This is not a book about the environment. Rather, it is an effort to improve existing models of bureaucratic behavior, especially in regard to how bureaucrats relate to the larger political environment. It relies on data from questionnaires from the U.S. Environmental Protection Agency and the New Mexico Environment Department.

Richard Waterman, Amelia Rouse, and Robert Wright are especially concerned with presumptions made by principle-agent models, but they also include a discussion of the theory of budget-maximizing bureaucrats (p. 10). In regard to the principal-agent model, they argue that it is improved by relaxing “the assumptions of an information asymmetry and goal conflict. We seek to treat those concepts as variables rather than as constraints” (p. 22). Unfortunately, the authors never clarify if these are dependent, independent, or interceding variables. This severely limits the explanatory power of the theoretical changes they propose. For example, the authors “hypothesize that many different relationships can exist between principals and agents” (p. 31). However, they do not present a hypothesis that specifies an independent and dependent set of variables. The closest they get is on page 65: “… [O]ne can hypothesize that the greater the level of expertise involved in an agency's functions, the harder it will be for principals to monitor a bureaucracy's behavior, and the more likely it will be that an information asymmetry exists.” They operationalize expertise by counting the degrees held by bureaucrats. This hypothesis is tautological; both variables are measures of the possession of specialized knowledge.

The authors make another attempt to present a hypothesis on page 68: “[A] high level of discretion also is likely in situations where agents take advantage of an information asymmetry.” Again, there is no effort to indicate which variables are independent and dependent. Later, they discuss the “causes of this discretion” (p. 74), but causation cannot be identified without an understanding of which variables are independent. The authors find evidence that the bureaucrats who responded to their questionnaire “see themselves as having discretion” (p 77). However, their data do not reveal if this level of perceived discretion is typical of most agencies, or even if it is higher in agencies with information asymmetry. Without a comparative context, the authors cannot explain the role of discretion and how it might affect the theoretical assumptions they are discussing. They argue that “we need to model a variety of political principals at both the state and federal levels when we examine the question of who controls the bureaucracy” (p. 92). This is certainly true, but without an identification of causative factors, we cannot reach conclusions about who controls whom.

After a lengthy discussion of the data, Waterman, Rouse, and Wright offer two sets of conclusions regarding the principal-agent model and the budget-maximizer model. In regard to the former, “we need to consider not only different types of political principals, but also the different ways bureaucrats are likely to respond to them” (p. 111). Or in other words, the models variously labeled as issue or policy networks, subsystems, or advocacy coalitions are more useful than dyadic principal-agent models (see p. 128). In regard to budget-maximizing, the authors conclude that their “findings undermine the basic budget-maximizer assumption that personal and bureau utility is perceived as the same thing” (p. 125). This conclusion is based on data that indicates that many of their respondents did not know if their agency's budgets had increased or decreased in recent years. It is somewhat of a leap of faith to interpret such data to mean that bureaucrats are not interested in how their agency's funding affects them personally.

Despite the authors' failure to adequately specify independent and dependent variables, their overall conclusion is an important one: “[W]e recommend the development of theories that provide a more dynamic interaction between multiple principals, multiple agents, and agents motivated by a variety of goals. We are especially critical of theories that include normative bureaucrat-bashing assumptions” (p. 130). This book does not make a significant contribution to the development of such a theory, but it does point out avenues of future research.

The last section presents what the authors call “a list of twenty commonly advanced characteristics of the bureaucratic process, which we present as ‘myths’” (p. 130), and they posit that these myths have been “commonly expressed in the bureaucratic literature” (p. 133). Unfortunately, none of them is accompanied by citations, and so we do not know who is perpetrating them. But it is presumptuous to make such claims based on evidence from just two agencies. On the basis of the evidence presented in this book, we have no idea if these agencies are widely representative of bureaucracies in general, or they are two outlying aberrations. Thus, it is not possible to generalize from their findings and use them to declare large segments of the bureaucratic literature as mere myths.

The authors ask some very important questions that have puzzled many scholars, and they attempt to answer some of them. I admire them for setting such a difficult goal for themselves. However, Bureaucrats, Politics, and the Environment provides precious few answers to these questions, and ends by making assumptions that are far beyond what can be sustained by the evidence presented.