MSU RETURNING MATERIALS: P1ace in book drop to LIBRARIES remove this checkout from , your record. FINES wil] be charged if book is returned after the date stamped be10w. W“ m mam m4gwqguuuu _M¢7 kP:A~:7 I F.";:.:"1’r “"35”.” l r 29% ““1 Mi - 'W b '83 1-» PRIVATE INFLUENCE AND THE NATIONAL LABOR RELATIONS BOARD BY Ronald Carl Heacock A DISSERTATION Submitted to Hichigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 1983 ABSTRACT PRIVATE INFLUENCE AND THE NATIONAL LABOR RELATIONS BOARD BY Ronald Carl Heacock The ability of private interests to influence the decisions of regulatory agencies has long been a subject of keen interest. However, studies have focused on commissions where a single interest dominates the policy space. This study looks at the National Labor Relations Board, a commission in which two strong private interests dominate: the policy space. Conventional wisdom founded on pluralist thought has; long maintained that two opposed interests will balance each Iother eliminating the problem of private influence. This study seeks to explore this balanced state and to further understand private influence under such conditions. Three channels of influence utilized by private interests are investigated in the study - direct influence, budgetary influence, and appointment influence. These are the most commonly studied channels when exploring private influence. In the general case where one private interest exists the following results prevail. Direct influence is usually'considered the most important with groups controdling information and expertise. Influence exerted through the appointment of commissioners with policy Preferences like those of the interest is second. Influence obtained through the control of the commissions budget is third. The study relies heavily on the use of quantitative data to test hypotheses concerning private influence. Data on regional level NLRB decisions is employed to test propositions on direct influence. Data on the NLRB's budget and workload are employed to test the budget influence. Data on major board decisions is used to look at appointment influence. The findings indicate that private influence does exist in the balanced environment with both interests using cost effective channels to further their policy positions. However, the channels are different in terms of their importance. Obtaining favorable appointments is the most important channel with groups lobbying both the President and the Senate oversight committee. The budget process becomes a cooperative effort with both groups supporting the NLRB's budget and forcing the public to pay for the resolution of conflict between them. Direct influence aPpears also to be somewhat important although the findings are mixed. Some regions where one group is stronger than another are open to direct influence. However, most regions are not open to direct influence. To My Parents ii ACKNOWLEDGMENTS I would like to express my gratitude to the members of my dissertation guidance committee, Professor Gary Miller, Professor Jack Knott, and Professor John Aldrich. Their assistance in the preparation of this dissertation and the great amount of time they devoted throughout my doctoral program deserves special thanks. I would also like to express my thanks to Professor Richard Block for his assistance with labor law and Harriet Dhanak whose computer expertise made the work possible. My wife Carol deserves my special gratitude for her patience and understanding throughout, and her moral support along the way. iii TABLE OF CONTENTS List of Tables List of Figures Chapter I. INTRODUCTION The NLRB - A Deviant Case Footnotes to Chapter I II. REVIEW OF THE REGULATORY LITERATURE Macro-Theories of Private Influence Summary of Macro-Level Theories Influence Channels Regulatory Appointments Budgetary Oversight Direct Influence Summary Footnotes to Chapter II III. THE NATIONAL LABOR RELATIONS BOARD History of the NLRB Current Structure of the NLRB Macro-Level Theories and the NLRB Summary Footnotes to Chapter III IV. INFLUENCE CHANNELS AND THE NLRB Direct Influence Budget Commission Appointments Summary Footnotes to Chapter IV V. RESEARCH DESIGN METHODOLOGY AND DATA Statistical Methodology Hodels Direct Influence Budget Process Commissioner Appointments Footnotes to Chapter V iv Page vi vii 157 VI. RESULTS Direct Influence Budget Commissioner Appointments NLRB Results Final Results Footnotes to Chapter VI VII. FUTURE CONSIDERATIONS Footnotes to Chapter VII APPENDIX A APPENDIX B APPENDIX C BIBLIOGRAPHY 160 160 172 181 192 199 208 210 225 227 229 232 236 LIST OF TABLES Table Page III-1 NLRB Workload 1953-1980 70 VI-l Percent Pro-Management ULP Rulings At The Regional Level - 1966 Cases 161 VI-2 Percent Pro-Management ULP Rulings At The Regional Level — 1970 Cases 162 VI-3 Percent Pro-Management ULP Rulings At The Regional Level - 1974 Cases 163 VI-4 Percent Pro-Management ULP Rulings At The Regional Level - 1977 Cases 164 VI-S NLRB Regional Level Summary 168 VI-6 NLRB Budget Model Results 174 VI-7 Presidential and Congressional Budgetary Allocations for the NLRB 180 VI-8 Commissioner Voting Averages 182 VI-9 Presidential And Senate Appointment Influence On NLRB Appointments 185 VI-lO Appointment Impact Hodel 188 VII-1 U.S. Union Membership 224 vi LIST OF F IGURES Figure Page IV-l Cooperative Budget Model 103 IV-2 Case Decision Model 114 V-l Ordinal Dependent Variable Error Terms 152 vii I INTRODUCTION The purpose of this dissertation is to explore the problem of private influence on the National Labor Relations Board (NLRB) during the post Taft-Hartley era. This period covers the commission from approximately 1952 to the present. Specifically, this research will investigate a number of the nonlegislative channels of influence which private interests are purported to utilize as a means to achieve their policy preferences. Nonlegislative refers to any action other than the passage of statutory legislation. These influence channels can be divided into two general categories- direct and indirect. The former refers to immediate contact between private interests and the commission, that is through agency hearings, supply of information, etc. The latter refers to those influences which flow through the formal political institutions, such as the Congress and the Presidency. Examples of these are appointments of commissioners, appropriations, and oversight. While this research is concerned primarily with the question of influence as it relates to the alteration of agency decision making, and is only secondarily concerned with the ways in which private influence is used to obtain legislation which establishes or alters the statutory 2 authority of a commission, it is conceptually impossible to separate the two. Problems of private influence and the origin and amendment of enabling legislation are important concerns and they will be addressed. However, this will be done only for the purpose of establishing the context of the research. There will be no attempt to resolve these controversies. The research design and methodology of this study will rely heavily on the use of quantitative research techniques to test empirically the propositions outlined regarding private influence. The data and measures created from them upon which the analysis is based will be performance measures of agency or other institutional decision makers. Over the last thirty years a number of macro-level theories have been developed, and numerous studies have been conducted on the various linkages upon which these theories are founded. However, these studies have for the most part relied on impressionistic case studies to verify their hypotheses. This has led Bernstein, Stigler and others to call for more quantitatively oriented studies to provide empirical evidence for the confirmation or rejection of competing theories and hypothesesR' In addition, the majority of the quantitative studies performed to date have relied heavily on impact data or attribute data to determine the degree of private influence. The former is especially true of the studies performed by economists, the latter by political scientists and governmental studies. This study 3 relies on the use of quantitative data built on measures of agency or institutional performance. This research obviously falls within the context of a large literature on federal regulation. Currently the subject of federal regulation is generating a tremendous amount of interest in academic and political circles. The costs of regulation, the impacts of regulatory programs, as well as calls for deregulation, are important issues of the day. This renewed interest and reconsideration of prior attitudes on regulation has followed a phenomenal period of growth in federal regulatory programs. During the decade of the 1970's Congress enacted one-hundred and thirty regulatory statutes. Federal budgetary allocations for regulatory programs increased by over five-hundred percent in nominal dollars and the number of employees engaged in regulatory programs tripled}2 These costs represent only a fraction of the total costs of regulation for society, for the bulk of the costs are related to compliance with federal regulation. It has been estimated that in 1976 alone the costs of compliance were approximately twenty dollars for every dollar budgeted for regulatory programs;3 The current interest in regulation and the conflicting positions which arise are by no means a unique phenomenon. The subject has arisen periodically since the creation of the Interstate Commerce Commission in 1887. In this period of almost one-hundred years the study of federal regulation has been subjected to voluminous treatment by economists, 4 political scientists, jurists, journalists, and government analysts. Despite this massive effort, few if any of the controversies surrounding federal regulation have been resolved. Closely related to the general subject of federal regulation is the govermental institutional form known as the independent regulatory commission- the insitutional form of the NLRB. This is not to say that all federal regulation is administered by independent commissions, or that all independent commissions are strictly regulatory in their missions and functions. Rather, it is to say that the independent commission form has developed as a major institutional form in conjunction with federal regulatory programs. Further, the bulk of the studies on federal regulation have centered on this institutional form and many of the questions surrounding federal regulation involve questions concerning the independent commission. One controversy surrounding federal regulation, and the general subject and concern of this dissertation, which is inextricably intertwined with the subject of independent commissions, is that of private influence. This subject is identified by a variety of other labels, such as constituency influence, industrial influence, capture, producer protection, and clientele influence to name just a few. Generally, private influence has come to mean the ability of a regulated group (usually industry) to achieve favorable policy outputs from the regulatory commission 5 which has jurisdiction over it. The debate over private influence, especially that of organized groups, runs deep in American political thought and in the political ideology of the American public. The discussion of factions by James Madison in Federalist 10 and 51, and the development of philosophical pluralism demonstrate the intensity of these views and concerns. Madison's definition of factions as inherently bad, selfish and Opposed to the interests of the public or community is present even today.4 President Carter's frequent attacks on what he termed the special interests echo Madison's commentary almost two-hundred years later. The subject of private influence is closely linked to the independent commissions owing to the promise they originally held for halting the influence of unwanted special interests relative to other governmental forms. The creation and development of the independent regulatory commission form was strongly legitimized as an institutional means to eliminate undue private influence in the public policies of the federal government. The structural features, organizational placement and statutes of these commissions were seen by the Progressives, reformers, and classical public administration scholars as an institutional answer to powerful economic and partisan political interests.5 Although most of the commissions were created after this era, the logic of these earlier thinkers lived on in the minds of those who later created the commissions.6 6 First, most of the commissions require that the board of commissioners be bipartisan. This feature no doubt arises from the essential distrust of political parties as exemplified in the writings of the Progressives.7 Second, the commissions reflect the fundamental Progressive and classical public administration notion of the politics/administration dichotomy}3 The commissions are institutionally separated from Presidential influence in that presidents cannot remove commissioners once they are appointed. Finally, the independent regulatory commissions represent the Progressive faith in administration by neutral experts through an application of the laws of administration. Free from political influence and outside private sector pressures, and in some sense responsive to public opinion, the commissions with expert knowledge would be able to pursue the public interest? .As McConnell notes: “In the high tide of Progressivism confidence in impersonal expertise took on an almost milennial tone...LaFollette asked rhetorically, How has it been possible that both the people of Wisconsin and investors in public utilities have been so greatly benefited by this regulation? Simpley because regulation is scientific Herbert Croly argued, I'The administrative commissions only do right. Just as soon as they go astray the bonds tighten on them. They derive their authority from the knowledge, and their peculiar relation to public opinion."10 Likewise, economists writing during this period saw the advent of regulation as a means to ensure the social responsibility of business in the market place. The market was viewed as I'extremely fragile and apt to operate very 7 inefficiently."ll Government intervention through regulation of the economy was designed primarily to protect consumer interests. In the economics literature this is generally referred to as the consumer protection hypothesis.12 The widely held public interest conceptions of the origin and administration of regulation has largely disappeared. The dominant public interest conception has been replaced at worst by the juxtaposed position of capture- the total domination of an agency's policy preferences, or at best by a high degree of private influence. The argument is so pervasive in the literature that one author recently stated that, 'No industry offered the opportunity to be regulated should decline it.‘13 That agencies are subject to external influences, especially from the regulated interests is a given. The debates over private influence are primarily concerned with the degree, the origins, the channels of influence and the actors (variables) involved in the process. A Deviant Case - The NLRB Given the plethora of work which has been accomplished in the area of private influence and the independent regulatory commission, it is at first surprising that so little of this work and effort has been directed towards the NLRB. A recent massive study on federal regulation and the independent regulatory commissions performed by the Congress 8 mentioned the NLRB only in passing.14 This disinterest is especially true for political scientists. The research domain in the study of the NLRB has been dominated by two groups of scholars, lawyers and labor industrial relations specialists. The only area that has interested political scientists has been the study of the passage of the primary labor acts and legislation, and the role of labor and management groups in this process. Virtually no work has been directed at influence on the agency proper. The primary reason for this disinterest other than by those with a substantive interest in the agency is probably that the NLRB is considered a deviant case. Unlike other independent regulatory commissions it has a reputation which is considered exemplary. It is often mentioned as an example of a commission that works. Its perceived success in some sense has relegated it to the position of an uninteresting organization. The point of interest then becomes the underlying reasons for this perception. The deviant case status which has led to such little notice is that the NLRB is an example of an independent regulatory commission in which Opposing interests exist, those of management and labor. In this sense, the NLRB is considered to be in a state of balance or political equilibrium. According to the traditional notions on private influence developed by pluralist thought everything should run smoothly. Therefore, greater interest should be focused on commissions where this balance does not exist or on those 9 which attempt to artificially recreate this natural system of checks and balances provided by opposing interests. A second reason for this lack of interest by political scientists probably relates to the fact that there is no real question of the public interest to be explored. Grant McConnell has refered to the labor management policy space which contains the NLRB as an example of the politics of autonomy.15 In other words, the general public has little involvement in the policy space. However, it is for these very reasons that the NLRB deserves further study by political scientists and others concerned with the role of the federal government in labor management relations in particular, and regulatory commissions in general. The assumption that the NLRB is in some sense in a state of balance deserves investigation. For what do we really know about this perceived state and does it really exist? Are the channels of private influence which labor and management groups employ to alter policy decisions somehow less interesting simply because they may tend to counter balance one another? Such attitudes are the product of our assumptions regarding the influence of groups in this society resulting from a long tradition of pluralist thought. Should these assumptions prove to be completely correct, this study would at the least provide further corroboration for this belief. There are also a number of other benefits from a study of this type which will prove useful for the study of the 10 common pattern of private influence and the commissions. A good example of this is the reform literature on the independent regulatory commissions. In the past ten years a number of reform proposals have arisen which seek to eliminate undue private influence by the regulated industries through the creation of so called public interest groups. These groups, it has been argued, will balance the influence from the regulated industries and protect the public interest. Recently, the state of Michigan passed legislation requiring public funding of these types of groups in the area of utility regulation. Similar proposals at the federal level also exist under a number of different formats.16 The result of these reforms may be better understood if more is known about this state of balance. An additional benefit concerns the opportunity this study presents to investigate the question of private influence in an environment in which the problem of defining the public interest does not exist. In other studies the researcher is forced to make value judgments when defining the public interest. This must be done in order to determine the degree to which the commission is meeting its obligations to serve the public interest. The problems this poses are obvious. Two researchers investigating the same commission with identical results can differ completely in their interpretations based upon their prior definitions of the public interest. In this study,the Opposing interests are well defined in advance and the problem is alleviated. 11 The following section outlines the remaining chapters of the dissertation and provides a brief description of each chapter. CHAPTER 11. REVIEW OF THE REGULATORY LITERATURE: This chapter outlines the general literature on private influence and the regulatory commissions. It describes the theoretical perspectives which dominate the literature and draws from its discussion the major factors which enable us to describe the NLRB. Finally, it reviews the findings which cover the three channels of private influence to be investigated within the context of the NLRB. CHAPTER III. THE NATIONAL LABOR RELATIONS BOARD: This chapter provides the reader with the history of the NLRB and provides information on its structure. The reader with a working knowledge of the commission may want to skip the first part of the chapter. The second part of the chapter describes the commission in terms of the major factors gleaned from the general theories of private influence outlined in the first chapter. This provides the context for testing the channels of private influence as they relate to the NLRB. CHAPTER IV. INFLUENCE CHANNELS AND THE NATIONAL LABOR RELATIONS BOARD: This chapter investigates each of the influence channels within the context of the NLRB. It develops hypotheses regarding each influence channel and outlines the models to be tested in the next chapter. 12 CHAPTER V. RESEARCH DESIGN METHODOLOGY AND DATA: This chapter provides information on the research design to be employed in the dissertation. It Operationalizes the models developed in the previous chapters and outlines the statistical methodologies to be employed in the testing of the models. Finally, the data sources employed, and the methods for gathering the data are considered. CHAPTER VI. EMPIRICAL RESULTS: This chapter will outline the results obtained from the statistical tests performed on the models generated earlier in the dissertation. Each of the models will be discussed, as well as the overall findings generated by considering the results as a whole. CHAPTER VII. FUTURE CONSIDERATIONS: This chapter will consider the implication of the results obtained in the last chapter for future study. Simply, what questions should be considered as a result of our findings. Further, the problems which occured in this study may require additional work. APPENDICIES: This section will contain any pertinent information which could not be easily integrated into the text. FOOTNOTES 1. The call for more quantitatively oriented studies is the norm in the regulatory literature. Virtually all of the major theoretical studies over the past thirty years have issued this call. For example, see Marver Bernstein, "The Regulatory Process: A Framework for Analysis," Law and Contemporary Problems, 26 (Spring,1961), pp. 329-346. 2. Ronald Penoyer, ed., Directory Of Federal Regulatory Agencies (St. Louis: Center for the Study of American Business, 1981), pp. 4-14. 3. M. Wiedenbaum and DeFina, The Costs of Federal Regulation of Economic Activity (Washington,D.C.: The American Enterprise Institute, 1978), p. 3. 4. James Madison, “Federalist 10' and "Federalist 51' in The Mind of the Founder: Sources of the Political Thought Of James Madison, ed. by M. Meyer (Hanover, New Hampshire: University of New England Press, 1973), pp. 88-95, 129-133. 5. Grant McConnell, Private Power and American Democracy (New York: Alfred A. Knopf, 1966), pp. 35—48. 6. Harold Seidman, Politics, Position and Power (New York: Oxford University Press, 1975). 7. Most federal independent regulatory commissions (approximately seventy percent) still require bipartisan boards, as do most state regulatory commissions. 8. Woodrow Wilson, ''The Study of Administration,” in Classics of Public Administration, ed. by Shafritz and Hyde (Oak Park, 111.: Moore Publishing Co., 1978). 9. Carl Friedrich, I'Public Policy and the Nature of Administrative Responsibility,’ in Bureaucratic Power and National Politics, ed. by Rourke (Boston: Little, Brown and Co., 1978). 10. McConnell, op. cit., p. 43. 11. Richard Posner, “Economic Theories of Regulation,“ Bell Journal of Economics and Management Science, 5(2) (Autumn,l974), p. 36. 12. William Jordan, "Producer Protection, Prior Market Structure and the Effects of Government Regulation," Journal of Law and Economics, 15(1) (April,l972), p. 165. 13 l4 13. Owen and Braeutigan, The Regulation Game (Cambridge, Mass.: Ballinger, 1978), p. 2. 14. 0.8., Congress, Senate, Committee on Governmental Affairs, Study on Federal Regulation, 95th Congress, 1st and 2d sessions, 6 vols., 1978. 15. McConnell, op. cit., pp. 298-335. 16. Public Participation in Regulatory Agency Proceedings, Vol. III of Study on Federal Regulation, op. cit.. II REVIEW OF REGULATORY LITERATURE The literature which deals with the subject of private influence and the independent regulatory commissions can best be described as eclectic. In terms of its accomplishments, the literature is best summarized by Paul Quirk. '... the literature on the sources of industry protection constitutes an embarrassment of riches. There is a plethora of suggested explanations, causes, and contributing factors. The reason why this outpouring does not decisively relieve our ignorance is that we have very little idea which, if any, of them are true, and to what extent." The apparent reason for this state of affairs in the regulatory literature is that numerous experts in various disciplines, each with their own implicit perspective of the world, have approached the subject of regulation and written extensively on it.2 What emerges when these strains of thought are combined is a collage of thought that is nearly impossible to categorize. This chapter will first outline the macro-level theories of private influence. These theories have attempted to draw meaningful generalizations which apply to all independent regulatory commissions. Secondly, the factors which these theories explicate will be considered. Finally, the non-legislative influence channels will be considered. 15 16 MACRO-THEORIES OF PRIVATE INFLUENCE The theoretical framework most often associated with private influence and the regulatory commissions is the capture or life-cycle theory of Marver Bernstein.3 This theory is essentially a general description of the developmental process of the regulatory commissions, from nescency to decay. The cycle is divided into four phases: gestation, youth, maturity and old age. The life-cycle describes the “general pattern of evolution more or less characteristic of all (commissions)."4 Gestation characterizes the period prior to the establishment of the regulatory commission. It is a period of social struggle with ensuing legislative action and bargaining which results in the formulation of legislation authorizing the commission. The period is characterized as one of mounting pressure for reform from the general public and organized public interest groups for the regulatory form, while those who are the objects of regulation oppose it. Catalysts such as scandal, economic crisis, etc., can randomly spark the process. Only after the problem becomes severe will the legislature take action. However, the resulting legislation is often vague and incorporates both the desires of the organized interests as well as those of public interest groups and general public opinion. 'The statutory mandate lacks clarity. Although it climaxes a prolonged struggle for reform, the policy which it establishes rarely provides clear directions for a new agency...agitation for l7 regulation rarely produces a first statute that goes beyond a compromise between the majority favoring and the powerful minorities opposing regulation.'5 The remaining three phases of the life-cycle describe the development and decay of the formal commission. While the development is broken into three distinct phases, they collectively describe an incremental process of an agency which is slowly but steadily captured by its organized regulated interests. In youth the commission is characterized as zealous and aggressive. However the commission finds itself without political support, without experience and in a hostile environment dominated by strong organized regulated interests. Maturity is a period of adjustment to conflict in the environment. The agency becomes more passive in its enforcement and regulation and becomes more of a manager of the regulated industry. Old age finds the agency in decline. Finally recognized as a protector of the industry it faces budgetary cuts from Congress. The last three phases of the life-cycle differ from the first phase in that the conflict has moved from the purely legislative level to one of bureaucratic conflict and development. That is, while the legislative battles have ended, the attempts to control agency policies through different channels persist. While Bernstein does not provide a distinct model for the process, he does mention a number of factors which interact to result in the incremental process of capture. 18 The commission finds itself politically isolated from the other major institutions in the government. This isolation makes it dependent on the regulated industry for survival. The President has been institutionally separated (as noted earlier) and the Congress reinforces this isolation by viewing regulatory commissions as agencies of Congress and not part of the executive branch as is the case for other bureaucratic institutions. The Congress however cannot (or will not) provide support for the pursuit of the public interest. To do so would be to anger the strong concerted interests of the regulated industries, and this is Obviously politically unwise. That is, to actively oppose these organized interests would result in political costs without any corresponding benefits. The constituency configuration of the regulatory commission is one of strong organized interests (primarily the regulated industry) which control channels of information and monopolize expertise needed by the commission. Public interests are largely unorganized and decline after the passage of the commissions enabling legislation. Finally, Bernstein includes a number of organization variables, i.e., internal organization factors, which add to the process. These include problems with expertise, budgets, rising workloads, backlogs, and highly judicialized proceedings. Organizational rigidity results from these factors and the procedures adopted to deal with them. 19 While not Offered as complete studies on the subject of capture, a substantial amount of research in the political science literature has explored parts of the process as described by Bernstein. Anthony Downs' work Inside Bureaucracy, provides a life-cycle for bureaucratic agencies in general which is similar to and strongly supportive of Bernstein's theory of capture.6 Studies on Congress and its relationship to the bureaucracy in general and to the commissions in particular have likewise provided support for Bernstein's ideas of political isolation. The Congressional Study on Regulation, Volume II, Congressional Oversight of The Regulatory Agencies, argues that there is little public interest oversight.7 The majority of the oversight relates specifically to constituency casework for Congressmen. That is, Congressmen serve the needs of a particular constituent and not the interest of the general public. The factors which underlie Bernstein's life-cycle theory have been extended and developed by other scholars. Two factors have received special attention. The first is the constituency pattern and the second is the nature of the statutory mandate. Sabatier has argued that the rise or maintenance of other organized groups to protect the public interest will slow the process of capture.8 This theme will be developed extensively in the following macro-level theories. Other scholars such as Jaffe, Lowi and Sabatier have placed additional importance on the statutory mandate.9 Lowi 20 has developed arguments which hinge on this factor and has called for the rule of law to prevent capture.lo A recent Congressional study has likewise pursued the theme, this time focusing on the goals of the commission as defined by the statute. For example, whereas Bernstein argued for the combination of policing and promotional functions in an agency as a natural and effective organizational structure, the Congressional study has urged that the functions be separated. This logic is based on the fact that, when combined, promotional goals always tend to overshadow policing functions.ll While Bernstein's capture or life-cycle theory has undergone much criticism, there is a certain intuitive appeal to the idea of life-cycles. Samuel Huntington, for example, in his study of the Interstate Commerce Commission brings a life—cycle argument to bear, citing the dependence 12 Other authors on external groups as the primary factor. prior to Bernstein also employed life-cycle arguments, although their arguments are concerned with the life-cycles of industries and not of a regulatory commission. See for example the works of Glaser and Farris and Sampson.l3 A number of criticisms regarding the life-cycle theory on purely theoretical grounds, i.e., the criteria of good theory, are rather damaging. That is, as a description perhaps Bernstein's ideas are relatively interesting, but as a testable theoretical framework they suffer from a number of deficiencies.14 21 First, the theory is far too general and too descriptive to be testable. For example, Bernstein and others fail to note the length of each phase or cycle. Any test of the model must therefore make a number of assumptions to operationalize the components Of the model. Only one attempt has been made to test the model empirically, by Meier and Plumlee, and this was done without much success.15 Second, while Bernstein notes a number of variables or factors which are supposedly involved in the process, he fails to explicate the relationships among them. This results in serious specification problems for the model, especially with regard to the functional form of the model. The second major theoretical framework on private influence and the regulatory commissions is the economic theory of regulation developed by George Stigler.16 Although originally posited by the Nobel Laureate, it has been further developed, refined and generalized by Posner, Jordan, and Peltzman.17 The elegance of this perspective rests on the essential arguments of supply and demand from economic theory. Simply stated, the theory argues that there is a supply and demand for regulation. According to the economic theory of regulation, there exists a demand for regulation in the industry. Government, it is argued, can provide a supply of regulatory controls or mechanisms for the industry based on the government's power to coerce- to authoritatively allocate values, goods and 22 resources. Through the power of coercion the government can confer benefits on the regulated industry and distribute the costs to the rest of the society. These benefits according to Stigler are direct government subsidies to the industry, control of entry into the market, limitation on substitutes, 18 The industry purchases the and direct price fixing. original regulation with votes and resources which democratic politicians require to maintain their hold on public office. The politicians who compose the government in the model supply regulation for the industry. From the economic theory perspective the main conflict is purely legislative. Once the legislation is passed the regulation has been effectively purchased. For the life-cycle theorists, or the public interest perspective the battle has just begun at the bureaucratic level, and it is here that capture results. Control of a policy area at the legislative level by organized interests is explained by Stigler in the following manner. Political decisions are universal and infrequent. Therefore, 'The expressions of preferences in voting will be less precise than expression of preferences in the market place because many uninformed people will be voting in the election."19 The political system does not require positive inducements to acquire knowledge about the decisions for the ordinary voter. The costs of acquiring information are greater than the benefits for the individual voter, a fact well supported by the voting behavior literature in 23 political science. Organized interests, on the other hand, do have positive incentives to inform themselves. Therefore, he argues that I'the system is calculated to implement all strongly felt preferences of minorities but it disregards the lesser felt preferences of majorities."20 Revisionist historians such as Gabriel Kolko have argued a line of logic similar to that of the economic theorists.21 In his study of the Interstate Commerce Commission (ICC), Kolko argues that the ICC was strongly supported and manipulated by the railroad industry. The existence of the ICC was only supported by the strong railroad interests when they realized it was in their favor to do so. The argument is diametrically opposed to that offered by Huntington, although the commission is the same. 'It was business control over politics (and by business I mean the major economic interests) rather than political regulation of the economy that was the significant phenomenon of the Progressive Era.... The regulation itself was invariably controlled by the regulated industry, and directed towards ends they deemed acceptable or desirable."22 Peltzman has generalized the model in two ways. First the model has been generalized beyond just a concern with regulatory commissions to a concern for all government institutions and policy areas and private influence. Secondly, he argues that the regulated industry need not be the only interest to control regulation. That is, 'regulatory agencies will not exclusively serve a single economic interest."23 This latter development is important 24 for our purposes in that it generalizes the model beyond the consideration of one single group to a consideration of two or more organized interests. The proponents of the economic theory of regulation have emphasized the use of data which focuses on the impact of the regulation on the regulated industry or policy area. Stigler's studies in occupational licensing, transportaton, and electric utility rates focus on the impacts of the regulation on the regulated industry, vis-a-vis, those results which would prevail under the free market structure.24 The preferences of the regulated industry, as noted earlier consist of price fixing, limit of market entry, subsidies, and substitute limitations. The public or consumer interest, for Stigler, is implicitly those policies which would reproduce the benefits of the free market. Finding that industries' preferences are met by regulatory policy, Stigler argues that regulation must have been "acquired and designed by the industry to operate for its benefit."25 Jordan also uses impact data in his analysis, while in addition incorporating the prior market structure of the industry into the preference structure of the industry. However, Jordan and others note that while the impact data verifies that regulation operates in the favor of the regulated industry, this in no way invalidates the position of the public interest hypothesis and its counterpart the 26 capture hypothesis. Both perspectives argue for the Operation of the regulatory agency in favor of the regulated 25 industry, the key question is how this state of affairs comes about. As conducted, the studies do not answer this key question. Criticisms of the economic theory of regulation focus primarily on its rather incomplete analysis of the institutions of government. Much like Easton's systems theory, the economic theory constructs a black-box for 27 Stigler government and looks only at inputs and outputs. merely notes in passing that the regulation desired by the industry often creates a new actor in the environment, the regulatory commission. Bureaucrats are in effect seen as politicians maximizing their ability to obtain re-election. Government is thus entirely passive in response to demands from external sources of influence. One of the primary disagreements between the life-cycle school and the economic theory of regulation concerns the origin of the regulatory commissions. For the economic theorists capture exists from the initial creation of the commission. That is, the initial legislation is designed to benefit the regulated industry. Industry does not oppose the original legislation, as the life-cycle theory argues. The industry is in fact the primary force behind the original legislation. Posner goes so far as to argue that the legislation is symbolic of the public interest only to satisfy the public, and its true motivations are hidden between the lines.28 26 Jordan provides an excellent analysis of this difference. The life-cycle theory is essentially a supplementary hypothesis to the original consumer protection hypothesis (what has been called here the public interest perspective). The latter states “that regulation will protect consumer interests,“ while the former argues that “despite the real purpose of regulation, the regulated industries have managed to pervert their regulators until the commission becomes the protector of the industry.“29 The producer protection hypothesis (what has been called here the economic theory of regulation) argues that “the actual effect of regulation is to increase or sustain the economic power of an industry.