Leaders in the Field

The Public Administration of James Q. "Wilson:A Symposium on Bureaucracy ^̂ ^John J. Dilulio, Jr., Princeton UniversitySteven Kelman, Harvard UniversityChristopher H. Foreman, Jr., The American University and The Brookings InstitutionRobert A. Katzmann, The Governance Institute and The Brookings InstitutionPietro S. Nivola, The Brookings Institution

Among contemporary scholars who study the role of

bureaucracy in American government, Professor James Q.

Wilson stands out for the outstanding contributions he

and his students have made to the field. His latest work,

Bureaucracy: What Government Agencies Do and Why

They Do It, has received widespread acclaim among both

academics and practitioners. In this symposium, five of

Wilson's students offer their critical insights on their men-

tor's most recent book, fohnf. Dilulio, fr. focuses on

Wilson's "contrarian " approach and highlights his view of

the role of leadership in bureaucracy. Steven Kelman

notes the prescriptive message Wilson provides his readers

interested in bureaucratic reform, while Christopher H.

Foreman, fr. highlights the central role street-level "opera-

tors" play in Wilson's perspective. Robert A. Katzmann

outlines Wilson's message to the courts about their interac-

tion with bureaucracies. Finally, Pietro S. Nivola shows

how Wilson's work takes into account the role of "external"

factors on the operations of government bureaucracies.

T he publication of James Q. Wilson's treatise,Bureaucracy, What Government Agencies Do AndWhy They Do It (Basic Books, 1989), marks a mile-

stone in the public administration literature. To underscorethe importance of Professor Wilson's contribution to the field,a roundtable discussion of the book was held at last fall'smeeting of the American Political Science Association in SanFrancisco. The comments of the participants—five ofWilson's former Harvard graduate students—are summarizedin the following articles.

Notes on a %Contrarian Scholarby John]. Dilulio, Jr., Princeton University

For nearly three decades, James Q. Wilson has been anextraordinarily insightful, original, and (public choice theo-rists among others would hasten to add) contrarian scholar ofcomplex public organizations. Bureaucracy displays anenviable mastery of the relevant theoretical and empirical lit-eratures, an uncommon common sense, and genuine literarygrace.

Wilson argues something that the classic public adminis-trationists pretty much took for granted; namely, that "organi-zation matters." For him it remains an important and open-ended question whether there exists a nontrivial relationshipbetween how public agencies are structured, led, and man-aged, on the one hand, and how (and how well) they per-form, on the other.

Wilson's career path has also differed from that taken bymany authors of the classic works. They had significantexperience in government, and while there is no mistakingtheir intellectual fascination with their subject, their researchclearly had prescriptive aims. His three years in the Navy inthe 1950s aside, Wilson has never held a full-time govem-

Public Administration Review • May/June 1991, Vol. 51, No. 3 193

ment post. And in Bureaucracy, as in virtually everythingelse he has written in this field, Wilson's research motivationis staunchly nonprescriptive. He studies human behavior inpublic agencies and other complex settings because, as hewould say, it is "intellectually interesting." His Varieties ofPolice Behavior (1968), for example, analyzes how differentpolitical environments engender different policing practices;but the book is silent on the question of how to organizepolice so as to reduce crime.

It is unsurprising to find, in one so motivated by the intrin-sic joys of intellectual puzzles, a strong contrarian strain, animpatience with oversimplifications, and a cautionary, evenpessimistic outlook on the practical payoffs (if any) of generalknowledge about complex human affairs. This contrarianstrain is evident in many places. Consider the following para-dox. In one of his most important books in the criminal jus-tice field. Thinking About Crime (1975), Wilson takes seriousissue with criminologists, mainly sociologists, who focus onthe complex "underlying causes" of criminal behavior. Inplace of this sociological perspective, he offers what may bestbe described as an economic perspective: It is not so muchbad homes as bad incentives that cause criminal behavior.The typical criminal, he suggests, is best understood not as abyproduct of complex social forces but as a rational, self-interested actor. He thus embraces the common sense of thesubject. If crime pays, then more crimes will be committed.If the penalties for crime are made more swift, certain, andsevere, then fewer crimes will be committed. For decades,the dominant perspective in this field was sociological;Wilson bucked it brilliantly with a perspective drawn mainlyfrom economics.

But now turn to his public administration. Here Wilsondefends the rampans against those who are inclined to viewbureaucratic behavior as merely rational and self-interested, offer-ing instead a heavy dose of sociology, in particular organizationalsociology, to counter the idea that bureaucratic behavior isany sort of simple maximizing (e.g., budget-maximizing)behavior. Easily the most pointed example is chapter 10 ofhis 1980 edited volume. The Politics of Regulation. Of course,a fuller example is Bureaucracy itself, though in this bookWilson picks no fights with anyone; public choice theoristsare simply not directly addressed.

Which is not to say that Wilson is entirely happy with thepublic choice perspective. Undeniably, to understand gov-ernment bureaucrats as budget-maximizers is to understand agreat deal about bureaucratic behavior. To assume thatbureaucrats behave as i/they were engaged in such maximiz-ing behavior is to make a very simple and powerful assump-tion, one that not only comports with the facts of numerousstudies and the anecdotes of innumerable practitioners, butthat lends itself to formal models that are (depending on yourtaste and your math-readiness) intrinsically interesting in theirown right and capable of yielding other, often highly counter-intuitive, leads, clues, and empirically-testable hypothesesabout the behavior in question.

