INFORMATION TO USERS The most advanced technology has been used to photo­ graph and reproduce this manuscript from the microfilm master. UMI films the original text directly from the copy submitted. Thus, some dissertation copies are in typewriter face, while others may be from a computer printer. In the unlikely event th at the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyrighted material had to be removed, a note will indicate the deletion. Oversize m aterials (e.g., maps, drawings, charts) are re­ produced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each oversize page is available as one exposure on a standard 35 mm slide or as a 17" x 23" black and white photographic print for an additional charge. Photographs included in the original manuscript have been reproduced xerographically in this copy. 35 mm slides or 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A ccessing th e World’s Information since 1938 300 North Z eeb Road, Ann Arbor, Ml 48106-1346 USA Order Number 8807126 M u n icip al charter change in M ichigan: T esting altern ative ex p lan ation s Sych, Lawrence, Ph.D. Michigan State University, 1987 UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106 PLEASE NOTE: In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark V . 1. Glossy photographs or p ag es_____ 2. Colored illustrations, paper or print______ 3. Photographs with dark background____ 4. Illustrations are poor copy______ 5. Pages with black marks, not original copy______ 6. Print shows through as there is text on both sides of p a g e _______ 7. Indistinct, broken or small print on several pages 8. Print exceeds margin requirements_____ 9. Tightly bound copy with print lost in spine_______ / 10. Computer printout pages with indistinct print______ 11. Page(s)___________ lacking when material received, and not available from school or author. 12. Page(s)___________ seem to be missing in numbering only as text follows. 13. Two pages num bered 14. Curling and wrinkled pages______ 15. Dissertation contains pages with print at a slant, filmed a s received_________ 16. Other______________________________________________________________________ . Text follows. UMI MUNICIPAL CHARTER CHANGE IN MICHIGAN: TESTING ALTERNATIVE EXPLANATIONS By Lawrence Sych A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 1987 MUNICIPAL CHARTER CHANGE IN MICHIGAN: TESTING ALTERNATIVE EXPLANATIONS By Lawrence Sych AN ABSTRACT OF A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 1987 ABSTRACT MUNICIPAL CHARTER CHANGE IN MICHIGAN: TESTING ALTERNATIVE EXPLANATIONS By Lawrence Sych This dissertation addresses development of local institutional structure as an outcome of political competition. Rules and structure have played various roles in the study of local politics and decision­ making. After the Progressive reform period, scientific study of institutions largely neglected formal structure as a theoretical component of municipal development. Positive theorists have revived interest in institutions. Recognized for their influence on organizational behaviors, institutions again became theoretically important. Several explanations have been offered to explain institutional development and their changes. Municipal charters define local institutional rules and structure. This dissertation develops a model with three sets of variables to explain charter amendment between 1950 and 1980 in sixty Michigan Home Rule cities. The first set specifies charter change as a function of citizen preferences. function of political cleavages. A second set specifies change as a A third set specifies incremental charter change based on its prior status. The model uses the National Municipal League's Model Charter as a basis for comparison. The model is derived and estimated with probit techniques and the variable sets are tested singularly and competitively. The major findings reveal an overall trend of charter reform toward the League's Model. The charter's prior status in relation to the Model strongly influences the likelihood of charter change in the Model's direction. This is conditioned by strong influences exerted by political factors. Charter change must appear bi-partisan to be successful. The preference measures have a marginal effect on the probability of charter amendment. Change in population preferences, as measured by denographic variables, affects the degree of charter amendment. TABLE OF CONTENTS LIST OF TABLES .......................................... vi LIST OF FIGURES I. II. ..................................... vii INTRODUCTION............................................ 1 Background and Rationale ................................ Statement of Purpose .................................... Outline of Chapters .... 2 4 6 REVIEW OF THE LITERATURE .......................... 8 Conceptions of Charters ................... The Municipal Charter as Compact ......................... The Municipal Charter as Contract .......... Limits on Municipal Charters ................ Types of State Charters..................... Home Rule Charters ............... Characterizations of Municipal Structure .................. The Community Power Schools ......................... The "Reformism" School .................................. Diffusion of Reformed Charters ............ The Urban Policy School ................................. Reformed Structure Effects on Policy..................... Structure Effects on Mobility............................ The Policy Effects of Structural Components ............... Conclusions ............................................. Notes .............................................. 10 10 11 12 15 17 19 20 24 33 35 35 39 42 46 48 III. THEORETICAL ORIENTATION ................................. 50 Definition of Municipal Charter.......................... 50 An Element of Bargaining ................................ 54 Limitations on Bargaining ................................ 55 Charter Characterization Along a Single Dimension......... 57 Early Reform Strategies ......................... 57 The National Municipal League and Structural Reform....... 58 The Ideal Charter ........ 61 Three Pillars of the Ideal Charter ........................ 62 The Charter's Relationship to Preferences .................. 68 Traditions of Charter Change ............................. 72 The Rational-Economic Tradition................... 