evidence that in public agencies, the uncoupling of policy and operative goals can leave staff confused and demoralized. Staff at the employment agencies in these three sites were generally more positive about their work than were those in the traditional welfare agencies. Low morale was not inevitable in the welfare agencies, however. While none of the front-line welfare workers we interviewed described their jobs as easy, the extent of goal congruence was closely associated with how staff experienced their work and the advent of welfare reform.
Enthusiasm for welfare reform was palpable in Summer County, Georgia, particularly among staff who provided more specialized and employment-related services. Staff crowded into our site-visit interviews in order to describe their components of the program. They displayed the local slogan for welfare reform on T-shirts. They credited their managers with supporting their efforts. "If you're a caseworker and you see the motivation at the top trickling down, that really makes a difference," is how one front-line welfare worker described it.
For the Michigan FIA workers, in contrast, welfare reform has brought new titles and new job responsibilities but few rewards and little satisfaction. Workers describe themselves as overwhelmed and largely disappointed that they lack the tools, the time, and the authorization to pursue the social work aspects of their jobs. Some have elected to demote themselves to less demanding positions within the agency.
In the Texas DHS, front-line staff dismiss new welfare policy initiatives as political symbolism and hold to their beliefs that political officials are primarily concerned with reducing welfare caseloads and saving money. One result is widespread cynicism among staff, which is freely expressed to clients. Another is high turnover in the TWA staff. "Pay is low," one supervisor commented, "and the job market is good in Texas. People can go get another job that pays better and is less of a hassle."
These differences suggest that the congruence of policy and operative goals may matter, not only for the achievement of policy objectives but for the achievement of effective and innovative public organizations. Differences in staff morale at the three sites in this study cannot be fully attributed to their understanding of the link between their work and policy achievement. These sites also vary in their personnel systems (e.g., extent of unionization); in the talent and leadership capacity of their administrators; and in other aspects of working conditions (e.g., remuneration and opportunities for advancement). Nor do all staff at each site express the same attitudes. These cases do suggest, however, that conditions that support greater congruence between formal policy and operative goals at the agency level may have payoffs in terms of staff morale and commitment.
An earlier version of this article was presented at the fifth national Public Management Research Conference, College Station, Texas, December 3-4, 1999. This project is funded by a grant from the U.S. Department of Health and Human Services to the Rockefeller Institute of Government.
(1) Along with this authority, however, Congress gave states specific new performance requirements. States are still required to match each dollar of federal funding. States are also directed to decrease clients' welfare dependency by restricting TANF assistance to a lifetime maximum of sixty months and by imposing more stringent work requirements on aid recipients. By the year 2002, states are required to have one-half of all TANF recipients placed in work or community services activities.
(2) "The purpose of this part is to increase the flexibility of States in operating a program designed to (1) provide assistance to needy families so that children may be cared for in their own homes or in the homes of relatives; (2) end the dependence of needy parents on government benefits by promoting job preparation, work, and marriage...."
(3) The state for each site is identified, but to protect the anonymity of our respondents the name and some details of the local sites are changed.
(4) Our sample includes both state administered and county administered welfare sites. In state administered sites, we assume that the formal goals of welfare reform are those of state policy officials. In county administered sites, we consider the goals articulated by both state and county officials.
(5) Formerly the Department of Social Services.
(6) The MJC was first given responsibility for a welfare-related demonstration project in 1994 and was given responsibility for the JOBS program (formerly run by the DSS) and the EDGE program (formerly run by the Department of Education) in FY 1994-95.
(7) Efforts to privatize remaining welfare services were finally dropped only when federal officials ruled that private contractors could not be hired to determine food stamp and Medicaid eligibility.
(8) The JOBS (Job Opportunity and Basic Skills) program was the employment and training component established by the 1988 Family Support Act. The Local Workforce Development Board contract with Jobs Plus specifically directed Jobs Plus to hire DHS staff who were laid off by the program transition.
(9) We are indebted to an anonymous reviewer for suggesting this formulation.
Bane, Mary Jo, and Ellwood, David T. 1994 Welfare Realities: From Rhetoric to Reform. Cambridge, Mass.: Harvard University Press.