As I view it, Wilson's argument against this perspective onbureaucratic behavior revolves around several points, chiefamong them that, while the theory is powerful, and whilethere is no competing theory that explains "as much with aslittle," what the theory captures is less interesting and impor-tant than what it misses. In essence, public choice theoristsknow just enough facts about "what government agencies do"to misunderstand "why (and how!) they do it."

Wilson's key gripe with public choice notions of bureau-cratic behavior seems to me about the "unassimilated data":He thinks there are more such data than these theorists gen-erally allow, and that the discoverable patterns of real-woddbureaucratic behavior that lie outside the theory are moreinteresting and consequential than those that lie within it.Yet there is no logical or other contradiction between view-ing his gripe as 100 percent legitimate (as I do) and viewingwith great favor the work of public choice theorists in thisfield and in others (as I do).

Two classics of organizational theory seem to have helpedshape Wilson's theoretical perspective: Chester I. Barnard'sThe Functions of the Executive (1938), and Phillip Selznick'sLeadership in Administration: A Sociological Interpretation(1957). It is impossible here to say more than a few wordsabout the Barnard-Selznick-Wilson nexus. In brief, Barnarddefines organization as a system of consciously coordinatedactivity of two or more persons. Barnard's highly abstracttheory of how organizations arise, persist, and change is com-prehensive and compelling, and lends itself well to empiricalanalyses of large bureaucracies, both public and private. It isreflected in the three-pronged scheme that Wilson used inThe Investigators (1978); it helped organize key parts of hisHarvard bureaucracy courses; and it has now been elaboratedin Bureaucracy's framework: operators (part II), managers(part III), and executives (part IV). Essentially, operators per-form the central or critical tasks of the organization (line-levelemployees, "street-level bureaucrats"), managers coordinatethe activities of persons within the organization (foremen,supervisors), and executives are responsible for the mainte-nance of the organization (CEOs, commissioners).

In Bureaucracy, Wilson treats operators, managers, andexecutives in that order, and the order is not incidental.Rather, the order reflects one of his most fundamental andpersistent ideas about the determinants of agency perfor-mance; namely, that it is mainly the nature of operators'tasks—as opposed to the styles of the agency's managers, thebehavior and preferences of its executives, and the goalscarved into statutes by those who legislate and lobby over itsexistence—that ultimately determines "what governmentagencies do and why they do it."

As Wilson claims, until quite recently, Selznick's book onleadership made waves among analysts of business organiza-tions but was largely ignored by students of public adminis-tration. Essentially, Wilson takes from Selznick the idea thatto attract and retain persons capable of doing the work,bureaucratic organizations, public and private, must developa "sense of mission" or "distinctive competence," defined as a

194 Public Administration Review • May/June 1991, Vol. 51, No. 3

widely shared view among organization members (the bulkof whom—the operators—are on the lowest rungs of theadministrative hierarchy) as to the nature, rightness, andimportance of the organization's principal tasks (arrest crimi-nals, prevent escapes, comfort the sick). Many organizations,especially public ones, have multiple, vague, and contradicto-ry missions (prisons punish, deter, incapacitate, rehabilitate).Strong organizations, however, can be defined as ones thatresolve (in operational, if not logical, terms) any such con-flicts by how they recruit, train, reward, and manage workers.

By my reading, Wilson argues that public agency leaderscan be efficacious, but only by more or less dutifully obeying,not by bravely commanding, the unchanging administrativeand ever-changing political imperatives between which theyare sandwiched. Efficacious executive leadership may requireworkaholics with vision, but if Wilson is to be heeded—and Iknow he is not asking to be—then their administrative andpolitical vision had better be more pragmatic than transforma-tional, more logical than inspirational, more incrementalistthan fundamentalist.

But let me offer a partial dissent from what I take to be hisgeneral views on the leadership and administration of gov-ernment agencies. I believe that the annals of American pub-lic administration furnish many examples of efficacious lead-ers who transcend the limits imposed by tasks, on one end,and political context, on the other.

The conventional view is that the problems of governmentagencies are problems of goals, resources, and leadership,and can be fixed accordingly. I believe that there is less tmthin this "top-down" view than is commonly supposed, butmore truth in it—especially with regard to leadership—thanWilson allows. Tasks are "to an important degree the mostenduring, hard-to-change elements of the situation; adminis-trative procedure must adapt to tasks, not the other wayaround" (Wilson, 1978, p. 8). But under certain conditions-conditions which can be manipulated into being or fosteredby plan, and that do not depend on the accident of leaderswith unique and nontransferable talents—public er.;cutivescan alter tasks, and improve organizational performance inthe bargain. My mind runs to the perhaps insignificant stockof exceptional cases in which agency heads have been morethe political ventriloquist and less the political dummy; and, Isuspect that the ways and means of achieving this state ofaffairs in the public interest go above and beyond Wilson's"rules of thumb" for achieving agency autonomy.

