Public Administration Review; July 1, 2000

When Bill Clinton embraced national service as one of his administration's priorities, he took 6 step forward on two of his key initiatives. Not only was national service a new initiative in its own right, but Clinton also held it up as a model of his efforts to reinvent government. It would be an exemplar of government that is catalytic, competitive, decentralized, and results oriented. This case study examines the theory and reality of reinvention. The Corporation for National Service and its programs have come under fire for being more political than catalytic, being simultaneously too centralized and too decentralized, and pursuing too many unclear goals. This article seeks to identify, discontinuities between the rhetoric and the reality of reinvention in this instance and draw lessons for public-sector reform.

Bill Clinton is often referred to as a "new" Democrat. It is not surprising, therefore, that he brought to his presidency some unconventional ideas about government, among them reinventing government and national service. In fact, Clinton boldly joined these two sets of ideas in the National and Community Service Trust Act of 1993. The act simultaneously became an experiment in changing how government does its business (Osborne and Gaebler 1992) and in rekindling the spirit of public service.

This case study of national service addresses the relationship between reinvention and national service as it has played out since 1993. We are interested in examining how the reinvention model actually took shape in a particular political and historical context. We begin with a discussion of how reinvention principles were applied to national service programs. We then explore the realities of what transpired in the aftermath of the 1993 act. We conclude with some lessons about the reinvention movement from its application to national service programs.

The phrase "reinventing government" came into popular usage with the publication of David Osborne and Ted Gaebler's book Reinventing Government (1992). The central theme of the book is that government in America is restructuring in response to citizen dissatisfaction, competitive pressures of a global marketplace, and revolution in information technology. Reinvention achieved movement status with the Clinton-Gore National Performance Review (Gore 1993) and the National Commission on the State and Local Public Service (Thompson and Radin 1997; National Commission on the State' and Local Public Service 1993).

Government reform initiatives are known for being short lived, owing in part to the attention spans of politicians (March and Olsen 1983), so the longevity of the reinvention movement is notable. After all, President Reagan's Grace Commission recommendations received only fleeting public recognition and little legislative or executive branch attention. Even Jimmy Carter's high-profile civil service reform, legislated in 1978, lasted only until Reagan's election, when its provisions were dismantled, ignored, or forgotten.

The most recent attention to reinvention has been longer lived than many of its predecessors, for a variety of reasons. One is the longevity of the Clinton administration, which, unlike two of its three predecessors, ex tended to a second term. Another factor is Vice President Al Gore's commitment to and identification with reinvention. Gore's political future rests, in part, with images associated with his role as leader of federal reinvention. The centrality of reinvention in the ideology of new Democrats also looms large in its longevity. In contrast to the Reagan Republicans, new Democrats believe in government and, therefore, are compelled to seek ways to make it work better.

What does the reinvention movement have to do with national service? The answer is that Bill Clinton "decreed that the national service program would be the model of his efforts to 'reinvent government'" (Waldman 1995, 102). In 1993, Congress passed the National and Community Service Trust Act, creating the Corporation for National Service (CNS), which unified all domestic service initiatives under one roof.

President Clinton and implementers of the legislation envisioned national service as an exemplar on several levels. For the president, service was quintessentially American. Service was a means to create community, an enduring strand in America--political philosophy (Sandel 1996). To build on this history was itself exemplary. In his speech unveiling the legislation, Clinton said, "National service is nothing less than the American way to change America. It is rooted in the concept of community: the simple idea that every one of us, no matter how many privileges with which we are born, can still be enriched by the contributions of the least of us.''[1]

The president's reinvention themes were developed and expanded by the CNS after its creation in 1993. Community building was a central concern for the CNS, but it was joined by other reinvention principles--among them mission and competition--that more directly guided the development of organizational philosophy and administrative practices.

The preeminent reinvention principle the CNS embraced was that it is driven by results. This choice was a product of its flint chief executive, Eli Segal, who decided during the legislative process that results were necessary for get ting and maintaining political support (Waldman 1995). The influence of the reinvention philosophy is clearly evident in the CNS's first quarterly report:

During our first quarter, we have moved quickly to design systems and strategies that will help communities get things done. We are building what we hope will be a model agency, one that can make well-in formed decisions without delay, keep staffing lean and flexible, multiply resources, and delegate authority and responsibility in-house and in the field .... In this way, following the principles of reinventing government and working in partnership with communities across America, we will create a national movement, driven by mission and dedicated to results.

The CNS's emphasis on reinvention has persisted. For example, the Corporation's 1997 strategic plan (CNS 1997, 3) referred to familiar reinvention themes, including entrepreneurship. To achieve its vision, the CNS sought to be

• a catalytic, coordinating, and creative force in realizing this vision for service in America;

• a valuable resource to and a partner with national, state, and local organizations that seek to solve problems through service; and

• entrepreneurial, innovative, effective, and efficient in utilizing its resources, influence, and activities.

President Clinton's rhetoric, the act itself, and the Corporation's strategies clearly embody symbols of reinvention. But at a deeper level, how are these principles manifested in the Corporation's administrative policies and actions? Specific design and administrative features are discussed next to illustrate operationally the meaning of reinvention.