Barnard, Chester I. 1938 The Functions of the Executive. Cambridge, Mass.: Harvard University Press.
Behn, Robert. 1988 "Management by Groping Along." Journal of Policy Analysis and Management 7:4:643-63.
1991 Leadership Counts. Cambridge, Mass.: Harvard University Press
1995 "The Big Questions of Public Management." Public Administration Review 55:4:313-24.
1999 "Do Goals Help Create Innovative Organization?" In H.G. Frederickson and J.M. Johnston, eds. Public Management Reform and Innovation, pp. 70-88. Tuscaloosa: University of Alabama Press.
Bohte J., and Meier, Kenneth. 2000 "Goal Displacement: Assessing the Motivation for Organizational Cheating." Public Administration Review 60:2:173-82.
Bozeman, Barry, and Kingsley, G. 1998 "Risk Culture in Public and Private Organizations." Public Administration Review 58:2:10918.
Brodkin, Evelyn Z. 1987 "Policy Politics: If We Can't Govern, Can We Manage?" Political Science Quarterly 102:4:571-87.
1990 "Implementation as Policy Politics." In Dennis J. Palumbo, and Donald J. Calista, eds. Implementation and the Policy Process. New York: Greenwood.
1992 "The Organization of Disputes: The Bureaucratic Construction of Welfare Rights and Wrongs." Studies in Law, Politics, and Society 12:53-76.
1997 "Inside the Welfare Contract: Discretion and Accountability in State Welfare Administration." Social Service Review 71:1:1-33.
Considine, M., and Lewis, J. 1999 Governance at Ground Level: The Frontline Bureaucrat in the Age of Markets and Networks. Public Administration Review 59:6:467-80.
Cyert, R.M., and March, James G. 1963 A Behavioral Theory of the Firm. Englewood Cliffs, N.J.: Prentice-Hall.
Ferman, Barbara. 1990 When Failure is Success: Implementation and Madisonian Government." In Dennis J. Palumbo, and Donald J. Calista, eds. Implementation and the Policy Process. New York: Greenwood.
Golden, Olivia. 1990 "Innovation in Public Sector Human Services Programs: The Implications of Innovation by `Groping Along.'" Journal of Policy Analysis and Management 9:2:219-48.
Hagen, Jan L., and Lurie, Irene. 1994 Implementing JOBS: Progress and Promise. Albany: State University of New York, Nelson A. Rockefeller Institute of Government.
Hall, Richard. 1996 Organizations, Structures, Processes, and Outcomes. Englewood Cliffs, N.J.: Prentice-Hall.
Hasenfeld, Yeheskel. 1983 Human Service Organizations. Englewood Cliffs, N.J.: Prentice-Hall.
1992 Human Services as Complex Organizations. Newbury Park, Calif.: SAGE.
Heclo, Hugh. 1994 "Poverty Politics." In S.H. Danziger, G.D. Sandefur, and D.H. Weinberg, eds. Confronting Poverty. Cambridge, Mass.: Harvard University Press.
Hindmoor, Andrew. 1998 "The Importance of Being Trusted: Transaction Costs and Policy Network Theory." Public Administration 76:25-43.
Hjern, Benny, and Porter, David O. 1981 "Implementation Structures: A New Unit of Administrative Analysis." Organization Studies 2/3:211-227.
Kettl, Donald F. 1993 Sharing Power: Public Governance and Private Markets. Washington, D.C.: Brookings.
Lipsky, Michael. 1980 Street Level Bureaucracy: Dilemmas of the Individual in Public Services. New York: Russell Sage.
Matland, R.E. 1995 "Synthesizing the Implementation Literature: The AmbiguityConflict Model of Policy Implementation." Journal of Public Administration Research and Theory 5:2:145-74.
Mazmanian, D.A., and Sabatier, Paul A. 1989 "Framework for Implementation Analysis." In Mazmanian and Sabatier, eds. Implementation and Public Policy. New York: University Press of America.
Mead, Lawrence M. 1999 "The Decline of Welfare in Wisconsin." Journal of Public Administration Research and Theory 9:4:597-622.