Let me sketch one brief example. The Federal Bureau ofPrisons (BOP) is a far-flung federal bureaucracy that, over a60-year period, has maintained a remarkable degree ofadministrative uniformity and control at the line level in theface of enormous counterpressures of many kinds. For mostof its history, the BOP—the nation's safest, most programmat-ic, most cost-effective penal agency—has been an extraordi-narily close-knit organization with a strong sense of missionand a highly developed organizational culture. But that wasnot achieved by managing the agency around the tasks that

the operators most preferred; nor was it achieved by yieldingregularly to the constraints imposed by a generally hostileand unsupportive political environment. Instead, it wasachieved largely through transformational acts of executive,political, and administrative leadership, some of which tookthe form of highly visible public words and deeds, others ofwhich effected silent revolutions on the relevant behavior ofeveryone from workers in the cell blocks (who substitutedone sense of mission for another) to officials in the Congress(who went from indifferent or anti-agency positions to lavishsupport for the bureau).

Again, in believing that the BOP story and others like itmay be more than exceptions to the rule, I may be guilty ofletting the wish be father to the thought, a charge to whichWilson would never need to plead guilty. I recognize as wellthat my belief here may be explicable in the very terms oforganizational analysis that Wilson sets forth. After all, publicpolicy and public administration programs are predicated onthe assumption (unrebuttable, of course!) that public leaders,including agency heads, can make a major, positive, andeven transformational difference.

The Prescriptive Message oby Steven Kelman, Harvard University

With characteristic modesty, and reflecting a viewpoint hetraditionally has espoused regarding the usefulness for practi-tioners of his work, James Q. Wilson states in the preface toBureaucracy that "though what follows is not very theoreti-cal, neither is it very practical." The reader, he goes on towrite, "will not learn very much—if anything—about how torun a government agency" (p. xii). At the very end of thebook, Wilson manages to present only what he calls "A FewModest Suggestions That May Make a Small Difference" (pp.369-376).

As an academic engaged in training young people forpublic service at a public policy professional school, I believethat Wilson's book, and in particular his few modest sugges-tions, contains much wisdom about how to improve the per-formance of public organizations. It is also interesting thatWilson associates himself in the suggestions he makes withthe (I think growing) number of scholars and reflective practi-tioners of public management who argue that the manage-ment of public organizations relies excessively on rules andexcessively fears the exercise of discretion or judgment byexecutive-branch public officials—that we need, in theexpression of Constance Homer that Wilson uses to start hisdiscussion, to "deregulate the government."

According to this view, an important part of the problemwith the performance of public organizations derives fromexcessive reliance on rules as a way of organizing the workof government agencies. Despite their virtues and indeed thepractical necessity of their use in many situations, rules can

The Public Administration of James Q. Wilson: A Symposium on Bureaucracy 195

produce anomalous results when they are applied to inappro-priate situations or when the environment for which they hadoriginally been developed changes.

There is an important reason why we rely so extensivelyon rules in the management of public organizations: we fearthe exercise of discretion by those working for those organi-zations. That fear, in tum, has a number of sources. It partlyderives from standard democratic ideology that suggests thatthe coercive power of the state ought orily be exercised bydemocratically responsible officials, leaving civil servants toapply rules that have been developed by those closer to thedemocratic process. It partly derives from a worry that in theabsence of rules, public officials will do their jobs inappropri-ately. As Fred Schauer has written (1991), we may be willingto sacrifice poor decisions in some individual cases to avoidabuses we fear in other cases.

There are also reasons to challenge each of these argu-ments. Many political science and public administration schol-ars have made a career out of challenging, on empiricalgrounds, the notion that it is possible to avoid frequentlylocating significant discretion in the hands of the bureaucracy.This is either because, as students of regulatory delegationhave noted. Congress and the White House cannot cope withthe large number of specific and often technical decisionsthat government makes or because, as students of "street-level" bureaucracy have noted, it ends up being impossible todevelop rules to handle all the contingencies that people atthe operating levels face. Wilson himself, in eadier bookssuch as Varieties of Police Behavior (1968), has made impor-tant contributions to making clear the empirical problemswith the suggstion that it is possible to remove discretionfrom the actions of public organizations.

The empirical challenge is, of course, consistent with thenormative proposal that discretion be eliminated as far as fea-sible from the behavior of public organizations. Indeed,many of those scholars who have been at the forefront of theempirical demonstration of the unavoidability of discretionhave tended to regard their findings as evidence of a prob-lem, or even of a "challenge" to democratic theory. Wilsonhimself, in a subordinate clause in Varieties of PoliceBehavior, refers to "controlling discretion" as "the problem ofadministration" (1968, pp. 96-97). "Irreducible discretion" ispresented in this view as if it were the name of some dreaddisease. The empirical demonstration that it is often difficultto eliminate discretion is perfectly consistent with the systemwe have now, where discretion is widespread, but illegiti-mate, and we are always looking for new ways to fight it.

But there are challenges to the doctrine on grounds otherthan impossibility. From the point of view of democratic the-ory, it may be argued that citizens might appropriatelychoose to delegate significant infiuence (although not totalcontrol) over the process of developing specific practices in apolicy area to people who are both expert in the area andshare the basic values that undergird the organization's

enabling statutes. In principle, a democratic meta-decision toso delegate need be no more problematic than any otherdemocratic decision—indeed, it would be undemocratic toforbid such delegation.