“30 Bernstein's life-cycle or capture thesis and Stigler's economic theory of regulation are by far the most widely known of the macro-level theories. However, a number of additional macro-level theoretical perspectives have been developed which advance the ideas of these perspectives. Unlike the previous theories these are not system-wide deterministic theories of regulation and private influence. The term deterministic here refers to the single path of develOpment which these theories predict for all regulatory agencies. That is, all regulatory agencies according to Bernstein should be captured over time. For Stigler, the ecomonic theory predicts that all regulation is purchased by the regulated industry from the outset, although as noted earlier, some develOpments by others regarding this theory 27 do not make this prediction. These system-wide deterministic theories have been attacked as being too general and too simplistic. In the early 1950's Louis Jaffee labeled Bernstein's theory as pseudo-sophisticated.31 Others such as James O. Wilson have also attacked the general theories of private influence.32 However, two points should be raised in defense of these earlier theories. First, they appear even today to cover the predominant mode of regulatory development and private influence. That is, regardless of their imperfections, they still provide alternative explanations for the general case. Secondly, these alternative explanations explicated a number of factors which are significant for understanding the regulatory environment. From them a number of more sophisticated and less deterministic theories can be developed based on the variables and processes that they made apparent. These new theories can then develop alternative paths based on the presence or absence of these factors, and trace the way in which these variable values affect the underlying processes. The following presents a composite of these new perspectives developed from the prior theories of private influence and the regulatory commissions, as well as work accomplished in other studies of private influence and American bureaucracy. That is, in many ways there has been a joining of two separate literatures in recent years, the first concerned with private influence and regulatory 28 commissions and the latter concerned with private influence and the American bureaucracy in general. This latter theoretical work has led to the develOpment of theories of 33 As it is a vast subgovernment, policy types, etc. literature it will not be covered here, however when strains of it are integrated into our discussion the reader should be aware that its development did not occur solely within the context of the question of private influence and regulatory agencies. James O. Wilson presents a cogent analysis of this type in his article, “The Politics of Regulation.“34 Wilson along with others builds his analysis on the distribution of costs and benefits and the concentration of those costs and benefits in a policy area. While numerous combinations are possible, Wilson outlines the primary patterns as noted below. The perceived costs and benefits of the regulation can be seen to be widely distributed or narrowly concentrated. Thus a simple taxonomy can be constructed which is a matrix of the costs and benefits and their concentration or dispersion. Wilson considers three of these possibilities although other authors have added the fourth.35 They are (l) concentrated benefits - diffused costs, (2) diffused benefits - concentrated costs, (3) concentrated benefits - concentrated costs, and (4) diffused benefits - diffused costs. The last category is not covered by Wilson. 29 Regulatory commissions with these different patterns will have different origins and different patterns of development. As Wilson notes, the primary pattern which Stigler considers in his economic theory is one of concentrated benefits and diffuse costs. As for the Bernstein argument the logic is somewhat different. Under the life-cycle hypothesis and notion of capture over time the regulation was originally designed as one of diffuse benefits and concentrated costs. However, over time the commission changes through its policies to one which offers concentrated benefits and diffuse costs. The regulated industries have essentially changed the cost and benefit structure of the agency over time.36 Wilson further argues that the concentration and dispersion of benefits affects the politics of regulation in a second way. He argues that groups are more sensitive “to sudden or significant decreases in their net benefits than they are to increases in net benefits.“37 Groups with interests in a particular policy area are therefore likely to form in response to regulatory legislation. This prOpensity would obviously depend on the concentration of costs and benefits. For example, in the concentrated costs and concentrated benefits mode the opposition of strong organized interests would be the likely outcome. Wilson adds yet another factor to the modern political environment which goes beyond the simple interest group explanation and adds even more uncertainty to the regulatory 30 environment. This is the area of the mass media and its affect on the creation of regulatory statutes. As he notes, “to obtain any regulation at all, it (is) necessary in each case to get legislators to take the problem seriously, to forge a winning coalition among legislators with diverse interests and perceptions, and to overcome the arguments and influence of Opponents.“38 These arguments suggest that even within the described policy types there is no single and sure path of development. The macro-level perspective has thus become highly probabilistic. This is true even when the concentration of costs and benefits are known as well as the constituency configuration in terms of organized interests. A recent paper by Fiorina clearly attests to this fact. Investigating the question of legislative delegation from the perspective of a legislator and developing a formal calculus he notes that: “the extensive delegations of authority to the Older regulatory agencies suggest that at least in the beginning legislators were not Obviously ripping off their constituents for the benefit of the capitalists, but were instead seeking re-election by benefiting a broad range of constituents.“39 The ideas of Wilson, especially his taxonomy, are well known to political scientists interested in policy areas in general. The works of Theodore Lowi and his arguments that policy types determine politics are also well known.40 Lowi's typology of policy types is similar in many ways to the one outlined by Wilson. Here we refer to Lowi's policy 31 types of redistributive, distributive, constituent policy, etc.41 Additional authors have also pursued this theme, such as Salisbury and Heinz. Their typology of policy types is also similar to that of Wilson.42 The develOpment of macro-level theories have thus become even more probabilistic in their approach. Rather than arguing a definitive and predictable end they now refer to general patterns and have begun to recognize a plethora of intervening environmental and political variables.43 SUMMARY OF MACRO-LEVEL FACTORS We now turn to a summary of the major factors which can be gleaned from the macro-level theories. It should be noted that these factors do not constitute an exhaustive list of all the points considered by scholars of regulation. The purpose is to explicate the primary factors which are important for understanding the NLRB. Further, it should be noted that it is not the purpose or intent of this study to build a general theory for regulatory commissions. Rather, we are interested in these factors as they apply to and describe the NLRB and provide a context within which hypotheses concerning private influence patterns can be developed. These theories offer five primary factors which are relevant for our study. These are the constituency configuration, the statutory guidelines, the process of 32 action, the process of development and the primary channels of influence. First, the constituency configuration in terms of the groups involved is highly important. This has been demonstrated throughout the previous discussion. In each case a different constituency configuration is linked to a different probable pattern of development. As Paul Quirk has noted, “one can hardly doubt the cogency Of this line of analysis, involving the configuration of interests making up the political environments of regulatory agencies.“44 The second important factor which deserves consideration is the statutory guidelines of a commission. This factor is not the focus of our study, but rather an assumption, as we are interested in non-legislative influence channels. The statutory guidelines determine the cost-benefits distribution pattern, the organizational structure of the agency, the goals of the agency and the means of policy implementation. The importance of the cost-benefits has been noted in detail and requires little further discussion. It is closely related to the constituency configuration and further is a primary factor in distinguishing the different patterns of development. Its importance for our purposes will be in describing the deviant case status of the NLRB. The statute does more than just determine the cost-benefit distribution. It also is an extremely important factor in determining the organizational structure of the 33 agency. For example, the independent regulatory agencies employ a combination of quasi-judicial and quasi-legislative functions. Those that employ the former are likely to be different than those that employ the latter as the primary means of implementing policy. Another important factor is the goal structure that the commission receives from the statute. The agency may have a purely policing function, or a combination of policing and promotional functions. Sabatier has shown how different combinations of these goals can lead to different levels of private influence.45 Also the statute will determine the way in which the goals are to be implemented by detailing the enforcement powers of the agency. The third important factor developed from the prior theories is the logic of action they provide. The rationalist perspective made explicit by Stigler and the economic theorists, and implicit in Bernstein, provides us with a means of determing how individuals will behave in the environment. Thus the logic underlying the bargaining and influence relationships can be understood. Simply stated we expect individuals to act in their own self-interest as a general rule, given the numerous constraints offered for this principle developed in the rationalist literature.46 Actors should weigh the costs and benefits of an action and act to maximize the benefits based on their perceptions. This logic when applied to the actors in the environment helps us understand probable courses of action. 34 The final factor concerns the develOpmental process specifically developed by Bernstein over time.47 While rationalism provides a logic of action, to assume that this is the only process would result in a static model of the world. The residue of actions at one time should result in a different environment at a later time. These changes may be minor or incremental, such as the alterations in the procedures of an agency, or major structural changes as in the case of major legislative amendment. Finally, these theories employ a number of channels of private influence either explicitly stated or implicitly ignored. These may be divided into two distinct categories, legislative and bureaucratic (non-legislative). The legislative influence channels refer to the ability of private interests to obtain statutes which are favorable to their policy preferences. This is an extremely important channel, and the disagreement over the initial intent of regulatory statutes clearly separates the public interest and economic theory interpretation. Likewise, the importance of including promotional goals or policing goals in the statute has also be shown. As the importance of this channel has been clearly demonstrated in earlier discussions we now turn to the consideration of the bureaucratic channels of influence. 35 INFLUENCE CHANNELS There are two types of non-legislative or bureaucratic influence channels, direct and indirect. The former involves attempts by private interests operating through the formal political institutions of government to influence the policy decisions of the regulatory commissions. This category would involve appointments to the commissions, budgetary allocations, etc. The latter involves attempts to influence the agency Operating directly on the agency itself. This would involve attempts to control information required by the agency, participation at hearings, etc. Bernstein has spent more time on these channels of influence than other macro-level theorists, as his theory is highly dependent upon them. In this section we shall review some of the major channels of influence developed by the middle level theorists on private influence. These are the appointments process, the use of the budget as an oversight mechanism, and the broad category of direct influence. REGULATORY APPOINTMENTS Of all the channels of private influence, the appointment of regulatory commissioners has received by far the greatest amount of study. Studies have explored numerous questions concerning the regulatory appointments process such as the quality of the appointments, the characteristics 36 of the appointees, the post-service employment patterns of the appointees, and the role of the President and the Congress in the appointments process. There are two reasons for this interest. The first concerns the high positions which these appointees hold. They make the true high level policy decisions for the agency, what Stone calls middle level decision making surpassed in importance only by the enabling statute.48 The more vague the statute, the more their power should increase. A second reason for this interest can be traced to the unique character of appointments to the federal independent regulatory commissions. Appointees to these positions unlike other members of the federal executive establishment cannot be removed at the discretion of the President. This principle was firmly established by the Supreme Court in the case of Humphrey's Executor.49 Further it has been written specifically into a number of the statutes of independent federal regulatory commissions. In one sense their protection from arbitrary removal is similar to that afforded to federal judges, however unlike judges their appointment is not for life but for a specified term. A large portion of the literature deals with the quality of the appointees. This interest goes back to Herring's studies in the 1930's and it has not waned.50 A major study of the quality of federal executives in general and regulatory commissioners in particular was completed by Stanley, Mann and Doig just a few years ago.51 In the past 37 eight years alone two Congressional committees have conducted studies of the federal appointment process for the independent regulatory commissions, one by the Senate Commerce Committee, and the second by the Senate Committee on Government Operations (part of a multivolume work on the independent regulatory commissions).52The concern that underlies these studies and their search for good men to protect the public interest was best stated by Landis. “The prime key to the improvement of the administrative process is the selection of qualified personnel. Good men can make poor laws workable: poor men will wreak havoc with good 1aws.'53 The studies have shown however that the regulatory appointments have been made for numerous other reasons, primarily dictated by political concerns. Expertise and knowledge of the regulated policy area are not important criteria as they should be according to the public interest perspective. Political concerns can mean a number of things, among them appointment for partisan political reasons, or due to some clientele influence. In their study of the appointments process of the Federal Trade Commission and the Federal Communications Commission, Graham and Kramer note the following. “Partisan political considerations dominate the selection of regulators to an alarming extent. Alarming in that other factors such as competence, experience, (etc.)...are secondary considerations. Most commission appointments are the result of well stoked campaigns conducted at the right time with the right sponsors, and many selections can be explained in terms of powerful political connections and little else.“5 38 Two considerations arise in the study of commissioner attributes that are of interest to the subject of private influence channels. The first concerns the way in which the appointments process is manipulated in order to obtain the nomination of persons who share the preferences of those in the regulated industry. The second concerns the ways in which the appointees once confirmed and sitting on the commission boards will have their preference structure altered by private interests. Paul Quirk has labeled these the pro-industry appointments hypothesis and the industry jobs incentive hypothesis respectively.55 The former states that “individuals selected to fill high regulatory offices tend to hold policy attitudes on agency issues that support the regulated industry interests and preferences.“56 The latter states that “regulatory officials have personal career incentives to favor industry supported policies.“57 Although the latter hypothesis is related to what we term direct influence we will discuss it briefly in this section. The results obtained in many studies of these hypothesis are hard to separate because they conclude that a combination of the two processes leads to private influence of high appointees. This is especially true for those who adopt a public interest perspective and employ attribute data to confirm their results. The attribute data related to these hypotheses concentrates primarily on the employment backgrounds of 39 commissioners and the post-employment patterns of commissioners.58 The former employed for the pro-industry appointments hypothesis, and the latter for the industry jobs incentive hypothesis. In both cases the preferences and actions of commissioners are inferred based on the attribute characteristics of the commissioners. Bernstein, Graham, Kohlmeier, and a major Congressional study lend credence to the idea of private influence on the appointments process arguing that industry seeks sound men who represent their positions.59 The Congressional study reports that: “present commission membership appears to be top-heavy with members with a disposition towards the industries subject to regulation“60 The appoinments literature on the NLRB has a singular focus. Here the appointments process is more closely related to the prior attitudes hypothesis. The study of the appointments process has been a concern to the labor law community, many of whom see it as a key problem in the development of labor relations law. Reforms have been suggested such as the creation of a labor court with members appointed for life to alleviate the problem of private influence on the process. The bulk of the NLRB appointments literature (Scher, Welborn: and DeLorme, Hill and WOod) concludes that in the post Taft-Hartley period private interests have in effect led to the creation of pro-labor and pro-management boards.61 Further, they argue that Republican presidents have supported moderately 40 pro-management boards, while Democratic presidents have favored strong labor boards. The general legal writings and study of NLRB cases also indicate the existence of boards with a definite pro-labor and pro-management bias.62 A number of cases on critical policy areas have been decided by the board only to be reversed when a new board comes into being. However, a closer investigation would indicate that the results on the issues of private influence are mixed for the general case. Perhaps part of the reason for this dilemma is due to the fact that in the general case one must in some manner determine what the public interest is in order to determine the degree of private influence. This process has obvious problems and can lead to varied interpretations. While the recent federal studies argue in their general conclusions that the appointments process is open to private influence, an examination of some of the survey data included in the report would lead one to the opposite conclusion. The data clearly indicates that the commissioners are considered to be “fair and impartial.“63 In a question concerning the responsiveness to consumer interests (the public interest) or industry interests the report concluded that “panelists were happy with the amount of responsiveness“ and that in the case of the Federal Trade Commission (FTC) there was too much responsiveness towards consumer interests.64 41 Likewise Paul Quirk's study, which concerns both of the hypotheses on private influence and the appointees to the regulatory commissions, indicates that private groups do not control the appointments process. “the pre-appointment policy attitudes of our regulatory officials were least in accord with the pro-industry incentives hypothesis. Instead of having generally held attitudes supporting industry-favored positions, the officials... held attitudes or positions Opposing industry policies or held no previous policy inclinations.“ 5 In summary, there appears to be no single pattern of industry or private domination of the appointment process. We now turn to the consideration of the second indirect influence channel- budgetary oversight. BUDGETARY OVERSIGHT Another way in which private interests can have an impact on the policy preferences of regulatory commissions is through the oversight functions of the Congress and the President. Investigations, agency hearings before the congressional oversight committee and other mechanisms have been seen as a means for private interests to flex their muscles through their congressional allies. The Congress has been chided by scholars and its own investigations for not employing these oversight mechanisms for the public interest. Rather, the studies have argued that they are used to serve the particular interests of individual constituents.66 Such activities would include obtaining a 42 license for a constituent, speeding up a hearing, etc. Normally, when the oversight functions are employed for a concerted interest or group interest it is the preliminary stage for a legislative amendment. As such hearings are concerned with legislation we will not consider them here for they are part of the process of legislative influence and should be considered along with other statutory guideline factors. This is not to say that all oversight hearings are part of the legislative amendment process, but most are either directly or indirectly involved with this process. One of the major means of oversight concerns the power of the purse. The yearly (and sometimes more often) process where a commission receives its resources to continue operation for another fiscal year would seem ripe for private influence. This power has long been deemed crucial as a method for exerting control over institutions. However, in the regulatory literature it has received scant notice and little serious direct attention.67 Only one direct empirical study exists which tests the degree to which private interests use this process to achieve favorable policy outcomes.68 Given the ability of private interests to employ their influence with their Congressional allies in obtaining legislation or appointments, it follows that the budget process can and should be used as a means to achieve favorable policy outcomes. Or at the least, one might expect 43 that it could be used in the negative, that is to prevent unfavorable policies through the elimination of funding. The bulk of the budgetary literature has not been concerned with the issue of private influence, especially as it relates to the regulatory commissions. For the most part, it has become an area of interest with an independent focus all its own. This focus has centered on the attainment of general theories of the budgetary process, the expansion of agency budgets, budgets as control mechanisms, etc.. The literature in this area is quite voluminous, with scholars such as Wildavsky focusing on the theoretical questions of the bugetary process, and Niskanen focusing on the issue of budgetary expansion.69 The logic of the argument for the influence of private interests has been developed by Paul Quirk based on the implied arguments of others interested in private influence on the regulatory agencies. As he notes, pro-industry. considerations are the result of two factors- “dependence of the agency, on having sufficient political support for survival,“ and the fact that the regulatory agency normally only has one organized supporter in the environment- the regulated interest. The conclusion he infers from these factors is that the “regulatory agencies have strong budgetary incentives to adopt the policies favorable to industry views and interests.“70 This sentiment is echoed by Wildavsky in his consideration of the strategies which agencies should pursue 44 in the attainment of their budgetary goals. He argues that an agency should find a client to support its programs, and serve that client.71 This is not the only strategy, but it is an important one considering the plethora of agencies and their programs attempting to achieve funding. Depending on the validity of the macro-level theories of private interest presented the most likely way to test this proposition or seek empirical evidence would be in the negative. That is, if an agency supports a client with a particular program there is no way of being certain that it was the budgetary process that led to this. The most appropriate way is to find programs to which the regulated interests are highly or moderately Opposed and to see what happens to the funding in these instances. This is exactly the contention of Quirk in his study. When budgetary considerations are noticable they are directed towards specific programs.72 This is also the indication that one gets when looking at the budgetary process of the Wagner Act NLRB. One section of the agency which raised a good deal of controversy among the opponents of the agency was the Economic Section. Its funding was eliminated after just a few short years.73 In summary, it is clear that there is a lack of literature on the use of the budgetary process as a channel of private influence. However, the NLRB provides us with an excellent Opportunity to expand our information on this potential channel of direct influence. It was important in 45 the era before the passage of the Taft-Hartley Act, and we can now monitor its behavior in the post Taft-Hartley era. DIRECT INFLUENCE The role of direct private influence on the regulatory commissions to obtain policy preferences is well established and for many it is the essential channel of influence. The argument can be made that given good commissioners, well meaning bureaucrats, etc., the decisions they make can still be pro-industry if the regulated interest controls the expertise and information surrounding the policy space. As Kohlmeier states: “The men and women who regulate industry in the public interest deal occasionally with the White House, frequently with select members of Congress and constantly with the executives of regulated industries.“ Conventional explanations of the importance of direct influence revolve primarily around the ideas of information and expertise. Information provided by the regulated private interest plays an essential role in the development of policy. This is especially true when one industry dominates the policy space. Agencies such as the FDA and FTC depend on the regulated industry for much of the information they employ in rendering policy decisions. For example, data on new drugs and the tests which verify them are conducted by the industry and not by the agency.75 This influence channel 46 is further amplified when other conditions are present, such as complex technological questions where the industry controls the pool of expertise available. Thus one would expect that the Atomic Energy Commission (ABC) and its heir the Nuclear Regulatory Commission (NRC) would be highly dependent on the regulated industry for information and expertise. Other factors are also important and can further increase the commissions' dependence on the regulated interest. Mitnick provides a comprehensive list of these factors such as impactedness of information, information 76 Porter and Sagansky argue that information and costs, etc. the costs of obtaining it play a key role in determining policy outputs in a complex environment. They ask, “is it (the commission) imprisoned by its dependence on the regulated industry for life giving cooperation, for the data it needs to do its job, even though its heart remains true to the public interest?“77 That is to say, the role of information is so important in the policy process that its control can lead to the implementation of pro-industry policies even when the commissioners seek the public interest. The role of information and expertise in determining policy ouputs is also enhanced based on certain statutory considerations. Where the mandate is general and the discretion open to officials is broad there is a greater propensity for influence. Lowi has made this argument famous 47 with his call for the rule of law and the implementation of strict policy guidelines.78 Promotional goals contained in the statute along with policing functions can further cause the commission to see the goals as conflicting when information is controlled and properly manipulated by private interests. The result is that they may be led to pursue the former in lieu of the latter set of goals. A recent study on the role of the regulated interests in the policy formation process has demonstrated the dominance of their position. In rulemaking proceedings, the industry is dominant in the process almost without exception. The report found: “At agency after agency, participation by the regulated industry predominates- often overwhelmingly.....public interest representation accounts for a very small percentage..... In more than half of the formal proceedings, there appears to be no such participation whatsoever, and virtually none at informal agency hearings.“79 The logic which underlies this process can be found in the rationalist perspective forwarded by Stigler and other economic theorists. When it is developed by those investigating individual or micro-level behavior, such as Noll and Eckert, the arguments are indeed persuasive.80 Further, the extent to which direct influence can penetrate the commissions is extraordinary. The argument is that assuming rational behavior on the part of the commissioners and other agency bureaucrats (that is utility maximization) and a set of preferences; such as status, ease of working conditions, and expectation of future rewards, a great deal 48 of direct private influence can be expected. The mere existence of a regulated industry and the opportunities it affords for officials may alone be enough for the industry to achieve its preferences without any premeditated action on its part. Eckert has argued that, “An individual seeking to maximize his lifetime income might regard a regulatory position as an investment that would yield subsequent employment in... the regulated industry."81 This argument was introduced as the post-industry jobs hypothesis in the earlier section on regulatory appointments. Quirk found in his study that, “incentives relating to subsequent industry employment were for the most part consistently pro-industry among the incentives studied.“82 S UMMARY In summary it appears as if the most important channel of private influence according to the general literature is direct influence. The appointments process appears to be the next most important channel, and the budget channel appears to be the least important. We can now turn to a consideration of these channels within the context of the NLRB. The NLRB can be understood in terms of the general theories put forth in the beginning of the literature review so as to set the context for our consideration of the channels of private influence. We will now turn to a 49 consideration of the NLRB and the way it can be understood by these general theories. FOOTNOTES l.Paul Quirk, Industry Influence in Federal Regulatory Agencies (Princeton, N.J.: Princeton University Press, 1981), p. 21. 2. The most comprehensive review of the regulatory literature is that of Barry Mitnick, The Political Economy of Requlation: Creating, Designing, and Removing Regulatory Forms (New York: Columbia University Press, 1980). 3. Marver Bernstein, Regulating Business by Independent Commission (Princeton, N.J.: Princeton University Press, 1955), pp. 74-102. 4. Ibid., p. 74. 5. Ibid., p. 76. 6. Anthony Downs, Inside Bureaucracy (Boston: Little, Brown and Co., 1967), Chapters 3 and 13. 7. Congressional Oversight of Regulatory Agencies, Vol. II of the Study on Federal Regulation, op. cit.. 8. Paul Sabatier, “Social Movements and Regulatory Agencies: Toward a More Adequate- and Less Pessimistic- Theory of Clientele Capture,“ Policy Sciences, 6(3) (September,l975), pp. 301-342. 9. Sabatier, op. cit.: Louis Jaffe, “The Effective Limits of Administrative Process: A Reevaluation,“ Harvard Law Review, 67(7) (May,1954), pp. 1105-1135. 10. Theodore Lowi, The End of Liberalism (New York: Norton, 1969). 11. Regulatory Organization, Vol. V of Study on Federal Regulation, op. cit., pp. 93-110. 12. Samuel Huntingdon, “The Marasmus of the ICC: The Commission, the Railroads, and the Public Interest,“ Yale Law Journal, 61(4) (April,l952), pp. 467-509. 13. Martin Glaeser, Public Utilities in American Capitalism (New York: Macmillan, 1957). MartIh Farris and Roy Sampson, Public Utilities: Regulation, Management and Ownership (Boston: Houghton Mifflin, 1973). 50 51 14. For a more comprehensive list of criticisms see Barry Mitnick, op. cit., p. 73: and Richard Posner, op. Cite, pp. 335-3580 15. Kenneth Meier and J. Plumlee, “Regulatory Administration and Organizational Rigidity,“ Western Pglitical_9uarterly, 31(1) (March,l978), pp. 80-95. 16. George Stigler, “The Theory of Economic Regulation,“ Science, 2(1) (Spring,197l). PP. 3-21. 17. These articles represent the essential developments to the economic theory of regulation. See William Jordan, op. cit.,: Richard Posner, op. cit.,; and Samuel Peltzman, “Toward a More General Theory of Regulation,“ Journal of Law and Economics, 19(2) (August,l976), pp. 211-240. 18. George Stigler, Op. cit., pp. 2-3. 19. Ibid., pp. 11-120 20. Ibid.) p. 11. 21. Gabriel Kolko, W (New York: The Free Press, 1979). 22. Ibid., p. 3. 23. Samuel Peltzman, op. cit., p. 211. 24. Stigler, op. cit., and George Stigler and Claire Friedland, “What can Regulators Regulate?: The Case of Electricity,“ Journal of Law and Economics, 5 (October,1962), pp. 1-16. 25. George Stigler, Op. cit., p. 3. 26. William Jordan, op. cit., pp. 152-153. 27. Although the black-box criticism is still valid, a number of scholars have attempted to bridge this gap in recent years as we shall see later in the text. 28. Richard Posner, Op. cit., p. 681. 29. William Jordan, op. cit., p. 152. 30. Ibid., p. 153. 31. Louis Jaffe, “The Independent Agency- A New Scapegoat,“ Yale Law Journal, 63 (1956), pp. 1069-1076. 52 32. James Q. Wilson, “The Dead Hand of Regulation,“ The Public Interest, 7(25) (Fall ,1971), pp. 39-58. 33. See for example Harold Seidman, Politics, POsition, and Power (New York: Oxford University Press, 1975) and Theodore Lowi, “Four Systems of Policy, Politics and Choice,“ Public Administration Review, (July/August, 1982), pp. 298-310. 34. James Q. Wilson, “The Politics of Regulaton,“ in James McKie, ed., Social Responsibility and the Business Predicament (Washington, D.C.: The Brookings Institutiofi, 19741. pp. 135-168. 35. Ibid., pp. 142-146. 36. Ibid., pp. 157-159. 37. Ibid., p. 139. 38. Ibid., p. 151. 39. Morris P. Fiorina, “Group Concentration and the Delegation of Legislative Authority,“ (unpublished paper, California Institute of Technology Working Paper, 1982), pp. 27-28 0 40. Theodore Lowi, “Four Systems Of Policy, Politics and Choice,“ Public Administration Review, (July/August, 1982), pp. 298-310. 41. Ibid. 42. Robert Salisbury and John Heinz, “A Theory of Policy Analysis and Some Preliminary Applications,“ in Ira Sharkansky, ed., Policy Analysis in Political Science (Chicago: Markham Publishing Co., 1970), pp. 39-61. 43. For an excellent listing of the numerous variables which have been considered see Barry Mitnick, op. cit., pp. 206-214. 44. Paul Quirk, op. cit., p. 15. 45. Sabatier, Op. cit., pp. 307-309. 46. The literature which develops the logic of rational behavior is quite extensive and can only be briefly cited here. Numerous constraints on the principle of pure rational behavior (utility maximization) have been developed. A good review of this literature is: Robert Abrams, Foundations of Pgégtical Analysis (New York: Columbia University Press, 53 47. Marver Bernstein, ess b Inde endent Commission (Princeton, N.J.: Princeton University Press, 1955). pp. 74-102. 48. Alan Stone, Regulation and its Alternatives (Washington,D.C.: Congressional Quarterly, 1982). 49. C. Herman Pritchett, The Amerigan Cghstitution (New York: McGraw-Hill, 1968), pp. 346-347. 50. Pendleton Herring, W A Study of Their Careers and Qualifications (New York: McGraw-Hill, 1936). 51. Stanley, Mann and Doig, Men Who Govern (Washington,D.C.: Brookings Institution, 1967). 52. The Regulatezy Appointments Process, Vol. I of the W. 0P- Cit- . 53. James M. Landis, Repert 29 the Regulatory Agencies to the President-Elect (Washington, D.C.: Committee on the Judiciary, 86th Congress, 2nd Session, 1960), p. 66. 54. 0.8., Congress, Senate, Committee on Commerce, .APpointments to the Regulatory Agencies, 94th Congress, 2nd Session, 1976, p. 391. 55. Quirk, op. cit., p. 23. 56. Ibid. 57. Ibid. 58. See for example William Gormley, “A Test of the Revolving Door Hypothesis,“ American Journal of Political ‘Seienee, 23(4) (November,l979), pp. 665-683. 59. Bernstein, Regulating Business by Independent Commission. 0p. cit.. p. 84. Louis Kohlmeier, The Regulators: Watchdog_Agencies and the .Pnhlie_1ntere§t (New York: Harper and Row, 1969), p. 284. 60. The Regulatory Appointments Process, VOl. I of the Study on Federal Regulation, op. cit., p.7157. 61. Seymour Scher, “Regulatory Agency Control Through Apppointment: The Case of the Eisenhower Administration and the NLRB,“ Jourgal of Politics, 23 (1961), pp. 667-688. David Welborn, “Presidents, Regulatory Commissioners, and Regulatory Policy,'.lQnLnal_Qfi_£HDlis_L§!. 15(1) (1966), pp. 3-39. DeLorme, C., Hill, R. and Wood, N., “The Determinants of Voting by the NLRB on Unfair Labor Practice Cases: 1955-1975,“ Public Choice, 37(2) (1981): pp. 207-218. 54 62. See Chapter III on the NLRB. 63. The Regulatory Appointments Process, Vol. I of the Federal Study 93 Regulation, op. cit., p. 289. 64. Ihid., p. 291. 65. Quirk, op. cit., p. 175. 66. ’ of the Re ulatory Agencies, Vol. II of the Study on Federal Regulation, Op. cit.. 67o QUiIk, Op. Cite] p. 96-970 68. Ibid., p. 176. 69. Aaron Wildavsky, The Politics of the Budgetary Process (Boston: Little Brown, 1979). William Niskanen, Bureaucracy and Representative Government (Chicago: Aldine-Atherton, 1971). 70. Quirk, op. cit., p. 96. 71. WildaVSky' Op. Cit.) PP. 65-740 72. Quirk, op. cit., p. 133-142. 73. James Gross, The Reshaping of the National Labor Relatiens Board 1937-1947, (Albany, New York: State University of New York Press, 1981), pp. 180-187. 74. Kohlmeier, Op. cit., p. 69. 75. Stone, op. cit., pp. 210-213. 76. Mitnick, op. cit.: Pp. 208-216. 77. Michael Porter and Jeffrey Sagansky, “Information, Politics, and Economic Analysis: The Regulatory Decision Process in the Air Freight Cases,“ Pgblic Policy, 24(2) (Spring,1976), p. 264. 78. Lowi, End of Liberalism, op. cit.. 79. Public Participation in Regulatory Agency Proceedingg, Vol. III of the Study on Federal Regulation, Op. cit., p. vii. 80. R. Eckert, “On the Incentives of Regulators: The Case of Taxicabs,“ Public Choice 14 (Spring,1973), pp. 83-99. Rodger Noll, “The Economics and Politics of Regulation,“ Virginia Law Review, 57 (1971), pp. 1016-1032. 55 81. ECkert' Op. Cit.’ pp. 97-98. 