72 iv IV. V. VI. The Political Cleavage Tradition.............. Charter Status Influence ................................. The Structural Model .................................... Theoretical Propositions .............................. First Set of Inferences — Conditions for Change ....... Second Set of Inferences — Directions of Change....... Notes .................................................. 74 77 79 81 81 84 86 THE EMPIRICAL TEST ...................................... 88 The Dependent Variable .................................. The Independent Variables .......................... The Data Set ........................................... The Statistical Model .................................. A Set of Methods ....................................... ..................................... Model T e s t s Notes ............................ 88 97 109 113 116 117 120 EMPIRICAL RESULTS ...................................... 121 Empirical Results for the Model ......................... Goodness of Fit for the Model .......................... Testing the Theoretical Propositions ................. A Parsimonious Model ................................... Comparing Goodness of F i t .......... Municipal Cases as Illustrations of Change ........ Notes ................................................. 122 126 130 137 140 142 148 CONCLUSION ............................................. 149 Summary of Model's Substantive Performance ............... General Conclusions .................................... Major Conclusions .............. Future Research ....................................... 149 157 160 163 APPENDIX A Charter Characteristics ..................... 167 APPENDIX B Characteristic Coding ....................... 169 APPENDIX C Characteristic Ranking...................... 176 BIBLIOGRAPHY........................................... v 178 LIST OF TABLES Table Table Table Table 1 2 3 4 List of Cities ........ Model Estimation Results ........................... Final Version EstimationResults ..................... Hypothetical Municipal Cases ....... vi 112 123 139 144 LIST OF FIGURES Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 1 2 3 4 5 6 7 8 9 10 11 The Federal Plan of Municipal Government ............ 66 Structural Model .................................... 80 Summary of Charter Characteristic Ranking ......... 92 Distribution of Charter Base Scores .................. 94 Distribution of Charter Change Scores ................ 95 Statistical Model .................................. 115 Model Summary Measures ............................. 129 Final Version Summary Measures...................... 141 National Municipal League's Model Charter (6th ed.) ... 167 Charter Characteristics and Coding .................. 169 Characteristic Ranking and Weights ................. 176 vii CHAPTER I INTRODUCTION The study of political institutions has been a central focus in political science since the founding of the discipline. In the late 1800's and early 1900’s, the Progressive Reformers were developing, experimenting with and creating new rules and structures for local governance. Indeed, not only at the local but also at the state and national levels of government, the Progressive period represented the greatest institution building period in American history. These newly created urban political structures were studied by political scientists prior to 1940. But in the post-war period, the scientific study of politics entered a behavioral period. Individuals rather them institutions became the focus of political inquiry and analysis. The study of local politics moved away from an exposition on rules and structure to the analysis of actors and context. Such a change in the units of emalysis also affected the kinds of scientific studies undertaken. Structural effects were at best considered given and not up for examination or at worst neglected. In recent years, a new examination of institutions -rules and structure- 1ms emerged in political science. This "neo­ institutionalism" harks back to the earlier study of local institutions but with a new awareness based on behavioralist and empiricist legacies. With this kind of scientific heritage, theoretical and empirical researchers are re-evaluating the earlier work and products of the Progressive reformers. This new effort respects rules as important elements in determining political outcomes. No longer are individuals and contextual factors the only- important elements used to understand politics. third element in study of local politics. Rules comprise a This dissertation is a product of a new "institutionalism" in political science. It seeks to examine one rule component in local politics, the municipal charter. Background and Rationale One area of local politics undergoing reexamination is development and change in government structures. Structural rules are considered to be important in helping determine local political outcomes. They outline and describe the operation and functions of governmental organizations. effects. Yet these rules are not neutral in their Schattschneider, in his "realist's" view of democracy, writes, "all forms of political organization have a bias in favor of the exploitation of some kinds of conflict and the suppression of others because organization is the mobilization of bias. Some issues are organized into politics while others are organized out"(1975, p.69). Local political structures are thought by some scholars to be a product of cleavages within the constituent populations. to advantage one side while disadvantaging another. They tend Thus change in political structures becomes vital in directing change in other structural-dependent outcomes. Other scholars follow a Progressive ideal and argue that such bias is not present. They say one may not necessarily impute bias in municipal reform from the composition of its proponents. If there was a bias, it was against the corruption and political machine dominating many localities in the nineteenth century. The municipal charter is the instrument that defines structural rules. This document is often likened to a local "constitution" operating in much the same fashion as the U.S. Constitution adopted two hundred years ago. It spells out the officers, offices, and duties imposed on local government along with its attendent powers and obligations. Since 