From the point of view of effective organizational perfor-mance—a concern closer to Wilson's in Bureaucracy—it maybe argued in criticism of the anti-discretion arguments thatthe cost in forgone opportunities for excellence and achieve-ment that come from the fear of discretion simply outweighany gains in terms of discouraging imcompetence or abuse.In an age when In Search of Excellence preaches the impor-tance of initiative and responsibility, government must notcontinue to be locked into state-of-the-art management prac-tice from the tum of the century.

Wilson himself is probably not overly optimistic that analternative conception of how one should manage in thepublic sector can be realized. While I think everyone wouldagree that wild optimism on this score would hardly be calledfor, I am inclined toward somewhat more optimism than Ithink Wilson is. I have several reasons for this.

First, I believe there exists an alternative program that canbe presented to those who worry that increasing the freedomof bureaucrats to exercise discretion would deprive citizensand elected officials of the opportunity to hold bureaucraciesaccountable for their behavior. The program is: "judge orga-nizations by results, not by whether they followed bureau-cratic rules." Wilson presents this as one of his proposals forreform, but I think it's necessary to undedine that this is notsimply another proposal but a crucial feature of the overall

Strategy. I am of course aware of the problems with develop-ing results-oriented performance measures for many publicactivities. But I believe that, first, these problems are some-times made to appear bigger than they are and, second, thatwe can begin in areas where it is relatively easy to developmeasures and save the hard problems for later.

Second, I believe that people in government organizationsmight be disinclined to embrace the notion that their substan-tive performance be held up to scaitiny, except that the alter-native may be worse—mind-deadening reliance on rules,focus on scandals, and so forth.

Last, I believe that at least a significant part of the reasonfor the excessive reliance on rules involves an intellectual fail-ure to understand that the fear of discretion extracts a highprice in terms of poor substantive government performance,especially as the work force becomes more educated and thesubstantive problems government faces, more complicated. Idiscovered in a study of public-sector procurement that mostof the major political participants in debates about procure-ment (erroneously) believe that the rule-bound governmentprocurement system serves not only to safeguard againstabuse but also, by lowering the prices government pays, toimprove substantive performance as well. A part of the battleagainst the fear of discretion in government managementmust be waged in the realm of ideas.

196 Public Administration Review * May/June 1991, Vol. 51, No. 3

operatorsby Christopher H. Foreman, Jr.,The American University and The BrookingsInstitution

In the preface to Bureaucracy, James Q. Wilson disclaimspractical relevance. Even so, near the end, he makes a per-haps unavoidable stab at thinking about reform, remarks hecarefully announces as "A Few Modest Suggestions That MayMake a Small Difference." I want to argue that a good dealof what comes earlier can be of practical interest as well, par-ticulady if one believes, as I do, that educating people abouthow to regard the structures and processes that affect theirlives may be at least as important as concrete reform. Mymain focus here is part II, the section of the book dealingwith what Wilson calls "operators."

One valuable aspect of the book is that it compels anyintellectually honest reader to reexamine some common prej-udices in the blinding glare of hard evidence. In chapter 4,on "Beliefs," Wilson exhibits a pronounced skepticism towardthe linkage between "generalized attitudes," as he calls them,and individual operator behavior. He then shows himselfvery supportive of the notion that more precise and germanebeliefs may be very important indeed as behavioral influ-ences.

The thesis that the attitudes of officials do not necessarilydictate their behavior is, interestingly, both ancient andprovocative. It is ancient in being supported, as Wilson him-self notes, with reams of data on noncorrelation that haveaccumulated over decades of social science research. It is aprovocative idea not only because it is somewhat counterin-tuitive but because it invites us to question some widespreadpopular impressions of a highly controversial set of agencyinteractions—namely, encounters between officials of onerace and clients of another.

Let us take a recent example. It is interesting that, in all ofthe commentary in the local and national press on the recenttrial of Washington, D.C., Mayor Marion Barry on drug-relatedcharges, almost no one has thought to mention the ratherobvious point that one possible factor behind the zealouspursuit of the mayor by federal prosecutor Jay Stephens wassimply that prosecutors generally face strong incentives toprosecute. Their professional training and peer-group cuesmay incline them to hunt assiduously for "prosecutable" per-sons and, having found them, to invest considerable effort inbuilding and presenting the strongest possible cases againstthem. For a prosecutor with higher political ambitions, or forone who craves merely the respect that comes with havingcaught a "big fish," it may make sense to go after a high pub-lic official who seems vulnerable . And even thoughStephens' effort ultimately failed to yield a felony conviction,there can be little doubt, in retrospect, that the mayor'slifestyle had left him vulnerable. One does not have to con-struct an elaborate paranoid theory of a white power struc-ture driven toward unhinging Black officialdom to explain

why the prosecutor did what he did. By the same token, itwould probably come as a surprise to some Black organiza-tion employees, clients, and customers to learn that thebehavior of those whites with whom they must deal may beaffected by important forces other than deep-seated racialattitudes.