82. Quirk, op. cit., p. 176. III THE NATIONAL LABOR RELATIONS BOARD The National Labor Relations Board represents the deviant case within the set of independent regulatory commissions. This fact has been noted explicitly by Wilson, McConnell and other scholars, and admitted implicitly by others through their ommission of the agency in their studies of private influence.1 The NLRB is the only commission in which two stong and organized material interests oppose each other in the policy space of the commission. In this chapter information on the history and current structure of the commission will be presented. Further, those factors which explain the agency as well as distinguish it from other independent regulatory commissions will be explored. HISTORY OF THE NLRB There are four primary pieces of legislation which constitute the foundations of the basic labor law, policy guidelines and jurisdiction of the NLRB. These are the National Labor Relations Act of 1935 (NLRA), the Labor Management Relations Act of 1947 (LMRA), the Labor Management Reporting and Disclosure Act of 1959 (LMRDA) and the Postal Reorganization Act of 1970 (PRA).;2 The first 56 57 three pieces of legislation are commonly referred to by their sponsor names, the Wagner Act, the Taft-Hartley Act, and the Landrum-Griffin Act, respectively. A discussion of these acts in a legalistic fashion alone would provide us with the current labor law of the United States as it concerns the NLRB, but such a discussion would be rather sterile. The order in which these pieces of legislation were passed (especially the first two) as well as their historical context is paramount to our understanding of the development of the NLRB. The NLRB was established in its present commission form by the National Labor Relations Act of 1935. Although referred to as “the first genuinely regulatory federal statute dealing with relations between labor and capital,“3 it was not the first legislation designed to ameliorate the differences between labor and management. Since the early part of the century a number of pieces of federal legislation as well as executive orders had been drawn up to address the problem of labor management relations. One of the first major developments occured in 1918 when President Woodrow Wilson established the National War Labor Board to settle disputes between labor and management.4 These disputes were interfering with the nation's war production. At the close of the war Wilson attempted to continue the work and practices of the war Labor Board through the National Industrial Conference.5 However these efforts failed when employer groups refused to support the 58 principles of the conference. Although Wilson failed in his attempts to continue the work of the War Labor Board, the major principle which guided it was to prove invaluable later in the establishment of the NLRB and its predecessor agencies. The most important principle of the War Labor Board concerned the rights of workers to organize. “The right of workers to organize in trade unions and to bargain collectively, through chosen representatives, is recognized and affirmed. This right shall not be denied, abridged, or interfered with by the employers in any manner whatsoever.“6 The Railway Labor Act of 1926 as amended in 1934 likewise provided valuable precedence and experience for the establishment of the NLRB. The primary importance of the act was the establishment of a board to achieve peaceful settlement of labor disputes in the transportation industry.7 With the advent of the great depression in the 1930's the move to establish boards to reconcile labor management disputes became an important issue of the day. Two boards immediately preceded the Wagner Act NLRB, the National Labor Board and the '016' National Labor Relations Board which grew out of the former.8 These boards were charged with the administration of Section 7a of the National Industrial Recovery Act (NIRA). Section 7a gave employees the right to organize, bargain collectively, and select their representatives without interference. It further prohibited employers from requiring employees to join company 59 controlled unions or refrain from joining a union as a condition of employment.9 However, in May of 1935 the Supreme Court invalidated the NIRA in the Schecther Poultry Case, and with it the labor boards.10 The passage of the Wagner Act followed closely on the heels of the invalidation of the NIRA. When the act was finally passed into law after numerous attempts by Senator Wagner, and without much Presidential support, it borrowed much from the experiences of the prior labor boards, especially the latter two. This is not surprising considering the fact that Senator Wagner had been the chairman of the National Labor Board and that members of this board had helped him draft the act.11 “almost every provision of the Wagner Act from the definition of the words “employee“ and “representative“ to each of the employer unfair labor practices, is rooted in the experiences of these two pre-Wagner boards. The personnel of the old NLRB played an active role in the writting of the Wagner Act.“12 The passage of the Wagner Act however did not guarantee the success of the new commission or immediately usher in a new era in labor management relations. Employer opposition to the new agency was extreme and the NLRB found it nearly impossible to Operate effectively.13 The Supreme Court had overturned a number of other New Deal programs and this was the hope of the employers again in this instance. Further, many prominent management lawyers assured employers that the act was undoubtedly unconstitutional.l4 “bitter opposition to the law made it impossible for the NLRB to function as Congress 60 intended.....until the act was sustained by the Supreme Court....the act remained essentially inoperative.“15 In April of 1937 the Supreme Court sustained the constitutionality of the Wagner Act in the now famous (or perhaps infamous) case, NLRB vs. Jones Laughlin Steel Corporation: 301 0.8.1 (1937).16 This five to four decision was to change the basic nature of labor relations in the United States. While the NLRB had borrowed much from the experience of prior labor boards it was not a carbon copy. The writers of the act incorporated those factors which they found valuable, and incorporated changes which experience suggested were vital. One of the primary changes in the NLRB relative to the previous labor boards was the establishment of a government board which ruled on labor cases without attempting to incorporate direct representatives from labor and management into the decision making process. Whereas previous boards consisted of representatives of labor, management, and government who attempted to reach decisions through compromise, the new commission was structured in the fashion of a labor court.17 The NLRB was modeled overall on another independent regulatory commission, the Federal Trade Commission. It was designed as “a quasi-judicial tribunal with defined legal authority and power to have its orders enforced by court decree.“ 18 61 The Board (as outlined in Section 3 of the Wagner Act) consisted of three members with the length of appointment set at five years. The original appointments were set for terms of one, three, and five years so as to create a staggered appointment process. Appointments were to be made by the President with the advice and consent of the Senate. Unlike many other independent commissions there was no bipartisanship requirement concerning the composition of the Board.19 The organizational structure of the NLRB was as follows.20 The organization consisted of a central office in Washington, D.C. with a number of regional offices throughout the United States (as outlined in Section 5). The Board sat in the main office and supervised the major organizational subdivisions or sections. The Litigation Section and the Review Section were both under the direction of the General Counsel. The former section prosecuted unfair labor practice cases heard by the Board. It was further charged with representing the NLRB before the federal courts. The latter section reviewed cases for the Board and composed preliminary decisions for their final consideration. The Economics Section performed economic analyses and related research for the Board in the area of labor matters. The Information Section was charged with maintaining documentation on the operations of the NLRB and handling press relations. Finally, the Office of the Executive 62 Secretary was charged with the oversight and administration of the entire agency. A few years later the Division of Trial Examiners was established by the Board. As a judicial division it was separated administratively from the other sections of the NLRB . The NLRB under the Wagner Act could truely be said to be an agency charged with promoting the growth and development of labor unions in the United States. Section 1 of the act which outlines the purposes and justifications for the act makes this abundantly clear. The legal justification for this new role of the federal government is based on the commerce clause in the 0.8. Constitution. However, the political reason for federal intervention is clearly based on the inequality of power between the two Opposing forces in this policy area, labor and management. This is evident upon a reading of the stated purpose of the act. “Experience has proved that protection by law of the right of employees to organize and bargain collectively safeguards commerce....by restoring equality of bargaining power between employers and employees...The inequality of bargaining power between employees who do not possess full freedom of association or actual liberty of contract, and employers who are organized in the corporate of other forms of ownership association substantially burdens and affects the flow of commerce...“21 The NLRB had two primary functions under the Wagner Act, which remain largely intact today. These were the settlement of unfair labor practice cases and the settlement of issues concerning employee representation. The unfair 63 labor practices under the Wagner Act are outlined in section 8. All of the unfair labor practices were employer unfair labor practices. There were no union unfair labor practices cited. This section of the act deals solely with employer misconduct. These employer unfair labor practices cover a wide 22 Section 8(1) makes it unlawful for number of items. employers to restrain or coerce employees in their rights to organize, choose representatives and bargain collectively as outlined in section 7 of the act. Section 8(2) prohibits company or employer dominated unions. Section 8(3) makes it unlawful to discriminate in hiring or tenure based on an employees pursuit of section 7 rights. Section 8(4) provides protection for employees who testify or bring charges to the NLRB. And Section 8(5) makes it unlawful for employers to refuse to bargain with union representatives. The Wagner Act thus created what may be called a set of “protected activities“ for employees, allowing them to pursue their rights of self-organization without fear of harrasment by employers or losing their jobs. Representation issues are covered under section 9 of the act. The Board attained the general authority to determine if employees wanted a union to represent them. This power involves two essential points: first, do the employees desire representation by a union, and second, which employees are to be included in the bargaining unit. 64 In 1947 the Eightieth Congress began its session and for the first time in a number of years the Republican Party gained control of both the Senate and the House of Representatives. As the Republican Party had been the supporter of employer interests throughout the life of the Wagner Act, calls for its repeal began to surface quickly. The fact that the Wagner Act was in serious jeopardy was readily apparent when Senator Taft, an ardent Opponent of the Wagner Act NLRB, became chairman of the Labor and Public Welfare Committee, the Senate oversight committee for the NLRB. Taft, who was the ranking member on both the Senate Finance Committee and the Labor and Public Welfare Committee chose the latter, the less prestigious committee and thus signaled his intentions which were quickly put into action.23 After a series of bitter legislative battles and a Presidential veto the Taft-Hartley Act was passed into law.24 The Taft-Hartley Act amended the Wagner Act in a great many ways. What follows is a non-exhaustive list of the major changes brought about by its passage.25 Structurally, the Taft-Hartley Act achieved a number of important changes. First, the number of Board members was increased from three to five members. Second, and most importantly, the General Counsel became an independent prosecutor no longer under the control of the Board. Therefore, where the functions of prosecution and adjudication had been combined under the Wagner Act they 65 were now separated. Another major structural change was the elimination of the Economics Section. This section had generated a great deal of debate, with Congress refusing funding at one point in time. It was argued that this section had caused the Board to consider non-legal criteria in its adjudication of cases. In the area of unfair labor practices a great number of changes were instituted. While the existing employer unfair labor practices were retained in the act, a new series of union unfair labor practices was added to Section 8 of the act. It became unlawful for labor organizations to restrain or coerce employees in the pursuit of their Section 7 rights. It also became unlawful for unions to cause employers to discriminate against employees. The act further outlawed secondary boycotts, thus isolating all labor disputes. Finally, Section 8(c) ensured employer free speech after numerous complaints that employers had been denied their First Amendment rights under the wagner Act. In the area of representation issues a number of significant changes also took place. First, the treatment of certain classes of employees was spelled out. Supervisors were specifically eliminated from the bargaining units, and guards were to be separated from all other bargaining units. Professional employees were likewise afforded special treatment under the new act. 66 In addition a new series of election types were added to the general employee petitioned elections. First, employers were allowed to petition for elections. Second, decertification elections enabled employees to oust incumbant unions. Finally, elections in the area of union shop arrangements were written into the act. The overall impact of the Taft-Hartley Act on the NLRB was immense. In many ways it changed the basic character of the agency and its role in American society. A number of employees felt so strongly about the new law that they left the agency. For while the government had seemed to say to workers under the Wagner Act “we want you to join unions,“ under Taft-Hartley the government seemed to say “we don't care if you join unions or not.“ “Taft-Hartley had a profound effect on the NLRB as an institution. Perhaps the most important change was pyschological. From the outset, the Board has been charged with promoting unions and collective bargaining. There were only employer unfair labor practices under the Wagner Act, and inevitably the Board's work - however fairly and objectively performed - was perceived as pro-labor. This perception was shared by labor, management, and the Boards own staff. But Taft-Hartley was unquestionably conceived as a series of restraints on unions, and in this sense was designed to favor management. Thus, the Board's job after the Taft-Hartley Act was passed more nearly in the middle of the ongoing conflict between labor and management ... The NLRB had lost its clear identity as an agency committed solelyzgo promoting the organization of unions.“ Although there were a number of attempts to alter the NLRA as amended by Taft-Hartley in the mid 1950's these were unsuccessful. In 1959, the third act affecting the NLRB was passed, the Landrum-Griffin Act. While this act did not have 67 the profound impact that the prior acts had on the NLRB it did make a number of minor changes worth noting. As McCulloch and Bornstein observe, “the new legislation made only relatively minor changes in the fundamental labor laws erected by the Wagner and Taft-Hartley acts.“27 The legislative process that led to the Landrum-Griffin Act began as a series of hearings on internal union abuses. What began as a mild bill to curb these abuses led to the final act after a series of intricate legislative maneuvers with management and labor organizations deeply involved. Unlike the legislative struggles that preceded the passage of the Taft-Hartley Act, the Landrum-Griffin Act was for the most part free from bitter legislative struggle.28 When it finally passed the Congress it had support from labor and management groups, and the final vote was nearly unanimous. The final form of the act contains seven titles, six of which relate to internal union abuses (administered by the Department of Labor), and the last title which amended the Taft-Hartley Act.29 The act altered the structure and jurisdiction of the NLRB in two ways. First, the process for handling representation cases was altered. Prior to the act only the Board could make decisions “to determine bargaining units and direct representation or other kinds of elections.“30 The act now allowed the Board to delegate these powers to the regional directors as the delays in representation cases had become enormous with an increasing workload. The Board, 68 however did not delegate these responsibilities until 1961. With regard to jurisdictional issues, the Landrum-Griffin Act dealt with the “no-mans-land“ problem created by court rulings. The Supreme Court ruled that states were pre-empted from acting in areas of labor dispute where the NLRB had the power to act. However, the NLRB could refuse to act, and thus a set of cases existed where no agency could act. Landrum-Griffin allowed the states to act in these cases, while at the same time stating that the Board had to accept those cases it would have taken under its jurisdictional standards existing in 1959. The Landrum-Griffin Act also amended some basic provisions of labor law under the Taft-Hartley Act.31 First, the act amended Section 9 to allow economic strikers the right to vote in representation elections. Under Taft-Hartley replacement workers could vote and economic strikers could not. Second, it amended Section 8(b) to strengthen the secondary boycotts provision and outlawed “hot cargo agreements.“ Third, it added a new unfair labor practice section to detail regulations concerning recognitional and organizational picketing. Finally, it allowed pre-hire agreements in the construction industry. Since the passage of the Landrum-Griffin Act in 1959 the statutory guidelines of the NLRB have remained constant with one minor exception. In 1970, the Postal Reorganization Act extended the jurisdiction of the NLRB to cover employees of the Postal Department.32 If one views the last two acts 69 as relatively minor adjustments to the statutory guidelines of the NLRB, the agency has existed in legislative quiet for over twenty-five years. CURRENT STRUCTURE OF THE NLRB Today the structure of the NLRB remains largely unchanged from that imposed by the passage of Taft-Hartley in 1947. While the actual organization chart may change from time to time the functions have not. With a few minor exceptions incorporated under the Landrum-Griffin Act, the procedures have likewise remained fairly constant. The major change concerns the handling of representation cases by the regional directors rather than by the board. However, it should be noted that when the Board delegated these powers it did so under strict guidelines.33 The NLRB still relies almost exclusively on the adjudication of cases to establish the basic policy guidelines of the agency.34 The guidelines are published in a case history method in the Annual Report by major case type and policy area. Further, the rulings of the NLRB are constantly updated in law journals along with the appropriate federal court rulings. While the Board has used its rulemaking powers a bit more in recent years it is still the exception and not the rule. The workload of the NLRB has increased dramatically in the last twenty-eight years as shown in Table III-1. In FISCAL YEAR 1953 1955 1957 1958 1960 1961 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 70 Table III-1 TOTAL CASES 14756 13391 13356 16748 21527 22691 25371 27403 28025 28993 30425 30705 31303 33581 37212 41039 41077 42373 44923 49335 53943 53261 54907 57381 ULP CASES 5460 6171 5506 9260 11357 12132 14166 15620 15800 15933 17040 17816 18651 21038 23770 26852 26487 27726 31253 34509 37828 39652 41259 44063 NLRB Workload 1953-1980 REP CASES 9243 7165 7797 7399 10130 10508 11116 11685 11989 12620 12957 12307 12107 12077 12965 13711 14032 14082 13083 14189 14358 12902 12905 12400 71 fiscal year 1953 for example the total number of cases filed stood at 14,756, of which 5,454 were unfair labor practice cases and 10,247 were representation cases. By 1960 the total number of cases filed had risen to 21,527, an increase of about fifty percent. Likewise, the total number of cases filed in 1970 was 33,581, a fifty percent increase in ten years. In 1980 the number of cases filed stood at $5,587 an increase of over sixty percent in ten years. Over the thirty year period the increase was over five-hundred percent.35 As noted earlier the NLRB cannot initiate action but must wait for a petitioner to file charges in an unfair labor practice case (ULP) or file a petition for representation. Petitioners may be employers, employees, unions, etc.. As an example of the adjudicatory procedures involved, the typical procedure for a ULP case will be outlined.36 1. To initiate the ULP case procedure a charge must be filed by the petitioner with one of the regional offices. 2. The charge is investigated by a field examiner who makes a recommendation to the regional director as to whether a complaint should be issued or the charge should be dismissed. If the case is dismissed the decision can be appealed to the Office of the General Counsel. 3. If a complaint is issued the respondent can settle the case informally and on numerous occasions an informal settlement is achieved before the formal complaint is issued. 72 4. If not settled informally a formal hearing is held before an Administrative Law Judge (ALJ). This is a highly judicialized proceeding with the General Counsel prosecuting the charge on behalf of the petitioner. The respondent is represented by his own counsel. At the end of the hearing the ALJ reports his finding either dismissing or upholding the complaint. 5. If any party dissents with the decision the case can be appealed to the Board. The Board then issues its decision on the case. It should be noted that Board orders are not self enforcing and an order must be obtained from the federal courts. Likewise, any party may appeal the decision of the NLRB to the federal courts. It should also be noted that the litigants are allowed to settle the dispute at any point in the proceeding. Only a small portion of the cases filed ever reach the Board and these are the most important cases. The vast majority are settled, dismissed or withdrawn in the early part of the procedure. Although the procedures are much different there are also complex and well defined procedures for the handling of representation issues. A complete guide and explanation of these may be found in A Guide to Basic Law and Procedures Under the NLRA.37 73 MACRO-LEVEL THEORIES AND THE NLRB Based upon the substantive information outlined in the previous section on the NLRB and the theoretical perspectives of private influence outlined in the previous chapter a clearer and theoretically informed description of the NLRB can be constructed. First, we have noted that the NLRB represents a deviant case among the other independent regulatory commissions. The primary reason for this status is the distribution of the costs and benefits in the policy space and the related constituency configuration. A second reason for the deviant case status concerns the statutes of the commission. Whereas most regulatory agencies have promotional statutes the NLRB clearly has a purely policing statute. Most agencies fall within the diffused costs and concentrated benefits pattern and are subject to the pressures of one organized interest. The NLRB, unlike other regulatory agencies, has a concentrated costs and concentrated benefits distribution pattern. Its constituency configuration, to use Wilson's and Olson's logic, is one in which two well organized material groups oppose one another in the policy space.38 This constituency configuration makes the NLRB different from all of the other regulatory commissions and this logic would seem to be consistent with all of the macro-level theories in their developed forms. Bernstein's capture or life-cycle theory may at first seem 74 inappropriate, but as it is developed by Sabatier it fits the deviant case explanation. While Bernstein indicates a similar pattern of development for all commissions, Sabatier shows that the presence of a second organized public interest group can slow the process of capture.39 One can infer from this that when opposing material interest groups exist, the process of capture will be changed dramatically, if not eliminated altogether.40 Wilson's arguments likewise lend support to this prOposition, and he is one of the few authors who fit the NLRB into their general model. Whereas most others avoid the subject of the NLRB, Wilson uses it as the primary case to provide evidence for his different cost benefit patterns.41 Finally, the economic theory of regulation as generalized by Peltzman also confirms the notion of the deviant case. One of Stigler's key arguments concerned the nature of organized interests, information gathering, and the electoral process. As Peltzman notes, there may be more than one organized interest. In the case of two organized interests the pattern of development is likely to be quite different as each side gains control or superior amounts of influence in the legislature.42 It should be noted that the constituency configuration and the cost-benefit distribution pattern of the NLRB can be found in other regulatory commissions along different sub-policy dimensions. However, in the case of the NLRB it dominates the policy space of the entire regulatory 75 commission. For example, one might find the Federal Communications Commission (FCC) faced with this situation when ruling on issues which impact commercial broadcasting and cable television broadcasting interests. Likewise, the ICC would face a similar problem when trucking and railroad interests clash. The statutory development of the NLRB has produced a pure policing statute without promotion for one group or the other. This appears to bear out some of the hypotheses of James Q. Wilson for the macro-level theory pattern described above. Wilson has argued that in the concentrated costs and concentrated benefits pattern the statutes will develop a certain way. He argues that, “a charter will be adopted that contains a definition of the competing rights and obligations of each party.“43 This hypothesis seems to be born out over the statutory history of the NLRB. The NLRA clearly spells out the rights and obligations of labor and management as noted previously. However, as noted earlier the basic charter was not created overnight, but rather over a period of fifteen years, and refined even after this period. After labor's initial victory in 1935, the Wagner Act was passed through the efforts of labor's Congressional allies. The act was clearly a promotional statute which aided labor. During this period the number of union members grew by almost five-hundred percent. In 1947, when management and its allies gained control of the legislature, 76 the Taft-Hartley Act was passed. Only then did the charter of rights and obligations for both sides result. This scenario is in general agreement with the economic theory of regulation as generalized by Peltzman. As each opposing interest was able to gain influence in the legislature through the election of candidates it endorsed and supported, it secured legislation favorable to its position.44 However, one caveat must be inserted. Simply stated, the exchange relationship is not a precise one between two groups which excludes all other parties and factors. In 1935 the labor movement was very small, only three million 45 It is unlikely that organized labor on its own members. was strong enough to make it rational for a Congressman to support such a measure and thereby obtain the legislation it desired over the opposition of management interests. However, as some survey evidence indicates the public had 46 Further, the strong pro-union sentiments at this time. economic conditions were severe, and labor has always done well in terms of securing favorable legislation during periods of crisis. This is demonstrated by the passage of labor legislation in this century, and adds credence to Bernstein's notions of catalysts being important in securing the passage of legislation. When the Taft-Hartley Act was passed, management and labor were both very strong, organized interests; however, public opinion was strongly anti-union.47 Also, the labor 77 movement itself was split. The NLRB under the Wagner Act had supported policies which were favorable to the Congress of Industrial Organizers (CIO), and unfavorable to the American Federation of Labor (AFL). In the period which preceded the Taft-Hartley Act, the AFL was not supportive of the NLRB. As Gross notes, “despite intensive hostile pressure from powerful employers (and others).... the NLRB was in the main invulnerable to political attack until the AFL... entered into an alliance with powerful anti-NLRB employer interest groups.“48 Although, the AFL supported the wagner Act in the legislative confrontation of 1947, previous policies had seriously weakened the act's position. Since the passage of Taft-Hartley the opposing interests have reached a legislative impasse. Wilson has argued that there will be continuing efforts to alter the basic charter.49 Although there have been a number of attempts to alter the charter, neither side has been successful. Only minor changes have been secured, primarily by the Landrum-Griffin Act, as noted earlier, and these were supported by both sides, or at least the final act was the result of compromise. As noted in a recent study on labor and the House of Representatives, labor is simply “trying to retain the status quo“ and maintain the freedoms granted under the Wagner Act.50 There are a number of reasons for this legislative balance and impasse, although none guarantees that the balance will continue indefinitely. 78 First, it may be impossible for either interest to gain enough influence and leverage to secure a major legislative change. Ripley and Franklin, among others, have argued that when two concerted interests come to grips the legislative conflict will be escalated beyond the committee system to a consideration by the entire legislature.Sl Both sides will as Wilson argues, seek to enroll allied interests in their cause. Further, in both of the previous legislative struggles public opinion had been strongly allied to the victor. Unless there is a stong shift away from the status quo in public opinion both sides may be unable to secure alterations in the act. Even when one side is in the minority it may still be able to offset major threats from the other side. Second, the NLRB has existed for a number of years and has established a place in American society and politics. As with most institutions its role has not been completely passive. Its existence has fostered a number of changes. As we noted, during the period of the Wagner Act NLRB, the union membership of the United States grew tremendously. In a sense, the balance which the Wagner Act sought to induce structurally was created by the commission through union growth. Another result of the existence of the NLRB has been the development of a strong labor-management relations community. This community is strongly supportive of the NLRB and its approach for dealing with the conflict between labor and management. Any attempt, then, to alter radically the 79 fundamental role of the NLRB by either opposing interest would require enough influence to overcome both the organized opposition and the supporters of the NLRB. A third consideration which is closely related to the first two points concerns the judicialized procedure of the NLRB. This dependence on an adjudicatory procedure should stabilize or even mobilize public support for the NLRB. Likewise, it provides the underpinning for the strong support that the labor management relations community provides for the NLRB. Spaeth and other scholars interested in the American judicial system have noted the high degree of faith that the general public has in judicialized proceedings.52 These observations have also been supported by survey data which indicate that Americans have a great deal of faith in the judicial process.53 As Spaeth notes: “No aspect of American government is more suffused with myth than judicial decision making. Judges are viewed as bloodless incarnations of human rectitude......Americans believe (or at least hope) that judicial decisions are value-free, essentially nondiscretionary, and objective....No one disputes that legislators, executive officials, and bureaucrats make policy...Americans, however, have resisted the idea that judges are also governmental policy makers.“ This same faith in judicialized proceedings and the role of the NLRB in this process is echoed in the labor management relations literature. Most labor law experts view the passage of the Taft-Hartley Act as a point where balance was obtained in the system.55 The faith in this judicialized system is so strong that a number of labor law experts have 80 called for life tenure for NLRB Board members to eliminate any political influence whatsoever.56 The aforementioned reasons for stability have assumed that either side would prefer a change in the current relationship. However, both labor and management benefit from the existence of the NLRB and therefore may in fact prefer the current arrangement. It provides a system wherein conflict has been institutionalized, and this has introduced a great deal of stability into the environment and a high degree of certainty. Labor has achieved a definitive set of procedures for establishing unions, and a set of protected activities under the NLRA. Mangement also gains from the institutionalization of this conflict as it provides for a peaceful settlement of labor disputes and discourages their spread. In addition, both sides benefit from the role of the government in absorbing the cost in maintaining the system. As noted, when a complaint is filed the government provides the prosecutor at government expense. Likewise the government bears the costs of conducting elections. In effect, the public shoulders the cost of settling labor management disputes.57 These developmental ideas seem to be in agreement with the life-cycle process which Bernstein has set forth. This is not to say that the life-cycle has followed the path which he has predicted, rather that the process of a life-cycle is correct. This is likewise the case if we look at the institutionalization of the agency with regards to 81 its procedures and the development of statutory guidelines. Sabatier has argued that an opposing interest can prevent rigidity in an agency which benefits an organized interest. The NLRB has become institutionalized but in a very different manner. In an environment of opposing interests the commission, in order to avoid conflict and attack from either side, has employed highly judicialized procedures and has eschewed the use of its rulemaking ability. This has been demonstrated in numerous studies of the NLRB.58 Bureaucrats in the agency, who are primarily lawyers, are also likely to desire a specific set of rules to follow so as to reduce uncertainty.59 This argument and its implications will be explored more fully in later sections. From a legislative perspective then, the traditional pluralist arguments would appear to be correct. There has been a balance between the opposing interests. However, before concluding that there are no channels of influence which can be employed, we should look to the less visible means of influence, the bureaucratic influence channels. For as Scher notes: “Regulated interests use whatever means they find available and consider useful to affect, first, the character of the regulatory legislation itself :zd'ggen the direction of the agency that applies If Scher is correct, and we strongly believe that he is, then even in an environment of balanced interests groups will employ whatever means remain available to achieve 82 favorable policy outcomes. If the benefits outweigh the costs, groups will continue to employ certain influence channels to achieve their policy preferences. S UMMARY In this chapter we have shown that the NLRB is truely a deviant case compared to other federal regulatory commissions. First, its constituency configuration consists of two strong material interest groups as opposed to the general case where only one organized interest exists in the policy space. Second, the statute governing the commission is clearly one of rights and obligations for both parties and not a promotional statute favoring the development of one group at the expense of the other. These characteristics provide for a condition wherein one might expect a perfectly balanced environment without any private influence being exerted by either group. However, we have maintained that groups will continue to attempt to influence the policy process. In the next chapter we will explore the influence channels which groups may try employ in order to achieve their policy preferences even in this balanced environment. FOOTNOTES 1. Wilson, “The Politics of Regulation,“ op. cit.. Grant McConnell, op. cit., pp. 298-335. 2. There are other statutes which apply to the NLRB, as well as some minor amendments which are not covered here. These other statutes, such as the Civil Rights Act of 1964, would require too much time and space to discuss and are not primarily pieces of labor legislation. For a complete reference on the statutory guidelines of the NLRB see walter Oberer, Kurt Hanslowe, ad Jerry Andersen, Statutory Supplement to Cases and Materials on Labor Law and Collective Bargaining in a Free Sociéty (St. Pau17 Minn.: West Publishing Co., 1979). 3. Robert E. Cushman, The Independent Regulatory Commissions (New York: Oxford University Press, 1941), p. 345. 4. Archibald Cox, Derek Bok, and Robert Gorman, Labor Lew: Cages and Materials, 9th Edition (Mineola, New York: The Foundation Press, 1981), p. 41. 5. Ibid., p. 42. Also see Cushman, op. cit., pp. 346-354 for a general historical overview of labor legislation prior to the passage of the NLRA. 6. Ibid., p. 41. 70 Ibid., pp. 69-71. 8. For a complete analysis of the National Labor Board and the old NLRB see James Gross, The Making of the National Labor Relations Board: A Study in Economics, Politics, and the Law- Volume I (1933-1937) (Albany, New York: State University of New YOrk Press, 1974), pp. 1-148. 9. Walter Oberer, Kurt Hanslowe, and Jerry Andersen, Labor Law: Collective Bargaining in a Free Society (St. Paul, Minn.: West Publishing Co., 1980), pp. 109-110. 10. Herman Prichett, op. cit., pp. 267-268, and Oberer, Hanslowe and Andersen, Labor Law, op. cit.. PP. 110-116. 11. J. Joseph Huthmacher, Senator Robert wagner and The Rise of Urban Liberalism (Forge Village, Mass.: The Murray Printing Co., 19687, pp. 160-164. See also, Gross, op. cit., pp. 130-138. 83 84 12. James Gross, op. cit., p. 3. These sentiments are echoed elsewhere in the literature. 13. Ibid., pp. 170-183. 14. Ibid., and John Madden, “Origin and Early Years of The National Labor Relations Act,“ in Oberer, Hanslowe and Andersen Labor Law, op. cit., p. 120. 15. James Gross, op. cit., p. 3. 16. Herman Prichett, Op. cit., pp. 269-271. Cox, Bok, and Gorman, op. cit., pp. 69-71. Oberer, Hanslowe and Andersen, Labor Law, Op. cit., pp. 110-116. 17. James Gross, op. cit., p. 132. 18. Madden, op. cit., p. 117. 19. See text of The National Labor Relations Act (49 Stat. 449,29 U.S.C. 151). Text of the act as amended by the Taft-Hartley Act can be found in U.S., National Labor Relations Board, 0 o t Nat'ona Labor Releriene_fieerd (Washington, D.C.: Government Printing Ofc., 1948), PP. 163-176. 20. The following discussion cites sections of the NLRA, op. cit.. Details of the organization and procedures of the Wagner Act NLRB can be found in Gross, op. cit., pp. 157-171 and Frank McCulloch and Timothy Bornstein, The National Labor Rergtions Board (New York: Praeger Publishers, 1974), pp. 20-24. 21.NLRA, Section 1. 22. Oberer, Hanslowe, and Andersen, Statutory Supplement, op. cit., pp. 48-49. In 1947 with the passage of the Taft-Hartley Act the Section 8 unfair labor practices were recoded under Section 8a. 23. Congressional Quarterly, 1947 Congressional Qgerterly Almanac (Washington, D.C.: Congressional Quarterly Press, 1947). PP. 32-33. 24. See 1947 Congressional Quarterly Almanac, op. cit., pp. 279-296, for a detailed account of the legislative history of the act and pp. 301-308 for the voting records of members on the bills. For a detailed discussion of the NLRB and the political and legislative events leading up to Taft-Hartley see James Gross, The Reshaping of the Natonial Labor Relations Board: National Labor Policy in Transition 1937-1947 (Albany, New York: State University of New York Press, 1981). I." '5 u‘ " ' ..' ”\h r 4 '5 1. 85 25. See the National Labor Relations Act as Amended, in the NLRB Twelfth Annual Report, op. cit., pp. 163-176, for the exact changes to the act. For analyses and discussions of the changes see McCulloch and Bornstein, op. cit., pp. 42-53: Oberer, Hanslowe and Andersen, Labor Law, op. cit., pp. 151-153. See also, 1947 CongressionallQuarterly Almanac, op. cit., p. 292 for a list of changes. 26. McCulloch and Bornstein, op. cit., pp. 56-57. 27. Ibid., p. 69. 28. Congressional Quarterly, 1959 Congressional Quarterly Almanac (Washington, D.C.: Congressional Quarterly Press, 1959), pp. 156-172, 382-383, 392-393, 404-407. 29. Ibid., pp. 156-172, and McCulloch and Bornstein, Op. Cit.’ pp. 65-69. 30. McCulloch and Bornstein, op. cit., p. 67. 31. Oberer, Hanslowe and Andersen, Statutory Supplement, op. cit., pp. 42-67, 83-110. 32. Congressional Quarterly, Federal Reguletory Directory 1981-1982 (Washington, D.C.: Congressional Quarterly Press, 1982), pp. 359-360. 33. McCulloch and Bornstein, op. cit., p. 67. For a more detailed discussion of the changes and procedures see the NLRB, Tuentyzfiixth Annual Report of the NLRB, op. cit.. 34. William Murphy, “The NLRB- An Appraisal,“ Minnessota Law Review, 52(819) (1986), pp. 819-849. David Shapiro, “The Choice of Rule-Making or Adjudication in Policy,“ Heryerd Law Review, 78(5) (March,l965), pp. 921-972. 