Meyers, Marcia, and Dillon, Nara. 1999 "Institutional Paradoxes: Why Welfare Workers Can't Reform Welfare." In H.G. Frederickson and J.M. Johnston, eds. Public Administration Reform and Innovation. Tuscaloosa: University of Alabama Press.
Meyers, Marcia K.; Glaser, Bonnie; and MacDonald, Karin. 1998 "On the Front Lines of Welfare Delivery: Are Workers Implementing Policy Reforms?" Journal of Policy Analysis and Management 17:1:1-22.
Nachmias, D., and Rosenbloom, D.H. 1980 Bureaucratic Government, USA. New York: St. Martin's.
Nathan, Richard P., and Gais, Thomas L. 1999 Implementation of the Personal Responsibility Act of 1996. Albany: State University of New York, Nelson A. Rockefeller Institute of Government.
O'Toole, Laurence J. Jr. 1993 "Interorganizational Policy Studies: Lessons Drawn from Implementation Research." Journal of Public Administration Research and Theoly 3:2:232-51.
O'Toole, Lawrence, and Meier, Kenneth. 1999 "Modeling the Impact of Public Management: Implications of Structural Context." Journal of Public Administration Research and Theory 9:505-26.
Perrow, Charles. 1961 "The Analysis of Goals in Complex Organizations." American Sociological Review 26:854-66.
1986 Complex Organizations: A Critical Essay. New York: Random House.
Pressman, J.L., and Wildavsky, Aaron, 1984 Implementation. Berkeley: University of California Press.
Provan, Keith G., and Milward, H. Brinton. 1995 "A Preliminary Theory of Interorganizational Effectiveness: A Comparative Study of Four Community Mental Health Systems." Administrative Science Quarterly 40:1-33.
Rainey, Hal G. 1991 Understanding and Managing Public Organizations. San Francisco: Jossey-Bass.
Rainey, Hal G., and Steinbauer, P. 1999 "Galloping Elephants: Developing Elements of a Theory of Effective Government Organizations." Journal of Public Administration Research and Theory 9:1:1-32.
Riccucci, Norma M. 1996 Unsung Heroes. Federal Execucrats Making a Difference. Washington, D.C.: Georgetown University Press.
Sabatier, Paul A. 1986 "Top-Down and Bottom-Up Approaches to Implementation Research: A Critical Analysis and Suggested Synthesis." Journal of Public Policy 6:1:2148.
Salamon, Lester M. 1989 Beyond Privatization: The Tools of Government Action. Washington, D.C.: Urban Institute.
Sandfort, Jodi. 1999 "The Structural Impediments to Human Service Collaboration: Examining Welfare Reform at the Front Lines." Social Services Review 73:3:314-39.
Scott, W. Richard. 1992 Organizations: Rational, Natural and Open Systems. Englewood Cliffs, N.J.: Prentice-Hall.
Smith, Steven R., and Lipsky, Michael. 1993 Nonprofits for Hire: The Welfare State in the Age of Contracting. Cambridge, Mass.: Harvard University Press.
Thompson, James D. 1967 Organizations in Action: Social Science Bases of Administrative Theory. New York: McGraw Hill.
Vancouver, J.B.; Millsap, R.E.; and Peters, P.A. 1994 Multilevel Analysis of Organizational Goal Congruence. Journal of Applied Psychology 79: 5: 666-79.
Witt, L.A. 1998 "Enhancing Organizational Goal Congruence: A Solution to Organizational Politics." Journal of Applied Psychology 83:4:666-74.
Marcia K. Meyers Columbia University
Norma M. Riccucci and Irene Lurie University at Albany, State University of New York
|Printer friendly Cite/link Email Feedback|
|Author:||Meyers, Marcia K.; Riccucci, Norma M.; Lurie, Irene|
|Publication:||Journal of Public Administration Research and Theory|
|Date:||Apr 1, 2001|
|Previous Article:||Developmental Dynamics: Interagency Collaboration as an Emergent Phenomenon.|
|Next Article:||Designing State Clean Air Agencies: Administrative Procedures and Bureaucratic Autonomy.|