Catalytic

Separating policy decisions from service delivery is the essence of catalytic government (Osborne and Gaebler 1992, 35; Barzelay 1992). The capacity to steer programs was designed into national service by several mechanisms, including creation of a federal administrative corporation, relaxation of civil service rules, and allocation of service delivery to nonprofit organizations.

The CNS is an independent agency with a 15-member board of directors that is appointed by the president and confirmed by the Senate. By law, no more than eight members may come from the same political party. Its chief executive officer reports to the board, but is also a presidential appointee subject to Senate confirmation, as are other senior agency officials. Hence, in theory, the Corporation is meant to be a policy-making and funding body, at least partially insulated from administration directives--much as the National Endowments for the Humanities and for the Arts are--rather than a provider of services and a pro gram operator.

The CNS sought independence from traditional civil service roles by instituting an alternative personnel system. During the fight for national service legislation, Eli Segal, then head of the White House Office of National Service, was concerned with avoiding the fate of ACTION, a Great Society-era agency which housed Volunteers in Service to America and other federal programs to promote community service. Over the years, ACTION had developed a reputation as ossified by civil service and union rules (Waldman 1995). The Corporation was given authority to adopt more flexible personnel roles that permitted employees to be hired under renewable, fixed-term contracts rather than traditional job-security provisions. Since ACTION was folded into the new agency, initially 80 percent of the CNS workforce, and now about 50 percent remained under the old personnel system. Nonetheless, compared to other federal agencies, the CNS has greater ability to recruit non-career employees with specialized expertise (rather than only bureaucratic know-how) who can be re placed or reassigned when their assignments are completed.

Like Osborne and Gaebler's prototype steering organization, CNS delivers virtually no programs itself. Instead, most service programs are operated by local or national nonprofit organizations, or by consortia of nonprofits, under broad priorities set by the president and Congress (GAO 1997b). (One exception is the National Civilian Community Corps, which is managed directly by the CNS, but it absorbs only a small share of the CNS budget.) The CNS also encourages community service from programs run by other federal agencies (such as the Department of Education's federal work-study program) and by volunteers at the grassroots level.

Competitive

Unlike other federal agencies, which have traditionally designed only limited competion into their programs the CNS tries to use competition for much of the funding it dispenses (even when it is not required to do so by Congress). By law, one-third of federal monies for AmeriCorps (the Corporation's "signature" full-time service program) are allocated strictly on the basis of population (formula grants). At least another third of the federal funds avail able for AmeriCorps programs are awarded to organizations nominated by state commissions, but chosen on a nationally competitive basis (GAO 1997b, 3). The remaining one-third of federal AmeriCorps funds are directly ad ministered by competitively awarding grants to national nonprofit organizations, institutions of higher learning, or multistate organizations--so-called national direct grants. Though it is not required by statute to do so, the CNS also requires applications for part-time, school-based service learning programs, such as Learn and Serve America, to undergo peer review. The funding formula for AmeriCorps is representative of the CNS's overall efforts to achieve balance among promoting national service across the states, avoiding state entitlements, and encouraging high quality.

Decentralized

By mandating that two-thirds of AmeriCorps funds flow through state commissions (and, in the case of Learn and Serve America, that all funds generally go through state education agencies), the president and Congress sought to make the states partners with the CNS in the administration of national service. State commissions comprise 15 to 25 members representing labor, local governments, service programs, youth, senior citizens, the state education agency, service-learning advocates, and others (Waldman 1995). In addition to administering funds, state commissions are responsible for conducting a statewide planning process to identify areas where community service can have the greatest impact.

Decentralization, as Osborne and Gaebler point out, also includes relaxing hierarchical controls and encouraging innovative management and employee development. In this respect, the CNS seeks to implement its programs in a highly decentralized manner through myriad local organizations, many of which are exceedingly small. Even the so-called "national direct" grantees, to whom one-third of AmeriCorps funds are allocated, typically operate through local chapters. The CNS also manages, generally through third-party contracts, an extensive training and technical assistance program aimed at helping its grantees to run successful programs.

Results-Oriented

As noted earlier, a central reality of national service is the importance of "getting things done." For AmeriCorps, this means helping communities meet critical human, educational, environmental, and public-safety needs. For Learn and Serve America, it means helping participants do better in school. Consequently, local programs must have the capacity to identify program objectives, develop performance measures, and collect adequate data to document program outcomes. CNS funding guidelines emphasize the importance of measurable results, and the agency itself conducts an extensive research and evaluation program aimed at identifying "best practices" and assessing pro gram impacts.

The CNS's results orientation is also evident in its operating systems. In 1998, it completed detailed logic models for all of its programs as part of its initiative under the Government Performance and Results Act of 1993 (CNS 1998). This process entailed creating logic models for the seven major programs administered by the CNS and identifying inputs, activities, and intermediate and end out comes. The process of outcome identification is challenging because so much of the CNS's work is performed by third parties, particularly grantees and volunteers.