35. U.S., Department of Labor, Bureau of Labor Statistics, Handbook of Labor Statistics 1981(Washington, D.C.: Government Printing Ofc., 1981), pp. 548-551: and NLRB Annual Reports, FY 1979 and FY1980, op. cit., Statistical Appendices. 36. NLRB, A Guide to Basic Law and Procedures under the National Labor Relations Act, (Washington, D.C.: Government Printing Ofc., 1981). 37. Ibid., pp. 8-25. See information on representation elections. 38. Mancur Olson. WW (Cambridge, Mass.: Harvard University Press, 1971). See especially pp. 132-159 for a discussion of economic groups. .- \Ju;‘ "Hid: 86 See also James O. Wilson, Pplitical Organizatioge (New York: Basic Books, 1973), p. 31, and sections on material incentives. 39. Sabatier, op. cit.. 40. As Sabatier discusses only public interest groups with fewer resources it is reasonable to infer that two well organized material interest groups would result in even less capture for either side, if any at all. 41. Wilson, “The Politics of Regulation,“ Op. cit.. 42. Peltzman, op. cit.. 43. Wilson, “The Politics of Regulation,“ op. cit., p. 143. 44. These ideas are also supported by the group theory literature and the political science literature on lobbying. See for example, Louis A. Dexter, HQH_QLganizntinnE_ALe Represented in Washington (New York: Bobbs-Merrill, 1969). 45. Handbook of Labor Statlerlcs 1981, op. cit., p. 412. There are other sources of union membership in the United States, each calculated in a different manner. See for example, Leo Troy, Dierributign of Union Membership Aang_the_States, National Bureau of Economic Research, Occassional Paper 497 (New York: NBER, 1957). 46. Gallup Poll, 1935. Public opinion was clearly pro-union as indicated by questions on unions. 47. Gallup Poll, 1947. Public opinion had shifted after turbulent union organizing years. 48. Gross, The Reshaping, op. cit. p. 2. 49. Wilson, “The Politics of Regulation,“ op. cit., p. 143. 50. Marcus Pohlman and George Crisci, “House Support for Organized Labor,“ PoliticalfiScience Quarterly 97(4) (Winter, 1982-1983), p. 212. 51. Ripley and Franklin, Ppblic Policy (Homewood, 111.: Dorsey Press, 1980). 52. See Harold Spaeth, Spprepe Court Deci§ionlMgking (San Francisco: W. H. Freeman and CO. ) and Lon L. Fuller, “The Form and Limits of Adjudication,“ in American Court Systems: Reedipge in Judicial Process and Behavior ed. by Goldman and Sarat (San Francisco: W. H. Freeman gand Co., 1978). 87 53. Gregory Casey, “The Supreme Court and Myth: An Empirical Investigation,“ Law end Societleeview 8(1974), pp. 385-417. 54. Spaeth, op. cit., p. 3. 55. Oberer, Hanslowe, and Andersen, Lehpr_Leg, op. Cit. I pp. 149-158. 56.Booker and Coe, “The NLRB and Its Critics,“ Leher Lay Journal, 17 522 (1966): and “The Labor Board and Its Reformers“, Labor Law Journal 18 (67) (1967). 57. The budget for the NLRB has grown considerably and is now well over one-hundred million dollars. It is one of the most well funded of the independent regulatory commissions. 58. See Shapiro, op. cit., and Murphy, op. cit.. 59. For a good review of the desire of bureaucrats for rules especially in a complex environment see Donald Warwick, A Theory of Public Bureagcracy (Cambridge, Mass.: Harvard University Press, 1978). 60. Seymour Scher, Op. cit., p. 667. to “l 9 ~ 1% l\ H ‘55. IV INFLUENCE CHANNELS AND THE NLRB Now that the general context of the NLRB has been established based on the general or macro-level theoretical perspectives on private influence we can consider the specific channels of private influence. In this chapter the private influence channels outlined for the general case in chapter two will be discussed within the context of the NLRB in the post Taft-Hartley era. Three specific channels will be investigated: direct influence, the budgetary process, and the appointment of NLRB commissioners. Further, the general hypotheses derived from these discussions as well as the models to be tested later in the dissertation will be presented. Information on where cooperation replaces conflict, where conflict results in stalemate, and where private influence still flows, as well as the logic which underlies this behavior, will better enable us to test propositions regarding the flow of private influence in the NLRB's balanced environment. In the general case, a regulatory commission is faced with one strong organized interest group in the environment. The only opposition is normally an unorganized interest. In this situation all of the influence channels are effective to one degree or another as a means to promote private interests. When the statutory guidelines of the commission 88 an 89 are such that they further promote the interests of the organized group, private influence increases even more, resulting in what many have termed capture. The debate over which channel of influence is the most important has raged on in the general literature, with different factions arguing one position or the other. Our review indicated that the findings were mixed. However, it appears that in the general case the direct influence channel is the most important (if one holds statutory variables constant), followed by the appointment of commissioners. The evidence on the impact of the budgetary process is still too limited to draw any meaningful conclusions. The NLRB, as outlined in the previous chapter, is a deviant case in which the group environment is balanced by two Opposing organized interests. Further, this balance provided by the groups is maintained and furthered by the neutral statutory guidelines of the commission. This has been the case since the passage of the Taft-Hartley Act. It is our general contention that private influence will not cease completely in this balanced environment. While there is a legislative balance and there exists a fairly stable environment, as evidenced in the previous chapter, organized interests will still attempt to influence the commission and achieve their policy preferences. Whenever the benefits of the action are greater than the costs incurred, groups will continue to use the channels of I}: v A 90 private influence open to them. Private influence will still flow through the influence channels, but the way these channels function will change. We now turn to a discussion of the three channels of private influence which are of interest and show how they will function in the NLRB in the post Taft-Hartley era. DIRECT INFLUENCE Direct influence on an agency by private interests through the control of information and expertise is an important if not sufficient condition for high levels of 1 This at least is the private influence in the general case. major view of a number of authors who study private influence. The basic reason is that only one organized interest group dominates the policy space and controls information and expertise which the commission needs to accomplish its mission.2 Even when other channels have been closed, that is, they are used for political patronage or other favors, control of information can lead to high levels of private influence. However, in the case of the NLRB direct influence is likely to be the least successful avenue for private interests to obtain influence . We therefore state the following hypothesis. HYPOTHESIS I: Labor and management groups will both be unable to influence policy decisions, to the benefit of one 91 and at the expense of the other, through the use of direct influence on the NLRB. It should be emphasized that this argument for noninfluence concerns the question of labor or management influence at the expense of the other. It does not imply that the two groups, either singularly or together, cannot influence policy at the expense of some third party. Thus these two groups could employ this channel of influence to gain benefits at the expense of the general public. In this sense, McConnell's arguments for the politics of autonomy are in agreement with our arguments.3 There are a number of reasons for formulating the hypothesis and for believing that it will be supported by the evidence. First, the environment of the NLRB consists primarily of two material interest groups, labor and management, which oppose each other in the policy space.4 Both groups have the resources and the interest to expend them in the pursuit of their policy goals. It is therefore highly unlikely that one group will be able to influence the agency directly as neither has control over information in the policy area. Likewise, neither group has expertise, i.e., labor law experts, which cannot be obtained by the commission or the opposing group. Therefore, the two resources which are at the heart of the issue of direct influence in the general case, information and expertise, are not dominated by any one organized interest, and as such the direct channel of influence cannot be dominated. ‘0’- I“§t— -d a.-- u-J “A. Jac- a.‘ Jr; a. J-..a,~ ...]- 92 Second, the NLRB is a highly judicialized agency which relies on a case method to make policy and not on rulemaking. It is primarily a quasi-judicial agency which has allowed its rulemaking powers to wane.5 This approach allows both sides to present their evidence and reasons for their preferred decision in a highly structured judicial proceeding. It further provides a clear set of precedences and rules for the regulators at all levels to make decisions in accord with commission policy. Further, the judicialized approach allows for numerous reviews or appeals of lower level agency decisions.6 These factors should lead to a situation where the rules and procedures of the agency are important enough to thwart the impact of private influence. Related to this point is the fact that in the post Taft-Hartley era the statutory guidelines of the commission are neutral between the two Opposing groups. They merely put forth the rights and Obligations Of each group. Therefore, the bureaucrats in the agency should not feel predisposed towards the position put forward by either group. In general, the combination of the judicialized process with numerous levels of appeal and the presence Of two opposing interests makes it rational for bureaucrats at all but the highest decision making levels to pursue agency policy goals and not be swayed by pressure from either side. That is to say, intendedly rational bureaucrats in the NLRB (most of whom are lawyers) if they desire promotion will find it rational to follow the law and procedures which "r' .-.-“- \ H-I 93 govern the agency. To Openly favor one side over the other would result in the constant reversal of the bureaucrat's decisions and most probably non-promotion.7 Bureaucrats who are not inclined to follow agency rules and precedence would further find the world outside the NLRB, either the labor or management camp, more to their liking than that Offered within the NLRB. Over time one would then expect that rule followers would come to dominate the commission.8 Third, in all cases the NLRB bears the cost for part of the proceedings if not the entire case. In unfair labor practice cases the NLRB bears the costs for prosecution if a charge is filed. In the area of representation cases the NLRB bears the cost for conducting elections. Therefore, even when a small union might Oppose a large business, or vice-versa, the commission will offset this difference even at the lowest level of agency decision making. Given these arguments it seems unlikely that direct channels Of influence would remain vital or even mildly important. Finally, we should consider where private interests would be most likely to directly influence the policy decisions of the agency, if at all. There are essentially two levels of decision making - the high level policy decisions made by the board and the lower level decisions made by the regions. The vast majority Of the cases which come before the NLRB are settled long before they reach the board.9 Every year only a few hundred of the many thousands Of cases filed with the commission reach the board. A great 94 number Of cases are settled at the regional level both formally and informally. If influence can be exerted it would most likely be present at this lower level. This is the area where the greatest degree of direct contact takes place as rulemaking hearings are not held at higher levels. High level decisions made by the board are the least likely to be directly influenced to a great degree by private interests. First, these decisions have received the most attention and review by the commission. That is, the numerous levels of decision making have provided an adequate Opportunity for review and appeal and it is likely that cases settled solely on the basis of private influence, and not on the basis Of the law, would be overturned. Second, these cases are the most likely to be the cases where the groups will invest their resources. The decision to appeal the case all the way to the board is an indication Of its importance to the group involved. Further, it has long been recognized that the most important cases reach the board.10 Finally, a number of cases which reach this level are automatically appealed to the federal courts, once again indicating the importance of the case and providing yet another level in the review process. It can therefore be argued that if direct private influence does exist in this environment it will most probably occur at lower levels of the agency. For example, it has been shown that the length of time it takes to order a representation election can alter the outcome.11 As the f 95 time increases, managements' probability of thwarting a union victory increases. There may be additional ways in which the groups can affect policy decisions at this level. Therefore, a model which looks at this decision level should inform us as to the degree of private influence which can be directly exerted on an agency in a balanced environment. Although it is extremely difficult to test the impact Of direct private influence in a quantitative fashion, it is possible to construct some statistical data which will add evidence to support our contention. If our arguments are correct, we would not expect a significant amount of variance in the decision patterns reached between different regions within the NLRB. That is, other things being equal, the decision patterns reached in one region of the NLRB should not be significantly different from the overall decision patterns of the agency. If for example we look at the charges brought by unions as unfair labor practice cases, the number dismissed, settled, and withdrawn as a percentage of the total number of charges should be approximately the same from region to region. However, it must be emphasized that this percentage will only be the same if the regions are similar in terms of the types of cases filed, etc.. We would expect variance in the figures on decision patterns from year to year due to changes in the law and changes in the members Of the board. As the law changes over time groups may be slow to respond. That is , what was once 96 legal is no longer legal and groups may find that a charge filed in the past which resulted in the issuance of a complaint is now dismissed. Therefore, one way to investigate the impact of private influence is to investigate the between regions variance in decision patterns for given years. If our contentions concerning the importance of the rule of law and a balanced group environment are correct, the patterns between regions should be highly similar, other things being equal. BUDGET The budgetary process of the NLRB is the second channel of influence that will be investigated. As with direct private influence, we would not expect the budget process to be employed successfully as an influence channel in the attainment of policy preferences by either group at the expense of the other. However, unlike the direct influence channels this condition of balance should not be one which results in spite of attempts by the groups to alter policy. Rather, we would expect that there would be cooperation between the Opposing groups concerning the budget of the NLRB, in lieu of conflict resulting in stalemate from equal and opposing interests. In other words, while the outcomes regarding the impact of private interests on policy decisions are the same in both cases, the underlying political logic which leads to these similar outcomes is 97 very different. The general hypotheses for the budgetary process for the NLRB as a channel of private influence are therefore as follows. HYPOTHESIS II: The budget of the NLRB will not be a channel of private influence which either group will employ to attain its policy preferences at the expense of the other. HYPOTHESIS III: The budget of the NLRB will be based on cooperation between the two opposing interests. The NLRB will be funded to a level which benefits both interests. In the post Taft-Hartley era the budgetary process of the NLRB should be one in which private influence is not exerted to achieve policy preferences of labor over management or management over labor. Rather, we would expect that the budget would be a cooperative process between management and labor. That is to say, in the budgetary process the two groups have become allies, up to a point, in support of the NLRB. As such, the NLRB has a very strong clientele, composed of two otherwise arch rivals, to support its budget requests.12 This was not always the case. In the Wagner Act era there was a continual battle over the budget of the NLRB.13 Management interests and some labor organizations sought to reduce the budget, or at least control the agency programs which were funded. As noted earlier, funding of the Economics Section, which was originally part of the NLRB, 4'4 98 generated heated battles during the budget hearings. Its funding was finally eliminated, although the organization was not specifically eliminated until the passage of the Taft-Hartley Act. Until the passage of the Taft-Hartley Act the NLRB was a commission which promoted labor at the expense of management. Under these circumstances one would expect attempts by management to curtail if not eliminate the funding for the NLRB. Thus in the era before the passage of Taft-Hartley the two organized clients of the NLRB were at odds, and as both attempted to use the budget process to attain their policy goals, conflict resulted. With the passage of the Taft-Hartley Act the NLRB allowed both labor and management to bring charges against the other in this labor court. It established rights and obligations for both, and not just for labor. Under these conditions the NLRB became a mediator between labor and management which settled their disputes regarding the rights and obligations of each party. As it no longer promoted labor, the policy goals which management attempted to achieve were no longer tied to reducing the budget of the commission. In addition to the mediation role, the NLRB also bears a large portion of the costs for the settlement of disputes. In unfair labor practice cases it supplies the prosecutor for the petitioner, and bears the cost of settling the dispute. In representation cases it bears the cost for 99 conducting representation elections. It also bears the costs for conducting investigations in all types of cases. The total cost to the taxpayer for these services was over one-hundred and twenty million dollars in fiscal year 1981.14 As has been noted previously, it is in the best interest of management and labor to settle disputes which arise between them especially as they concern the rules of the game. The costs of continued strife are damaging for both sides. Labor loses income for its members during strikes and cannot achieve the benefits of unionization for its workers ( i.e., higher wages, job security) during protracted representation cases or unfair labor practice disputes. Management likewise cannot gain from costly strikes, especially when the strike is over an unfair labor practice and management cannot hire replacement workers. Representation disputes likewise cause friction among employees, thereby causing problems for employers, and are best settled in a reasonable period of time. The NLRB serves as a means whereby the conflict between management and labor has been institutionalized. It provides for what could be called a court dedicated solely to the settlement of disputes between labor and management. Settlements can be reached between the parties through the NLRB, and in addition the government bears the burden for the cost of the settlements. It is therefore in the interest Of both groups to cooperate and ensure that the NLRB is 100 funded at a level where it can settle disputes between labor and management. This willingness to institutionalize conflict can be seen elsewhere in the labor management community. The National Mediation Board (NMB), binding arbitration, etc. all add additional evidence to support this point.15 In each case the cost of continued conflict and strife between the parties is worse than the institutionalization of the conflict. This is especially true if the government is willing to bear the costs Of institutionalizing the conflict. However, while both sides desire to see the NLRB carry out its mission as it is to their benefit, they do not desire to see the commission overfunded. They prefer to see the agency receive adequate funding, but not so much as to allow it to alter its current role in the process. With more than adequate funding the NLRB could possibly begin to expand its role and exert more control over labor management relations. It might even desire to begin making rules, and to consider delving into areas it has ignored in the past. Such a situation would not appear to be a benefit for either labor or management. For while they benefit from the institutionalization of conflict that the agency provides, they are not likely to gain from more governmental control of their affairs. The status quo Operation of the NLRB is a known quantity. There is certainty in the relationship between the 101 parties and the commission. To overfund the agency would result in a highly uncertain environment in which the commission's power over the parties would increase. While this might result in short term gains for either side, the long term prospects are filled with only one certainty: the power of the commission over their affairs would increase. For these reasons, both labor and management should desire to see the NLRB funded in a manner which allows it to accomplish its present mission of resolving disputes between the parties. If this is the case, then one might expect that the NLRB will be funded so as to complete its workload. This alternative explanation is viable without worrying about attempts by the NLRB to expand its role as the agency does not generate its own workload. As noted earlier the NLRB cannot initiate a case but must wait for labor or management to do so. A model of the budgetary process of the NLRB that would provide an adequate test for our hypotheses is one which builds on the notion of funding based on workload. The budget of the NLRB should be based on an estimate of the workload that the agency expects to receive in the coming fiscal year. Further, the model should include other factors relevant to the budgetary process as they impact the workload. That is to say, backlogs, changes in the case handling process which may reduce or increase costs, and economic conditions should also be included in the model. 102 The chart at Figure IV-l represents the model of the budgetary process for the NLRB in the post Taft-Hartley era. The estimated workload of the NLRB can be broken into two distinct types- unfair labor practice cases and representation cases. As noted earlier these two case types are different and therefore a model of the budgetary process based on workload should include both types. The second set of impacts are those of the workload backlog of the commission for both types of cases. Underestimates in the estimated workload in the previous fiscal year or a greater percentage of cases requiring high level settlement can lead to unexpected backlogs in the commissions workload. These backlogs should also be taken into account when considering the commission's budget. The next consideration is the impact of the institutional changes in the processing of the commission's workload. The Landrum—Griffin Act had such an impact on the case handling procedures for the commission. As noted earlier, the act delegated the handling of representation cases to the regional directors under strict guidelines. This institutional change in the way representation cases were handled would obviously impact the cost estimates for representation cases. Most probably, the cost of handling the cases would increase and the backlog associated with representation cases would decrease. The act had no bearing on the manner in which ULP cases were handled, and as such it should have no impact on the cost estimates or backlogs 103 FIGURE IV-l Cooperative Budget Model WORKLOAD ESTIMATES /_ ULP REP ESTIMATED ESTIMATED WORKLOAD WORKLOAD NLRB BUDGET CONSUMER PRICE INDEX ¥ 4 ECONOMIC FACTORS ULP CASE BACKLOG - \ INSTITUTIONAL CHANGES LANDRUM GRIFFIN ‘ ACT PROCEDURES REP CASE BACKLOG CASE BACKLOGS 104 of ULP cases. The final factor to be taken into account concerns the impact of economic conditions on the budget of the NLRB. Specifically, the impact of inflation and the falling value of the dollar must be taken into account. As the inflation rate has increased, the value of the dollar has fallen and therefore the dollar is not a constant measure over time. The inclusion of this variable in our model will provide us with a constant dollar over time. Each of these variables should be positively related to the budget enacted for the NLRB. That is, as each variable increases in value the budget of the NLRB should increase in value. COMMISSION APPOINTMENTS The third and final channel of private influence involves the appointment of commissioners or board members to the NLRB. Unlike the other channels of private influence which lead to either an unwilling balance such as direct influence, or balance through cooperation as in the case of the budget, the appointments process should remain an active channel for private influence. The hypotheses concerning the appointments process can be stated as follows. The first hypothesis offered is a general statement while the others which follow from it will verify the first. Therefore our efforts will be directed at developing these latter two 105 hypotheses. HYPOTHESES IV: The appointments process will remain an active channel for private influence with both groups attempting to secure appointees who hold policy preferences similar to their own. HYPOTHESIS V: Influence on the appointments process will be exerted by labor and management through both the Congress (the Senate oversight committee) and the President. HYPOTHESIS VI: The impact of private influence through the appointments process will be an important factor when considered in conjunction with institutional and economic factors in determining the policy decisions of the NLRB. The appointments channel should remain an important avenue for private influence for many reasons. Among the most obvious of these is the fact that the channel should increase in importance as other channels begin to close, either through a balance due to conflict or through cooperation and accommodation. The reason for believing that the appointments channel will remain open while most of the other channels of private influence begin to close is derived from the general level theory. That is, when two organized groups Oppose each other in the policy space a balance or stalemate results. However, some influence channels should remain open where the organized interests find it in their benefit to apply pressure and are capable 106 of doing so. It follows that those areas which remain open will become more important, and perhaps more structured. Structured here refers to a pattern of bargaining relationships which can be expected to develop. In the general case the appointments process of different commissions exhibited different histories and no specific determination was made concerning its impact on the attainment of private preferences. In some instances the appointments are used as political patronage. In others, although less frequently, pro-consumer officials are appointed. In still other instances, the dominant organized interest group is able to secure favorable appointments. In the case of the NLRB the pattern and discussion of pro-management and pro-labor boards is well known.16 When labor controls the means of appointment, a pro-labor board is expected to be appointed. The same situation holds true for management. One would expect that this would be the situation and that the likelihood of the positions being used as patronage positions would be extremely low. The positions are much too valuable to strong constituencies to be used for any other reason than to maintain the support of those constituencies. Related to the above argument is the consideration that the direct channels of private influence have been closed. If this is the case, then the preference structures of commissioners will not be altered after their appointment. The post-employment hypothesis which is so popular in the 107 general literature is inoperative in the case of the NLRB. That is to say, in the general case it is argued that the preferences of commissioners can be altered by the post employment opportunities presented by the organized industry.l7 Commissioners, it is argued, will seek employment in the organized sector after they leave the agencies and will therefore rule in their favor. However, in the case of the NLRB a commissioner can find post-employment opportunities in either camp. Thus, there is no pressure on a commissioner to change his/her preferences. It therefore becomes imperative to control the appointments process so as to select individuals who are initially in agreement with the position of an organized group as the probability of changing their preferences later is extremely low. Another major reason for the importance of the appointments process of commissioners concerns the types of decisions that the board members make. As noted previously, commissioners make the high level or policy decisions of the agency when they rule on major cases. These decisions are in fact the precedents and policies upon which the commission operates.18 This is especially important if our notions concerning the preferences of lower level bureaucrats in a judicialized agency are true. That is, if agency bureaucrats are rule followers, the impact of a board decision will increase. Given their key position in the commission and the high level policy decisions they make, an organized interest can gain a great deal of influence by influencing the 108 appointments process and dominating the board. The above general argument on the importance Of the appointments is a hypothesis that is well accepted in the literature on the NLRB. It has long been an accepted fact that boards are either pro-labor or pro-management. The boards appointed by Truman, Kennedy, Johnson, and most recently Carter have all been reputed to be pro-labor. Along the same lines, the boards appointed by Eisenhower, Nixon, and now Reagan have a reputation for being pro-management. Democrats have long been associated with organized labor and management groups have long been associated with Republicans and these findings appear consistent with this knowledge. This position has been corroborated by studies on NLRB appointments, and is also verified in many areas of labor law case history.19 In all of these studies as well as the legal writings the President has been seen as the primary force behind the appointments process. That is to say, it has been argued that the President determines the policy positions of board members. While we agree with the general view on the existence of pro-labor and pro-management boards, we would argue that not only does the President play an important role in the appointments process but so does the Senate. The role of the Senate to provide “advise and consent“ has been too long overlooked. Therefore, in line with our fifth hypothesis we would argue that both the President and the Senate will play a key role in the appointment process. 109 Therefore, organized interests will find it beneficial to lobby strongly both at the legislative level and the Presidential level when seeking to obtain the appointment of commissioners who hold policy preferences similar to their own. The traditional view of the strong role afforded to the President and the limited role afforded to the Senate oversight committee is based on a number of different points. First, it has long been accepted in the appointments literature that the President is the major actor. There is mixed evidence on this point, however as a generalization it would seem to hold. The extension of this situation to the NLRB therefore seems to be rather straightforward. Second, the evidence, at least the written record, would seem to support this position. The confirmation hearings on the Presidents' appointments to the NLRB are extremely brief and almost without controversy. Further, the appointees of the President have been confirmed by the oversight committee without exception. Third, the investigators have been able to explain the pro-management and pro-labor leanings of the different boards without the need to involve the Senate in the process. Democratic Presidents with the strong backing of labor have been said to appoint very strong pro-labor boards. Labor is so important to them as an organized constituency, and management support is so unimportant that this would be the expected outcome. Republican Presidents, 110 on the other hand, while they receive and desire strong management support cannot afford to alienate labor. The union movement has too many members and cannot therefore be totally ignored. Therefore, these boards, while they are pro-management are only moderately so. This contention has been supported by all of the studies on the appointments process of the NLRB to date.20 An alternative explanation, and one which we find to be more plausible, involves the role of the Senate oversight committee in the appointments process. In other words, we would argue that the so called moderately pro-management boards appointed by the Republican management supported Presidents in recent years are the result of labor control of the Senate oversight committee and not due to the attempts of the President to molify labor. The allies of labor have controlled the legislature for many years, maintaining control from 1955 through 1979, forcing the appointment of these so-called moderate boards. The reasons for the active role afforded the Senate in the appointments model are as follows. First, the Senate Labor Committee has long played an extremely active role in the policy area concerning the NLRB. All of the major pieces of legislation concerning the NLRB originated in the Senate Labor Committee. The Wagner Act had little Presidential support from Roosevelt, and the Taft-Hartley Act which was also born in the Senate committee was vetoed by Truman. In both of these cases the oversight committee in the Senate lll played the leading and the dominant role. Likewise, the Landrum—Griffin Act was born in the legislature. A second reason for the active role of the Senate Labor Committee concerns the nature of the political conflict which surrounds the appointment process. The conflict does not escalate and end up on the floor of the Senate for consideration. Escalation of conflict leads to higher costs for the legislators as both sides attempt to enroll more allies in their cause.21 As the conflict remains localized during the appointments process, it is easier for interests to apply pressure and also easier for legislators to respond to it. A third reason concerns the prominent role the committee has for members from strong labor states. While the Senate Labor Committee is generally regarded as a low prestige committee, members from strong labor states have chosen to stay and remain on the committee. Likewise, when management endorsed candidates control the Senate the committee becomes very valuable for them. The decision of Taft to assume the chairmanship of the committee over a much more prestigious appointment is a good example of this.22 The importance of the post has been reinforced when one notes that the chairman of the Senate Labor and Public Welfare Committee retained control of the Labor Sub-committee after the committee reorganization in 1974. This subcommittee assumed the control over the NLRB that had been held by the entire committee prior to the 112 reorganization.23 From the perspective of a Senator the use of the appointment power to make political profit is also sound. Those on the committee usually come from strong labor states where the ratio of labor union members to eligible voters is approximately .18.24 As the conflict remains localized, a Senator can obtain political favor from a large supportive constituency without incurring the wrath of other constituencies. Based on the above reasoning we would argue that a bargaining process between the President and the Senate Labor committee will determine the preferences of the board members. When labor controls the Senate committee and the Presidency, i.e., candidates they endorsed are in control, the preferences of commissioners should be highly pro-labor. When management controls both the Presidency and the Senate committee the preferences of commissioners should be highly pro-management. When there is a split over control, i.e., labor Senate and management President, or management Senate and labor President, the preferences should be closer to a more median position. The last hypothesis in the study concerns the importance of the appointments process given the institutionalized nature of the commission. As has been argued throughout, the NLRB is a highly judicialized agency with a great deal of legal precedence. As such, it is important to consider the impact of the appointments process 113 within the context of these factors. In addition to the institutional factors noted, economic conditions should also have an impact on cases which come before the board. These will also be considered in the model. A diagram for this model can be found at Figure IV-2. The NLRB is a commission which depends on labor law and a case method to determine its policies. As such, one would expect that the prior rulings of the commission would have an impact on the cases which come before the board. It is highly unlikely that the board starts at ground zero with every case that comes before it for consideration. In fact, if one approaches the study of the NLRB from the perspective of a legal scholar, the law and the precedents established in prior cases are the paramount reasons for the decisions reached by the board in almost every case.25 Legal scholars recognize the importance of the political nature of the appointments to the board, but for many it is at best a nuisance which on occasion rears its ugly head. Some have found it to be more troubling and have argued that the board members should be appointed for life terms along the same lines as federal judges.26 One law student, majoring in labor law at a prestigious law school, remarked that the only thing wrong with the NLRB was the political nature of the appointments to the board. If this problem could be solved, the tenets of labor law would take care Of themselves.27 114 FIGURE IV-2 Case Decision Model POLITICAL APPOINTMENT FACTORS ll, PRESIDENTIAL SENATE APPOINTMENT APPOINTMENT NLRB v VOTE DECISION L ECONOMIC CONDITONS AGENCY GENERAL DECISION COUNSEL DECISION \i INSTITUTIONAL FACTORS 115 Such attitudes towards the law in general and the role of judges are common in the United States, although political scientists such as Spaeth have done much to dispel them in recent years.28 Still, one cannot deny that the impact of the law created by the commission in previous rulings or contained in the enabling statutes should have an impact on the rulings of the board. Previous studies have not considered these factors and we consider this to be a great shortcoming in these works. Two major factors should capture the impacts of the law and the prior rulings of the board on current board decisions. These are the rulings of the commission on a case before it comes to the board and the ruling of the prosecutor in the case. Almost all cases are considered elsewhere in the commission before they reach the board, and as such there is a ruling on almost all cases when they comes before the board. As the ruling may change as the case is processed or appealed through the agency, some consistent prior level of ruling is neccessary. For this reason, we can look at the last ruling on a case before it comes before the board. These would be either the rulings of the Trial Examiner or the Regional Director depending on the type of case.29 The second factor concerns the ruling of the prosecutor in the case. The decision of the commission to issue a complaint is an indication of the fact that it has decided that the charge has merit under the existing laws of the 116 commission. The General Counsel is an independent officer who has the final decision as whether or not to prosecute a case based on the initial evidence. It follows that if the prosecutor follows the law as laid out by the commission in previous decisions, his views would have a positive bearing on the case and impact the rulings of the board. Under provisions contained in the Taft-Hartley Act the General Counsel became an independent officer.30 Prior to this point in time the board controlled the Office and thus the functions of prosecution and judgment were combined. As such, the rulings reached by this Office are not merely rubber stamp decisions issued by the board. However, if the law has an impact in line with the reasoning laid out earlier, the support of the General Counsel should have a positive impact on the rulings of the board. Therefore, the support of the Trial Examiner or Regional Director constituting the last major ruling of the commission and the support of the General Counsel should both have a positive relationship to the rulings reached by the board. Summary evidence would also lend support to this view. The number of decisions accepted by the board without change from the Trial Examiner as reported in the aggregate data of the Annual Reports of the NLRB is quite high.31 Also, the number of cases in which the petitioner wins is also quite high as reported by the aggregate data.32 The final factors which should be considered are the economic conditions which exist in the environment. These 117 are by far the most difficult factors to incorporate into our model. For while it seems to be a generally accepted truth that economic conditions impact labor management relations and government policy decisions, the actual ways in which these impacts occur is far from understood. Previous studies on the NLRB and the impact of economic factors on the decisions reached by the commission have assumed that the conditions impact the commissioners directly. That is, when making decisions on the cases which come before it, board members will consider the economic conditions of the day in reaching their decision. The general arguments are confusing, but one may assume generally that when times are bad for unions that the commissioners will be more predisposed to rule in their favor. One study recently looked at a large number of economic factors and reached inconclusive results at best.33 This same view is reinforced by the historical role of economic factors as criteria for decisions by the NLRB. While it was only in existence for a few years, the Economics Section caused a great deal of controversy and has had a lasting impact in terms of the way scholars view the role of economic factors as a basis for decisions. The view taken in this study is somewhat diferent. While we would agree that economic conditions will impact the decisions of the board, the way in which this impact flows is much different. Major board decisions, as noted earlier, comprise only a small portion of the total number 118 of cases. The decision to fight and appeal a case to this level and not settle earlier is an indication of the importance of the case. In difficult economic times, we would argue that parties are predisposed to fight more cases on the fringe of the current law. That is, cases in the gray zone which would be more likely to alter the current law in some way are more likely to reach the board. In good economic times, such cases may be of less importance and a settlement may be more desirable. Given this argument, the economic conditions which exist at the time a case is filed, or in its early stages are likely to impact the decisions of the board. As unions bring the majority of the cases before the board, the overall impact of these economic factors should be to increase the number of union victories. SUMMARY The propositions put forth in this chapter regarding channels of private influence can be summarized as follows. First, it should be very clear that the change in the environment of the policy space and the statutory guidelines of the agency has led to a dramatic shift in the relative importance of the channels of private influence. In the general case all of the specific channels investigated were important. The relative importance Of each channel in the general case is difficult to ascertain but the following ll9 generalizations seem appropriate. First, the direct channels of influence appear to be important if not sufficient to account for high levels of private influence. The appointment process is likewise an important channel but if it is not controlled and is used as a source of political patronage other means can be used to overcome it. The budget as a tool for oversight and private influence has not been readily tested but there are indications that it is also important. In the case of the NLRB, the direct channel of private influence is closed to the groups involved and this forces other channels to become more important. Further, as long as the commission is a mediator of disputes and not a promoter of one side or the other the lower level officials will not be predisposed to support either side. The budget and its control is radically different and ceases to become a channel of private influence, but rather one of cooperation between the two opposing groups. This situation should continue as long as both of the interests find it to their benefit to institutionalize the conflict between them. Finally, the appointments process becomes by far the most important channel of private influence. It has become a highly structured channel which is not used for political patronage, but for the exertion of private influence by both groups. Further, it will remain important even when considering the institutional and economic factors which impact the commission. That is, even when the commission 120 becomes highly structured and institutionalized the appointments process will remain important. FOOTNOTES 1. For an example of this view see Micheal Porter and Jeffery Sagansky, “Information, Politics, and Economic Analysis: The Regulatory Decision Process in the Air Freight Cases,“ Public Policy, 24(2) (Spring,1976), pp. 263-307. See also Louis Kohlmeier, The Regulators: Watchdog Aqencies and the_2ublic_lnteresf Op- cit-. pp- 69-82- 2. Prggeedings, Vol. III of the op. cit. This study makes clear the massive one-sidedness of participation. 