1899, the National Municipal League has promoted its idea of the Ideal-type charter. This is contained in its publication, the Model Charter (1964), a document currently being updated and revised by League members. A great deal of literature and study has been generated to explain the initial founding of city institutions like the charter. These have usually taken one of two tacks in recent years. One stream of examination has studied the political culture, the actors and the context to determine rules as a product particular to the city at the time of adoption. It holds that the charters cannot be extracted from the context and times at which they were created. institution, it is in part a product of its times. Like any These scientists have looked to the backgrounds and character of the reformers to gain some insight into their purpose in instituting reform (Judd, 1984; Hofstadter, 1956; Hayes, 1964) They see a class bias and suggest that the structures were created according to class beliefs in democracy and American life. Such institutions were largely a reaction to societal changes of industrialization, immigration and urbanization. Another stream of study was a political science adaptation of several conceptual and theoretical developments in the economics discipline. These scholars examined local structural rules as a product of forces competing over public goods. Individuals are considered rational decision makers seeking intrinsic and instrumental goals. Their pursuit of instrumental goals leads to conditions where desired goods are not provided. Provision of these collective goods are the responsibility of governments created principally for that task. The design of such a government reflects the summation of the preferences of individuals participating in the adoption of municipal charters. The structural design would follow their set of preferences for a government. A majority vote decision rule allows the structure to reflect the median voter’s preferences. goals are dictated by preferences. Rules as instrumental Most scholars of this school believe that these preferences originate in individual class, economic and social backgrounds. According to this line of thinking, the Progressive reformers were rational actors designing a government to provide collective goods in an efficient manner. Statement of Purpose Each of these general schools of thought have developed over the last several decades. They try to reexamine and explain the development of local political institutions. However, for the most part many of these local institutions have been founded decades ago. The suburbanization of American populations has generated new municipalities with entirely new municipal charters. A focus on incorporation holds immediate utility for these new developing cities. Yet when comparing the numbers of these new cities to those already established, one finds relatively fewer local entities to comprise a sample. ago. Clearly most cities have now incorporated several decades The study of incorporation efforts would not have immediate benefit for these existing units of government. What has not been as fully examined yet, and which is closely related to the incorporation of rules, is structural change to existing rules. Just as the U.S. Constitution has changed with the times over two hundred years, so have municipal charters changed. We can extend the focus on original incorporation of municipal charters to change in existing charters. To assess change, we have to reference it to something other than itself. The municipal charter changes are therefore measured in relation to the national standard, the Model Charter. This model represents an ideal charter, one each municipality is encouraged and exhorted to comply with. This dissertation is an attempt to develop an explanation of municipal charter change. It reviews the major published works relating in some way to explanations of municipal incorporation, or to structural rules or their implications. It distinguishes three traditions that offer alternative explanations of charter change. These traditions, however, are not necessarily mutually exclusive. From each tradition, we develop a model of change and deduce various propositions. Using data from sixty Michigan municipalities, we estimate our model and thereby test our hypotheses. Given these results, we are then able to develop an explanation of change that modifies the conditions for change imposed by each tradition. Outline of Chapters Chapter Two, Review of the Literature, presents a summary of the major works pertaining to local governmental structure, its development, and its effects on policies. Several different bodies of literature are discussed, the common thread of which is their relationship to change in local political structures. Chapter Three, Theoretical Orientation, provides a specification of a model of charter change. It tries to incorporate into a single model the elements of promoting change that were outlined in the literature review. From this model, several theoretical propositions are developed and presented. Chapter Four, The Elnpirical Test, presents and discusses the methodology and the data measures and sources used to estimate and test our theoretical model. It describes the research design, its data needs, and variable operationalizations as well as the statistical techniques later employed for the test. Chapter Five, The Empirical Results, provides detailed analysis of the statistical results generated from the model estimation. It then presents conceptual analysis of results in theoretical terms. Based on the original model specifications, a second hybrid model is developed in both theoretical and empirical terms. This model is presented, estimated and its results evaluated in theoretical terms. This chapter also provides illustrations of charter change and an explanation of conditions promoting them. Chapter Six, Conclusion, seeks to develop a path for further analysis of charter change and, more importantly, its implications for local democracy and consequences for local governmental action. It presents a refined analysis of the problems confronting researchers and suggests future methods of overcoming than. It also provides directions for future research and scientific inquiry. CHAPTER II REVIEW OF THE LITERATURE Contemporary political research has often