As part of its efforts to achieve results, CNS embraced quality-management principles. The logic of quality management is evident in regulation implementation, grant review procedures, and guidance issued by the CNS. One prominent example of the quality-management emphasis is Principles for High Quality National Service Programs (CNS 1994), "part of an ongoing collaboration among the Corporation for National and Community Service and community-problem solvers across the country to attempt to define what 'high quality' national service means" (1). A review of initial implementation experience in AmeriCorps by Smith and Jucovy (1995) suggests that quality initiatives were having an impact. They observed: "Thus, in a manner not commonly found among federal grant-making organizations, the Corporation, rather than establishing minimum standards for acceptability, set ambitious quality criteria, emphasized them consistently, and assessed the funding proposals submitted by states against them" (Smith and Jucovy 1995, vi).

Good government is an enduring theme in American public affairs. It is even more compelling when coupled with the appeal of community service, another hallowed tradition of American public life. How do the realities of national service measure up to expectations created by its role as a reinvention exemplar?

Catalytic Is in the Eye of the Beholder

David Rosenbloom (1993) observed at the outset of federal reinvention in 1993 that "the congressional view of public administration does not draw a dichotomy between politics and administration" (505). The right mix of steering (politics) and rowing (administration) is likely to be governed by Miles's Law: Where you stand depends on where you sit. Although catalytic government is most of ten associated with steering, political imperatives may re quire the kind of directives characteristic of rowing. More over, whether an agency operates catalytically or bureaucratically may have less to do with what it actually does than with how its activities are perceived.

The CNS discovered this early on, when it faced fierce congressional criticism over the scope of political involvement by CNS grantees. When the law creating the pro gram was enacted, it contained a provision prohibiting fund recipients from taking part in partisan political activities. However, shortly after the Republican Party took control of Congress in 1995, ACORN, an antipoverty group known for political action which had a legally independent (but closely tied) affiliate that was a CNS grantee (ACORN Housing Corporation), disrupted a speech by House Speaker Newt Gingrich. (Following an audit, the CNS inspector general concluded that it was impossible to role out commingling of CNS funds between the two groups.) Along with accusations of political activities by other CNS grantees, this episode triggered a renewed emphasis in the agency on enforcing the separation of activities it supported from political activism. Since many nonprofit groups regard advocacy as an essential part of their mission, the CNS's heightened concern moved the agency toward con trolling not just the direction its grantees moved, but also how they conducted their activities.

Another early casualty was the CNS's efforts to steer other government programs. In its first year, the CNS pro vided grants not only to state, local, and nonprofit organizations, but also to 13 federal agencies such as the Departments of Agriculture, Defense, Labor, and the Interior. The grants, authorized in the statute, were designed to encourage these agencies to rely more on community service in implementing their programs. But to congressional critics, such grants looked more like back-door funding of federal operating programs to help them do more "rowing" than their appropriated funds allowed. In response to such criticism, the Corporation voluntarily chose to suspend the grants and eventually eliminated them by its third program year (GAO 1997a). Ironically, Aguirre International evaluators found that in AmeriCorps' first program year, federal agencies that were granted AmeriCorps funding proved to be some of the most creative sponsors in the development of new institutions in local communities (Aguirre International 1997, 26).

Notwithstanding its institutional independence, the CNS also found itself doing more rowing than steering, principally at the behest of the White House. Having declared "the era of big government" over, President Clinton saw stimulating community service an increasingly valuable way of attaining administration objectives. Thus, en route to the 1996 Democratic convention, he pledged to mobilize one million young people as tutors to improve school children's reading scores. Along with the Department of Education, the CNS was given the responsibility of implementing the President's initiative through its programs. In a sense, this epitomized how the federal government might steer, but not row, especially since local interest determined participation in the program. Yet the progression from presidential speech, to agency programming, to grantee response bears such a close resemblance to traditional relationships between government and nonprofits that-the distinction seems a fine one, at best.

Nor has the CNS's new personnel system succeeded in enabling the agency to change traditional staffing patterns. This was only partly due to the requirement that former ACTION employees be covered by the old personnel sys tem. In addition, as a National Academy of Public Administration study of the agency discovered, the CNS has been slow to develop the procedures and criteria needed to make the new system work (NAPA 1998). Moreover, it turned out that the skills and experience the agency required were often possessed by other government employees who proved reluctant to sacrifice the job security they had acquired under existing federal roles to accept fixed-term contracts under the new system (NAPA 1998).

At the board level as well, separating steering from rowing proved more difficult than might have been expected. In keeping with the practice of other federal corporations, the CNS board should have principally had a policy-making role. However, NAPA (1998) found it had acquired other responsibilities as well. According to NAPA, the board was "a strong, policy-setting, decision making entity and an advisory body" (34). At the same time, the statute creating the CNS required the board to be heavily involved in the approval process for AmeriCorps and Learn and Serve America grants. The NAPA panel recommended that both the board's effectiveness and the Corporation's operating units would be enhanced if the detailed grant approval role were removed from the board. But both congressional and board concerns about accountability militate against such a change.