3. Grant McConnell, Private Power and American Democracy, op. cit., pp. 298-336. 4. See Mancur Olson, The Logic of Collective Action, op. cit.. Olson cites business associations and unions as the primary examples of material interest groups. See especially pp. 132-159. See also James O. Wilson, Political Organizations, Op. cit., p. 31 and sections on business and union interest groups. 5. William Murphy, “The NLRB- An Appraisal,“ Minnesota Law Review, 52(819) (1976), pp. 819-849. 6. See Chapter III for a complete discussion of the hearing procedures for ULP cases. For further information see also NLRB, A Guide to Basic Lagyand Procedure§_unggr_thg National Labor Relatons Act, (Washington, D.C.: Government Printing Office, 1981). 7. Rodger Noll argues that one of the primary preferences of top level bureaucrats is not to have their decisions overturned. It follows that this preference can be attributed to lower level bureaucrats. See Rodger Noll, “The Economics and Politics of Regulation,“.Yirginia_Law_Rexiawp 57 (1971), pp. 1016-1032. 8. This argument is somewhat similar to that provided by Anthony Downs. He argues that over time zealots will be replaced by conservers in an agency. Here, we argue that rule followers will come to dominate the agency for different reasons. In each case the agency becomes dominated by a type of bureaucrat that finds the benefits of remaining in the agency akin to the preferences that they hold. See Anthony Downs, Inside Bgreaucracy, op. cit.: PP. 5-24, 88-91 0 121 122 9. A review of the Annual Reports of the National Labor Relations Board, op. cit., for all of the years of the agency makes this point abundendtly clear. For example, in Fiscal Year 1976 only four percent of the Unfair Labor Practice cases closed reached the Board. This year is not atypical from other years. 10. See Frank McCulloch and Timothy Bornstein, The National Labor Relations Board, pp. 90-93. 11. Myron Roomkin and Richard Block, “Case Processing Time and the Outcome of Representation Elections: Some Empirical Evidence,“ Uniyersity 9f Illinois Lew Reyieg, 1981 (l) (1981). pp. 75-97. 12. Aaron Wildavsky, 2W Process, Op. cit., pp. 65-74. 13. NLRB budget hearings before the Senate and the House of Representatives. See also James Gross, The Reshaping of the National Labor Relations Board, Op. cit., pp. 132-147, 168-173. 14. Office of Management and Budget, Budget of the United States 1981, (Washington,D.C.: Government Printing Office, 1981). 15. The National Mediation Board is an agency of the federal goverment designed to break the deadlocks which occur in labor management negotiations. 16. This proposition is well accepted in the literature on the NLRB, as well as by those who participate in the day to day activities of the board. See for example Frank McCulloch and Timothy Bornstein, TheyNetiongl Labor Relations Board, Op. cit., pp. 60-77. 17. Paul Quirk, Industry Influence in F edegal We, op- cit.. pp- 143-174- 18. A study of labor law in the area of the NLRB and the National Labor Relations Act makes this point abundently clear. The NLRB in its rulings goes to great lengths to cite precedents from previous cases employed in reaching a decision. This is also clear from the hearings conducted by Trial Examiners as reported in The Decisions and Orders of the National Labor Relations Board, op. cit.. This is also true for the preceedings at the regional level. 19. Seymour Scher, “Regulatory Agency Control Through Appointment: The Case of the Eisenhower Administration and the NLRB,“ op. cit., pp. 667-688. David Welborn, “Presidents, Regulatory Commissioners, and Regulatory Policy,“ op. cit., pp. 3-39. C. DeLorme, R. Hill, and N. 123 Wood, “The Determinants of VOting by the NLRB on Unfair Labor Practice Cases: 1955-1975,“ op. cit., pp.207-218. 20. Seymour Scher, “Regulatory Agency Control Through Appointment: The Case of the Eisenhower Administration and the NLRB,“ op. cit., pp. 667-688. David Welborn, “Presidents, Regulatory Commissioners, and Regulatory Policy,“ Op. cit., pp. 3-39. C. DeLorme, R. Hill, and N. Wood, “The Determinants of Voting by the NLRB on Unfair Labor Practice Cases: 1955-1975,“ op. cit., pp.207-218. 21. See Randall Ripley and Grace Franklin, Congress, the Bureaucracy and Public Policy, op. cit., pp. 123-130. 22. For information of the prestige of different committees see Donald R. Matthews, U.S. Senators and Their Werlg (New York: W. W. Norton and Co., 1973), pp. 147—176. The mean number of years Democratic Senators remain on the committee is eight as opposed to three and one-half for Republican Senators. For information on the decision of Taft to assume the chairmanship of the committee see Congressional Quarterly, 1947 Congressional Quarterly W0 OP. Cite! 990 32-330 23. U.S., Congress, Senate, Cengessional Directory (Washington, D.C.: U.S. Government Printing Office, 1974). 24. The ratio reported here is that of the total number of union members divided by the number of eligible voters in a state. Union membership figures were obtained from the Bureau of Labor Statistics, Handbook of Labor Statistics, op. cit., p. 412. Voting pOpulation figures were obtained from the Bureau of Labor Statistics, Statistical Abstract of the United States (Washington, D. C.: Government Printing Office, 1980). 25. See legal texts on labor law concerning the National Labor Relations Act. For example see Oberer, Hanslowe, and Andersen, Labor Law, Op. cit.. For information on NLRB cases before they go to the federal courts see individual decisions of the NLRB in Decisions and Orders of .the_Netignel_Labor Reletions Board, op. cit.. 26. Booker and Coe, “The NLRB and Its Critics,“ op. Cite I Pp. 175-193. 27. Conversation with Richard Polley, law student, University of Indiana Law School. 28. Harold Spaeth, Supreme Court Decision Making, op. cit.. 124 29. Normally unfair labor practice cases are last heard by a Trial Examiner (called Administrative Law Judges after 1974) although in some rare cases decisions can go directly to the board. Regional Directors are often the last decision making authority on representation cases. For more details on this point see NLRB, A Guide to Basic Law and Procedures under the National Labor Relatons Act, (Washington, D.C.: Government Printing Office, 1981). 30. NLRB, Twelfth Annual Report of the National Labor Relatone Board, op. cit., PP. 163-176. 31. See the Annual Reports of the National Labor Relations Board, op. cit.. Figures are reported in the statistical indexes. Approximately have of all the decisions are accepted by the board without any changes. This is true for most of the years in the post Taft-Hartley era. 32. Ibid. These aggregate figures are analyzed by case type. This pattern of petitioner victories holds true for all case type categories. 33. C. DeLorme, R. Hill, and N. Wood, “The Determinants of Voting by the NLRB on Unfair Labor Practice Cases: 1955-1975,“ op. cit., pp.207—218. A number and wide variety Of economic variables are included. These are unemployment, inflation, and strike days lost, etc. V RESEARCH DESIGN METHODOLOGY AND DATA The purpose of this chapter is to provide information on the research design and methodology utilized in the dissertation. The models employed to test the hypotheses outlined in the last chapter will be operationalized, and the statistical methods and techniques to be employed will be detailed. Further, information and documentation on the data sources and sampling procedures will be provided. The chapter begins with a general discussion of the research design and statistical methods and then explores each of the three areas of private influence in detail. This study employs what has come to be known as an “outside“ research design, as opposed to the more popular inside research design. This terminology was developed by scholars studying Congress and is useful also in the study of bureaucratic institutions. The latter approach normally consists of working within an agency or institution, conducting interviews with those directly involved in the process, engaging in participant observation, conducting an investigation of agency records, etc. This technique is normally qualitative in its approach, making little use of quantitative data and advanced statistical techniques. In the congressional literature Richard Fenno is the master of this technique, and in the study of bureaucratic 125 126 institutions Herbert Kaufman has employed this same technique with great success.1 The outside design normally makes extensive use of quantitative data, employs a wide array of statistical techniques, constructs measures of institutional action and develops models of agencies and their environments, constrained by the data available. Studies of this type have been used extensively in the congressional literature in roll-call studies, and in studies of congressional institutionalization.2 In the study of bureaucratic agenices much less use of this technique has been made, although a number of very fine studies do exist.3 Many studies utilize both techniques either explicitly or implicitly. This study makes wide use of the information of numerous inside studies on regulation in general and on the NLRB in particular. The model of the world offered here is therefore highly dependent on the substantive information contained in numerous prior pieces of research. In addition, this study employs a number of agency and Congressional documents to provide additional evidence in support of its arguments. No study, even the so-called outside research design, can be properly conducted without a complete analysis of the agency involved. This study could therefore be considered a theoretically informed quantitative case study. The models, their specifications, functional forms, and the variables utilized in their construction all depend on prior 127 information and research. To utilize only statistical results without prior information and a theoretical base would be an exercise in futility. The result would be a large number of meaningless statistical correlations producing nothing, and adding even less to our knowledge about the NLRB. Unlike many studies of regulatory agencies and bureaucratic institutions this study focuses where possible on performance measures of agency decision making rather than relying on the analysis of impact data and infering motivations or actions based on outcomes. This should eliminate a number of serious validity threats from external factors. As has been noted previously, a great deal of the literature in the area of regulation relies on the use of impact data. The studies by economists are especially prone to this type of data analysis.4 Some recent studies have begun to utilize performance measures of agency activity, among them Moe's study of the impact of presidential administrations on regulatory agencies.5 STATISTICAL METHODOLOGY The statistical methodology employed in this study for the most part is centered on the general linear model (otherwise known as the classical linear regression model), although other statistical methods such as descriptive statistics are employed where appropriate. Within the 128 context of the general linear model a number of different estimation techniques are also employed. These include Ordinary Least Squares (OLS), Generalized Least Squares (GLS), and Maximum Likelihood Estimators (MLE) employed in Probit Analysis. Before continuing with the discussions of each model we will first outline the general linear model and the estimation technique of OLS so that the other techniques will have a common point of comparison. The general linear model and the assumptions upon which it rests can be written in the following manner. (1) y=XB + u Assuming there are n observations, i = 1,2,...n, and there are k-l exogenous variables, y is a column vector of the observed y values, of order n x 1. x is a matrix of order n x k of the observed values of the exogenous variables with a column of ones for the intercept term. B is a column vector of the parameter values of order n x l, and u is a column vector of the disturbance terms of order n x l. (2) E(u) = 0 (3) E(uu') = 021 (4) x is a set of fixed numbers (5) X has rank k < n 6 The first assumption is that the model is correctly specified. In this instance this means that the endogenous 129 variable can be computed as a linear function of a set of specified exogenous variables plus an error or disturbance term. The second assumption states that the expected value of the disturbance term is zero. The third assumption is that the disturbance terms have a constant variance and are not correlated with one another. The fourth assumption is that the observations of the independent variable are fixed in repeated samples. Finally, the fifth assumption states that the number of observations is greater than the number of exogenous variables and that there are no exact linear relationships between the exogenous variables. Normally, the ordinary least squares (OLS) estimation technique is the preferred estimator, and by far the most commonly used among socal scientists. Some of the reasons for this widespread use are as follows. First, the OLS estimator has a number of desirable theoretical properties. It can be demonstrated that the OLS estimator is a Best Linear Unbiased Estimator (BLUE). Unbiased meaning that the expected value of the estimate is equal to the parameter value, and best meanig that the .._.~-: ‘ ""— estimate has minimun variance. These prOperties can be shown through the use of the Gauss-Markov Theorem, assuming a number of the assumptions Of the classical linear regression model are not violated. Second, the computational costs and availability of program packages make this technique relatively inexpensive. This is especially important when considering the analysis 130 of large data sets. Third, the technique is widely known and widely used by social scientists making the results interpretable to a great number of individuals. Finally, the OLS estimator is a robust estimator and thus is a logical choice when one faces unknown problems that may appear during data analysis. Here the term robust is used to describe an estimator “as one whose properties are insensitive to departures from the assumptions under which it is derived.“7 The popularity of the OLS estimation procedure is perhaps too great. There are numerous instances when the assumptions necessary to produce the desirable properties of the estimator are violated leading to poor estimates and even more importantly false conclusions when performing hypothesis tests. The hypotheses tested in this study are such that serious violations of the assumptions can occur and therefore alternative estimation techniques have been employed. These are Generalized Least Squares (GLS) and Maximum Likelihood Estimation (MLE) employed in Probit Analysis. We now turn to the consideration of the three areas of private influence as outlined in the last chapter. MODELS In our discussion of each of the following models we will outline the operational hypotheses and the models to be 131 tested. Next the statistical methods and estimation procedures employed will be considered. Those which are complex or not widely known will receive more indepth treatment. Finally the data sources and sampling techniques employed will be discussed. Although the order of the presentation may change in our discussion of each model all of the above information will be included. DIRECT INFLUENCE It is extremely difficult to provide exact measures and statistically test direct influence on the NLRB by private interests. Our approach here is to provide quantitative evidence in support of our contentions concerning a lack of private influence. If these ideas are correct, rulings at lower agency levels within distinct regions should not be different than those for the agency as a whole. To provide this quantitative evidence in support of our ideas, we will employ simple descriptive statistical data. Decisions reached at lower levels of the agency can be divided into four types. Cases may be withdrawn by the petitioner (normally at the urging of the agency), dismissed by the regional director, settled by the regional director (the parties come to an agreement out of court so to speak), or finally, a complaint may be issued by the regional director. Employing NLRB data we can obtain information on rulings for the agency as a whole at this decision level and 132 information on the rulings within each region. These data can then be compared and the similarities and the differences in the patterns of the rulings can be noted. The data sources employed in this portion of the dissertation were obtained from the records of the NLRB. While the commission has reported aggregate figures on the disposition of cases since its inception in the Annual Report of the National Labor Relations Board, detailed reports have never been made available. That is, the information on individual cases has not been provided in a format that would be useful to those interested in political questions. Nor for that matter have the lower level decisions been reported. Cases have been reported only on the basis of the final ruling. Beginning in 1964, the NLRB began to record the information on individual cases on magnetic tapes. Prior to this time, it was impossible to obtain the data as it was not recorded in a format which was readily available. The data employed here are taken from the period 1964 to 1980. The data are from tapes containing information on all unfair labor practice cases closed during this period. These tapes provide information on the disposition of all closed unfair labor practice cases, including the respondent, the petitioner, the level at which the case was closed, etc.. A total of 437,920 cases closed during this period are recorded on these tapes. The tapes and the accompanying documentation, i.e., codes and codebooks, were obtained from 133 the NLRB Media Library. Due to the great number of cases over this period and the high costs of utilizing and transforming this data, we will confine our investigation to four calendar years, 1966, 1970, 1974 and 1977. Using a fortran program a number of variables were created from the tapes for use in our descriptive statistics. These four years are spread over the entire period to be investigated and should give us a good approximation of the patterns which exist over all of the years. REGION - This variable denotes the region in which the charge was filed, and the region in which the initial ruling by the agency was made. As the number and the area of the regions has changed over the history of the agency a code will be provided in Appendix A. The number of regions during the period 1964 to 1980 is 31. Sub-regions were not included in the analysis. CASE - This variable notes the type of ULP case filed. 1 - is a CA charge. 2 - is a CB charge. 3 - is a CC charge. 4 - is a CD charge. 5 - all other charges are included in this category. RULE - This variable represents the initial ruling of the regional director. It is not reported in the annual data published by the NLRB, and may be different from the final ruling on the case. The final ruling is the ruling under which cases are normally reported in the annual statistics. 134 0 - denotes the case was dismissed by the regional director. 1 - denotes the case was withdrawn with the approval of the regional director. 2 - denotes that the case was settled, i.e., an agreement was reached among the parties through the regional director. 4 - A complaint was issued by the regional director. YEAR - This variable is for the calendar year in which the case received its initial ruling in the region. Cases reported by the NLRB are reported when the case is closed and not when the initial charge is made. Four years of data will be employed in the study, calendar year 1966, 1970, 1974 and 1978. These four years will allow us to look at decisions of the agency over the entire period. It will allow us to see how the NLRB functions under different boards and political administrations without looking at every year. PETITIONER - This variable denotes the party bringing the charge. It is a binary variable. The variable has a value of one if the party is a union or a member of a union, and the value is zero if the party is an employer, or an employer association. RESPONDENT - This variable denotes the party who is charged in the case. It is a binary variable. The variable has a value of one if the party is a union or a member of a union, and the value is zero if the party is an employer, or an employer association. 135 DECISION - This variable denotes a pro-union decision. If the variable RULE is a 0 or 1 and the RESPONDENT variable is a union then the decision is pro-union. Also, if RULE is a 2 or 4 and the PETITIONER is a union the decision is pro-union. The logic is reversed for a pro-management decision. These variables will allow us to consider the data of the NLRB in a new way. While there will most likely be differences among the regions with regard to their initial ruling patterns on cases due to other reasons, we expect that most of the regions will exhibit relatively similar patterns. The four years involved in the study will also enable us to look at the agency when it is controlled by both a pro-management board and a pro-labor board. BUDGET PROCESS The hypotheses concerning the budgetary process of the NLRB argue for a workloading model as an alternative explanation. The lack of political conflict and resulting cooperation between the opposing interests makes this a plausible explanation. The budgetary process can be modeled employing the general linear model (with one restriction) and can be written as follows. 136 NLRB Budget - Estimate of ULP cases + Estimate of the Representation cases + Backlogged ULP cases + Backlogged Representation cases + Institutional Impacts + Inflationary Impacts The model takes into account the estimated workload for each major type of case, the backlogs of each major type of case which exist due to an underestimation of the estimated workload, the impacts of rule or institutional changes in the case handling procedures, and the impact of the changing value of the dollar over time. The operationalized model to be tested can be specified as follows. The reader should note the change in the functional form of this model relative to the general case written above. The impacts of inflation have been incorporated into the budget variable. RBUD 8 blESULP + szSREP + b3BKULP + b4BKREP + bSLGR + bGLGB + E The variables in the model, the data sources and the functional form can now be explained and outlined. RBUD - Real Dollar NLRB Budget. RBUD is the actual budget enacted for the NLRB corrected according to the Consumer Price Index. This enables us to use constant or real dollars over the twenty-eight year period covered by the model. The inflationary impacts argued for in the general model are thus taken into account in a mathematical function rather 137 than in the statistical calculations. RBUD is thus equal to the actual budget of the NLRB divided by the consumer price index corrected dollar value, where 1967 dollars are used as the benchmark. The sources of data for the variable RBUD are twofold. The CPI figures were obtained from the Bureau of Labor Statistics as reported by the Department Of Commerce. The figures for the actual budget of the NLRB before being corrected are the figures for the enacted NLRB budget as reported in the Budget of the United States for the years covered by the study.8 ESULP - Estimated Unfair Labor Practice (ULP) caseload. While it is impossible to predict with precision the actual formula the agency employs in estimating its workload for the coming fiscal year it is possible to find a variable which can serve as a measure for this concept. The variable employed in this study as an operationalization of this term is the number of cases filed with the commission in the previous fiscal year. The rationale for this variable is relatively straightforward. First, as noted above, the estimation procedure actually employed by the agency would be difficult to predict. While the commission may use the available figures plus a straight lining procedure to estimate the workload there is no guarantee of this. Also, the logic used to determine the workload estimation may change with different executives. The variable employed provides us with 138 a consistent and parsimonious means of measuring the estimated workload over time. In addition, the figures used may also be considered a good estimate if the agency uses its current rate of case intake for the previous fiscal year at the time of budget preparation as the best estimate of is anticipated workload. The source of the data employed is the Annualrgeport of the NLRB for the years covered by the study. These figures are also available in compiled form for the period up to 1978 in the Bureau of Labor Statistics' Handbook of Labor Statistics.9 ESREP - Estimated Representation (REP) caseload. This variable represents the second major case type workload category. The rationale for this variable and the actual measure employed is identical for that of ESREP with one minor exception. There are a number of types of representation cases and they have been summed into this one figure. Likewise, the sources of the data are the same for this variable as noted for ESREP. BKULP - Backlog ULP Case Level. This variable provides us with a measure of the backlog of the commission in terms of its ULP caseload. It is calculated as a ratio of the ULP cases pending at the beginning of the previous fiscal year and the ULP cases closed in the previous fiscal year. This figure provides us with a constant measure of the backlog of the agency overtime. As the workload of the NLRB has 139 increased dramatically in recent years and cases are filed throughout the year the simple number of cases pending would not give us a constant measure over time. The backlog under such a measure would appear to increase over time and this would be incorrect. That is, as the total caseload increases the acceptable level of backlog at the time would also increase. Measured as a proportion, the variable corrects for this providing us with a figure of the percentage of the cases not completed in a given year. The data sources for this figure are again the Annual Reperte 9f the NLRB for the years covered by the study. BKREP - Backlog Representation Case Level. The logic which underlies this variable is the same for the variable BKULP only representation backlog figures are employed. The sources of data are also the same. The variable is equal to the ratio of REP cases pending at the beginning of the previous fiscal year and the number of representation cases closed in the previous fiscal year. Before turning to the consideration of the final two variables in our model we must digress for a moment and consider another variable which is employed in their construction. This is the Landrum-Griffin variable, LG. LG - Landrum-Griffin Binary. The Landrum-Griffin variable is a binary variable which is included to capture the institutional change in the case handling procedures brought about by the passage of the Landrum-Griffin Act in 1959. As 140 noted previously the act altered the procedure for handling representation cases brought before the NLRB. Prior to the act all cases went to the board for final settlement. However in 1960 the board delegated its powers to the regional directors under strict guidelines. The purpose of the change was to speed up the case handling time for representation cases and to relieve the board of some of its ever growing workload. The variable has a value of 0 prior to 1960, and a value of 1 for 1960 and every year after. This variable is employed in conjunction with the representation case variables to create the last two variables in our model. LGR - Landrum-Griffin Representation. This variable is an interactive variable with LG multiplied by ESREP. It provides for the change in case handling costs for the estimated representation workload in the post Landrum-Griffin period. The data sources are the same as those for ESREP. LGB - Landrum-Griffin Backlog. This variable is an interactive variable with LG multiplied by BKREP. It provides for the change in backlog handling costs for the representation case backlog in the post Landrum-Griffin period. The data sources are the same as those for BKREP. The hypothesis concerning the budgetary process of the NLRB and the model used to test it requires that time series data be employed. As noted the period covered by the model 141 is from 1952 to 1981. The use of time series data can lead to serious violations of the assumptions of the general linear model which produce estimation problems when OLS is employed. These violations are especially troubling when conducting classical hypothesis tests. Specifically, when employing time series data the third assumption of the general linear model is apt to be violated. The problem arises most often because the error terms are correlated with one another. This violation is known as autocorrelation. The impact of the violation for the OLS estimation procedure concerns the best criterion. When autocorrelation is present the estimates remain unbiased, however they no longer have minimum variance.10 This result can obviously lead to severe if not fatal problems when attempting to conduct classical hypothesis tests. “when the disturbances are autoregressive, the conventional formulas for carrying out tests of signifigance or constructing confidence intervals with respect to the regression coefficients lead to incorrect statements. That is, the calculated acceptance regions or confidence intervals will be either narrower or wider than the correct ones, depending 33 whether the bias is negative or positive.“ As the problems associated with autocorrelation are severe, tests for violations of the assumption of no autocorrelation can and should be conducted. The Durbin-Watson statistic is one of a number of such tests and readily available in most computer packages. If the null hypothesis that the value of rho is equal to zero is rejected then the GLS estimation procedure will be employed. 142 Rho can be understood to represent the prOportion of the disturbance term in time t caused by the disturbance term at t minus one.12 GLS estimation procedures while BLUE in the presence of autocorrelation require that the value of rho be known. As this is not the case a technique which Ostrom and others refer to as a pseudo-GL8 technique will be employed. A number of pseudo-GL8 techniques are available, and this study makes use of one of the more readily available, the Cochrane-Orcutt procedure. This procedure is relatively straightforward and related to the more conventional OLS estimation procedure. Further, the technique is now readily available in a number of computer program packages. The procedure as outlined by Kmenta is as follows.13 First, one obtains OLS estimates of the equation and calculates the residuals. Second, these residuals are employed to obtain an estimate of rho. Third, new variables (both endogenous and exogenous) are constructed employing the estimate of rho contained in the second step. These new variables are equal to the value of the original variable minus rho times the lagged value of the variable. The intercept is recomputed as its original OLS value times the value one minus rho. Fourth, these new variables are then run in a conventional regression procedure employing OLS. As the procedure is iterative, one continues the process until the estimated values of rho converge. That is, steps two through four are repeated until the estimates of rho 143 converge. While it is possible it is readily apparent that this procedure can become quite involved employing only a program package which performs the traditional OLS estimation procedure. For these reasons the Time Series Processor (TSP) program package is employed in the study. This package performs the Cochrane-Orcutt procedure for the investigator. The model of the budgetary process for the NLRB as specified employs a functional form which deletes the intercept term common to the general linear model. This type of model is often called a restricted form. It does not present us with any great problems, however as it is an alteration of the general linear model it will be discussed. Before proceeding with the mathematical implications of the model we should first explore the theoretical reasoning which underlies the choice of this model specification. The theoretical reasons for the choice of the restricted model are rather straightforward yet they are the primary reasons for the choice. Workloading models, as the one presented for the NLRB, as a general rule include overhead costs (administrative costs, overhead, etc.) in the cost factor for cases. That is, for x amount of workload an agency receives a specified amount for overhead costs. They are calculated as part of the workload, i.e., the cost factor is built into each case, rather than as a separate base independent of the workload figures. As such, a workload model of this type should not include an intercept 144 term. This argument is given more weight by the fact that the NLRB had been in existence for a number of years prior to the point where our model begins and as such the commission should have a good idea of the costs associated with different levels of workload. A restricted model involves the use of prior information which is incorporated into the model. As indicated by the above discussion the value for the intercept term is known, i.e., it is equal to zero. As such, the intercept term should not be included in the model specification. As Kmenta points out this restricted form causes no problems. The estimates for the coefficients can be derived with this form and we can procede as though we were employing the classical linear model.14 COMMISSIONER APPOINTMENTS The third and final channel of private influence to be considered involves the appointment of the commissioners or board members to the NLRB. As noted in the last chapter this channel of private influence was expected to remain open for private influence. There are two related hypotheses which must be investigated. The first concerns the impact of the Senate and the President as channels of private influence on the preferences of the commissioners appointed. The second concerns the relative impact of these political factors in the context of the institutional factors and environmental 145 factors which surround the NLRB. As the first hypothesis is relatively straightforward to test, we shall not spend a great deal of time detailing it here. The second hypothesis however deserves a good deal of explanation and as such we shall concentrate on it. The first model calls for the test of the impact of the Senate and the Presidental appointment on the policy decisions of NLRB commissioners. The actual votes of NLRB commissioners on major cases brought before the board is used as a measure of the policy preferences of the board members. Binary variables are used to indicate whether labor or management supporters control the Senate Labor and Public Welfare Committee and the Presidency. Both variables have a value of one when labor supporters are in control, and a value of zero when management supporters are in control. (For a more detailed explanation see the next section on the extended model and the explanation for the variables SAPP and PAPP) . The dependent variable employed in this model is VOTAVG. This variable is the average voting score for a commissioner during his entire tenure on the board. Decisions on individual cases (VOTE) are scored between one and four, the former score pro-management, the latter pro-labor. The variable VOTAVG is the average voting score for all of a commissioner's votes. (See the next section for a detailed explanation of the variable VOTE). The model can 146 be specified as follows: VOTAVG - b0 + bISAPP + bzpapp + E As the dependent variable is a continuous variable within the range one to four and the independent variables are binary variables the use of the linear model employing an OLS estimation procedure would seem appropriate. Since this model and the estimation procedure were discussed in great detail earlier in the chapter they will not be discussed further here. The second model to be tested involving the appointments process is an attempt to determine the impact of the appointments process given the institutional factors and economic factors which impact the agency. The model can be specified in the following manner employing the linear model. VOTE I b0 + blPAPP + DZSAPP + b3AGDEC + b4GENCON + bSREAP + bGLMIS + 3 VOTE - Board Member Case Vote. This is the vote of an NLRB board member based on a union versus management scale. The scale has a range of l to 4. Votes are coded to this scale based on the written decision of the commissioners. 