neglected institutional roles in politics, but recent developments in political science reveal scholarly attention refocused on institutions. Institutions are beginning to reemerge as central elements in theories of politics. March and Olsen (1984) discuss evidence and trends in this new "institutionalism." They highlight an absence of a guiding, coherent theory and seek to stimulate more rigorous theoretical development. March and Olsen remark that "institutional thinking emphasizes the part played by institutional structures in imposing elements of order on a potentially inchoate world"(p.743). They review several structures deserving theoretical attention including historical, temporal, normative, demographic and symbolic orders. Each order relates organizational elements of institutional structure to decision making. The authors call for "simple models of institutional expectation, choices, and post decision assessments [to] clarify the occasions for expecting positive or negative surprises from deliberate action"(p.747). These models by necessity would reflect contextual matters. At their core, such theories would specify first, a context, and second, conditions for making decisions. The positive theorists have also expanded their interests to encompass institutions (Ostrom, 1986; Miller, 1985; Riker, 1980). Like other political scientists, these theorists focus on decision making. Riker defines institutions as "rules about behavior, especially about making decisions"(p.432). even further. Ostrom narrows the focus She notes multiple uses of the term "institution" and tries to clearly define one major part of it: rules. Ostrom defines rules as: potentially linguistic entities that refer to prescriptions commonly known and used by a set of participants to order repetitive, interdependent relationships. Prescriptions refer to which actions are required, prohibited, or permitted. Rules are the result of implicit or explicit efforts by a set of individuals to achieve order and predictability within defined situations by: (1) creating positions; (2) stating how participants enter or leave positions; (3) stating which actions participants in these positions are required, permitted, or forbidden to take; and (4) stating which outcome participants are required, permitted, or forbidden to affect (p.5). Rules provide behavioral incentive for most conceivable situations. They operate within and on the structure of a situation. While they are usually conceptually separable in most studies, rules in practice combine in configurative and interactive manners. Therefore, Ostrom argues that completely or nearly completely delineated subsets of rules are better units for examination them rules individually. Using Ostrom’s definition as a base, a municipality’s "political structure" can likewise be defined by sets of rules. Rules create positions, state how they are filled, denote actions required of municipal participants, and provide outcomes given alternative actions. Rule change, Ostrom claims, is one important defining characteristic of rules(p.6).[1] This dissertation will examine charters as collections of rules defining municipalities as both legal, corporate entities and 10 "political structures" as contexts for rule-based political conflict. By these definitions, charters embody a major component of the larger municipal institution. Because we focus on rules, other conceptual links to "institutions" will not be examined here. analyzes municipal charters as rule-defined This dissertation political structures. Conceptions of Charters There are two basic conceptions of charters found in the literature. The first is one of charters as compacts. The second is a classical economic view of charters as contracts. Each conception is developed in turn. The Municipal Charter as Compact The first conception holds charters to be compacts through which a higher sovereign authority grants specific rights and powers to an individual or group. In the United States, the concept of a charter came to refer to local constitutional compacts drawn up by individual States (White, 1973). These compacts were grants of State governmental powers and authority to municipalities. they defined a city’s formal political structure. Called charters, Adrian and Press (1977) define a charter as "the fundamental law of a corporation which establishes: (1) the structure or form of government; (2) the powers that may be exercised by it; and (3) the general manner in which the powers granted may be exercised. The charter is almost never a single document, but includes all state laws and judicial opinions that affect the structure, powers, or manner of exercising the powers of the corporation" (p.129). In this way, all municipal powers are 11 implicitly expressed in a charter. A charter legally defines a municipal corporation as an artificial person like any other private corporation. It establishes sets of procedural and substantive rules governing behaviors of the municipality as an artificial person, its public officials and constituent citizens. Taken together, these rules construct "action situations" as contexts for decision making (Ostrom, p. 17). Charters form one part of a two part model of local government action. define a structured action situation. First, rules help Second, this structure together with a decision maker creates outcomes. The municipal charter, by defining sets of rules, helps create structure for participants’ decision making. In this way the compact view fits into institutional tradition by helping structure state-granted powers of local government decision making. The Municipal Charter as Contract An alternative and not entirely incompatible view sees the charter in terms of a more traditional economic contract based on propertyrights (Miller, 1981; Maser, 1985; Heckathom and Maser, 1987). Maser considers a municipal charter to be: a contract is an agreement among people stipulating, first, actions by each to be carried out at some time in the future, and second, rewards and penalties to be meted out following compliance. A constitutional contract stipulates a mechanism for making political choices about who gets what, when, where, and how in the face of scarcity. Presumably, each citizen can express a preference about these outcomes. The mechanism described in the contract, the government, translates individual preferences into social choices and enforces them. Thus, a constitutional contract describes the relationship between the citizen and their government (p.129). 