Even the CNS's efforts to "leverage" government re sources by catalyzing public-private partnerships--a major "reinventing government" objective (Rainey 1997, 365)--have been problematic. Evidence that AmeriCorps has been able to decrease program costs by stimulating other resources from the private sector is largely anecdotal. City Year's founder Alan Khazei, for example, claims that between 1992 and 1997, City Year was able to use the $8 , million obtained through federal national service funding to leverage the same amount from the private sector (Cantigny Conference Series 1997, 95). But the Government Accounting Office (1997a} could not document over all trends in private contributions in their study of AmeriCorps reform efforts. In fact, critics charged that much of the money required to "match" CNS funds actually came from state or local government sources, as al lowed by law, if not necessarily by the goals of "reinventing government" champions.

Furthermore, the mix of federal and private (or other governmental) money has fueled a dispute over how much the program really costs. To critics of AmeriCorps, the amount required to support a full-time "volunteer" in the field was the total spent, which typically exceeded $25,000 per participant per year (GAO 1996). To its defenders, the cost was the federal share, while the matching funds-typically about 25 percent of the total--were a "benefit." Thus, by applying reinvention principles, CNS complicated--or from its detractors' viewpoint, obfuscated--how it accounted for the resources it used.

Examination of oversight hearings and the CNS's strategic plan for fiscal years 1997-2002 clearly demonstrates, however, the CNS commitment to reduce federal costs and to strengthen partnerships with the private sector, as well as with state and local governments. Indeed, a larger pro portion of nonprofits are expected to participate in the pro gram without receiving grants to pay living expenses for the AmeriCorps members they use. The members will continue to receive an educational award from the CNS on completion of their service. How much this will affect the CNS's ability to steer, but not row, should provide a good test of how dependent reinventing government is upon grant relationships, and how much federal leadership can be influential without them.

Creating Competition in Non-Market Contexts

The effort to improve quality by creating cross-state competition for nonformula AmeriCorps funds has fallen short of expectations. The amount of money involved has been less than planned because of action by a House-Senate Conference Committee that capped FY1998 national direct appropriations at about 50 percent of the level permitted by the authorizing legislation. The selection process has also been time consuming and cumbersome, requiring approval of winners by the CNS board (NAPA 1998) for some programs.

States have learned how to "game" the competitions as well (Smith and Jucovy 1995). Because they were given the latitude to categorize the programs they submit as either formula or competitive, states tended to put their strongest programs in their competitive application. States were also allowed to adjust their funding priorities after the CNS's decisions about competitive programs, leading them to move unfunded competitive grant, applications into the formula grant, either displacing or reducing funding for other programs (Smith and Jucovy 1995). This tactic may have lowered the number of weak projects in each state, but it also meant that projects judged of lesser value nationally were still able to obtain support.

Indeed, the competitive grant-making process raises issues about equity among different-sized states. In its study of 24 AmeriCorps programs administered by seven diverse state commissions, the General Accounting Office found that state commission officials differ on the extent to which the CNS should control the grant-allocation process (GAO 1997b). California officials, for example, claimed that a competitive grant-making process results in redundancies because states select their own projects for funding prior to the competitive process (15). Rhode Island officials, not surprisingly, argued that eliminating the competitive grant-making process in favor of formula grants only penalizes small states and rewards large states like California (15). For them, the competitive grant-making process did result in rewarding quality national service programs, because it took into account the variation among state commissions in their capacities to administer local AmeriCorps programs. But for larger states whose formula allocations were already sizable, competition seemed like an extra step, delaying them from supporting good projects until the possibility of funding from the national allocation was determined.

Whatever the merits of the competitive grant-making process, the reality is that, apart from President Clinton himself, the states have emerged as the principal 'political supporters of the CNS in Washington. The agency cannot afford to alienate powerful governors because they are its strongest and most consistent allies. As a result, despite the preference in the reinventing government movement for allocating funds through competitions rather than formulas, the key requirement for the CNS is to ensure the states are generally satisfied with the results.

How Much Decentralization Is Enough?

As Dwight Waldo noted in The Administrative State (1948), his classic study of the political theories of American public administration, the debate about centralization versus decentralization is as old as the modern state. The case of national service is no different. Robert Gordon, a young aide during the creation of national service legislation, expressed concern that the program could be too de centralized (Waldman 1995, 108). From the perspective of many grant recipients, the program is too centralized and bureaucratic.

Do these opposing perspectives suggest that national service has reached a "Goldilock's equilibrium," where there is a balance between centralization and decentralization? The reality may be that national service is simultaneously too decentralized and too centralized. This is the general thrust of recommendations in a recent National Academy of Public Administration (1998) analysis of the CNS.

Empirical evidence from Keith Provan and H. Brinton Milward's (1995) research on four community mental health networks suggests that "centralized integration and direct, non-fragmented control" proved to be more effective in clients' eyes than programs with highly fragmented, decentralized program designs (23). Furthermore, decentralization often leads to serious accountability problems (Kettl 1995; Romzek 1996).