1 is a vote for the management position on the case in full. 2 is a vote largely in favor of the positon held by management but with some of labor's positions held to. 3 is a vote for labor in major part with some of management's positions held 147 to. 4 is a vote for the labor position in full. PAPP — Presidential Appointment Binary. This variable represents the appointment impact of the President on the commissioners. The binary has a value of 1 when the President has labor endorsement and support, a value of 0 when the President has management endorsement. Labor support was determined from the vote cast for President by union members. Endorsements for labor were determined from reviewing the endorsements made by the major labor organizations, the AFL-C10, UAW, etc. These endorsements were found in convention hearings, and various news sources such as the New York Timee, Management endorsements were obtained from statements issued by the National Association of Manufacturers, and Chamber of Commerce and from various news SOUICGS o 15 SAPP - Senate Appointment Binary. This variable represents the appointment impact of the Senate Labor and Public Welfare Committee on the commissioners. The binary has a value of 1 when the majority of the Senators on the labor committee have labor endorsement and support, a value of 0 when the majority of the Senators have management endorsement. Endorsements for labor were determined from reviewing the endorsements made by the major labor organizations in the state, the AFL-CIO, UAW, UMW, etc. These endorsements were found in state convention hearings, and various state news sources. Management endorsements were 148 Obtained from statements issued by the state chapters of various business associations, the National Association of Manufacturers, the Chamber of Commerce and from various news sources . 16 AGDEC - Agency Decision. This variable is the first of two institutional variables. It scores the decision of the agency on a labor management scale as it comes before the board. The decision can be either that of the Administrative Law Judge or the Regional Director depending on the type of case. In some instances cases come before the board without any position being taken by the agency. The scale has a range from 1 to 5. 1 is a ruling for management's position in full. 2 is for a ruling largely in favor of management but with some of labor's position held. 3 is for no prior decision. 4 is a ruling for labor's position in large part but with some of management's position being held. 5 is for a ruling for labor's position in full. GENCON - General Counsel Binary. The second institutional variable is the support of the NLRB prosecutor, the General Counsel. The variable has a value of 1 when the General Counsel supports the labor position in the case and prosecutes the case for labor. The variable has a value of 0 if the support and prosecution is on the side of management in the case. REAPP — Reappointment Binary. This variable has a value of 1 when a commissioner is reappointed to the NLRB, otherwise it 149 has a value of 0. This variable is intended to capture the impact of the appointment process beyond the initial appointment. As the vast majority of those who are reappointed are reappointed by an equally pro-labor or more pro-labor Congress and President this variable should have a positive relationship to the variable VOTE. LMIS - Lagged Misery Index. The misery index is the sum of the rate of unemployment and the Consumer Price Index (CPI) multiplied by 100. It is lagged one year as cases take a year or more to reach the board and we wish to understand the economic conditions at the time the case was filed. As there is no definitive logic on the relationship of economic factors as they impact unions or management with regard to the bargaining positions they take relative to one another this general statement serves to capture the impact of economic effects on the type of cases brought before the NLRB. The individual components of the measure were obtained from the Bureau of Labor Statisitcs. The analysis and appropriate test of hypotheses concerning the voting decisions of NLRB Board members for the general model requires the use of a model which substantially alters the general linear model. This model is known as Probit Analysis. The model requires both a new estimation procedure, known as maximum likelihood estimation (MLE), and the transformation of the general linear model.17 150 The primary and overriding reason for the use of Probit Analysis in the analysis of Board member voting decisions is the nature of the endogenous variable. As coded the endogenous variable consists of a fourfold categorization, and when collapsed the endogenous variable is a dichotomy. In the dichotomous categorization Board member decisions are coded as either pro-labor or pro-management. In the fourfold categorization they are coded as: (1) compliance in full with labor's position, (2) partial compliance with labor's position with some management dissent, (3) partial compliance with management's position with some labor dissent, and (4) full compliance with management's position. This coding creates an ordinal scale along a labor management dimension. The use of ordinal level data presents a number of problems when used as endogenous variable in the general linear model, although this is not an explicit assumption of the model. As McKelvey and Zaviona note. “The assumptions underlying the multivariate linear model require interval level measurment of the dependent variable. Because of this, the linear model is not appropriate for many social science applications. In general, even if the dependent variable of theoretical interest is appropriately conceptualized as interval level, measurement theory in the social sciences is simply not refined enough to generate an interval level operationalization of this variable. The best that can be hoped for, in most cases, is a rather crude ordinal scale which purports go represent this true underlying variable.“1 Two major problems arise with the use of a dichotomous endogenous variable which makes the implementation of Probit analysis appropriate. First, two assumptions of the general 151 linear model are likely to be violated. They are the second and third assumptons of the model, that the disturbance term has an expected value of zero and that the variance of the disturbance term is constant. The major problem is that the relationship is non-linear. Therefore, the specification assumption is violated, which implies that the second and third assumptions are violated. The second problem involves the problem of obtaining predicted values outside of the proper range. In the dichotomous case values should only be within the zero to one range. It is a relatively straightforward procedure to demonstrate that the assumptions of constant variance and expected value of the error term are violated when the general linear model is used in conjuncton with dichotomous (or trichotomous, etc.) exogenous variables. Pindyck and Rubinfeld present a formal proof of the violation of the former assumption for the dichotomus case assuming normality of the disturbance term. They conclude that: “the error term is heteroskedastic ...Observations close to 0 or 1 will have relatively low variances, while observations ...closer to .5 will have higher variances.“19 McRelvey and Zavoina likewise demonstrate the problem of heteroskedasticity and the fact that the disturbance term does not have an expected value of zero graphically. Figure V-l shows these points clearly. They conclude that, “there does not seem to be any possible linear model which could have generated the data and maintained an error term with mean zero and constant variance.“20 152 FIGURE V-l Ordinal Dependent variable Error Terms E(Y/X) Least Squares Line 153 The net result of these problems is that the traditional estimation procedures will produce both biased and inefficient estimates of the parameters. As such, classical hypothesis tests are no longer appropriate and inferences based on them can lead to serious misinterpretations. A second problem concerns the range of the predicted values. With the general linear model it is easy to obtain estimates beyond the bounds of the dichotomous endogenous variable. A number of correction procedures for this problem each involving a method to squeeze the predicted values within the proper range. One solution involves constraining predicted values to the range of the endogenous variable, either by eliminating the outlying observations or reinterpreting the results. A second solution involves restricting the model such that the exgenous variables can only produce predicted values within the proper range. However, each of these solutions is likely to produce more problems than they solve. These problems suggest that an alternative model specification is required. A number of these transformed models exist, including Logit and Probit Analysis, and this study employs the latter. Specifically the study employs the N-Chotomous Probit package developed by Mcxelvey and Zavoina. This package allows the researcher to deal with ordinal level data at both the dichotomous and increased levels. 154 The Probit Model as developed by McKelvey and Zavoina assumes that the true dependent variable underlying the observed and measured ordinal level dependent variable, is an interval level variable (although one need not make this assumption to derive the model). This assumption is well suited to our investigation, as it is easily assumed that there is an interval level scale along a labor management decision dimension that cannot be adequately measured. Previous studies have employed a dichotomous scale for the voting decisions of NLRB commissioners.21 Ours expands the number of categories to four, and there is no reason to believe that the scale could not be increased even further although the task would become increasingly difficult and expensive. The Probit Model is also a smoothing technique which essentially squeezes the estimated responses into the proper range indicated by the ordinal scale of the dependent variable. This eliminates the problem of estimates which are beyond the bounds indicated, a problem which was not eliminated by the other aforementioned methods. The model is a probability model, hence, by definition, it is constrained to estimate probability numbers.22 The estimation procedure for the Probit Model is also somewhat different than the OLS procedure which is traditionally employed. The estimation procedure employed here is that of Maximum Likelihood Estimation (MLE). As Rmenta notes, “this method is based on the relatively simple 155 idea that different populations generate different samples, and that any given sample is more likely to have come from some populations than from others.“23 The MLE can further be shown to have a number of desirable asymptotic properties (large sample properties). The MLE is asymptotically unbiased, efficient, and consistent.24 Finally we should consider the data sources for the appointment model. These data are also employed in the simple model on the impact of the Senate and the President in the construction of VOTAVG. First, the data source and the way the sample was obtained will be explored. Then the construction of the variables from this data will be explained. The votes of commissioners, and rulings of agency bureaucrats were obtained from a random sample of NLRB decisions over a twenty-five year period. These decisions of the board come from the Decisions and Orders of the National Labor Relatons Board, Volumes 100 through 239. Each case has a separate volume and case number. As such it was necessary to sum the total number of cases and provide each case with a unique ordered case number. A random sample of the total population was taken utilizing a computer with a random number generator. The sample size was 1500 out of a total population of 25,207 cases. The raw case numbers were then reconverted back into the original volume and case numbers with the use of a computer program. Cases which could not be coded as either pro-union or pro-management were dropped 156 from the sample. This situation arose most often when unions were on the opposite side of a case. The reader should refer to Appendix B for a complete copy Of the program used to generate this data set. The data coded on each case which is germane to the above model concerned the decision of each commissioner either for compliance, or dismissal. These were then converted into either pro-labor or pro-management decisions depending on the petitioner and respondent in the case. The same logic was used to construct the variables for the decisions of the general counsel, the regional directors, and the administrative law judges. Appendix C contains the coding sheet used to gather the data, and outlines the variables gathered. We now turn to the results obtained from the models outlined in this chapter. The results in the next chapter will hopefully provide the information required to allow us to evaluate the hypotheses outlined and operationalized here, and to draw conclusions regarding their validity. FOOTNOTES 1. See for example the works of the following authors: Richard Fenno, Home Style (Boston: Little Brown and Co., 1978). Herbert Kaufman, The Forest Ranger (Baltimore: The Johns Hopkins Press, 1960). 2. See for example John W. Kingdon, Qongessmen's Voting Decisions (New York: Harper and Row, 1973). 3. For example, William Gormley, “A Test of the Revolving Door Hypothesis of the FCC,“ Americen Journal of Politicel Science, 23(4) (November,l979), pp. 655-683. 4. See for example, George Stigler, “The Theory of Economic Regulation,“ Bell Journel of Economics and Management Science,2 (l) (Spring,197l), pp. 3-21; and George Stigler and Claire Friedland, “What Can Regulators Regulate?: The Case of Electricity,“ Jenrnel_ef_Lew_end, Eeegemiee, 5 (October,1962), pp. 1-16. 5. Terry Moe, “Regulatory Performance and Presidential Administration,“ WW, 26(2) (May,1982), pp. 197-224. 6. J. Johnston, Econometric Methods (New York: McGraw-Hill, 1972), pp. 121-126. This representation of the general linear model is in matrix form and can be found in a large number of econometric texts. For those unfamiliar with the matrix notation, Johnston also provides the more conventional notation. For the more conventional notation see also John Neter and William Wasserman, Appliedgginear Statistical Models (Homewood, Illinois: Richard Irwin, Inc., 1974). 7. Peter Kennedy, A Guide to Econometrics (Cambridge, Mass.: M. I. T. Press, 1979), pp. 24. 8. U.S., Office of Management and Budget, Bedget of the United Stetes, 1952-1982 (Washington, D.C.: Government Printing Office, 1952-1982). 9. National Labor Relations Board, Annual Report Of the_ at ona ons Bo 1 2- , (Washington, D.C.: Government Printing Office, 1952-1981). Statistics are reported in the statistical appendix. HandDQQK_Q£_LahQL. Statistics, op. cit.. See appropriate pages for each volume. For information on the way cases are counted and handled see U.S., Department of Labor, Bureau of Labor Statistics, BLS_ Handbook of Methods (Washington, D.C.: Government Printing Office, 1976). 157 158 10. Charles Ostrom, Time Series Analysie_(Beverly Hills, Cal.: Sage Publications, 1978), PP. 25-31. 11. Jan Kmenta, Elements of Econometrics (New York: McMillan Publishing Co., 1971), p. 281. 12. Jan Kmenta, op. cit., p. 288. See also Charles Ostrom, op. cit., p. 38. It should be noted that these tests are for first order autocorrelation only. 13. Jan Kmenta, op. cit., pp. 287-289. For a more detailed explanation of the procedure see D. Cochrane and G. Orcutt, “Application of Least Squares Regression to Relationships Containing Autocorrelated Error Terms,“ Journal of the American Statistical Association, 44 (March,l949), pp. 32-61. 14. Jan Kmenta, Op. cit., pp. 430-433. One possible problem can occur when the intercept is not really equal to zero. In other words the unconstrained model is appropriate. This can be, and was, tested using traditional specification. However, it should be noted that the theoretical reasoning for the chosen functional form is more important. 15. Endorsements for the President to determine whether he was supported by labor or management came from national sources. Articles from the New York Times or other papers served as one major source. Other sources for labor were the Proceedings of the American Federation of Labor or the Congress of Industrial Organizations. For management support the Proceedings from the National Association of Manufacturers were used along with statements issued by the Chamber of Commerce. 16. Newspaper articles for the states involved were used as a primary source for labor or management support within a state. Also the Proceedigge_from state labor conventions were used to determine labor support. For example Proceedings for the California Federation of the AFL-CIO were used along with those of the Michigan Federation of Labor. 17. For some applications of the Probit Model in social science research see the following articles. John Aldrich and Charles Cnudde, “Probing the Bounds of Conventional Wisdom: A Comparison of Regression, Probit and Discriminant Analysis,“ American Journal of Political Science 19 (1975), pp. 571-608. Steven Rosenstone and Raymond Wolfinger, “The Effects of Voter Registration Laws on Voter Turnout,“ American Political Science Reyieu 72 (1978), pp. 22-45. 159 18. Richard McKelvey and William Zavoina, “A Statistical Model for the Analysis of Ordinal Level Dependent Variables,“ Journal of Mathematical Sociology, 4 (1975), p. 103. 19. Robert Pindyck and Daniel Rubinfeld, Econometric Models and Economic Forcasts (New York: McGraw-Hill, 1981), p. 276. 20. McKelvey and Zavoina, op. cit., p. 105. 21. Delorme, Hill, and Wood, “The Determinants of Voting by the NLRB on Unfair Labor Practice Cases: 1955-1975,“ Public Choice, 37(2) (1981), pp. 207-218. 22. Pindyck and Rubinfeld, op. cit., pp. 276-282. 23. Peter Kennedy, op. cit., p. 174. 24. Jan Kmenta, op. cit., p. 184. See also Eric Hanushek and David Johnson, Statistical Methods for Social Scientists (New York: Academic Press, 1977), pp. 344-48. VI RESULTS The purpose of this chapter is to report the results obtained from the models outlined in the previous chapter. First, each of the three sets of results will be reported and discussed. Additional empirical evidence will also be provided where necessary to lend support to the arguments made by the models. Next, the overall picture that these results paint for the questions of private influence and the NLRB will be considered. Finally, we will consider the results in the light of what they say regarding private influence in a balanced environment and for the question of private influence and the regulatory agencies in general. DIRECT INFLUENCE The first set of results to be considered are those obtained from the data concerning direct influence. Four calendar years of data are presented on lower level regional decisions of the NLRB. These are for the years 1966, 1970, 1974, and 1977. These data can be found at tables VI-l through VI-4 respectively. The selection of years provides good coverage for the years included in the NLRB data on 1 tapes. Each table presents the percentage of decisions which were in managements favor.2 The mean and standard 160 161 TABLE VI-1 Percent Pro-Management ULP Rulings At The Regional Level - 1966 Cases REGION PERCENT Z'SCORE 1 56.8 .84 2 54.2 .28 3 47.2 -1.23 4 57.1 .90 5 53.5 .12 6 56.8 .84 7 46.0 '1.49 8 59.9 1.51 9 53.6 .15 10 55.8 .62 11 5206 -007 12 53.3 .08 13 55.4 .54 14 52.5 -.09 15 52.6 -.06 16 51.9 -.22 17 46.9 -1.30 18 41.0 -2.57 19 54.8 .41 20 56.6 .79 21 46.3 '1.42 22 55.7 .60 23 510‘ -o33 24 62.4 2.05 25 52.2 “.15 26 58.1 1.12 27 54.4 .32 28 55.2 .49 29 46.6 '1.36 30 46.9 -1.30 31 52.5 -.09 MEAN 52.9 STANDARD DEVIATION 4.64 REGION DEVIATIONS TOTAL FROM MEAN 21 0-1 8 1-2 2 2+ 162 TABLE VI-2 Percent Pro-Management ULP Rulings At The Regional Level - 1970 Cases REGION PERCENT Z‘SCORE 1 56.5 -.23 2 52.5 -1.05 3 48.7 -l.83 4 65.4 1.60 S 53.2 -091 6 53.6 -082 7 60.8 .65 8 63.5 1.21 9 58.7 .22 10 63.3 1.17 11 65.5 1.62 12 57.2 -.09 13 50.8 '1.40 14 58.3 .14 15 64.3 1.37 16 5700 -013 17 54.5 -.64 18 54.3 -068 19 50.9 -1.38 20 58.1 .10 21 60.4 .57 22 63.1 1.12 23 59.8 .45 24 65.0 1.51 25 58.6 .20 26 59.7 .43 27 5‘07 -060 28 51.7 -1.21 29 50.3 '1.50 30 60.3 .55 31 55.5 -.43 MEAN 57.6 STANDARD DEVIATION 4.87 REGION DEVIATIONS TOTAL FROM MEAN 18 0-1 13 1-2 0 2+ 163 TABLE VI-3 Percent Pro-Management ULP Rulings At The Regional Level - 1974 Cases REGION PERCENT Z'SCORE 1 51.0 -.83 2 51.0 -083 3 48.8 '1.25 4 59.6 .82 5 58.1 .53 6 53.4 -.37 7 55.3 .00 8 56.3 .19 9 54.5 -.16 10 57.8 .47 11 66.4 2.12 12 57.6 .44 13 49.0 '1.21 14 55.7 .07 15 63.5 1.56 16 56.5 .23 17 47.8 -1.44 18 54.7 -.18 19 51.0 -.83 20 53.7 -.31 21 53.7 -031 22 56.2 .17 23 51.0 -.83 24 53.3 -.39 25 64.7 1.79 26 53.9 -.27 27 58.9 .68 28 53.9 -.27 29 67.4 2.31 30 44.7 -2.03 31 55.8 .09 MEAN 55.3 STANDARD DEVIATION 5.24 REGION DEVIATIONS TOTAL FROM MEAN 24 0-1 4 1-2 3 2+ 164 TABLE VI-4 Percent Pro-Management ULP Rulings At The Regional Level - 1977 Cases REGION PERCENT z-SCORE 1 53.5 .05 2 45.8 '1.80 3 47.4 -1.42 4 50.1 -.77 5 54.7 .34 6 54.5 .29 7 52.3 -.24 8 52.1 “.29 9 56.2 .70 10 56.0 .65 11 59.0 1.37 12 55.5 .53 13 49.4 -.94 14 59.8 1.57 15 59.3 1.45 16 54.0 .17 17 50.2 -.74 18 53.8 .12 19 52.6 ".17 20 50.8 “.60 21 50.8 -060 22 49.6 -.89 23 58.2 1.18 24 54.3 .24 25 53.4 .03 26 61.4 1.95 27 59.5 1.49 28 51.2 '.50 29 44.2 '2.18 30 52.2 -.26 31 50.2 -.74 MEAN 53.3 STANDARD DEVIATION 4.15 REGION DEVIATIONS TOTAL FROM MEAN 22 0-1 8 1-2 1 2+ 165 deviation of pro-management decisions for each year, as well as the percentages and standardized scores for each of the regions of the NLRB are presented. Further, the number of regions which fall within one, one to two, and more than two standard deviations are listed. The hypotheses in the preceding chapter regarding direct influence argued for consistency between the regions in terms of their decision patterns. A lack of direct influence would allow the rules of the agency to dominate the decision making patterns of the commission. If this is the case we would expect that the decision patterns of the regions would be highly similar. While it is practically impossible to test this proposition statistically, descriptive statistics of the population of decisions can be employed to provide evidence in support of, or contrary to, this position. The results obtained from the population data are mixed in their support of our hypotheses. That is, there is enough variation in the results obtained to warrant different interpretations. Further, the perspective from which one approaches the data can also result in differing conclusions. Some of the more straightforward results obtained from the tables are as follows. The mean percent of pro-management decisions ranges from 52.9 to 57.6 for the four years investigated. The standard deviations for the years were also fairly consistent, ranging from 4.1 percent to 5.2 percent. Also, 166 the number Of regions which fall within plus or minus one standard deviation of the yearly mean is fairly consistent for three of the four years ranging from twenty-one to twenty-four. For the four years studied, the results are less consistent, ranging from eighteen to twenty-four regions falling within plus or minus one standard deviation. However, for the year with only eighteen regions falling within plus or minus one standard deviation, the remainder of the regions fall within plus or minus two standard deviations. Also, the number of outliers, those points which fall outside of the plus or minus two standard deviation range, is quite small for the years considered. Given the above results, we would argue that the distribution of the regions are roughly consistent from year to year. That is, there is great similarity in the percentage of decisions which are pro-management from year to year, and the distributions of the regions around the mean are fairly consistent from year to year. The point of confusion which can be interpreted in a number of ways regards the standard deviation for each year. While they are fairly close from year to year, there is no benchmark by which they can be compared. The question we must answer is whether a standard deviation of approximately five percent is a small enough difference to be attributed to numerous other factors in our ceterus paribus clause? It is clear from the results reported that our contention that the variation in decision patterns would be 167 greater over time than within any one year is false. It is obvious that the variation within any one year is greater than the variation over time. However, this does not mean that there is direct influence on the agency. To understand the results we must take a different approach. One way to consider the question of direct influence within this context is to look at the individual regions over the four years studied. That is, can the bulk of the variation be attributed to the same regions from year to year? Do some regions account for most or a majority of the outliers? Further, are regions from strong union or management areas more likely to rule in favor of unions or management respectively? Table VI-S has the results for the entire four year period. An investigation of the results along these lines demonstrates that eight Of the regions fall within plus or minus one standard deviation of the yearly mean for all of the years investigated. Thirteen regions fall within plus or minus one standard deviation of the mean in three of the four years investigated. Of the eighty-four total points which fall within plus or minus one standard deviation for the four years, these twenty-one regions account for seventy-one, or approximately eighty-five percent of the total. Of the thirteen regions which placed three of the four years in the plus or minus one standard deviation category, only one had an outlier - a year in which the number of standard deviations was greater than two. Of the 168 TABLE VI-S Regional Result Summary Standard Deviations REGION >-2 -2to-l -ltoO 0to+l +lto+2 >+2 H H lllll IlNlllllHlll-hNI lllli-‘llllllllll HH ma I I H )0 IIIIIIHII IIIHIHINI IIIHIIIII IlllHNHHI—H—‘lllIIIWI—IIINHINIIIHIII l wHINHHHHNHNNNNNNHHHHHIHHNNHHIHN HHIHNHNHHNHNHHININHwIwWHHNwNIHN N 0 IHHIIII IHNHIII Ill-0| 17 U 17 h C h :5 on Total ‘0 0‘ Active 2 13 10 3' Inactive 1 4 31 38 7 0 Pro- Union Pro- Managmt w 15 25 27 3 0 O N 15 17 14 3 169 remaining ten regions, six had two in the plus or minus one range, two only placed one in this range, and two placed none in this range. It is therefore clear that two-thirds of the regions are relatively consistent over time. These are the inactive regions found in Table VI-5. They are usually within one standard deviation of the yearly mean. This five percent variation, which is the normal standard deviation for any one year, is not very great. One-third of the regions account for the bulk of those regions whose decision results fall within the plus or minus one to two standard deviation range and the majority of the outliers. The question we should address is why these results are obtained in these particular regions? Also, do the outliers in the negative or positive ranges, which indicate pro-labor and pro-management decision patterns respectively, fall into any particular grouping? If we look at the performance of the regions over time in terms of their location in strong union versus weak union areas some interesting results are also Obtained. Strong union regions are those which are located in states with high union membership levels and weak union regions in states with low union membership levels. The total number of negative plus two standard deviation outliers is three, and the number falling within negative one to two standard deviations is seventeen. These are decision patterns which are pro-labor relative to the mean for a given year. The 170 positive plus two outliers total three, and those in the positive one to two range total seventeen. These are the pro-management decision patterns relative to the mean for a given year. If we look at these outliers with regard to strong union or weak union regions some interesting results are obtained. Of the negative outliers - greater than two standard deviations from the mean - strong union regions account for all three. Of the negative one to two total, strong union regions account for fifteen of the seventeen. Of the positive plus two outliers, weak union regions account all three, and of the positive one to two category, weak union regions account for thirteen of seventeen. When considered from this perspective, several points seem clear. First, a majority of the variation in any of the four years can be relegated to approximately one third of the regions. These regions are the same from year to year. Second, and perhaps even more importantly, the regions which are in strong union areas account for the bulk of the pro-union decision patterns, and the regions in weak union areas account for the vast bulk of the pro-management decision patterns. It is therefore highly possible that the variation which we find in the aggregate data is attributable to the strength of one group relative to the other in a given region. In summary, the following points can be made concerning direct influence and low level commission decisions. First, the distribution of the decision patterns is similar from 171 year to year. Second, one-third of the regions account for most of the variation and the outliers, and these regions are the same from year to year. Third, the majority of pro-labor decision pattern outliers occur in regions where unions are strong. Where unions are weak, pro-management decision pattern outliers are more likely to occur. Direct influence is not as important a channel of private influence in the balanced environment of the NLRB as it is purported to be in other unbalanced environments. However, our results as summarized above indicate that it is still present in some areas - or at least the possibility of it remains. This leads us to conclude that in most cases the rules and regulations of the commission predominate in low level agency decision making. In our study this seems to be the result in almost two-thirds Of the regions. In the other regions one of the groups may be stronger and therefore be able to effectively influence the decision patterns of the commission. A balanced environment appears to make it rational for the agency bureaucrats to follow the rules. Caught between competing interests is a no win situation for a bureaucrat unless he can justify his decisions based on some precedent. To openly favor one group or the other will lead to a situation which would most likely be intolerable. warwick has argued a similar point in his study of the State Department.3 Faced with conflicting interests the bureaucrats look for rules and regulations to guide their 172 decisions. When one group is stronger than another it appears as if the agency will bend to their demands and private influence will flow. In the case of our active regions one group may be stronger in the region than the opposing group. Aggregate balance in the environment is therefore no reason for balance within the regions. If the balance does not extend to the regions, then it appears as if direct private influence will still be important. These results however require further study. As noted later in the chapter there are many other possible reasons for the imbalance in the regions. BUDGET The second set of results to be considered are those regarding our hypotheses concerning the budgetary process of the NLRB and its role as a channel of private influence. As noted in the last chapter, the budget was not expected to be a channel of private influence for one group at the expense of the other in the post Taft-Hartley era. Instead, we expected that conflict would be replaced by cooperation. Arguing from this theoretical perspective, we constructed an alternative explanation which said that the budget of the NLRB would be determined by the workload of the commission. The model tested was specified as follows: 173 RBUD - bIESULP + szSREP + b3BKULP + b4BKREP + 13st + bGLGB + E The results obtained for this model can be found in Table VI-6. Included in the table are the beta coefficients, standard errors, and t-scores for each variable included in the model. Results reported for the overall model include the number of cases involved, the Durbin-Watson statistic, R-squared, and the adjusted R-squared. Overall, the explanatory power of the model is extremely high. The R-squared for the model is .990, and the adjusted R-squared for the model is .988. We can therefore state that the model explains ninety-nine percent of the variance. While these results are extremely encouraging one must remember that we are using time-series data, and the R-squared for time series data is often very high.4 However, even with this caveat the explanatory results of the model are very strong. As autocorrelation can be a serious problem with time series data, the Cochrane-Orcutt estimation procedure was applied. To test for problems of autocorrelation the Durbin-Watson statistic was obtained. The statistic has a value of 2.16 and this value is an indication that there are no problems with autocorrelation. Results obtained from an OLS regression (not reported here) for this model and data set produced a Durbin-Watson statistic of .46, an indication of high positive correlation between the residuals. 174 TABLE VI-6 Budget Model Results Workload Model Cochrane-Orcott Procedure Restricted Model Variable b s(b) ESULP .83 .08 ESREP .84 .21 BKULP 1886.85 956.16 BKREP 31.02 24.06 LGR .40 .15 LGB -7.17 71.04 Durbin-Watson 2.16 N = 29 R-Squared .990 Adj. R-Sq. .988 Workload Model (Political Variables Variable b s(b) ESULP .84 .08 ESREP .82 .25 BKULP 2077.87 1033.69 BKREP 30.21 24.71 LGR .43 .16 LGB 21.76 78.79 PRES -806.62 816.98 SEN 312.46 1119.97 Durbin-Watson 2.10 R-Squared .989 Adj. R-Sq. .985 t-score 10.46 3.94 1.97 1.28 2.71 -.10 Included) t-score 9.82 3.23 2.02 1.22 2.73 -.28 -.97 .28 175 The hypothesis tests for the individual variables in the model can now be considered. All of the hypothesis tests are one tailed tests, that is H1: b > 0, as indicated in the last chapter. Further, all of the tests were performed at the .05 significance level. Results obtained from the model for the individual variables indicate that the two major types of workload estimation, ESULP and ESREP, were both significant at the .05 level. Interestingly, the beta coefficients for the variables are extremely close in their values, .83 for ESULP and .84 for ESREP. This indicates that the costs for each case prior to the Landrum-Griffin era were extremely close if not identical. The results obtained for the backlog levels are not as clear. For the ULP case backlog, BKULP, the results obtained are significant at the .05 level. The beta coefficient is quite large, 1886.85: however, this is to be expected given that the value of the variable is a ratio. The results for the representation backlog, BKREP, are not significant at the .05 level. Here the t-score was only 1.28. Interesting, the beta cofficient for BKREP was also quite small, 31.02. The Landrum-Griffin interaction variables are the last set of variables to be considered in this model. These variables capture the institutional changes in the procedures of the NLRB. The impact of the act on the cost of representation cases, LG, was significant at the .05 level. The beta coefficient had a value of .40. If the value of 176 this variable is added to the value associated with ESREP, the cost factor for a representation case after the passage of the Landrum-Griffin Act is obtained. This value is 1.24, an increase of almost fifty percent in the cost of settling a representation case. The impact of the Landrum-Griffin Act on the backlog of representation cases, LB, is the final variable in the model. It is insignificant, with a t-score of -.10. The value of the beta-coefficient is also very small, -7.17, and its sign being negative is puzzling. It was anticipated that this variable would have a coefficient with a positive value. The overall results of the model indicate that a workloading model, one based on the cooperation of the parties, is an appropriate model. The significance of the workload estimation figures, the backlog figures for ULP cases, and the institutional impact of the Landrum-Griffin Act on representation caseloads all lend support to this view. The insignificance of the backlog ratios for the representation cases and the impact of the Landrum-Griffin Act on this backlog while at first troubling can be explained. Backlogs may be attributable to a number of different causes: underestimation of the cases to be filed in the coming fiscal year, too many cases settled at high levels of the agency, and finally procedural problems in settling the cases. The two former reasons, especially the first, would 177 seem to account for the backlogs in ULP cases. ULP cases make up a greater portion of the caseload, and are more easily handled in terms of the procedural requirements. Thus one might expect that a backlog in this area is due to unanticipated workloads and can be incorporated into the budget calculations. Representation cases on the other hand are quite a different matter. As noted earlier one of the reasons for the Landrum-Griffin reforms in the area of representation cases was the extreme amount of time it took for the cases to reach settlement. These causes were procedural in nature, requiring that the cases be heard by the board. Thus, backlogs in representation cases in the pre-Landrum-Griffin period could not be solved by increased budgetary outlays. If this is the case, then it is unlikely that they would be used in calculating the budgetary requirements for the agency. In the period after the passage of the Landrum-Griffin Act, the number of representation cases backlogged was extremely small. Again, these were most probably due to procedural problems and would therefore have no impact on the budgetary process. A model of the budgetary process based on workload might then remove from consideration the backlog of representation cases. Before considering the overall evaluation of the results of the model we must digress for a moment to discuss the impact of political factors on the budgetary process. In 178 order to test for political impacts we employed the binary variables for Presidential and Senate control by labor or management in the budgetary model. It should be noted that the House of Representatives is excluded from the model in that it was controlled by labor supporters throughout this period. The results of this model are given in Table VI-5. The results seem to confirm our hypotheses concerning the impact of political variables on the budgetary process. That is, control in terms of influence by either labor or management has no impact on the bugetary process of the NLRB. The variables from our original model remain fairly constant in terms of their beta coefficient values. The variables which were significant in the original model remain so in this model, and those which were insignificant remain insignificant. The variables for control of the Presidency and the Senate by labor or management are both insignificant. The beta coefficients are ~806.62 and 312.46 respectively. The interesting thing is that the Presidential binary in addition to being insignificant has a negative sign, indicating that a labor supported President would be likely to cut the budget. This is counterintuitive, for one would expect that if either side would want more money for the agency it would be labor as they bring a much larger number of charges before the agency.5 It is therefore clear that private influence to alter the budget does not exist. 