Now, having said this, one must add something thatWilson perhaps ought to have stressed more. For it is alsodoubtless true that there exist gaps in official structures ofrewards and penalties that may allow generalized racial atti-tudes to translate, in unsavory ways, into racially-drivenbehavior. Moreover, it is possible that the structure ofrewards and penalties may, in a given set df circumstances,be simply too weak to mute the impact of racial attitudes.And it is also possible that an organization's system ofrewards and penalties may operate symbiotically with pre-existing racial attitudes. (One supposes that the behavior ofthe South African police could be an interesting subject inthis regard.) Of course, if any or all of these things are true,there is no reason why they should not, ceteris paribus, applyequally to whites and Blacks alike. So Wilson's discussion, ifwidely disseminated, could have some practical use, if onlyby making its readers, of whatever race, more inclined tothink rigorously about where racially motivated behavior is,and is not, likely to crop up.

Wilson's book also has practical relevance in the consider-ation of administrative and public policy reform. Like manyothers, Wilson clearly believes that it makes a difference howyou structure a process, who executes it, and who overseesit. Following Richard Cyert and James March (1963), Wilsonsees large organizations as often composed of distinctive andsometimes competing subunits, perhaps with attachments todifferent tasks, cultures, and incentive systems. Regulatoryagencies are clearly not all dangling on the fine edge abovethe abyss of inevitable "capture" by business interests. As heargued a decade ago in The Politics of Regulation, patterns ofaccess and influence by private interests depend on manythings, and outcomes vary accordingly. He reprises some ofthis argument in the new book. Agencies, in Wilson's view,are not empty vessels into which attentive business interestscan simply pour their policy preferences for later reclamationin pristine form. But, alas, nor can anti-business reformershope to do so either.

One of the most important points in the book (as, Ibelieve, in the Harvard course from which it is heavilyderived) is the message that administrative "goals" are oftenno match for an array of other factors (incentives, institutionalcultures, situational imperatives, etc.) in shaping how agen-cies and the individuals in them perform. As a teacher, Ihave long thought that this is one of the most importantlessons that any student can learn in an introductory publicadministration course. The imponant role of incentives oftencomes as a genuine revelation to students. Though Wilson isclearly suspicious of self-interest as the sole mover of thingspolitical, he is fully prepared to grant it importance. Indeed,his application of this latter perspective in one particular fieldof policy (namely, crime and law enforcement) has earned

The Public Administration of James Q. Wilson: A Symposium on Bureaucracy 197

him a wide reputation as a conservative. But in noting that"crime pays," that efforts at penal rehabilitation yield uncer-tain payoffs, and that incarceration offers real benefits, Wilsonhas offered a point of view that has always seemed to me toderive less from prior ideology than from its opposite—apragmatic willingness to confront unpleasant realities head-on.

Wilson and the Study ofLaw and Policymakingby Robert A. Katzmann, Tbe GovernanceInstitute and Tbe Brookings Institution

James Q. Wilson has been a major influence on many whostudy law and policymaking. Chapter 15 of Bureaucracy isdevoted to an examination of courts. But Wilson's impact onthe field of law and policymaking extends beyond his con-cern with courts to his analyses of the politics of regulationand the workings of the administrative process—what wethink of as administrative law.

Wilson on CourtsIn an era in which the judicial role in public policy has

greatly expanded, the link between Wilson's work and thefield of law and policymaking should be clear enough.Courts have been actively involved in the administration ofpenal, welfare, education, mental health, and environmentalprotection policies. In remedying perceived wrongs, theyhave had to interact with the public bureaucracies operatingin the particular policy arenas affected. Judges, in so doing,have not sought merely to monitor the activities of these insti-tutions but have at times attempted to restructure them, tochange their processes and policies. Yet, as Wilson observes,although there have been countless essays on whether thecourts should intervene as they do, scarcely any describewhat systematic (as opposed to particular) changes in bureau-cratic life this intervention has produced. The challenge, ineffect, which Wilson has issued to those who study law andpolicymaking, is to not simply be concerned with legal doc-trines and constitutional development—though they areimportant—but also to understand the effects of those doc-trines on administrative behavior. I will return to that pointin a while. In that approach, he is allied with Martin Shapiro,who is as much concerned with administrative process aswith constitutional development.

Wilson identifies a number of effects of court interventionon agency management, in terms of cost, power, and policy.As to cost, intensive court intervention can increase the timeand money it takes to reach a decision. Judicializing agencyprocess takes time and money. The costs of the judicializa-tion of agency discretion are not limited to the internal man-agement of the organization—"when two organized groupsknow that it will be the court, not the agency, that has the

final say on the issue that divides them, the two parties haveno incentive to negotiate a settlement with the agency."

With regard to power, Wilson observes that court proce-dures, like any political arrangements, tend to give morepower to some and less to others. Access to the courts isexpensive and requires legal skills; thus as courts becomemore important to bureaucracies, lawyers become moreimportant in bureaucracies. That was certainly true, for exam-ple, in the case of transportation policy for people with dis-abilities. Court intervention, as Wilson writes, may sometimesempower professions other than that of lawyers. "Lawyersmay gain to the extent procedural requirements become com-plex, but other specialists gain to the extent the courts tend tofavor one substantive outcome over another." Witness thecase of the FTC and the role of economists. The point to bemade is not simply that courts can affect bureaucracies, butthat they can affect the power structure within a bureaucracy,sometimes aiding one part of an organization as opposed toanother. Indeed, the judiciary can infiuence outcomes bysupporting or altering the balance of forces, among bureau-cracies. Take, for example, the case of HEW's coordinatingrole with respect to disability policy, specifically, section 504of the Rehabilitation Act of 1973.