The CNS discovered this the hard way when an auditability study demonstrated significant problems with its financial statements and management controls for FY 1996. Although initially attributed to the difficulties of absorbing ACTION into the new agency, the problems now appear to be caused by the decentralized operation of its programs. Grantees, often inexperienced in generally accepted reporting requirements and relying on informal management practices, simply lacked the records to verify--to the satisfaction of government auditors-how they had used CNS funds. Taking advantage of its hiring flexibility, the CNS brought in experienced private-sector leadership that, working with officials from the Office of Management and Budget, has slowly made important progress in correcting these weaknesses (GAO 1997a; Wofford 1998). But how much these steps will reduce the discretion, at least on management issues, of CNS grantees, or create administrative barriers that would-be grantees cannot surmount, remains to be seen.

Folding ACTION into the new agency, however, did make decentralized operations more difficult. Dating back to the 1960s, when its key programs were created as part of the "war on poverty," ACTION had an extensive net work of regional and state staff members, who de facto comprised much of the CNS field staff. Yet, because the programs they had been running (especially VISTA) differed in important ways from the new ones (such as AmeriCorps), relationships with both Washington and the staffs of the new state commissions proved problematic. Indeed, according to the NAPA (1998) report, a cultural gap developed within the agency, hampering efforts at project coordination and administration. To overcome it, the CNS has begun to move ACTION's regional offices into the agency's headquarters (though even that step, the NAPA team judged, did not go far enough to assure better performance).

Another force for centralization has been the application process. In order to receive a CNS grant, applicants must submit a proposal that not only meets the agency's own criteria of merit, but also conforms to federal requirements. This entails compliance with a formidable list of procedures, antidiscrimination rules, and the like, as well as monitoring by the CNS, other federal agencies, and possibly state agencies, too. Moreover, depending upon the program, the process itself may. entail reviews at two levels of government, plus an expert panel. Hence, the lead time for obtaining CNS support is lengthy. Coupled with funding uncertainties created by congressional opposition to AmeriCorps in particular, it is not surprising that many CNS grantees are well established agencies (or units of local government, such as schools) with resources of their own to fall back on.

In a program that seeks to mobilize a wide range of local volunteers, these procedural requirements are a significant handicap. Grantees that might be especially close to the grassroots and adept at reaching underserved populations find themselves left out--at least from the substantial amount of funding allocated nationally--largely be cause they lack the wherewithal to comply with federal rules. To be sure, this would have been the case even if no effort to reinvent government had been made. (Some of the agency's punctiliousness also stems from the heightened political scrutiny it has received.) However, that red tape remains in place, even in a supposedly reinvented agency, testifies to the extent to which generally accepted principles of public administration ("gappa's") continue to command allegiance (or at least grudging acceptance) from public officials who seek to make government more efficient, flexible, and accountable.

Results of What? For Whom?

Results-oriented government requires the organizational capacity to adequately define program objectives, establish performance criteria and measurement tools, and iso late program outcomes. As Aguirre International discovered in its first-year report of AmeriCorps state/national programs, few of the programs they studied had experience with this kind of thinking (Aguirre International 1997, 30). They also found that despite the CNS's requirement that programs demonstrate success in meeting their objectives, many of them showed great difficulty understanding, collecting, and using data for this purpose (31). Al though results-oriented government represents an attempt to increase government responsiveness to its customers, in practice, the amount of staff time spent collecting and analyzing program data may produce the perverse effect of losing sight of program outcomes.

Furthermore, the problems inherent in distinguishing between program outputs, intermediate accomplishments, and end outcomes makes results-oriented government even more difficult to achieve in practice. At the most rudimentary level, CNS programs aim to increase the number of people who are providing service to their communities. This can be readily measured and tied back to CNS efforts. However, this civic energy is supposedly a means to other ends, such as improving programs for tutoring children in reading, or cleaning up blighted neighborhoods. Success at these tasks is more difficult to assess and more likely to be affected by factors other than CNS programs. Yet, even these accomplishments are not always ends in themselves, but merely steps toward a larger set of goals, such as raising reading scores on standardized tests or making an inner-city community more habitable. Judging when such outcomes which are often the most salient to the public--have been achieved is no easy (or inexpensive) matter, nor is it possible (or often reasonable) to assume that even the best CNS efforts are apt to have more than minor influence over them.2 But if the ability of the CNS to affect these kinds of "results" is hard to assess, how worthwhile is what it gets done, and what should it be held accountable for?

Moreover, where one sits often affects how one judges. Especially in an agency as decentralized as the CNS tries to be, differences in perspective at the national, state, and local levels about outcomes suggest that measuring performance can result in different assessments about national service achievements. What matters to "customers" on the streets of Boston may be quite different from what "customers'' in the Mississippi Delta value, let alone what looks like a desirable (and realistic) goal to CNS officials in Washington. (And what matters to CNS officials responsible for VISTA may not be the same as what agency staff running AmeriCorps value.) By whose star should the re invented ship of government steer?

Compounding this uncertainty is the ambiguity of purpose with which the agency was born. Partly because of the century-old debate over it and partly as a result of the Clinton administration's efforts to build political support for it, the idea of national service has rightly been dubbed "the Swiss Army knife" of American public policy (Waldman 1995). Its advocates have, at various times, portrayed it as an instrument for solving critical social problems, helping college students Finance their educations, building harmony among racial and ethnic groups, fostering community planning, invigorating retirees, and instilling a sense of civic spirit--or "gratitude," as William F. Buckley, Jr. (1990) once put it--among the young. Not withstanding its own effort to focus on "getting things done," consensus on the CNS's principal mission remains elusive. (Moreover, congressional oversight is split between two sets of committees, one rooted in the ethos of the ACTION programs, the other attuned to the goals of the national service movement.) And even "getting things done" calls for further explication Which things? At what cost?--if it is to usefully guide a results-oriented management.