179 The figures for the budgetary allocations reported at Table VI-7 also lend support to our notions of a cooperative budgetary process for the benefit of labor and management. These figures are the allocations for the NLRB from the Congress and the President. As the table shows, there is no pattern indicated. The figures are relatively close for both sides, with only a few discrepancies. Also, one does not find labor supporters attempting to increase the budget or management supporters attempting to decrease them. Some major increases do occur in years in which labor dominates the Congress and the Presidency, however the political implications of these increases can be explained. In each case when the workload increased the budget increased. Workload increases in these particular years were driven not by political considerations, but by the changes in the law concerning the jurisdiction of the NLRB. First, the Landrum-Griffin Act and then the Postal Reorganization Act increased the number of potential cases dramatically in the years indicated. Table III-1 shows the workload increase and it can be compared to the increases in Table VI-7 Finally, we should note that the justifications presented in the Budget of the United:§tates, and in the Congressional budgetary hearings for the period, rely almost exclusively on workload estimations to justify the NLRB's budget.6 In conclusion, we would argue that the results obtained from the workload model verify our notions of cooperation 180 TABLE VI-7 Presidential and Congressional Budgetary YEAR 1952 1954 1956 1958 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 Allocations for the NLRB CONGRESS 8,296 9,125 8,150 9,749 15,280 18,213 19,143 20,250 22,446 26,158 28,651 31,010 32,230 35,467 38,912 40,561 48,460 50,394 56,016 62,669 67,658 80,727 90,615 102,762 PRESIDENT 8,582 9,802 8,800 9,575 14,230 17,300 18,880 20,966 23,060 25,250 28,165 30,442 32,288 35,074 37,147 37,930 45,680 48,620 55,050 61,400 68,449 79,204 88,520 100,467 181 with regard to budgetary process of the NLRB. This channel of private influence which can be used in the general case as a means of eliminating unfavorable regulatory programs and supporting favorable ones is a cooperative effort in the balanced environment of the NLRB. However, if the laws which govern the NLRB were to become promotional in favor of one side or the other, i.e., pro-labor or pro-management, then we would expect that the outgroup would attempt to use the budgetary process as a means of curtailing the activities of the agency. COMMISSIONER APPOINTMENTS The final set of results to be considered concern the impact of private influence on the appointment of regulatory board members. We will first consider the simple model of the impact of private influence on this process through the institutions of the Presidency and the Senate oversight committee. We will then look at the impact of private influence in the context of the larger environment, considering institutional and economic factors as well. The overall voting scores for the commissioners can be found at Table VI-8. The commissioners have been grouped according to their initial appointment. There are three categories, those appointed by a pro-management Senate and President, those appointed by a pro-labor Senate and pro-management President, or a pro-management Senate and 182 TABLE VI-8 Commissioner Voting Averages SCALE: 1 - 4 1 - Full management support: 2 - Major management support: 3 - Major labor support: 4 - Full labor support. Overall Average: 2.71 Appointment: Management Senate and President: Average: 2.02 Farmer - 2.02 Rodgers - 2.35 Beeson - 1.68 Appointment: Management President and Labor Senate Management Senate and Labor President 4 Average: 2.69 Leedom - 2.63 Jenkins,J.- 2.33 Bean - 2.52 Fanning - 3.01 Kennedy - 2.61 Pennello - 2.86 Murphy - 2.52 Walther - 2.72 Kimberly - 2.89 Murdock - 2.69 4 Appointment: Labor President and Senate Average: 2.98 Houston - 2.97 Herzog - 3.08 Styles - 2.91 Peterson - 2.54 McCulloch - 3.08 Brown - 3.15 Jenkins,H.- 3.09 Zagoria - 3.02 183 pro-labor President, and finally those appointed by a pro-labor Senate and President. The voting scores reported for each commissioner are values of the variable VOTAVG, the average voting score for all of their votes while on the commission. From the scale it should be clear that the lower the score, the more pro-management the voting score for a commissioner. The table appears to verify our notions of the impact of the President and the Senate on the voting patterns of commissioners. The overall score for all commissioners is 2.71. Those appointed by a pro-management Senate and President have an average voting score of 2.02. Those with a split appointment pattern have an average score of 2.69. Those appointed by a pro-labor President and Senate have an average score of 2.98. These results are in line with what we would have expected and confirm our notions of the use of the appointment process as a means through which groups exert private influence. Considering that these commissioners are grouped solely by their first appointment pattern, the results are highly indicative of the importance of the appointment process. The scores of the members in each category are relatively close except for two outliers. In the split category, only Fanning has an extremely high voting score, 3.01. It is interesting to note that he is the only appointee of a Republican President to be reappointed by a Democratic President.7 In the pro-labor appointment 184 grouping only one commissioner has a low voting score, that of Peterson with a score of 2.54. These “errors“ in judgment are extremely small considering the post to which the individuals are appointed. After all, when appointing members to the Supreme Court, Presidents have made much greater mistakes, Eisenhowers' appointment of Earl Warren being the most famous. In order to go beyond descriptive statistics, and to also get some notion of the relative impacts of the Presidency and the Senate oversight committee on the appointments process, the results of the first model were tested employing a classical linear regression model with OLS estimators. These results can be found at Table VI-9. The dependent variable, VOTAVG, consists of the average voting scores of the commissioners as listed in Table VI-8. The independent variables are PAPP and SAPP. These are binary variables, 1 if the institution is supportive of labor, 0 if supportive of management. The model tests the impact Of private influence through these institutions on the voting patterns of commissioners, ceteris paribus. The overall results of this simple model are rather impressive. The F has a value of 18.03 and is significant at the .05 level. The multiple R for the model is .81. The R-squared is .65, and the Adjusted R-squared is .62. This means that the model explains sixty-two percent of the variance. 185 TABLE VI-9 Presidential And Senate Influence On NLRB Appointments Variable b s(b) t-score CONST 2.10 .11 17.73 PAPP .34 .10 3.36 SAPP .57 .13 4.41 R - .809 R-SQUARED - .655 ADJ. R-SQUARED - .620 N - 22 186 The results for the individual variables are also interesting. All of the hypothesis tests for the individual variables are one tailed, that is H1: b > 0, and are reported at the .05 level of significance. The beta-coefficient for SAPP is .57, with a t-score of 4.40. The beta-coefficient for PAPP is .34 with a t-score of 3.36. Both variables are therefore significant at the .05 level and both hypotheses are confirmed with the beta-coefficients being positive. It therefore appears that both the Senate oversight committee and the President are important actors in the appointment of NLRB board members. Both institutions serve as channels through which private influence can flow. When supporters of either group control the institution they are more than willing to use their authority to appoint board members whose preferences are similar to those of the group they support. The surprising factor concerns the relative impacts of the two institutions. As noted earlier, the President has normally been considered to be the key actor in the appointment of commissioners. Here we find that not only is the Senate oversight committee an important factor, but that its overall impact is greater than that of the President. The beta-coefficient for the Presidential appointment binary is .34, indicating that commissioners appointed by a pro-labor President will have a voting average which is .34 points higher than that of an appointee by a pro-management 187 President. The Senate oversight committee binary has a value of .57, indicating that an appointee by a pro-labor Senate will have a voting score which is .57 points higher than a pro-management Senate appointee. This is an impact which is approximately sixty-six percent greater. Therefore, we not only find that the Senate oversight committee is important, but even more interesting that its impact is greater than that of the President. The final model to be tested concerns the impact of private influence on the voting patterns of commissioners within the context of the institutional and economic factors which face the commission. The results of this model are contained in Table VI-10. The results obtained from the Probit Model indicate that all of the variables included in the equation are significant. Further, the estimated r-squared for the model is .41. While this is not nearly as high as that obtained from the budget model, it should be remembered that here we are using a pooled sample consisting of cross-sectional and time series data over a twenty-five year period. The MU coefficients are also reported in the table. These are the threshold points that separate the four values for the dependent variable. The values are 0.00, .226, and .75. That is, any predicted value of the independent variable which falls below 0.00 will fall into category 1 of the independent variable. Those values between 0.00 and .226 will have a value of 2, those between .226 and .752 will 188 TABLE VI-lO Appointment Impact Model Probit Results Variable MLE SE MLE/SE CONST -2.19 .10 -21.90 PAPP .16 .04 3.72 SAPP .37 .07 5.52 AGDEC .45 .01 30.90 GENCON .56 .06 9.91 REAPP .34 .05 7.33 LMIS .02 .01 3.76 COEFF. MLE SE MLE/SE MU(l) 0.00 MU(2) .23 .01 15.25 MU(3) .75 .02 31.65 ESTIMATED R-SQUARED - .415 PERCENT PREDICTED - 63.72 RANK ORDER CORRELATION - .54 LOG LIKELIHOOD FUNCTION - -3668.86 -2 X LIKELIHOOD RATIO CHI SQUARED 6 DEGREES OF FREEDOM - 1561.62 189 have a value of three and those greater than three will have a value of four. The standard errors for the coefficients and the MLE/SE values indicate that the threshold points are significant at the .05 level. Thus, our four fold categorization discriminates well along a pro-labor pro-management decision dimension. The impact of private influence through the appointments process is represented by three variables. SAPP and PAPP capture the impact of the Senate and the President on the initial appointment. REAPP captures the joint impact of these variables for those appointed to a second term. Of the three variables, SAPP has the greatest impact with a value of .37. This indicates that a Commissioner appointed by a pro-labor Senate oversight committee will vote on average .37 above that of a commissioner appointed by a pro-management Senate oversight committee. The impact of private influence through the Presidency while significant is much lower, at .16. On the surface then, the impact of private influence on an initial appointment through the Senate oversight committee is twice as great as that of the President. The final part of the influence pattern concerns the reappointment of commissioners. The beta coefficient of this variable is .34. It is almost as important as that of the Senate and shows the importance of oversight through reappointment of commissioners on their voting patterns. 190 In conclusion, the three measures of political impact of private influence are all significant, and all the relationships are in the direction theoretically predicted by the model. All are positive, and when summing the impact of all three variables the total impact for a commissioner due to private influence adds up to .87. The impact of the two institutional variables - GENCON and AGDEC - is quite interesting. Overall, these two variables are by far the most important, both individually and as a set. Both are significant, and each has a major impact. The impact Of GENCON is .56 which is greater than any of the political influence variables. It can be directly compared in magnitude to these political variables, as it is also a dichotomous variable. The impact of the decision of the agency before it reaches the commission is the greatest of all the variables. The beta-coefficient for AGDEC is .45, however it must be remembered that this variable is calculated on a five point scale. Thus, a decision which is totally pro-labor would have an overall impact of 2.25, and one which is totally pro-management would have a value of .45, the difference amounting to 1.8. The impact of the institutional variables would thus appear to be even greater than that of the political influence variables. This indicates that insitutional impacts are indeed important factors in board decisions. This point is often over-emphasized by legal scholars 191 studying the agency, but has been almost totally ignored by those who have pursued these questions from a political or economic perspective. The final factor to be considered is the impact of economic factors on the decisions which come before the board. Again, the variable LMIS is significant. Although the value of the beta-coefficient is very small, .02, the range of the impact can be quite great. The average value for the variable LMIS during the period of the study was 8.8, meaning that on average the impact of the variable would be almost .2. The percent of actual cases predicted correctly is 63.7. While this is not an extremely high level of prediction, two points should be made. First, the independent variables used are by and large simple binary variables. More sophisticated measures would most likely improve the level of correct predictions. Second, the dependent variable is a four-fold categorization and not a dichotomous categorization. In a dichotomous situation, simple guessing would give us a correct answer fifty percent of the time, and thus a value of 63.7 would not be a great improvement. In this case, simple guessing would provide us with the correct answer twenty-five percent of the time. In this light, a prediction probability of 63.7 is a great improvement. In summary, we can state the following points concerning private influence and the appointment of NLRB 192 board members and their voting decisions. First, it is clear that private groups - labor and management - do exert influence on the appointment of NLRB commissioners. Second, both the President and the Senate oversight committee are institutions through which this influence flows. Third, the Senate appears to have had a greater impact than the President on the voting patterns of commissioners over the years studied. Fourth, the institutional variables have a greater impact than the political variables on the voting decisions of NLRB commissioners. NLRB RESULTS What do these results say about private influence on the NLRB? Overall, the results are consistent with how private influence should act in a balanced environment with neutral statutes. This is especially true for the budgetary model presented and the appointments model presented. The evidence clearly indicates that private influence does impact the appointments process and that this influence plays a large role in the later decisions of the commissioners. Likewise, the cooperation between labor and management regarding the budgetary process to insure adequate funding of the agency is also well supported by the evidence. The information on direct influence is somewhat sketchy and can be interpreted in a number of ways. Depending upon 193 how we interpret the variance in the decision patterns among the regions, a number of different conclusions can be reached. If we attribute the variance in the decision patterns to other causes, i.e., economic conditions, industry performance, then the information provided is highly supportive of our overall contentions. If, however, the variance is not attributed to other causes, then direct influence is a possibility and our notions regarding the impact of direct influence and balance must be reformulated. For now it seems safe to say that while direct influence cannot be afforded the strong role for private influence that it receives in the general literature, it is still a possible channel of private influence. The results clearly indicate that the state of balance which is so often referred to in the group literature is not a static condition. One does not simply have a perfectly unchanging balanced condition when the proper ingredients are present. This state of balance is in fact a dynamic condition. The interests involved are constantly maneuvering to improve their position relative to the opposition. Wilson has argued that groups will attempt to dominate the administrative structure in such a situation, and that while both may be able to do so for short periods of time, neither will be able to dominate the commission for an indefinite period.8 This contention seems to be supported by the evidence, especially in the appointments process. While over time the influences of each group may balance the other at 194 some macro-level perspective, the internal dynamics which lead to this condition are far more conflictual than intimated by the macro-level perspective. In the case of the NLRB, groups have resorted to what may be termed peaceful coexistence, or the politics of accommodation. There are areas of cooperation, while at the same time, there are areas of conflict. Groups still act in their own best interest, however this action may result in cooperation or conflict between the parties. Where conflict results, it may either end with one group gaining short term advantage over the other, or it may result in stalemate. The stalemate of which we speak has provided the foundation for the consistent structure of the post-Taft-Hartley era. The Congressional stalemate, resulting from the inability of either group to gain a clear majority has provided the foundation for this period. Escalation of conflict has prevented either side from upsetting the.balance of the period at the legislative level. This same type of stalemate can also be found along some of the influence channels. Direct influence seems to follow this same pattern. In most of the regions studied, there was little evidence of direct influence. However, where direct influence could possibly be found based on our study results, the decision patterns were always in the direction of the stronger group. Thus it would appear as if both groups still try to directly influence the agency, but are only successful in regions 195 where they gain an advantage over the other group. In the other regions, stalemate has resulted with the agency opting to follow the rules of the commission. The contention that direct influence is important has a number of implications for our study. First, if it is possible to influence directly the regional decisions, then the rule of law has a lesser impact than we would have expected. If this is the case, then the decisions of the board have less importance and less impact on the decisions made at the regional level. Also, the finding that direct influence is able to exist even in a balanced environment lends support to those who argue that private influence is Of paramount importance in the general regulatory single private interest scenario. Cooperation results when both groups find it to their mutual advantage to pursue a particular policy. In the budget process of the NLRB, both labor and management have something to gain from seeing that the NLRB is adequately funded. In essence, these two material groups have deferred the costs for settling their conflict onto an unorganized third party, the general public. With regard to the appointment of the commissioners, both groups have used their influence to appoint commissioners who favor their policy positions. Here, there is open conflict with some compromise. Both groups desire to keep the conflict at a low level: they do not want it to escalate to the Senate floor and in open battle. However, 196 both groups realize the importance of gaining influence through the top echelons of the commission. As they cannot influence the decisions of commissioners after their appointment, they must exert influence during the selection process through their political allies. It is clear that the channels of influence are different in terms of their importance in the case of the NLRB than they are in the general case. Appointments are the most important, direct influence is second, and the budget has become an area of compromise and cooperation. In the general case direct influence was the most important. If this is a successful channel, then it appears that the others are not really needed. Direct influence allows the private interests to influence the grass roots decision making level and higher levels of agency decision making. Rules can be influenced at hearings, and if not successful here the groups can influence the outcomes during implementation. For example, a group may be unable to prevent a rule for inspections by an agency, but it may be able to prevent through direct influence meaningful inspections. The budget is not an important channel in the general case as it is used primarily to negate unfavorable actions by the agency. Again, if this attempt fails the private group can resort to direct influence to offset its loses. The same logic holds for unfavorable commission appointments. Groups can employ private influence to thwart 197 the control Of a new commissioner within his own agency. In the case of the NLRB, direct private influence is severely limited although not entirely eliminated. In this situation the groups use their resources to control other areas and influence the commission differently. The most important avenue now becomes the appointment of the commissioners. This is true because the groups can no longer impact those under the commissioner and he is able to control the actions of the agency. Nor can they induce the commissioner with post-employment rewards to which he is not already predisposed. Competition in the political environment has altered the ways in which private influence must be conducted and made some channels much more profitable than others. Another interesting finding concerns the importance of the institutional factors in the decisions and rulings of the agency. The impact of prior commission rulings, and the importance of the general counsel in the rulings of the board are clear indications of this fact. For too long social scientists have overlooked the importance of these types of factors, while legal scholars have stressed them to the exclusion of all other points. In the past few years it has become evident that both are important considerations when considering the performance of institutions, that is, the decisions that they make. For, example, prior studies of the decision patterns of the NLRB have focused entirely on political and economic variables and completely ignored the 198 institutional factors.9 These findings should not be understood as relationships that will endure forever, or for that matter even into the immediate future. The influence channels, the bargaining, and the cooperation which we have noted will last as long as the post-Taft-Hartley period endures. If for some reason this balance should be upset, then systemic level changes may come about. New legislation may be sought, and one side may come to dominate the other through the attainment of promotional legislation. Thus, the period which comes after the Taft-Hartley era may resemble the Wagner Act era with the agency predisposed towards one of the two groups.10 Therefore, in a sense, the notion of balance is a correct way of understanding the relationship between labor and management with regard to the NLRB. There must be some sense of balance between the two groups which allows the present system to continue. If, however, one group should push this state to far to one side or the other then conflict could replace cooperation at all levels. The escalation of conflict may then be a plausible strategy for the group which finds the current situation intolerable. Economic conditions in the present period and the unprecedented series of union givebacks in contract negotiations, as well as the strained state of the economy, may signal the beginning of a new era in labor-management relations which will fundamentally alter the role of the 199 NLRB. These are speculations only. Our knowledge is too limited to predict the future. This is especially true with regard to the relationship of economic factors and their impacts on regulatory agencies and the decisions they reach. With these results for the NLRB, we can now turn to the consideration of what these findings mean for questions of private influence and regulation in general. F I NAL RESULTS Finally, we must consider the meaning of these results for the question of private influence and the regulatory commissions in general. Clearly, the notions of a perfect state of balance are not verified by the data from our study. The arguments which have been presented as well as the data obtained to support these arguments show that even in a balanced environment groups will continue to influence the policy decisions of an agency. While this may not be a shocking revelation, the means by which groups proceed, and the institutions which they employ as means to achieve their policy preferences, are better understood. Ideas about the institutions involved as the means through which influence flows are also subject to revision. It is clear that the strong role afforded to the President in the appointments process and the weak role assigned to Congress has been overstated. Clearly, the ability and desire of a private interest group to influence the Congress 200 to achieve its policy preferences during the appointments process should receive more attention. Whether this result will hold for other regulatory agencies is a matter for more study. However, at least it is evident that private influence is important and does exist in a balanced environment and the Congress is an active target of the influence. This point has been made clearer in recent years with groups attempting to achieve legislative changes to over rule a commission. For example, a recent FTC ruling over information disclosure for used car dealers was overturned by the Congress. The legality of this action is still in the hands of the courts, but the point is that Congress can serve as a highly useful avenue for group influence in achieving their policy preferences. Recently, a number Of authors have begun to look more closely at the Congress and its influence in a more systematic fashion.ll Also, the role of the Senate in the appointment process in general may require some revision. The general wisdom places most of the importance on the President and affords little impact to the Senate. This study provides a first attempt in revealing when the Senate will play an active role in the process. Additional studies may provide further information on this subject, and further clarify the roles of the Senate and the President. The information provided on private interest in a balanced environment also has important implications for the 201 reformist position concerning the creation of public interest groups. The role of public interest groups to provide a balanced environment and thereby eliminate undue 12 The private influence must be brought into question. question of who public interest groups represent has been the most damning criticism of this position to date. We must now ask to what degree they will be successful in obtaining their intended goal of balancing the environment? First, they would normally not be as strong as the material interest groups located in the study of the NLRB. This would further weaken their ability to control direct influence which is still an open question with regard to the NLRB. Second, and even more damning is the simple fact that influence channels can shift. Even if the group is strong enough to stop direct influence, it is unlikely that it will be able to control the appointments process, or for that matter the budgetary process. These considerations limit the impact of public interest groups in a number of ways. It is likely that the strong interest normally associated with most regulatory commissions would shift its influence patterns. Further, our results as they relate to the impact of direct influence indicate that the groups may be unable to sufficiently block this line of influence. Sabatier's argument that public interest groups can slow the process of capture is not as important a finding as it was once thought to be.:[3 For this slowing, would probably be very minor given the results of 202 our study. The creation of public interest groups to initiate reform and prevent capture has a possibility of achieving success, in the short run. However, in the long run our study shows that organized interests will find alternative means of exerting their influence. The only hope for a public interest group is for it to be almost as strong as the organized interest, or have massive public support for its position. While these things may occur in the short run, it is unlikely that the countervailing power can be sustained over the long run. If we desire to create a state of balance, we must understand it more fully so as to make a success of the public interest group reforms. This study also highlights the importance of looking closely at group configurations and private influence. The general case, one organized interest versus an unorganized interest has been much studied and all of the ideas concerning private influence and the usefulnes of regulation have been drawn from it. This study has demonstrated that the balanced environment is also an interesting case and tells us something about private influence. It can be seen as a jumping off point for a more sophisticated look at private influence and regulation. Up to now most studies of this type have been atheoretical case studies. For example, the number of groups in the constituency configuration can be extended beyond two. This is a prevalent sitation in many of the sub-policy areas of many 203 regulatory agencies. For example, the FCC has recently been embroiled in a controversy concerning the ITFS network.14 These low frequency channels had been reserved for educational use in past years. Recently, commercial interests, and the public broadcasting network have entered the fray and are attempting to gain control of the channels. If a condition such as this exists over time it may be possible to apply more sophisticated techniques to trace channels of private influence Open to the groups. In turn, we can learn more about private influence channels in more complex environments. Another point of this study concerns its impact on the position of those who indicate that we should limit bureaucratic discretion and move towards a rule of law. Lowi's notion of juridical democracy is one of the best known treatments of this subject.15 Lowi has argued strongly for the rule of law to replace the liberalism which has allowed private influence to flow in the traditional governmental bureaucratic forms. The NLRB is a good example of this type of agency and is as specific and juridical as any agency can be. If we add to this fact that here the environment is balanced, it is obvious that even with specific laws and a lessening of bureaucratic discretion private influence will still flow. Groups merely shift their focus of influence. They do not attempt to bend the law, but attempt to control the positions of those who interpret the law. This is not to say that Lowi is fundamentally wrong, 204 but it does provide us with a more detailed look at this type of governmental design approach. The conclusions we are forced to accept concerning the rule of law as a means to replace private influence and interest group liberalism are not favorable. Again, it appears as if the strong interests will be able to overcome this type of structural reform. The NLRB has one of the best of all possible environments for the rule of law. If any environment should so little or no influence and a deference for the rule of law, this is it. However, if two groups which balance each other in most of the environment are still able to exert influence on a highly judicialized agency it is most likely that when only one strong interest is present their will be even more private influence. Lowi's reforms are another structural attempt to eliminate politics from administration. He does not simply part the two because their mixture is evil as the early writers in public administration did: however, the desire is the same. As our study shows though, this reform like all others is limited and must be understood to be so. It has trade offs and costs which must be more fully understood. If we are to chose this rule of law which Lowi argues for it deserves more study. Influence patterns with the NLRB show us how this type of approach will work in a balanced environment, we must now look to agencies without a balanced environment and a rule of law to see the situation more clearly. Only then can we draw more definite conclusions. 205 One problem which our study has not been able to overcome concerns the separation of the joint impact of neutral statutes and a balanced group structure. These two factors upon which our study is based must be disentangled. Some studies of other agencies may help us approach this question. A simple typology for further study of regulatory commissions is proposed to achieve such an understanding. The typology concerns a four fold categorization for regulatory commissions. Along one dimension we have the statutory considerations. Statutes are promotional or neutral. That is, the regulatory commission has the primary responsibility as a policing agency, enforcing government regulations. In the other direction, the commission is concerned with promoting the industry that it is supposed to regulate. Along the other dimension the constituency configuration is outlined. The two categories included are the general case with one strong organized interest opposed by an unorganized interest, and the other category is one in which two organized interests Oppose one another in the policy space. A number of commissions can be put into the typology to investigate the impact of the two factors. Some commissions can be included in both categories after major statutory changes. The NLRB is the most obvious example, indicating that our study should be directed back in time. The Wagner Act NLRB would be a good example of a commission with two organized interests and promotional statutes. Another agency 206 which comes to mind is the Atomic Energy Commission, and its successor the Nuclear Regulatory Commission. Here we have a commission which has one organized interest group Opposed by an unorganized interest group. However, the statute changes from promotional to policing during the reorganization of the ABC and the creation of the NRC. The implication of these concerns is that when agencies are designed the functions which are included in their mission should be carefully considered. The traditional planning perspective put forth by Berstein, Landis and others called for a combination of the functions - promotion 16 If we view this design issue from a and policing. perspective concerned with private influence, rather than coordinated planning, then it may be that the functions should be separated. Rather than seeking to eliminate private influence through the creation of a balanced environment or through the rule of law as discussed earlier, it may be better to attempt to curtail private influence through the structuring of organizations around specific types of objectives. The problem of combining policing and promotion functions which can exacerbate problems of private influence can be turned to as the basis for an organizational structure which will institutionalize the conflict with different agencies. This proposal is far from foolproof, for both commissions will still be subject to private influence. However, there are some distinct advantages. Public interest 207 groups can focus their efforts on the policing functions at least and use this organization as a tool to impact the organized interest through a governmental forum. Balance is in a sense achieved through the use of government institutions much like the balance achieved through the three branches of government. In conclusion, we would argue that this study is a beginning. It has provided us with information on the subject of private influence and the NLRB. We have seen how influence channels change and how private groups act in a balanced environment. But more than just a study of the NLRB has been provided. These results should be seen as part of a larger attempt to understand questions of private influence and regulatory agencies and to develop more sophisticated means and methods of understanding them. FOOTNOTES 1. These tapes were obtained from the National Labor Relations Board Media Library. They cover the period from 1964 to 1978. The tapes list cases closed during a fiscal year. As there are some cases which are not closed for a year until later years, the year 1977 was chosen in lieu of 1978. 2. Cases are coded in the following manner. If a case is dismissed or withdrawn then it is coded as a victory for the respondent. If the case is settled or a complaint is issued it is coded as a victory for the petitioner. Total victories for labor petitioners and respondents are the decisions in favor of labor, and victories for management petitioners and respondents are coded as pro-management decisions. 3. Donald Warwick, A Theory of Public Bureaucragy (Cambridge, Mass.: Harvard University Press, 1975). 4. Charles Ostrom, Time Series Analysis (Beverly Hills, Cal.: Sage Publications, 1978), p. 34. Jan Kmenta, Elements of Econometrics (New York: McMillan Publishing Co., 1971), pp. 279-281. 5. National Labor Relations Board, Annual Repogt ef the National Labor Relations Board, 1952-1981 (Washington, D.C.: Government Printing Office, 1952-1981). 6. U.S., Office of Management and Budget, Budget of the United Statee, 1952-1982 (Washington, D.C.: Government Printing Office, 1952-1981). See the specific sections on the NLRB budget. See also the Congessional hearings on the NLRB budget. These are found in the budgetary hearings for the Labor Department. 7. Fanning has been reappointed by both Democratic and Republican Presidents since his first appointment by Eisenhower. Kennedy reappointed him shortly after taking Office in 1961, shortly after the passage of the Landrum-Griffin Act. 8. James Q. Wilson, “The Politics of Regulation,“ in James McKie, ed., ' Egedicament (Washington, D.C.: The Brookings Institution, 1974), pp. 157-59. 9. DeLorme, Hill and Wood, “The Determinants of Voting by the NLRB on Unfair Labor Practice Cases: 1955-1975,“ Egblic Choice, 37(2) (1981), pp. 207-218. 209 10. For more information on periods of structural stability and influence see E.E. Schattschneider, The Semi-Sovereign Pegple (New York: Rinehart and Winston, 1960). 11. See for example Barry Weingast, Op. cit.. 12. Public Participation in Regulatory Agency Proceedings, Vol. III of the Study on Federal Regulation, op. cit.. 13. Paul Sabatier, “Social Movements and Regulatory Agencies: Toward a More Adequate and Less Pessimistic Theory of Clientele Capture,“ Policy,Sciences, 6(3) (September,l975), pp.301-342. 14. See FCC proceedings. 15. Theodore Lowi, The End of Liberalism, op. cit., especially pages 298-313. l6. Marver Bernstein, Regulating Business by Independent Commission, Op. cit., pp. 253-265. VII FUTURE CONSIDERATIONS As with most studies in the social sciences, it is often the case that they generate as many if not more questions than they answer. This is especially true if the study broaches a new subject area or approaches an Old area from a new perspective. Many of these questions arise from assumptions made in the earlier study, or are attempts to deal with problems for which the first study found no satisfactory solution. Others are questions which are suggested by the research. In this the final chapter we will consider a number of such questions. The specific area of interest which has generated the most problems and concerns is that of direct influence. A number of problems within this area can be investigated so as to improve the data analysis. The data are available, and a number of alternative variables can be added to the model to further aid our understanding of decisions at the regional level and the impact of direct influence on them. Further, we will be able to go beyond simple descriptive statistics and employ more sophisticated statistical techniques. The major factor which has prevented a detailed investigation of these propositions is a shortage of funds. The data tapes Obtained from the NLRB and utilized in the 210 211 study are extremely large with over four-hundred thousand cases. The expense involved in sampling these tapes and producing usable data sets is a primary stumbling block. In addition, the data for the period from 1947 to 1963 has not been converted to tapes. Converting this data to tapes would be a major undertaking. However, such a project would provide us with a usable data base for the entire period of interest. The problem is further exacerbated by the fact that the tapes contain alpha-numeric variables and zone-punches, as does the card-stored data for the period prior to 1964. The conversion Of these tapes and cards into a usable format will require a large amount of resources, extensive programming skills and a great deal of time. However, such an investment is worthwhile if one considers some of the factors which can be considered and the questions which may be addressed by such a project. For example, in the data presented on direct influence we were forced to use calendar years in lieu of fiscal years as the time period for grouping the data. While this is a minor point, the conversion of calendar year data into fiscal year data will provide a set of results which are comparable to the other data sources on the NLRB. One final problem also adds to the costs of such an undertaking. This involves the merging of other data sources with the NLRB data so as to create variables which are theoretically interesting. The tapes are agency data sources ..o. ,w r‘ a. . fa... A..