12 The charter contract is assumed to be voluntarily joined by both citizen and municipality. charter contract. In effect, each party "invests" in the As with other investment contracts, the charter may be bargained and arranged to satisfy both parties to the agreement. Constitutional "contracts ratify bargains that define duties and rights because people believe that there is some gain, to be had by contracting"(Heckathom and Maser, 1987,p.143). To introduce a concept of constitutional bargaining, Heckathom and Maser describe "limited" as opposed to purely classical contracts with implied absolute rights and powers. They recognize the classical contract view of municipal charters, if pursued to its logical theoretical implications, becomes untenable when contrasted with political and legal realities. They place four limits on actual contracts: (1) limited subjection to disorder and chaos with consequent limited costs; (2) a limited domain of issues that are problematic; (3) limited knowledge by contract participants; and (4) participants exercise limited autonomy in making contracts. This last feature means citizens are restricted by "the existing pattern of rights and other mutual understanding"(p.147). These limitations restructure theoretical explanations of constitutional contracts to bring their theory into greater conformity with an applied or "bounded" rationality. Limits on Municipal Charters There are two other important limitations on the view of charters as contracts. system. They relate to the city’s role in the American federal 13 The first restriction is that, in a pure legal sense, the charter is not a contract like those involving private corporations. People create private corporations on a voluntary basis and know the conditions under which they are formed and will operate. Their contract is binding and the State cannot alter it unless it contains provisions for such change. Municipal corporations, on the other hand, can be created and changed at the will of State legislatures. As Adrian and Press note, charter terms "may be quite different from what the people of the community desire, and even more important, the charter is not a contract and is hence subject to constant, involuntary, and sonetime arbitrary change. It can even be taken away without advance notice, unless the state's constitution specifically prohibits this"(p.130).[2] A charter is still a compact granting powers and not a contractual agreement transferring powers. This restriction has been upheld by federal and state courts. In 1819, the United States Supreme Court in Dartmouth College v. Woodward held cities to be "creatures of the state." State courts were left unencumbered by the U.S. Constitution in their dealings with municipalities as "children of the state"(Adrian and Press, 1977,p.129). In an important case involving an Iowan city and railroad, Iowa magistrate John F. Dillon explicitly stated a doctrine of state control. In 1868, Judge Dillon wrote: Municipal corporations owe their origin to, and derive their powers and rights wholey from, the legislature. It breathes into them the breath of life without which they cannot exist. As it creates so it may destroy. If it may destroy, it may abridge and control. Unless there is some constitutional limitation of the right, the legislature might, by a single act, 14 if we can suppose it capable of so great a folly and so great a wrong, sweep from existence all of the municipal corporations of the state, and the corporations could not prevent it. We know of no limitation of the right so far as the corporations themselves are concerned. They are, so to phrase it, the mere tenants at will of the legislature.(Judd, p.41) This doctrine later became known as "Dillon’s Rule." The second restriction applies to practical charter applications. State courts have had to make decisions about acceptable municipal powers. The "contract" has been narrowly defined. Again, Judge Dillon outlined the court’s rule on interpretation of the express powers contained in charters. He wrote: It is a general and undisputed proposition of law that a municipal corporation possesses and can exercise the following powers and no others: First, those granted in express words; second, those necessarly or fairly implied in or incident to the powers expressly granted; third, those essential to the accomplishment of the declared objects and purposes of the corporation - not simply convenient, but indispensable. Any fair, reasonable, subtantial doubt concerning the existence of powers is resolved by the courts against the corporation, and the power is denied.(Adrian and Press, p.132) State control and narrow interpretation effectively restrict municipal actions. The federalist principle insuring governmental independence does not exist within the unitary form of state government. Conceptions of charters as contracts between states and citizens fails to fully account for dominant state influence, regulation and interference in municipal affairs. Peterson (1981) effectively describes the limited nature of city government in policy making. Lack of sovereign power makes traditional comparison of cities to states and national governments inaccurate and misleading. City politics is limited politics. Of the three part capitalist production function of labor, land, and capital, cities may effectively control only land. This has policy 15 implications. Peterson argues cities can not engage in redistributional policies without threatening their survival. Land- based developmental policies are the only sensible ones a city can make to sustain itself in the American federal system. [33 By describing a charter as a contract, one implies it contains powers similar to those possessed by the state or national government. As Peterson notes, this is clearly not the case. Adrian and Press indicate municipal charters do not permit "contractual" obligations or rights as would a private corporation. Finally, Dillon’s Rule leads courts to expressly and narrowly prohibit a range of municipal functions. Both compact and contract conceptualizations do, however, indicate formally organized powers and defined application toward resolving problems and providing services at the municipal level. Charter limitations may be more clearly understood by the manner in which states dispense them. We can see from a description of this process that with the advent of Home Rule, some element of "bargaining" may exist, although with narrow restrictions and prohibitions. Types of State Charters While Judge Dillon wrote that states may impose charters on citizens, they very rarely do so in the United States. Such powers of creation and dissolution are rarely utilized when states grant municipal charters. Rather than imposing them unilaterally, state governments grant them on petition of its citizens. This process of granting charters has changed over time, and today most states rely on Home Rule (Judd, p.103). 