To its credit, the CNS has not backed away from these difficulties; instead, it has tried to employ a multifaceted approach to measuring its performance goals (CNS 1998, 23). It has also invested in technical assistance to local programs in program-evaluation techniques to help make results-oriented government successful. Still, whatever this effort may do to improve local capacity; problems of measurement and accountability will remain, because they are fundamentally rooted in the nature and complexity of CNS programs, as well as the political values that sustain them. Measuring results, in other words, is every bit as challenging for a reinvented agency as it is for a traditional one.

Three overarching lessons jump out from the experience of applying reinvention principles to national service. The first is that even reinvented agencies must adhere to some generally accepted principles of American public administration. Budgets must be balanced, accounting practices performed, merit roles honored, and federal contracting procedures obeyed. These principles have developed over time, often have their own constituencies (for example, civil rights groups, public-sector auditors), and, in effect, define "good government," at least procedurally. Despite repeated promises by political leaders to reduce such red tape, it has managed to persist, if not grow, long after the scandals, new management theories, or other generating events have been forgotten. The reason is simple: When the results of government activities are difficult to assess, a set of generally accepted principles of public administration offers at least a modicum of assurance of proper public-sector conduct, transparency, and discipline for poor performers. But inevitably, they limit the flexibility of public officials in implementing their programs.

The reinventing government movement aims to cut through this red tape. At least in the CNS case, it has had some success. Hiring procedures have been liberalized, greater competition injected into the grant-making process, and a relatively decentralized operating structure established. But even the leadership of a new agency committed to reinvention can bend traditional rules of public administration only so far. Especially in matters related to financial controls and awarding of grants or contracts, their desire for greater flexibility has not only proven futile, but has also engendered criticism. And even a reinvented bureaucracy eventually acquires an inspector general, chief financial of ricer, and congressional auditors who, no matter how committed they may be to an agency's mission, are themselves responsible for ensuring it abides by traditional rules. The public's interest in ensuring accountability by government agencies, expressed through these generally accepted principles of public administration, thus serves to restrict the steps necessary, at least in the minds of reinvention advocates, to achieve high-performance government.

The second lesson of the CNS experience is that the vagaries, arbitrariness, and unpredictability of democratic politics inevitably limit the possibilities for reinventing government (Rosenbloom 1993). Throughout its brief history, the CNS has been an agency immersed in political controversy. Its budget has been cut to zero and then restored; new pro grams have been unexpectedly added; priorities in its mission have changed; the tasks it sought to perform have grown more complicated; and quality control has emerged as a more critical issue. All of these developments have complicated the agency's ability to implement the principles of reinventing government, as did the political compromises initially made to bring the agency into being. At the core, no matter what administrative philosophy a government agency adopts, the political environment surrounding its programs will exert a decisive effect on any efforts to improve its performance. As Norton Long (1949, 257) observed, "the sources from which power is derived and the limitations they impose is as much a dictate of prudent administration as sound budgetary procedure.

To be sure, some reinventing government principles have aided the agency politically. Decentralization of the CNS's programs, for example, has helped create a constituency for them among the nation's governors, mayors, and non profit organizations. The criticism (and eventual elimination) of participation by other federal agencies in these programs suggests that if the CNS had not operated in a decentralized way, its survival would have been in even greater jeopardy.

However, other elements of the reinvention philosophy have contributed to the agency's problems. For example, by making the real cost of its programs harder to agree upon, or entrusting state agencies with more responsibility than some were able to handle, the CNS's administrative approach fostered questions and controversy rather than trust. Ironically, the agency's repeated promises to be a model of reinvented government created such a high standard for its programs that critics could readily find fault with them.

In contrast to traditional bureaucracies, advocates such as Osborne and Gaebler (1992) claim that a reinvented agency has an advantage in an environment that is unstable, complex in the tasks to be accomplished, and diverse in terms of what the public wants--in other words, in a politicized environment. Yet, the CNS experience casts doubt upon such a generalization. Some reinvention principles have been conducive to political success, while others have not. Ultimately, the success of a government program depends on the interplay of its administrative and political requirements.

Finally, despite the CNS's commitment to a new style in government, the third lesson that can be drawn from its history is how remarkably similar its programs are to past practices in assisting nonprofit groups. As Salmon (1995) and Smith and Lipsky (1993) show, the federal government has relied heavily, especially since the 1960s, on grants-in-aid to such organizations to deliver public services. The development of "third-party-government" (to use Salamon's phrase) supposedly permitted national initiatives to be tailored to the needs of local communities, while also expanding the base of political support for them. At the same time, it also increased the dependence of grant recipients on government funding and re-directed their priorities toward causes of interest to federal policy makers (Gronbjerg 1993 ). Notwithstanding efforts begun in the Reagan administration to reduce such spending programs, they remain a significant part of both the federal budget and the revenues of nonprofit organizations (though the types of aid and beneficiaries have changed).

