This text consists of papers originally presented at the Fifth Public Management Conference meeting at Texas A&M University in December, 1999. The editors in the introductory chapter aptly define the challenges facing contributors: “We know little about how bureaucratic (a) agencies make decisions … (b) structures affect responsiveness and performance … (c) agencies resolve goal conflict on a day-to-day basis, (d) external performance criteria affect bureaucracies and (e) organizations socialize their members” (p. 15).

The editors assert that embracing these formidable challenges collectively constitutes a “new organizations” approach to the study of public administration. This approach is “bureaucratic” and can be understood as an extension of the “new institutionalism,” which focused more on analysis of impacts of democratic institutions (the presidency, the Congress, and the courts) than on bureaucratic behavior and decision making: “Bureaucratic agencies … contain their own incentives, procedures and preferences for administering public policy” (p. 306).

As might be expected, the papers vary in focus and content. Any academic interested in the nexus between government and bureaucracy will find several of them interesting. The editors ordered papers under “theory,” “methodological technology,” and “empirical studies” labels.

In the theory section, Daniel Carpenter's analysis of the optimal stopping point in the Food and Drug Administration's drug approval regimen is well articulated and sophisticated. To this reviewer, George Krause's treatment of agency demands for discretion treats these demands as a unity, failing to recognize that each bureaucratic level seeks discretion from other levels. One of Carpenter's assumptions, that increasing task complexity is linked to decreasing demands for agency discretion, is counterintuitive. As Thomas Hammond states in his piece on “veto points,” “bureaucrats may know more than elected officials about what needs to be done and how to do it” (p. 74)—which supports a claim that greater task complexity leads to greater demands for agency discretion. The final paper in the theory section is David Spence's analysis of agency discretion from the public choice perspective. Public choice hewed to a simple principal-agent theory. The people and their elected representatives (principals) instructed agents (agencies), so public choice theorists initially had little sympathy for notions of agency discretion. Spence clarifies well those occasions when people might prefer agencies, rather than their elected officials, to solve public problems.

Under the methodological technology label, John Brehm et al. provide an enlivening break from high-level theory by reporting analysis of the 1977 Police Services Study using the Dirichlet distribution. The authors report that police spend more time on official duties when under direct observation by supervisors and fellow officers. Andrew Whitford develops a model of a “baseline agency” and poses three “levels of redirection: comparison, punishment and imitation” (p. 161). This piece is abstract and could better be placed in the theory section.

Several of the five papers in the empirical section merit at least a brief mention. Steven Balla and John Wright's study of federal-level consensual rule making asks if this procedure more quickly results in a rule, as compared with the APA's usual review and comment procedure. After examining 170 rules issued by “nearly forty agencies,” they conclude that consensual rule making does not yield quicker results. A careful reading of this paper will give readers a better understanding of the challenges facing scholars researching federal procedures.

Lael Keiser's explanation of the Supplemental Security Income (SSI) program is the clearest read by this reviewer. Even with refined formulation of the independent variable, only meager statistical support is given for logical and anecdotally supported relationships explaining differences in SSI approval rates by state bureaucracies.

Kevin Corder's study of the numerous federal credit programs finds that some are operated by cabinet departments, others by independent agencies, and others by “government sponsored enterprises”—and only the latter entities were designed to be substantially removed from partisan political influence. Generally, government enterprises can borrow from the private sector, while other entities must rely on appropriations for funding. Federal credit entities use their funds for direct lending and for loan guarantees. The Congress treats current loan guarantee costs as nondiscretionary and currently initiated guarantees as spending deferred to a future date. After careful formulation of the independent variable, Corder's conclusions are limited. He concludes that lending by all federal credit entities is responsive to election results and congressionally mandated structural changes.

Kevin Smith, analyzing data from a 1998 survey of teachers of kindergarten through grade 12 finds that when compared to teachers in public schools, private school teachers place a much higher priority on promoting moral/religious values and only slightly more emphasis on encouraging academic excellence.

Certainly the call for more rigor and a more empirical focus in public administration research is laudable. The editors' excellent overview of recent research leads this reviewer to conclude that their distinction between the “new organizations” approach and earlier approaches to public administration research is overdrawn and a bit of a straw man. Perhaps the fault lies more in the “unit of analysis problem” or a type of myopia within political science.

Empirical research in any discipline requires a clearly defined unit of analysis, linked to innovations in some units but not in otherwise comparable units. Decisions are made by individuals and can be aggregated, as in Supreme Court rulings and elections. By definition, a bureaucracy has multiple members and levels of authority. It is with good reason that bureaucratic decision making is sometimes referred to as a “black box.”

The editors of and contributors to Politics, Policy, and Organization are exclusively focused on American phenomena. No attention is given to earlier attempts to better understand bureaucentric decision making in a comparative context. As part of the Alliance for Progress, the U.S. Agency for International Development formented preparation of national development plans in most Latin American countries. Albert Waterston commendably compared these efforts. Fred Riggs led a Ford Foundation effort in the 1970s to develop a subfield of comparative administration.

Good bureaucentric research may also exist in other fields. This reviewer suspects that much research has been done by educationists comparing responses of school districts to state and federal mandates. Health care academicians may have compared the responses of hospitals and health maintenance organizations to federal requirements. Likewise, law enforcement research may have focused on varieties of local responses to the federal Law Enforcement Assistant Act and other mandates.

Compared to others interested in bureaucracy and governance, the contributors to this volume and to the new organizations approach are empirically oriented and methodologically sophisticated political scientists. Collectively, these studies support a claim that the move from case studies to middle-range theory in the study of agency structure on output will be challenging indeed.