16 Prior to the advent of Home Rule provisions in state constitutions, legislatures granted "special" and "general" charters to their cities. Special charters are individually designed for each city and no two are alike. from the legislature. Any changes to them must likewise crane A city receives what it desires in the form of a tailored charter, but only if the legislature is "amenable to their wishes"(Gosnell and Holland, 1951, p.459). The intent is to have a city’s charter individually suited to its needs. Serious problems accompany this manner of granting charters. First, a state legislature may not be "amenable" to a city’s (or cities') wishes. The legislatures were, prior to the 1960’s, severely malapportioned and dominated by rural district legislators who knew and cared little about the city and its attendant problems (Babcock, 1957, p.123). Second, citizens or elected officials cannot fashion their charters but must rely on the legislature which reduces any kind of implied "bargaining" among themselves. [4] After rapid urbanization and growth in the number and size of cities in the late nineteenth century, state legislatures developed "general" charters (Gosnell and Holland, p. 459). This plan provided for one charter applicable to all cities within a state. small cities were treated alike. Large and This form was developed "to counteract the abuses that have arisen under the special charter system, especially favoritism and lobbying connected with the amendment procedure"(Babcock, p. 123). and some simplicity into city charters. This plan introduces uniformity Municipalities’ ordinance power is enlarged to help cope with some differences in local conditions. However, general act charters tend to restrict either the 17 small or large cities, depending on the act. A compromise between the special and general charter plan is the classification system treating cities on the basis of population. Many state legislatures could abuse this plan and by introducing what amounts to special legislation. Further, the amendment process still relies on state legislative efforts and withholds popular appeal to change without legislative involvement. Under all these plans, the state legislature granted and amended municipal charters often without petition of and approval by city constituencies. Home Rule Charters The municipal "Home Rule" movement began in Iowa in the 1850’s and slowly spread across the United States. It was in large measure a reaction against the legislative abuses of city charters, followed a populist ideal of self-government at the local level, and indicated the new emerging role of cities in the American federal system. Three types of home rule charter provisions operate at the state level. The first is a semi-home rule "optional" system whereby the legislature provides a selection of charter plans. Cities may select charter features in a "cafeteria" style of choice (Adrian and Press, 1977). This plan provides an element of choice without altogether destroying the uniformity found under a general charter plan. It also inhibits the state from making special laws for particular cities. This type of plan meets arguments against home rule that say the city gets too much latitude and "leads to the creation of multiple legislatures"(Gosnell and Holland, 1951). compromise with Hone Rule proponents. It represents sane 18 The second type of home rule can be introduced by the state legislature itself. It can voluntarily renounce its powers to grant charters and, via legislation, give cities the opportunity to create and amend their charters. This is usually subject only to certain broad statutory guidelines (Babcock, p. 124). State legislatures can, however, repudiate earlier commitments to Home Rule and return to other plans. Constitutional home rule gives cities state constitutional guarantees to form, adopt and amend their municipal charters. Michigan, for example, adopted a Home Rule provision in its 1908 Constitution. New sections 20 and 21 of Article 8 required the legislature to "provide by a general law for the incorporation of cities," and provided that "under such general laws, the electorate of each city ...shall have power and authority to frame, adopt and amend its charter, and through its regularly constituted authority, to pass all laws and ordinances relating to its municipal concerns, subject to the constitution and general laws of this state”(Michigan, 1908). Constitutional provisions, while guaranteeing municipal home rule, nevertheless in most cases continue to give the legislature power to moderate such processes. The Home Rule plan has many advantages over other charter plans. Legislatures no longer have to consume time to create and amend charters. needs. Cities may adopt forms of government best suited to their It also makes citizens directly responsible for municipal structure and its effectiveness. This psychological value is perhaps the greatest benefit of constitutional home rule, providing ''encouragement and incentive" to local government activists (Adrian 19 and Press, p.141). This "attitude" allows change and amendment in cases of charter dissatisfaction. Constitutional home rule more closely approximates Heckathorn and Maser’s concept of charters as mini-constitutions. Bargaining for incorporation and amendment is undertaken by concerned, active citizens. They do not rely on pressing their demands on an indifferent legislature. The "bargaining process" culminates in a popular referendum on charter adoption or change. Characterizations of Municipal Structure Focus on charters and change leads to two basic