Although they supposedly provide support for staff to perform the functions that local nonprofit organizations deem important, federal program priorities have played an increasingly important role in allocating CNS funds. More over, there is some evidence that CNS grantees are switching funds from their own programs to those activities for which federal assistance is available (Thomson and Perry 1998). Indeed, with the CNS planning to reduce the amount it provides for each AmeriCorps participant, the incentives for this sort of displacement may grow.

Perhaps because it is so similar to the way Washington has supported them in the past, state and local officials, as well as nonprofit leaders, seem content to accept this kind of aid. But if the CNS were to move aggressively toward more steering and less federal funding, the political benefits of reinventing government may turn out to be far less substantial than they currently are.

We must offer a note of caution about generalizing from this case. The CNS is different from other federal agencies in several respects. It is new, relatively small, and its mission is largely defined by a major presidential initiative. But rather than diminishing the probabilities for reinvention's success, these circumstances may actually work in its favor. Newness and smallness are likely to be conducive to change. The political nexus of the CNS has no doubt placed it in the metaphorical fishbowl. At the same time, the presidential support it has received has off set political attack. Regardless of the representativeness of this case, our purpose has not been to create encompassing generalizations. Rather, our goal was to provide a detailed description of this particular case as a means for testing conventional wisdom and creating a detailed record that can be compared and analyzed against other cases where reinvention principles have also been tried?

The Corporation for National Service is a new agency that has had anything but a peaceful childhood. Its senior leadership has had to devote its primary attention not to implementing the principles of reinventing government, but to building political support for the agency and its key programs. For the most part, its principal partners-gubernatorially appointed state commissions--are also new and still establishing working relationships with local non profit organizations and other potential grantees, including state governments. Its main initiatives are more developmental than mature.

As a result, despite its self-conscious determination to be an exemplar, the experience of the CNS should not be taken as a definitive reflection of the possibilities for rein venting government. The agency itself remains committed to serving as a model for a new style of public administration. With reauthorization of its programs pending, 4 followed shortly by a presidential election and perhaps an administration not so closely identified with the agency, the potential to establish broader political support for the CNS may be close at hand. If that occurs, the agency's ability to deliver on its ambition to be a different kind of bureaucracy could be more adequately tested.

Even so, the CNS record should instill caution among advocates of reinventing government. Despite the sound management insights that underlie its key principles, efforts to incorporate them into the public sector must come to grips with the fact that improved performance is not the only measure of government effectiveness. Conformity with long-established rules of administration, the views of politically influential officials and groups, and expectations created by previous programs matter as well. Government is less a matter of invention or reinvention than of evolution and compromise.

This is a revised version of a paper presented at the Symposium on Nonprofits and Government, sponsored by the Institute for the Study of Government and the Nonprofit Sector, Indianapolis, Indiana, May 28-30, 1998. The authors thank Patrick McGeever, Eugene McGregor, and Gene Sofer for their comments on an earlier draft.

1. President Clinton's remarks from the 1993 announcement ceremony at Rutgers University appear in Stephen Bates (1996, 60).

2. In its FY1999 performance plan, the Corporation acknowledges that "the end results of national service and volunteer activities are, in large part, the work of our grantees and their members and volunteers," thereby creating major challenges for performance measurement. See CNS (1998).

3. Graham Allison (1983) suggests the knowledge base of public management should be developed by precisely this method, that is, an accumulation of case evidence about public managers. Yin and Heald (1975) offer an alternative method for systematically codifying and comparing cases. Although this case has immediate utility as a method for disconfirming theory, we believe it should also be viewed in this broader context.

4. President Clinton transmitted the National find Community Service Amendments Act to Congress on March 19, 1998. It died when Congress adjourned later in the year. A new set of amendments is under discussion.

Aguirre International. 1997. AmeriCorps State/National Pro grams Impact Evaluation: First Year Report. Report submitted to the Corporation for National Service. Washington, DC.

Allison, Graham. 1983. Public and Private Management: Are They Alike in All Unimportant Respects? In Public Management: Public and Private Perspectives, edited by James Perry and Kenneth L. Kraemer. Mountain View, CA: Mayfield.

Barzelay, Michael. 1992. Breaking through Bureaucracy. Berkeley, CA: University of California Press.

Bates, Stephen. 1996. National Service: Getting Things Done? Chicago: Robert R. McCormick Tribune Foundation.

Buckley, William F. 1990. Gratitude: Reflections on What We Owe to Our Country. New York: Random House.

Cantigny Conference Series. 1997. National Service: Getting It Right. Chicago, IL: Robert R. McCormick Tribune Foundation.

Corporation for National Service (CNS). 1998. Annual Performance Plan for Fiscal Year 1999. Washington, DC: Corporation for National Service.

_:. 1993. First Quarterly Report. Washington, DC: Corporation for National Service.

_:. 1994. Principles for High Quality National Service Pro grams. Washington, DC: Corporation for National Service.