-) l 0.4.:- 212 designed to aid them monitor agency perfomance and as such are not designed for the use of political scientists who are interested in a different set of questions. For example, while the tapes provide information on the unions involved in terms of their size and affiliation, they do not provide information on the degree of unionization of the industry or the economic conditions within the industry. The merging of these various sources will provide an extremely useful data set. A number of more theoretically interesting variables can be considered so as to improve our analysis of the causes for the variations in the decision patterns from region to region and perhaps from year to year. In other words, the ceterus paribus clause which looms so large in our study on direct influence can be slowly attacked. In turn, this approach will allow us to move beyond aggregate considerations of direct influence and better understand the process in greater depth and detail. A number of variables suggest themselves as possible factors in the determination of decision patterns at the regional level. The size of the unions involved, and the size of the businesses are two of these. Perhaps a large and wealthy organization has a distinct advantage over a smaller Opponent. Also, the type of union involved may have an impact. National unions may be able to provide more support for their locals than the independent unions and be more successful in presenting their case before the NLRB. The -—'u 4‘ ._‘u —\ .0 (1 213 industry involved may also prove to be important. Unrest in the industry due to economic conditions or the degree of unionization in the industry may have an impact. Industries undergoing a large upheaval in terms of poor economic performance or high unemployment may exhibit quite a different pattern of cases being brought before the NLRB. Likewise, industries undergoing unionization on a large scale for the first time may exhibit a different pattern with both sides bringing cases before the board with a great deal of regularity over the most minor issues. All of these factors and others provide future points for consideration. This approach will allow us to ascertain in much more detail the impact of direct influence in a balanced environment with neutral statutory guidelines. One problem of the utmost interest concerns the development of better measures for the variables in our study. This problem is inherent in most social science research. As McKelvey and Zavoina note, “even if (a) variable of theoretical interest is appropriately conceptualized as interval level, measurement theory in the social sciences is simply not refined enough to generate an interval level operationalization of this variable.“1 In the case of the dependent variable in our voting model we have been able to utilize a sophisticated methodological technique to overcome this problem. However, the use of binary variables especially as they relate to the appointment model deserve further study. The variables SAPP 214 and PAPP are sufficient for now but a great improvement would be achieved if a continuous variable could be substituted for the current binary variable. One such measure which may help us develop a measure of influence for the groups is that of union membership. A number of attempts were made to incorporate this measure into the study, but due to severe problems they were discontinued. Union membership figures have only been gathered on a state level since the late 1950's.2 While some measures were developed for the period prior to this they are not comparable and the data is limited to only a few years. It is possible to construct a ranking of the states in terms of union membership, but the measure needs to be further refined before it can be incorporated to a study such as ours. For example, the union membership variable for the nation and the states as measured by the Bureau of Labor Statistics only counts employed union members.3 A plot of the national data shows that membership figures normally increase during periods of Democratic or labor supported Presidential administrations and decrease during periods of Republican or management supported Presidental administrations. As such, any attempt to use this figure as a measure of the strength of the group would be inappropriate. These fluctuations may be attributable to the policies employed by the Presidents and one means of correcting it may be to include the unemployed union members -.‘--. v.- 215 in the figures. This can be accomplished by correcting the Bureau of Labor Statistic measures by the unemployment rate. The state unemployment levels can also be used to correct state union membership figures in a similar fashion so as to provide an indication of the number of active and inactive union members in a state. Once an appropriate measure of union membership has been obtained a ratio can be constructed to determine the relative strength of unions in the nation and in the states. Obtaining data from the Census Bureau on the number of eligible voters, which is often used to determine voter turnout levels, it is relatively simple to determine a ratio.4 This measure was developed in its uncorrected form. using the available union membership figures and was refered to in chapter four and six. Once such a ratio is developed it will serve not only to expand our study but will also provide additional benefits. For example, the usefulness of this measure on the votes of Senators can improve our understanding of the way constituency size impacts their voting patterns. One example of this involves the passage of the Taft-Hartley Act in 1947. Using an estimation formula to determine the strength of union membership in the states and constructing a ratio based on the number of eligible voters in the state, the votes of Senators on the Taft-Hartley Act were investigated.5 The usual explanation has been that the Northern Democrats vote for labor, while the Southern and 216 Western Democrats vote against labor. While these generalizations are generally correct, the ratio of union membership in the states provides a better explanation. First, we rank the states according to their ratios of union membership. Next, we divide the states into two groups - high union membership states and low union membership states. We find this to be a highly important factor in explaining the vote. Of the thirty-seven democrats who voted on the act, both before and after the Presidential veto, we find that nineteen came from states with low union membership ratios and eighteen came from states with high union membership ratios. Of those in the lower state union membership group, only six voted against the act, and thirteen voted for the act. Of those in the high group, two voted for the act, and sixteen voted against the act. It is clear from these preliminary results that the size of union membership in a state relative to the number of voters in the state had an impact on the voting of Democratic Senators. Those with low union membership ratios tended to vote for the Taft-Hartley Act and against labor, while those from states with high membership ratios voted against Taft-Hartley and for labor. This measure can be employed in other studies as well. For example, voting patterns of Senators on labor issues can be used to find the impact of union size on their votes. This is a more theoretically interesting way of investigating the votes of Senators than the older 217 explanation which looks at the geographic location of the Senators home state. An even more ambitious undertaking involves the extension of the above measure to the House of Representatives. This would require the estimation of union membership ratios within individual districts. These can be estimated from the state figures using sources on employment levels and industry locations available from the Bureau of Labor Statistics. The estimation technique would not allow us to produce an interval level scale as we propose for the states, but we may be able to build an ordinal level scale. Another way to improve the research is to test the assumptions of the dissertation. This will provide a more rigorous test of the conclusions put forth in this study. One key assumption of the dissertation is the consistency and stability of the post Taft-Hartley period. Throughout the dissertation we have assumed that this period is stable and that the essential relationships have remained constant. While we have argued and will continue to argue that this is a plausible assumption, it would be interesting to test this proposition. This is especially true for the models tested concerning the appointment of commissioners and the impact Of private influence on voting decisions. In the model on commissioner appointments the period known as the post Taft-Hartley era was considered to be a period of structural stability. The variables in the model have a consistent impact throughout the period. An 218 alternative explanation is also possible. It has been argued that the NLRB has gradually become more institutionalized over time.6 The implication of this view is that the impact of private influence through the appointments process on the voting patterns of commissioners is less in the latter part of the period than in the earlier part. Such a problem deserves further study and the opportunity for such an investigation will be available in a few years. This will be the case even if the binary measures are still employed. The problem that has prevented such an approach has been the domination of the Senate by the supporters of labor, i.e., the Democratic Party. From 1955 until 1980 the supporters of management were unable to gain control of the Senate. This prevents us from testing the impact of the Senate exclusively in this later period. With the election of Reagan in 1980 and the control of the Senate passing to management supporters for the first time in many years further study will be possible in a few years. With the present specification of the model we will have to wait until the present Reagan commissioners resign. At this point in time it will be possible to test the impact of the variables in both the early part of the post Taft-Hartley era and in the latter part. While there is no perfect point at which to split the period, the passage of the Landrum-Griffin Act would seem to be the most likely point at which to subdivide the period. As noted earlier, this act and the politics of its passage are quite different 219 from those of the earlier labor acts. Management and labor cooperated and supported its passage. Further, at this particular period in time, the first pro-management appointed commissioner ever to be reappointed by a labor supported president gained a second term. Another possible way to increase the explanatory power Of the model involves the division of the votes of commissioners into policy areas. In our study the votes of commissioners on all cases were considered together. It is possible, with the aid of NLRB documents and labor law experience, to break these cases down into issue areas. Other studies have employed a different data set, using the 7 cases listed in the Annual Report of the NLRB. These cases are broken down by issue area, however they only cover cases considered important by the NLRB. This will not allow a consideration of all the factors considered important in our model, that is institutional as well as political considerations. Also, this approach is limited to the time period in which the commission reported cases in this format - 1955 to the present. Further, this approach does not present the investigator with a sample of most of the major board cases. What is neccessary is the combination of the positive aspects of each approach. This can be accomplished by employing the data set from our study and breaking it into issue areas as with the NLRB reported cases. The budget model as presented could also be improved by investigations along a number of different lines. These 220 would involve the development of more sophisticated measures of workload estimation, a model which relates workload estimation to the President's estimated budget, and the testing of alternative hypotheses concerning the budget. A more sophisticated workloading estimation model can be created by employing additional factors which relate to the NLRB's workload. Our current model did not, for example, break down the percentage of cases settled at different levels of the agency. If these change greatly from year to year it could have an impact on the cost of settling the average case. Also, cases could be broken down into even more detailed categories. For example, the different types of representation cases could be broken out. Finally, there are other ways to calculate the estimated workload. For example, the caseload of two years prior plus an estimated increase could be employed. Also, the workloading model could be used to test the estimations presented by the Office of Management and Budget. The figures estimated two years prior and one year prior to the enactment of the final budget could help us understand more about the budgeting procedures of the OMB as they relate to the NLRB. Finally, the model should be juxtaposed to the general incremental model of the budgetary process to provide additional information on the validity of our model. The incremental model of budgeting, which is very widely accepted as a general theory of budgeting, will serve as a 221 good alternative hypothesis.8 If our model predicts the budget with greater accuracy than this model it will provide us with even more evidence in support of our workloading model. So far we have considered ways to improve the points and issues raised in the dissertation: improved measures, challenging some key assumptions, and testing other alternative hypotheses. The next point that should be raised concerns the paths to which this current work leads us. That is, beyond the straightforward improvement of the study, what directions does this research lead us to. In the last section of the last chapter we noted a number of ways to extend the study. These included a new and more sophisticated way of looking at the constituency environment, and the creation of a new typology for the purpose of studying regulatory agencies. Here we can suggest a few more avenues of research. The area of investigation which holds a great deal of promise concerns the investigation of the lower level agency decisions. The NLRB has available a great deal of detailed data in these areas including information on ULP and REP cases. Studies are being conducted in this area and can be expanded. For example, Roomkin and Block have conducted studies employing representation data, and others continue to employ ULP data in their studies.9 This data can also be used for an organizational study of a highly judicialized agency. It may provide us with a means to quantitatively 222 test numerous propositions put forth in the organizational theory literature. For example, the detailed data available on the decisions at various levels of the NLRB provides an Opportunity to study the impacts of high level policy decisions on the lower level rulings of the agency. By looking at individual policy areas and various rulings of the board as they relate to that area we may be able to better understand the impact of policy changes and the way they impact the agency. These results may tell us something about problems of agency control and rule following. The study also leads us to the general case for developing performance measures for other agencies based on the cases decided by the boards or rules issued by the commission. For example, Moe, in a recent article, developed performance measures on a number of regulatory commissions.10 This work can be extended to other agencies, and provide us with information that will enable us to quantitatively test issues surrounding the regulatory commissions. The next logical point in the progression of the research concerns the closing of the circle. To this point, we have been concerned with the impacts of the private interests on the agency and the way in which they obtain their preferred issue positions. However, the decisions of the agency also have an impact on the environment which surrounds them. One such impact should be the number of union members in the United States. Up until now, the primary factors under consideration in determining the 223 growth of union membership in the United States have been economic. However, this study points to the need to consider institutional factors in the study of the growth of union membership. One author has incorporated the Wagner Act in his model as a binary variable, but that is as far as the connection goes.11 The graph at Table VII-l shows the growth of union membership in the United States from 1930 to 1978. On the surface it appears that the growth of unions increased dramatically with the passage of the Wagner Act in 1935. This is consistent with our notion of pro-labor promotion by the NLRB. The growth rate levels off considerably in the period after the passage of the Taft-Hartley Act. While numerous other factors should be included in the growth equations, such as the impact of World War II, the initial evidence suggests that there may be a relationship between the passage of the Wagner and Taft-Hartley Acts and the growth of union membership in the United States. There are most probably other areas of research which can be pursued; however, the few we have suggested will take a great deal of time to complete. 224 TABLE VII-1 U.S. Union Membership (in thousands) YEAR MEMBERSHIP YEAR MEMBERSHIP 1930 3401 1955 16802 1931 3310 1956 17490 1932 3050 1957 17369 1933 2689 1958 17029 1934 3088 1959 17117 1935 3584 1960 17049 1936 3989 1961 16303 1937 7001 1962 16586 1938 8034 1963 16524 1939 8763 1964 16841 1940 8717 1965 17299 1941 10201 1966 17940 1942 10380 1967 18367 1943 13213 1968 18916 1944 14146 1969 19036 1945 14322 1970 19381 1946 14395 1971 19211 1947 14787 1972 19435 1948 14319 1973 19851 1949 14282 1974 20199 1950 14267 1975 19611 1951 15946 1976 19634 1952 15892 1977 19695 1953 16948 1978 20246 FOOTNOTES 1. Richard McKelvy and William Zavoina, “A Statistical Model for the Analysis of Ordinal Level Dependent Variables,“ Journal of Mathematicel Seeielegy, 4 (1975), p. 103. 2.The Bureau of Labor Statistics collected information on union membership figures for large unions, AFL-C10, UAW, etc., from 1959. In 1962 the bureau began to collect unformation on state union membership levels every two years. 3. U.S., Department of Labor, Bureau of Labor Statistics, BLS Handbook of Methods (Washington, D.C.: Government Printing Office, 1976), pp. 207-219, and other pertinent sections. 4. U.S., Department of Labor, Bureau of Labor Statistics, Statistical_Abstract_of_the_nnited_8tates (Washington, D.C.: Government Printing Office, 1960-1980). These figures are obtained from the Bureau of the Census. 5. Congressional Quarterly, 1947 Congpessional Quarterlx_A1manaQJ OP- cit-p PP. 32-35, 300-314- 6. This position has been presented by a number of legal scholars who have studied the NLRB. 7. See Delorme, Hill and Wood, op. cit.. This method involves the use of all cases mentioned in the Annunal Report. The cited article made use of the ULP cases only. 8. Aaron Wildavsky, The Politics of the Budgetary Process, op. cit.. See also Davis, Dempster and Wildavsky, “A Theory of the Budgetary Process,“ The American Politiqel Science Review, 50 (September,l966) pp. 529-547. 9. Myron Roomkin and Richard Block, “Case Processing Time and the Outcome of Representation Elections: Some Empirical Evidence,“ University of Illinois Law Review, (1) (1981). pp. 75-97. 10. Terry Moe, “Regulatory Performance and Presidential Administration,“ American Journal of Political Science, 26 (2) (Hay'1982)' pp. 197-2240 225 226 11. There have been a great number of articles written on the subject of union growth. For an excellent summary of the literature see Richard Block and Steven Premack, “The Unionization Process: A Review of the Literature,“ in Advances in Industrial Labor Relations: A Research Annuel, David Lipsky and Joel Douglas, eds., (Greenwich, Conn.: JAI Press, forthcoming). APPENDICIES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. APPENDIX A NLRB REGIONAL CODES Boston New York Buffalo Philadelphia Baltimore Pittsburg Detroit Cleveland Cincinnati Atlanta Winston Salem, North Carolina Tampa Chicago St. Louis New Orleans Fort Worth, Texas Kansas City Minneapolis Seattle San Francisco Los Angeles Newark, New Jersey 227 23. 24. 25. 26. 27. 28. 29. 30. 31. 228 Houston Santruce, Puerto Rico Indianapolis Memphis Denver Alberquerque, New Mexico Brooklyn Milwaukee Los Angeles APPENDIX B Random Case Sampling Program PROGRAM SAMPLE (INPUT,OUTPUT) C THIS PROGRAM SELECTS A RANDAOM SAMPLE OF CASES FROM C THE NLRB BOARD DECISIONS. THE VARIABLES ARE: C IRAN - RANDOM INTEGER CASE NUMBERS C 180 - SEQUENCED INTEGER CASE NUMBERS C VOLN - NUMBER OF CASES IN NLRB VOLUMES OF DECISIONS C VOLUM - NLRB VOLUME NUMBER C CASE - NLRB CASE NUMBER BY VOLUME INTEGER CASE,VOLN,VOLUM,IRAN,ISQ DIMENSION VOLUM(1500),IRAN(1500),CASE(1500), +ISQ(1500),VOLN(300) C THIS LOOP READS THE VALUES OF VOLN FROM THE DATA CARDS READ(1,10) (VOLN(J),J-100,239) 10 FORMAT(2013) D0 5 J-100,239 WRITE(2,8) VOLN(J) 8 FORMAT(1X,110) 5 CONTINUE D0 50 J'1,1500 IRAN(J)-0 50 CONTINUE C THIS NESTED LOOP SELECTS A RANDOM NUMBER, PUTS IT IN 229 230 C THE PROPER BPOUNDS (1-25206) AND ELIMINATES ANY C DUPLICATES. AT THE END OF THIS PROCESS IRAN CONTAINS C UNIQUE CASE NUMBERS IN INTEGER FORM 190 240 260 DO 260 I31,1500 YtRANF(N) YY=Y*25207 IYY-IFIX(YY) DO 240 J81,1500 IF(IYY.EQ.IRAN(J))GO TO 190 CONTINUE IRAN(I)8IYY CONTINUE C THIS LOOP CREATES VALUES FOR ISQ WHICH IS AN ORDERING C OF THE VALUES OF IRAN FROM THE LOWEST TO THE HIGHEST 280 290 300 320 D0 320 I81,1500 ISQ(I)=25208 D0 300 J81,1500 IF(IRAN(J).LT.ISQ(I))GO TO 280 GO TO 300 IF(I.LE.1)GO TO 290 IF(IRAN(J).LE.ISQ(I-1))GO TO 300 ISQ(I)-IRAN(J) CONTINUE CONTINUE C THIS LOOP CONVERTS THE VALUES OF 180 INTO VOLUME C AND CASE NUMBERS AS ASSIGNED BY THE NLRB DO 440 It1,1500 400 420 440 450 500 231 IDUMMY=0 DO 420 J'100,239 IDUMMY=IDUMMY+VOLN(J) IF(ISQ(I).GT.IDUMMY)GO TO 420 IF(IDUMMY.GT.256)GO TO 400 CASE(I)=ISQ(I) VOLUM(I)=100 GO TO 440 CASE(I)=ISQ(I)-IDUMMY+VOLN(J) VOLUM(I)8J GO TO 440 CONTINUE CONTINUE DO 500 I81,1500 WRITE(2,450)IRAN(I),ISQ(I),VOLUM(I),CASE(I) FORMAT(1X,4I10) CONTINUE END APPENDIX C CODE BOOK: MAJOR BOARD DECISIONS CASE NUMBER (1-4): Four digit code from 0001 to 1500. Each case has a unique user case number. ORDER (5): One digit code for each commissioner vote per C388. VOLUME (6-8): Three digit code corresponding to the NLRB Volume Number. Volumes run from number 100 to 239 inclusive. CASE (9-11): Three digit NLRB case number. Refers to the case number for a particular volume of cases. CASE TYPE (12-13): Two digit code for the type of NLRB case. Ol-CA' 02.CB' 03-CC, 04-CD, 05-RC’ 06-“, 07"”, OB-UD' MONTH (14-15): Two digit code for the month in which the case is closed at the board level. 01 is January and 12 is December. YEAR (16-17): Two digit code for the year in which the case is closed at the board level. Years for the study run from 1952 to 1977. 232 233 STATE (18-19): Two digit code for the state in which the charge was brought. The code runs from 01 to 50 for the states. States are coded in alphabetical order. 88 code is used when more than one state is involved, 99 is used when the state is not konwn. RESPONDENT (20-21): Two digit code which identifies the individual or organization which brought the charge. PETITIONER (22-23): Two digit code which identifies the individual or organization which is charged. TRIAL EXAMINER DECISION (24-25): Two digit code for the decision of the trial examiner or the regional director. 01- Full Compliance with the petitioner. 02- Partial Compliance with the petitioner. 03- Partial Dismissal in favor of the respondent. 04- Full Dismissal in favor of the respondent. 97- If there is no ruling required. 99- If the information is unknown. BOARD VOTE (26-27): Two digit code for the vote of the board. Those in the majority are followed by those dissenting. For example, a 50 is a unanimous decision, five votes for none against. A 32 is three votes for and two against. BOARD DECISION (28-29): Two digit code for the decision of the entire board. The decision of the vote of the board as a whole is coded from 1 to 4, from full compliance to dismissal. See the TRIAL EXAMINER DECISION variable. 234 BOARD MEMBER CODE (30-31): Two digit identification code for individual board members. Code is assigned in order of appointment to the board. 01- Madden, 02- Carmody, 03- Smith, E., 04- Smith, D.W., 05- Leiserson, 06- Millis, 07- Reilly, 08- Houston, 09- Herzog, 10- Reynolds, 11- Gray, 12- Murdock, 13- Styles, 14- Peterson, 15- Farmer, 16- Rodgers, 17- Beeson, 18- Leedom, 19- Jenkins, 8., 20- Bean, 21- Fanning, 22- McCulloch, 23- Brown, 24- Jenkins, H., 25- Zagoria, 26- Miller, 27- Kennedy, 28- Penello, 29- Murphy, 30- walther, 31- Truesdale, 32- Kimball. BOARD MEMBER DECISION (32-33): Two digit code for the decision of each individual board member. Range is from one to four. See Trial Examiner Decision for explanation of codes. BOARD MEMBER DISSENT (34-35): Two digit code for the vote of each board member in relation to the majority vote of the board, i.e., in relation to the variable Board Decision. 01- in full agreement with the board. 02- minor disent, in partial agreement with the board. 03- full dissent, in total opposition to the majority board decision. UNION CODE (36-37): Two digit code for the union involved as the primary litigant. 235 UNION CODE II (38-39): Two digit code for a union involved as an intervenor or the second union in a union verus union case. INDUSTRY CODE (40-41): Two digit code to identify the industry type in which the dispute took place. BIBLIOGRAPHY BIBLIOGRAPHY Abrams, Robert. Foundations of Political Analysis. New York: Columbia University Press, 1980. Abrams, B. and Settle, R. “The Economic Theory of Regulation and Public Financing of Presidential Elections.“ Journal of Political Economy, (April,l978), 86(2) pp. 245-257 0 Aldrich, John and Cnudde, Charles. “Probing the Bounds of Conventional Wisdom: A Comparison of Regression, Probit and Discriminant Analysis.“ American Journal of Political Science, 19 (1975), pp. 571-608. Bernstein,Marver. Regulating Business by Independent Commission. Princeton, N.J.: Princeton University Press, 1955. Bernstein, Marver “The Regulatory Process: A Framework for Analysis.“ Law and Contemporary Problems, 26 (Spring,1961), pp. 329-346. Block, Richard and Premack, Steven. “The Unionization Process: A Review of the Literature.“ Advances in Industrial Labor Relations: A Research Annual. Edited by David Lipsky and Joel Douglas. Greenwich, Conn.: JAI Press, forthcoming. Booker and Coe. “The NLRB and Its Critics“. Labor Law Journal, 17 (52) (1966), pp. 185-201. Booker and Coe. “The LAbor Board and Its Reformers“. Labor Law Journal, 18 (67) (1967), pp. 487-504. Cary, William. Politics and the Regulatory Agenciee, New York: McGraw-Hill, 1967. Casey, Gregory. “The Supreme Court and Myth: An Empirical Investigation.“ Law and Society Review, 8(1974), pp. 385-417. Congressional Quarterly. 1947 Congressional Quarterly Almanac. Washington, D.C.: Congressional Quarterly Press, 1947. Congressional Quarterly. 1959 Congressionel Querterly Almanac. Washington, D.C.: Congressional Quarterly Press, 1959. 236 237 Congressional Quarterly. Federal Regulatory Direetery 1981-1982. Washington, D.C.: Congressional Quarterly Press, 1982). Cox, Archibald; Bok, Derek: and Gorman, Robert. Leher_Law: Cases and Materiele. 9th Edition. Mineola, New York: The Foundation Press, 1981. Cushman, Robert. The Independent Regulatory Cemmiesiens. New York: Oxford University Press, 1941. Cyert, Richard, and March, James. A Behavioral Theory of the Firm. Englewood Cliffs, N.J.: Prentice-Hall, 1963. Davis, Dempster and Wildavsky. “A Theory of the Budgetary Process.“ The American Political Science Review, 50 (September,l966) pp. 529-547. DeLorme, C., Hill, R., and Wood, N. “The Determinents of Voting by the NLRB on Unfair Labor Practice Cases.“ Public Choice, 37(2) (1981), pp. 207-218. Dexter, Louis A. How Organizations are Repreeented in Washington. New York: Bobbs-Merrill, 1969. Downs, Anthony. Inside Bureaucracy. Boston: Little Brown, 1967. Eckert, Richard. “On the Incentives of Regulators: The Case of Taxicabs.“ Public Choiee, l4 (Spring,1973), pp. 83-990 Farris, Martin and Sampson, Roy. Public Utilities: RegulationyiManagemen;_end_Qynership, Boston: Houghton Mifflin, 1973. Fenno, Richard. Homestyle. Boston: Little Brown, 1978. Fiorina, Morris P. “Group Concentration and the Delegation of Legislative Authority.“ Unpublished paper, California Institute of Technology WOrking Paper, 1982. Fiorino, D. and Metlay, D. “Theories of Agency Failure.“ Unpublished paper presented at the American Political Science Association, 1977. Friedrich, Carl. “Public Policy and the Nature of Administrative Responsibility.“ Bureaucratic Power and National Politice, Edited by Rourke. Boston: Little Brown, 1978. 238 Fuller, Lon L. “The Form and Limits of Adjudication.“ American Court Systems: Readings in Judicial Procese_ and Behavior. Edited by Goldman and Sarat. San Francisco: W.H. Freeman and Co., 1978. Gellhorn, W., Byse, C., and Strauss, P. Administretive Law. New York: the Foundation Press, 1979. Gormley, William. “A Test of the Revolving Door Hypothesis of the FCC.“ American Journal of Political Scienge, 23(4) (November,l979), pp. 665-683. Greenstone, David. Labor in American Politics. Chicago: University of Chicago Press, 1977. Gross, James. The Making of. the.blatimmlJathRelatinns. Board. New York: State University Press, 1975. Gross, James. The Reshaping of the Ne;iepel_Leper_Reletiene_ Board. New York: State University Press, 1975. Hanushek and Johnson. Statistical Methods for Social Scientists. New York: Academic Press, 1977. Herring, Pendelton. Federal Commissioners: A Study of Their Careers and Qualifications. McGraw-Hill, 1936. Huntingdon, Samuel. “The Marasmus of the ICC: The Commisson, the Railroads and the Public Interest.“ Yele Leg ournal, 61(4) (April,l952), 467-509. Huthmacher, J. Joseph. Senetor Robert wagner and The Riee 9f Urban Liberalis . Forge Village, Mass.: The Murray Printing Co., 1968 Jaffe, Louis. “The Effective Limits of the Administrative Process: A Reevaluaton.“ Harvard Law Review, 67(7) (May,1954), pp. 1105-1135. Jaffe, Louis. “The Independent Agency-A New Scapegoat.“ Yale Law Journal, 63 (1956), pp. 1069-1076. Jaffe, Louis. “The Illusion of Ideal Administration.“ Harvard Law Review, 86(7) (May,1973), pp. 1183-1199. Johnston, J. Econometric Methods. New York: McGraw-Hill, 1972. Jordan, William. “Producer Protection, Prior Market Structure and the Effects of Government Regulation.“ Journal of Law and Economics, 15(1) (April 1972), pp. 239 Kaufman, Herbert. The Forest Ranger. Baltimore, Johns Hopkins Press, 1960. Kennedy, Peter. Elements of Econometriee. Cambridge, Mass.: M.I.T. Press, 1979. Kingdom John w. W. New York: Harper and Row, 1973. Kmenta, Jan. Elements of Econometrics. New York: McMillan Publishing Co., 1971. Kolko, Gabriel. W. New York: The Free Press, 1979. Kohlmeier, Louis. The Regulators: Watchdog Agencies and the Public Interest. New York: Harper and Row, 1969. Landis, James. The Administretive Process. New Haven, Conn.: The Yale University Press, 1938. Landis, James M. Report on the Regulatory Agenciee to the_ President-Elect. Washington, D.C.: Committee on the Judiciary, 86th Congress, 2nd Session, 1960. Leslie, David. Labor Law in a Nutshell. St. Paul, Minn.: West Publishing, 1979. Lowi, Theodore. The End of Liberalism. New York, Norton, 1969. Lowi, Theodore. “Four Systems of Policy, Politics and Choice.“ Public Administration Review, (July,August 1972). pp. 298-310. Madison, James “Federalist 10“ and “Federalist 51“. The Mind of the Founder: Sources of the Political Thought of James Madison. Edited by M. Meyer. Hanover, New Hampshire: University of New England Press, 1973. McConnell, Grant. Private Power and American Democracy. New York: Vintage Books, Alfred A. Knopf, 1966. Matthews, Donald R. U.S. Senators end Their WOrld. New York: W. W. Norton and Co., 1973. McCulloch, Frank and Bornstein, Timothy. The National Labor Relations Board. New York: Praeger Publishers, 1974. McKelvey, Richard and Zavoina, William. “A Statistical Model for the Analysis of Ordinal Level Dependent variables.“ Journal of Mathematical Sociology, 4 (1975i, pp. 103-120. 240 Meier, Kenneth and Plumlee, J. “Regulatory Administration and Organizational Rigidity.“ t Qgerterly, 31(1) (March,l978), pp. 80-95. Mitnick, Barry. T P ' E W. New York: Columbia University Press, 1980. Moe, Terry. The Organizaton of Interests. Chicago: University of chicago Press, 1980. Moe, Terry. “Regulatory Performance and Presidential Administration.“ American Journal of Politicel Seience, 26(2) (May 1982), 197-224. Murphy, William. “The NLRB - An Appraisal.“ innesote Ley Reviey, 52(819) (1968), pp. 819-849. Neter, John and Wasserman, William. A Statistical Modele. Homewood, Illinois: Richard Irwin, Inc., 1974. Niskanen. William. W Government. Chicago: Aldine-Atherton, 1971. Noll, Rodger. “Government Administrative Behavior.“ Unpublished paper, California Institute of Technology Working Paper. Noll, Rodger. “The Economics and Politics of Regulation.“ Virginia Leg Reyieg, 57 (1971), pp. 1016-1032. Noll, Rodger. Reforming Regulation: An Evaluation of the Ash Council Propposals. Washington, D.C.: The Brookings Institufion, 1971. Noll, Rodger. “The Behavior of Regulatory Agencies.“ Review of Social Economy, 29 (March,l971), pp. 15-19. Oberer, Walter: Hanslowe, Kurt; and Andersen, Jerry. Statutory Supplement to Cases and Materials on Labor Law and Collective Bargaining in a Free Society. St. Paul, Minn.: West Publishing Co., 1979. Oberer, Walter: Hanslowe, Kurt; and Andersen, Jerry. Labor Law and Collective Bargaining_in a Free Society. St. PauIiiMinn.: West Publishing Co., 1979. Olson. Hansur. MW. Cambridge, Mass.: Harvard University Press, 1971. !‘ 241 Ornstein, Norman and Elder, Shirley. Interest GroupsL Lobbying and Public Policymaking. Washington, D.C.: Congressional Quarterly Press, 1978. Ostrom, Charles Time Series Analysis. Beverly Hills, Cal.: Sage Publications, 1978. Owen and Braeutigan. The Regulation Game. Cambridge, Mass.: Ballinger, 1978. Peck, Cornelius. “The Atrophied Rule-making Powers of the NLRB.“ The Yale Lawyjournal, 70(129) (1961), pp. 729-761. Penoyer, Ronald, ed. Directory of Federal Regulatory Ageneiee. St. Louis: Center for the Study of American Business, 1981. Peltzman, Samuel. “Toward a More General Theory of Regulation.“ Journal of Law and Economics, 19(2) (August 1976), pp. 211-240. Pindyck and Rubinfeld. Econometric Models and Economic Forecasts. New York: McGraw-Hill, 1981. Pohlman, Marcus and Crisci, George. “House Support for Organized Labor.“ Political Science Quarterly, 97(4) (Winter,l982-1983), pp. 120-132. Porter and Sagansky. “Information, Politics, and Economic Analysis: the Regulatory Decision Process in Air Freight Cases.“ Public Policy, 24(2) (Spring,1976), pp. 263-307. Posner, Richard. “Economic Theories of Regulation.“ Bell Journal of Economics and Management Science, 5(2) (Autumn,l974), pp. 335-358. Pritchett, C. Herman. The American Constitution. New York: McGraw-Hill, 1968. Quirk, Paul. Industry Influence in Federal Regulatory A encies. Princeton, N.J.: Princeton University Press, 1981. Ripley, Randall and Franklin, Grace. Congress, the Bureaucracy and Public Policy. Rosenstone, Steven and Wolfinger, Raymond. “The Effects of Voter Registration Laws on Voter Turnout.“ American Political Science Review, 72 (1978), pp. 22-45. Rourke, Francis. Bureaucrady, Politics and Public Policy. Boston: Little Brown, 1969. 242 Roomkin, Myron and Block, Richard. “Case Processing Time and the Outcome of Representation Elections: Some Empirical Evidence.“ University of Illinois Law Review, 1981 (1) (1981), pp. 75-97. Sabatier, Paul. “Social Movements and Regulatory Agencies: Toward a More Adequate and Less Pessimistic Theory of Clientele Capture.“ Policy Sciences, 6(3) (September 1975), pp. 301-342. Salisbury, Robert. “An Exchange Theory of Interest Groups.“ Midwest Journal of Political Science, 13(2) (February,1969), pp. 1-32. Salisbury, Robert. “The Analysis of Public Policy: A Search for Theories and Roles.“ Political Science and Public Pgli y. Edited by Austin Ranney. Chicago: Markham Publishing Co., 1968. Salisbury, Robert. 'Interest Groups.“ Handbook of Political Science. 191. 1y; Nongovernmental Politics. Edited by Greenstein and Polsby, Nelson. Reading, Mass.: Addison-Wesley, 1975. Salisbury, Robert and Heinz, John. “A Theory of Policy Analysis and Some Preliminary Applicatons.“ olicy Analysis in Political Science. Edited by Sharkansky, Ira. Chicago: Markham Publishing Co., 1970. Scher, Seymour. “Regulatory Agency Control Through Appointment: The Case of the Eisenhower Administraton NLRB.“ Jgurnal of Politice, 23 (1961), pp. 667-688. Seidman, Harold. Politicey:Poeitien_end_£eyer, New York: Oxford University Press, 1975. Shapiro, David. “The Choice of Rulemaking or Adjudication in Policy.“ Harvgrd Law Reviey, 78(5) (March,l965), pp. 921-972. Sheperd, W. and Gies, T. Utility Regulation: New Directions in Theoryyand Policy. New York: RandOm House, 1966. Spaeth, Harold. Supreme Court Decision Making. San Francisco: W.H. Freeman and Co., 1980. Stanley, Mann and Doig. Men Whg_Govern. washington,D.C.: Brookings Institution, 1967. Stigler, George. “The Theory of Economic Regulation.“ Bell Journal of Economics and Manegement Scienee, 2(1) (Spring,197l), pp. 3-21. .v 243 Stigler, George and Friedland, Claire. “What Can Regulators Regulate?: The Case of Electricity.“ Journal of Law and Economics, 5 (October,1962), pp. 1-16. Stone, Alan. Regulation and its Alternatives. Washington, D.C.: Congressional Quarterly, 1982. Troy, Leo. Distribution of Union Membership Among the Stetes, National Bureau of Economic Research, Occassional Paper 497. New York: NBER, 1957. U.S., Congress, Senate, Committee on Commerce. Appointments to the Regulatory Agencies, 94th Congress, 2nd Session, Washington, D.C.: Government Printing Office, 1976. U.S., Congress, Senate. Congressional Directory. Washington, D.C.: Government Printing Office, 1974. U.S., Congress, Senate, Committee on Government Operations. Study on Federal Regulation. Vol. I. The Regulatory Appointments Process. Washington, D.C.: Government Printing Office, 1977. U.S., Congress, Senate, Committee on Government Operations. Study on Federal Regulation. Vol. II. Congressional Oversight of Regulatory Agencies. Washington, D.C.: Government Printing Office, 1977. U.S., Congress, Senate, Committee on Government Operations. Study on Federal Regulation. VOl. III. Public Participatibn in the Regulatory Process. Washington, D.C.: Government Printing Office, 1977. U.S., Congress, Senate, Committee on Government Operations. Study on Federal Regulation. Vol. IV. Undue Delay in Regulatory Administration. Washington, D.C.: Government Printing Office, 1977. U.S., Congress, Senate, Committee on Government Operations. Study on Federal Regulation. VOl. V. Regulatory Organizatigg and Coordination. washington, D.C.: Government Printing OffiEe, 1977. U.S., Congress, Senate, Committee on Government Operations. Study on Federal Regulation. Vol. VI. Framework for Regulationlyand Case Studies in Fedeeral Regulation. Washington, D.C.: Government Printing Office, 1977. 244 U.S., Department of Labor, Bureau of Labor Statistics. Statistical Abstract of the United States. Washington, D. C.: Government Printing Office, 1935-1982. U.S., Department of Labor, Bureau of Labor Statistics. B S Handbook of Methods. Washington, D.C.: Government Printing Office, 1976. U.S., Department of Labor, Bureau of Labor Statistics. Handbook of Labor Statistics 1981. Washington, D.C.: Government Printing Office, 1981. U.S., National Labor Relations Board. Annual Reports of the National Labor Relations Board. Washington, D.C.: Government Printing Office, 1935-1982. U.S., NLRB. A Guide to Basic Law and Procedures under the National Labor Relations Act. Washington, D.C.: Government Printing Ofcice, 1981. U.S., NLRB. The Decisions and Orders of the National Labor Relations Board. Washington, D.C.: Government Printing Office, 1950-1981. U.S., Office of Management and Budget. Budget of the United States, 1952-1982 (Washington, D.C.: Government Printing Office, 1952-1982. Warwick, Donald. A Theory of Public Bureaucracy. Cambridge, Mass.: HarvafaiUniversity Press, 1975. Weingast, Barry. “Regulation, Reregulation, and Deregulation: the Political Foundations of Agency Clientele Relationships.“ St. Louis: Center for the Study of American Business, 1980. Welborn, David. “Regulatory Commissioners and Regulatory Policy.“ Journal of Public Law, 15(1) (1966), pp. 3-39. Wiedenbaum and DeFina. The Costs of Federal Regulation of Economic Activity. Washington, D.C.: American Enterprise Institute, 1978. Wildavsky, Aaron. The Politics of the Budgetary_Process. Boston: Little Brown, 1979. Wilson, James 0., ed. The Politics of Reguletion. New York: Basic Books, 1980. Wilson, James Q. Political Organizatione, New York: Basic Books, 1971. Wilson, James Q. “The Dead Hand of Regulation.“ The Public Interest, 7(25) (Fall 1971), pp. 39-58. 2)...) I " .C. J 245 Wilson, James Q. “The Politics of Regulation.“ Social Responsibility and the Business Predicament. Edited by McKie, James. Washington, D.C.: Brookings Insitute, 1974. Wilson, Woodrow. “The Study of Administration.“ Classics of Pgblic Administration. Edited by Shafritz and Hyde. Oak Park, 111.: Moore Publishing Co., 1978. -n.~'~kfl