concerns. what is the nature of city charter change? First, In other words, what in the change’s character makes it desirable to be implemented in the existing charter? To help answer this question, one must be capable of characterizing municipal charters and changes made to them. The municipal government’s institutions must be given sane conceptual and empirical character. Second, why do charters change? If amendments require popular votes, then one might assume city constituents support or oppose change on the basis of sane preferred charter characterization. If charters can be characterized, then change in the charter may also be measured and assessed. Characterizations of structure in the study of urban politics have varied over time. Three general characterizations will be reviewed here covering several schools of thought. the Coranunity Power schools. The first one pertains to Second, the Reform school is reviewed. Lastly, the Urban Policy school is reviewed. Each review will focus on the school’s characterization of municipal political structure and 20 its change. The Community Power Schools One of the greatest debates in political science took place between two well-defined schools of thought over the nature and distribution of power in communities. These community power schools had opposing positions on government institutions generally and municipal structure specifically. Elite theorists downplayed the importance of formal structure. Following the behavioral creed, people became the principal units of political analysis. Formalized structures, as represented by sets of rules, were without significant meaning. Elites by definition possessed vast power in a community and could overcome rule impediments to direct and control political, social and economic institutions. C. Wright Mills (1956) succinctly expressed the evolution and standing of a tripartite power elite at a national level. Whereas other institutions like the family, church, and school "adapt" to modem life, Mills said the large, centralized institutions governments, armies, and corporations - "shape" modem life(p.6). He wrote "within each of the big three, the typical institutional unit has become enlarged, has become administrative, and, in the power of its decisions, has become centralized"(p.7). and centralized, so have they interlocked. circles of elites have arisen. As they have enlarged Within each domain, higher These circles form "the power elite." By Mills' definition, these elites "are simply those who have the most of what there is to have, which is generally held to include 21 money, power, and prestige"(p.9). They by necessity come to and retain power by their institutional positions. Institutions, Mills notes, "are the necessary bases of power, of wealth, and of prestige, and at the same time, the chief means of exercising power, of acquiring and retaining wealth, and of cashing in the higher claims for prestige"(p.9). Institutional command is a source and the end of the powerful elite. This is not, however, the sole source of power. Mills remarked "not all power, it is true, is anchored in and exercised by means of such institutions, but only within and through these can power be more or less continuous and important"(p.9). Institutions, according to Mills, represent a structure of powers, changes to which come by way of "institutional shift in the relative position of the political, the economic, and the military orders"(p.269). At a local level, the military order is non-existent, the political order is minimized, and the economic order is predominant in decision-making. The first studies of local elites were conducted by the Lynds when they looked at decision-making in "Middletown" (Muncie, Indiana). They found an economic elite of one family dominating city affairs. Floyd Hunter was the first to introduce more objective analysis into the study of local elites in Community Power Structure, published in 1953. A sociologist, Hunter turned from a formal institutional perspective of local government to examine who held power and made decisions in whose interests. He found a small, cohesive elite of business leaders as the basis for local power, capable of overcoming resistance from a political structure. This governing business elite could overcome adverse rules and control institutions. Political 22 opposition was ineffective. For example, Hunter discussed citizen and council participation in "Regional City" [Atlanta, GA] and claimed "the political organizations are ... so completely dominated by the power interests which have been identified that there is little hope of adequate expression being fostered by them at this time" (p.231). Power was described and defined by "social processes". As Hunter defined it, "power" described "the acts of men going about the business of moving other men to act in relation to themselves or in relation to organic or inorganic things"(p.2). These scholars felt that social forces were much more influential than political structures in decision making. This was especially so at local levels where structure is less "distinct." Some researchers found private, not public, power structures of social and economic elites dominating policy and directing it toward their own ends. Pluralists confronted the elite argument on many points. saw structure creating and aggregrating sources of power. They In his classic work Who Governs?. Dahl (1961) defined a "political system" composed of rules, procedures and essential characteristics (p. 315). Group consensus maintained its stability. These structurally-induced power bases could be captured and used by various groups to make important decisions in functional areas like schools and urban renewal. The pluralists described a variety of power resources individuals could use to influence policy or otherwise exercise power. In American society, such resources include votes, labor and time as well as money. Most citizens possess at least one of these resources. The pluralists argued that for the most part these political resources are 23 latent: in time of necessity, they can be utilized. As Polsby noted, "If a man’s major life work is banking, the pluralist presumes he will spend his time at the bank, and not in manipulating community decisions" (1963, p. 117). The banker’s decisions, while they might have substantial effects on local policy in