I would note that as we think about how courts affect thepower of bureaucrats, it is not always necessarily the casethat courts are positioned against bureaucrats. In some situa-tions, bureaucrats who seek increased funding from the chiefexecutive welcome court suits—indeed, sometimes arealigned in a de facto way with interest groups to engineercourt suits. If those bureaucrats, as defendants, lose the suit,they really win the suit. For then they can say to their politi-cal executives: "we have no choice but to spend moremoney, or change policy. The courts made us do that."Wilson's treatment of bureaucracy, not as a black box ormonolith, enables us to penetrate organizations, to see themix of strategies within agencies. In short, his frameworkallows us to understand the often complicated relationshipsbetween bureaucracies and courts.

Judicial intervention can affect not only costs and power,but also policy. A court's ruling can bear directly upon sub-stance—for example, when it either mandates that specificsteps be taken or holds that an agency is not required toundertake certain actions advocated by one or more of theparties. Judges can affect process, as they have in "hybrid"rulemaking cases, by imposing a wide range of proceduralrequirements on agencies engaged in informal rulemaking.When the judiciary interprets a statute or mandates orrestrains executive action, it perforce affects the administra-tive process. Just as the judiciary can bear upon the balanceof forces within Congress, so its decisions can influence therelationships among agencies or competing factions withinthe bureaucracy, or the capacity of the president to infiuencethe executive brahch. Policymaking can thus be thought ofas dynamic and complex, as a continuum of institutional pro-cesses—judicial, legislative and administrative—sometimesacting independently, but often interacting in subtle and per-

Public Administration Review • May/June 1991, Vol. 51, No. 3

haps not always conscious ways to influence the behavior ofother processes.

Wilson modestly eschews claims that his work has muchpractical value. I beg to differ with that assessment. • For Ivery much believe that Wilson's framework can be quite use-ful as a court seeks to fashion a remedy, or as a courtendeavors to understand both the opportunities for changeand the limits of organizational capacity. Let me allude tosome of these organizational concepts—organizational mis-sion, goals, power structure, tasks, autonomy versus coordi-nation, external forces (for example, legislatures)—and seehow they might be instructive to a judge.

A court that understands the concept of mission will rec-ognize that change will be difficult in an organization that hassuccessfully instilled a sense of mission: those who believe inthe bureaucracy's mission are likely to resist judicial efforts toalter it. Indeed, professionals whose behavior and policyperspectives may be constrained by norms derived fromexternal reference groups that allocate rewards such asadvancement and esteem, are particularly likely to resistattempts to reorient the organization in a way that conflictswith their conception of what the bureaucracy should bedoing. Moreover, a court should realize that it will more like-ly achieve its goals if it defines those ends unambiguously sothat the relevant actors will understand what is required ofthem. A court should make certain that the proposed remedyprovides the technology for attaining those goals.

A judge who understands a bureaucracy's power structurewill more likely fashion a decree that strengthens those forceswithin the organization that are more likely to comply withthe court's order. Appreciation of organizational tasks is alsoimportant. Routine tasks, for example, may be relatively easyto change because it may be necessary only to alter the pro-gram. Changing nonroutine tasks, however, could compli-cate implementation of a plan: the court, for example, mustestablish new monitoring devices to acquire performancedata and then must react to that information.

In addition, a court must be prepared to meet resistance ifit seeks to coordinate the behavior of bureaucracies that arefearful of losing autonomy. These difficulties will increase ifthe bureaucracies disagree about goals or the way in whichgoals should be achieved. In such circumstances, the courtmust invest its resources in an attempt to forge cooperativeactivity among the relevant actors. Finally, a judge who seeksto design a remedy should consider the extent to which itseffectiveness depends upon the support of external actors.Dependence upon external forces over which a court mayhave little control—for instance, a legislature that is not subjectto the suit—^may make it difficult to implement the remedy.

Wilson and Adoiinistrative LawI have concentrated on the relationship between courts

and bureaucracies. In conclusion, I would like to retum to apoint I made earlier: Wilson's work has much relevance tothose who, following Martin Shapiro, define the field of pub-lic law not simply in terms of courts or legal doctrines, but

the full range of variables within the administrative processwhich account for outcomes—^whether they be decisions toprosecute or not to prosecute, whether they be particularkinds of adjudicative decisions, rules, or regulations. TheSupreme Court's decision in Chevron v. NRDC (1984), calledthe most important decision in the field of administrative lawin the last decade, underscores the importance of Wilson'scontribution. Essentially, the Chevron line of cases holds thatwhere an ambiguity exists in a statute administered by a par-ticular agency, courts will defer to that agency's constructionso long as it is permissible. This deference is based on theSupreme Court's assumption that because Congress hasentrusted the administration of the law to the agency, itexpects the agency, rather than the courts, to resolve ambigu-ities. How an agency thus makes law in interpreting ambigui-ties can be understood by understanding the organizationaldynamics about which Wilson has written. It is a concem,not simply of Bureaucracy, but also the Politics of Regulation(1980).