Clinton, Bill, and Al Gore. 1992. Putting People First. New York: Times Books.

Gore, Al. 1993. From Red Tape to Results: Creating a Government that Works Better and Costs Less. Washington, DC: National Performance Review.

Gronbjerg, Kirsten A. 1993. Understanding Nonprofit Funding. T 'San Francisco, CA: Jossey-Bass.

Kettl, Donald F. 1995. Building Lasting Reform: Enduring Questions, Missing Answers. In Inside the Reinvention Machine: Appraising Government Reform, edited by Donald F. Kettl and John J. DiIulio, Jr., 9--83. Washington, DC: The Brookings Institution.

Long, Norton. 1949. Power and Administration. Public Administration Review 9(Autumn): 257-64.

March, James G., and Johan P. Olsen. 1983. Organizing Political Life: What Administrative Reorganization Tells Us about Governance. American Political Science Review 77(2): 281-96.

National Academy of Public Administration (NAPA). 1998. Corporation for National Service: Strengthening Organizational Effectiveness and Service Delivery. Washington, DC: National Academy of Public Administration.

National Commission on the State and Local Public Service. 1993. Hard Truths/Tough Choices: An Agenda for State and Local Reform. Albany, NY: Rockefeller Institute of Government.

Osborne, David. 1988. Laboratories of Democracy. Boston, MA: Harvard Business School Press.

Osborne, David, and Ted Gaebler. 1992. Reinventing Government: How the Entrepreneurial Spirit is Transforming the Public Sector. Reading, MA: Addison-Wesley.

Perry, James L., Ann Marie Thomson, Mary Tschirhart, Debra Mesch, and Lee Geunjoo. 1999. Inside a Swiss Army Knife: An Assessment of AmeriCorps. Journal of Public Administration Research and Theory 9(April): 225-50.

Provan, Keith G., and Brinton H. Milward. 1995. A Preliminary Theory of Interorganizational Network Effectiveness: A Comparative Study of Four Community Mental Health Systems. Administrative Science Quarterly 40(1): 1-33.

Rainey, Hal. 1997. Understanding and Managing Public Organizations. San Francisco, CA: Jossey-Bass.

Romzek, Barbara S. 1996. Enhancing Accountability. In Hand book of Public Administration, 2nd ed., edited by James L. Perry, 97-114. San Francisco, CA: Jossey-Bass.

Rosenbloom, David. 1993. Editorial: Have an Administrative Rx? Don't Forget the Politics! Public Administration Review 53(6): 503-7.

Salamon, Lester M. 1995. Partners in Public Service: Government-Nonprofit Relations in the Modern Welfare State. Baltimore, MD: Johns Hopkins University Press.

Sandel, Michael. 1996. Democracy's Discontent: America in Search of a Public Philosophy. Cambridge, MA: Belknap Press.

Smith, Steven Rathgeb, and Michael Lipsky. 1993. Nonprofits for Hire. Cambridge, MA: Harvard University Press.

Smith, T.J., and L.Z. Jucovy. 1995. Launching AmeriCorps: First Year Implementation of the National and Community Service Trust Act of 1993. Philadelphia, PA: Public/Private Ventures.

Thompson, Frank J., and Beryl A. Radin. 1997. Reinventing Public Personnel Management: The Winter and Gore Initiatives. In Public Personnel Management, edited by Carolyn Ban and Norma M. Riccucci, 3-20. New York: Longman Publishers.

Thomson, Ann Marie, and James L. Perry. 1998. Can AmeriCorps Build Communities? Nonprofit and Voluntary Sector Quarterly 27(4): 339-420.

U.S. General Accounting Office (GAO). 1996. AmeriCorps* USA--First- Year Experience and Recent Program Initiatives. Washington, DC: U.S. Government Printing Office.

_:. 1997a. National Service Programs: Status of AmeriCorps Reform Efforts. Washington, DC: U.S. Government Printing Office

_:. 1997b. National Service Programs: Role of State Com missions in Implementing the AmeriCorps Program. Washington, DC: U.S. Government Printing Office.

Waldman, Steven. 1995. The Bill: How Legislation Really Be comes Law: A Case Study of the National Service Bill. New York: Penguin Books.

Waldo, Dwight. 1948. The Administrative State. New York: Ronald Press.

Yin, Robert, and Karen A. Heald. 1975. Using the Case Survey Method to Analyze Policy Studies. Administrative Science Quarterly 20(June): 371-81.

~~~~~~~~

By Leslie Lenkowsky, Indiana University-Purdue University

is professor of philanthropic studies and public policy at Indiana University-Purdue University at Indianapolis and a research associate of the Indiana University Center on Philanthropy. He was appointed to the board of directors of the Corporation on National Service by President Clinton and had previously been appointed by President Bush to the Commission on National and Community Service. The views expressed in this article are his own

is Chancellor's Professor in the School of Public and Environmental Affairs, Indiana University-Bloomington. Among his honor are the Charles H. Levine Memorial Award for Excellence in Public Administration and the Distinguished Research Award, both given jointly by the American Society for Public Administration and the National Association of Schools of Public Affairs and Administration. He is a fellow of the National Academy of Public Administration.