Interbranch andInternational Dimensionsby Pietro S. Nivola, The Brookings Institution

On the jacket of the book. Senator Daniel P. Moynihanobserves that this is more than another treatise on bureaucra-cy; it's really a whole course, taught over a generation, spin-ning off (and spun from) some 50 Harvard Ph.D. disserta-tions. We could be having symposiums, not just on parts andchapters, but on paragraphs and sentences, even footnotesand asterisks. That said, the pages on which I'll comment areamong the most nettlesome conceptually. I think theyrequire special attention for the following basic reason: Thebook begins with the central thesis that bureaucratic organi-zation matters. But when we get to the discussion on therole of Congress, the president, the courts, and national dif-ferences, we come away keenly aware that organization mat-ters but often less than some other things.

I do not wish to imply that this amounts to a flaw orinconsistency in the book's main argument. To the contrary,I would suggest that Professor Wilson's contribution is inbeing careful not to insist that bureaucracy in the Americanregime is in any sense a "fourth estate." Put another way,while he wants us to get inside government agencies andunderstand their foibles from the bottom up, he also stressesthat much of what these agencies do is defined or infiuencedfrom the outside in—by "the context," that is, the politicalstructure and culture in which they operate. If this may seemlike a self-evident proposition, recall how often other scholarsseem oblivious to it. Ever since it became fashionable to takequintessential cases of congressional or presidential decision-making (episodes like the Cuban missile crisis, for instance)and interpret them as examples of bureaucratic activity, a lotof confusion has been generated about who actually doeswhat in the American policy process.

The Public Administration of James Q. Wilson: A Symposium on Bureaucracy 199

The bottom line is that certain cultural eccentricities andinstitutional peculiarities (like the formidable U.S. Congressand federal courts) sharply delineate bureaucratic behaviorand power in this country. To understand why particularfederal agencies do what they do, we can scrutinize thenature of their tasks and internal organizational characteris-tics, but ultimately we are left with the realization that othernational bureaucracies, with very similar organizations andsometimes identical task structures, often act altogether differ-ently.

Steven Kelman's comparative study of occupational safetyand health regulation in the United States and Swedengraphically illustrates the difference. Our OSHA inspectorstend to operate in a legalistic, adversarial mode, "going bythe book." Their Swedish counterparts exercise much morediscretion, enforcing the rules in a flexible, more consensualstyle. This is not because the Swedish regulators have a fun-damentally different task on their hands, or because their for-mal goals are not the same, or because their operational stan-dards are unique. In the final analysis, it is because bureau-crats in places like Sweden don't have to live with the hotbreath of congressional rulemakers and overseers on theirnecks, and because they do not have to take cover from theconstant threat of court action (or other contestation by oneor another interested party). Other studies in comparativepublic administration (f'or example, David Vogel's 1986 workon regulatory administration in Britain and Japan) come tomore or less the same conclusion.

Wilson, of course, is at pains not to overstate the influenceof Congress. He advances highly nuanced arguments abouthow Congress exerts more control over "production" agencies(with clear goals, tasks, and outcomes) than over so-called"procedural" and "craft" agencies, and how Congress takes agreater interest in agencies like HUD that distribute local ben-efits than in "presidential" agencies (like Treasury) that don't.The book critiques Barry Weingast and company for not rec-ognizing that bureaus must answer to multiple politicalactors, not just a dominant legislature. Readers are remindedthat congressional dominance is also delimited by divisionswithin Congress and by the tendency of legislators to tie theirown hands. And arguably, the modern Congress engages inless particularistic pork-barreling with the bureaucracy. Yet, Iwould submit, the burden of the book's evidence in both thischapter and the one on the presidency is that while thebureaucracy serves multiple masters, of these mastersCongress is first among equals.

I say this for several reasons. The first is that the newcongressional micromanagement represents, I believe, a defi-nite net loss of executive power (both presidential andbureaucratic) at the hands of Congress. It also represents anet gain in political security on Capitol Hill. By devisingdetailed rules and procedures (often attached to annualauthorizations or appropriations bills), the legislators continueto whittle away at executive discretion while still supplyingbenefits to particular constituencies. Granted, base closingsare no longer blocked on an ad-hoc basis by telephone callsto the Secretary of Defense but rather by generic rules (such

as environmental impact statements), and, yes, individualfirms or industries now get trade protection through univer-salistic-sounding administrative remedies rather than particu-laristic items in a legislated tariff. But the "pork" is still get-ting out there to specific groups—only now with the addedadvantage of political cover for Congress, since the favoritismis less transparent.

The same goes for various methods by which Congressputs the bureaucracy on automatic pilot. One way to con-template indexation, for instance, is by thinking of it as anexercise in congressional self-limitation. But a better way tothink of it, I suspect, is as a clever way of dictating to theexecutive exactly how to spend money while enabling mem-bers of the House of Representatives not to look like profli-gates every two years.

When it comes to programs not governed by formula,people often forget that Congress gets much of what it wantsthrough the anticipatory reactions of the departments andagencies. Just because Arthur Maass (1983) observes thatCongress does not normally name individual projects inauthorizing statutes, nor ordinarily earmark funds to them inappropriations measures, doesn't mean that those projectsand dollars don't wind up where the members want them.As R. Douglas Arnold (1979) and various others have shown,agencies build legislative coalitions for their budget requestsby dispersing benefits in ways that Congress appreciates.

Now, there is no question that Congress has an easier timedominating some bureaus than others. The Social SecurityAdministration is easy prey; the Food and DrugAdministration, the Office of Civil Rights, or the Federal TradeCommission are not. But I must confess, I've become some-what less persuaded by the contention that the relative easeor difficulty of interference has to do primarily with inherenttask characteristics of the agencies than I am by the implica-tions of legislative factionalism. Tasks, after all, are notentirely exogenous or independent variables. For better orworse, even complex and ambiguous administrative tasks canand do get defined by somebody (often, when all else fails,by the courts).

The question is: If judges are prepared to make thesedeterminations, why cannot Congress? To which the obviousanswer is: Judges are able to make up their minds, whereascross-pressured legislatures often are not. But from time totime, it is not unimaginable that Congress could resolve someof its differences and tell a body like the FTC: "Look, fromnow on, we want you to crack down on mergers of such-and-such a type and scale, and we want you to report so-and-so many prosecutions of these cases annually." In otherwords, I am just not sure how much the nature of administra-tive tasks necessarily conditions and constrains congressionalinvolvement. Disagreements in the legislature, it seems tome, may be the main underlying constraint. The extent ofdisagreement, however, is not immutable; it waxes andwanes depending on a variety of circumstances (not the leastof which are the intensity of partisan politics, the stage of theelectoral cycle, and the prevailing political Zeitgeist in

200 Public Administration Review • May/June 1991, Vol. 51, No. 3

Washington). Foreign policy, for instance, used to be the typeof hot (or unappetizing) potato that Congress would buckover to the executive. No more. Nowadays, there is a lotmore consensus on the Hill that micromanaging the StateDepartment, the Pentagon, the CIA, the U.S. TradeRepresentative and so on, is fair game.

There is one more point I want to make about the role ofCongress: Congress casts a lengthening shadow over moreagencies precisely because White House control is shaky.Most of the reasons for this shakiness are clearly laid out inthe book. Presidential appointees tend to "marry the natives"(often even after they've been subjected to Reagan-style ideo-logical litmus tests). Administrative reorganizations are lesseffective and feasible than ever. (Among other things, afterthe INSv. Chadha, they now require positive acts ofCongress.) Managerial reform has largely meant centraliza-tion of administrative details in the White House, not bettersupervision of agencies by the department heads of betterpolicy guidance from the president to those departmentheads. Interagency coordinating mechanisms (like theNational Security Council) have often had the same result.There is not much I would add to this baleful list of deficien-cies, except maybe for the following fundamental considera-tion: Unless we have seen the end of an era of one-termpresidencies (and of contested renominations), congressionalincumbents will continue to loom as the institutionalized

players inside the Beltway with more staying power than thechief executive, who, from the vantage point of folks in thebureaucracy, is somebody who could be here today and gonetomorrow.

As I stressed at the outset, one of Wilson's main messagesis that the federal bureaucracy is far from being a fourthbranch of government; it enjoys a lot less autonomy and flex-ibility than the administrative entities in some other advanceddemocracies. It may seem paradoxical, therefore, to leamthat American businessmen often seem to spend more timelooking over their shoulders at the IRS, the SEC, the DOJ, andvarious other regulators and potential litigants than theyspend, say, figuring out how to compete with the Japanese.(Surely, no other business community spends as much onvigilant corporate attorneys.) The main point to remember,however, is that this is not because the U.S. bureaucracy ismore powerful, differently organized, or has entirely distinc-tive tasks and functions to perform. Rather, it is the otherway around: our bureaucracy is frequently on the defen-sive—continually prodded by litigation, by congressionalintervention, and by other outside participants. This condi-tion, more than anything else, is what is truly unique aboutpublic administration in the United States. And James Q.Wilson has given a masterfully eloquent and cogent analysisof it.

• • •

References

Arnold, R. Douglas, 1979. Congress and the Bureaucracy. New Haven; YaleUniversity Press.

Barnard, Chester, 1938. The Functions of the Executive.Chevron V. Natural Resources Defense Council, 1984. 81 L. Ed 2d 694.Cyert, Richard and James March, 1963. A Behavioral Theory ofthe Firm.

EnglewocxJ Cliffs; Prentice Hall.Immigration and Narturatization Service v. Chadha, 1983. 462 U.S. 919Maass, Arthur, 1983. Congress and the Common Good. New York; Basic

Books.Peters, Thomas and Robert H. Waterman, Jr., 1983. In Search of Excellence.

New York; Harper and Row, 1982.

Schauer, F., 1991. Playing by the Rules. Oxford: Oxford University Press.Selznick, Phillip, 1957. Leadership in Administration: A Sociological

Interpretation. New York; Harper and Row.Vogel, David, 1986 National Styles of Regulation. Ithaca; Cornell University

Press.Wilson, James Q., 1968. Varieties of Police Behavior. Cambridge, MA; Harvard

University Press.,W5.Tbinkingy^out Crime NewYoifc Basic Books (revised ed, 1983).,1978. The Investigators. NewYork; Basic Books.,&i.,\9S0. The Politics of RegulaHon. NewYork; Basic Books., 1989. Bureaucracy: V!%at Government Agencies Do and Why

They Do It. NewYork; Basic Books.

The Public Administration of James Q. Wilson; A Symposium on Bureaucracy 201