terms of investment, risks and ventures, are still critically different in institutional context and in the explicit intentions and goals of the decision maker. Finally, these theorists had a major difference regarding potential and actual uses of power. Mills, Hunter and other elite theorists argued that their commanding circles in government, business and military had reputations for and potential to act in major decisions. Pluralists argued that while potential exists and influence may be great, the principal and actual uses of power are directed toward other means. The debate between the power schools could thus shift toward more empirically demonstrable illustrations of power. Pluralists held formal, rule-based political structures to be important power bases capable of being independent of social and economic power bases. Change in structure illustrates this position. Wolfinger’s (1974) description of New Haven charter change during Mayor Lee’s tenure is a striking example of pluralist relationships among groups, structure, consensus and decision making. The proposed charter revision of 1958 was formulated to create an autonomous Urban Renewal department far from a mayor’s ability to "interfere" in its administration. Interest groups both supported and opposed the proposal. Mayor Lee came out in support of it and lost, demonstrating limitations in mayoral leadership. Pluralist theorists saw structure 24 as a secondary feature creating contexts for important decisions. Decisions on rules involved majority consensus just as would decisions in any other functional area. Bachrach and Baratz criticized both sides of the community power debate (1962). These scholars claimed that "there are two faces of power, neither of which the (elitists) see and only one of which the (pluralists) see"(p.947). They contended that it is impossible to define "key political issues" because some issues are by design organized off the political agenda. Rules and structure can be used to exclude important issues from the political agenda. Their conception of a proper study of community politics would first assess its structured "mobilization of bias".[5] Once this is done, "one could conclude that any challenge to the predominant value or to the established 'rules of the game’ would constitute an ’important’ issue; all else unimportant"(p.950). theories. Structural change was a test of both By their criticism, Bachrach and Baratz turned scholarly attention toward explicit study of organization, rules and change as important elements in their own right. The "Reformism" School The second body of literature can be labeled the "reformism" school. A collection of ideas, theories and literature, it developed from the Progressive movement of the late 19th and early 20th centuries. This body of literature focuses directly on building political structures and institutions embodying Progressive values such as honesty, efficiency and economy. The reformist tradition prescribes various requirements for 25 government and creates institutional structures to achieve them. It has close links to implicitly and explicitly stated values. Its literature mixes values and preferences with institutions and political structures. The Progessive movement has been cast as a coalition of varied and often antagonistic groups that, on the subject of governmental structures, came to agreement (Knott and Miller, 1987). At one time or another, the Progressive coalition included populists, civil service reformers, an emerging middle class, urban merchant class, and social reformers. Their agreement rested on a roughly identical prescription for reform against the "evils" of politics. The coalition wanted to (l)separate politics from administration; and (2) guarantee their gains via structural change for efficiency, honesty and economy (p.38). A host of reformist associations formed to promote structural changes. Among these were the National Short Ballot Association, National Municipal League and the League of Women Voters. The National Municipal League promoted its Model City Charter since 1898 and the Council-Manager plan since 1915.[6] This school considers such a plan to be the "best" municipal form of government for efficiency, economy and executive leadership. It represents efforts to derive "One Best Way" of municipal management (National Municipal League, 1964). The League itself is now more concerned with "modernizing and simplifying" established municipalities and less on cleaning out corruption and machines (1964). More recently, reformist attention focused on altering the structure of the urban metropolis according to Progressive values (Campbell and Bahl, 1976; Zimmerman, 26 1972). The "reform" movement is often cast as a class-based movement (Judd, 1984; Hays, 1984; Hofstadter, 1955). Scholars and historians portray progressives as middle and upper class Anglo-Saxon Protestants seeking to gain control over government. The means to do so were contained in particular institutional rules. Structural reform was an instrument to promote Progressive values and control at the expense of working and lower classes (Hofstadter, 1955). One reformer of the Short Ballot Organization noted that members "have an eye for the effectiveness of their weapons and rather less for tradition or fine logic"(Gilbertson, 1912, p. 218). Other observers writing at the time of the Progressive movement questioned its causes and goals. Hollingsworth (1912) argued that the movement would eventually lead to "despotism" in the long run. In the short run, he found it to be a regressive and distinctly class-based movement aimed "at arbitrary control of other classes." It promoted reforms delegating greater power to single individuals or demagogues. The structures that Progressives hoped to create in effect' reshuffled existing power to a new set of individuals. Hollingsworth wrote, "its program of reforms does not provide for the active exercise of any real governmental power or functions by the electorate as a body, or by a majority of the electorate as a body, but only provides different and more direct means of delegating such functions to individuals"(p .33). Progressives sought to shift concentrations of power from legislatures (and economic and corporate interests) to the executive. This type of reform, according to Hollingsworth, opens the door to 27 demagoguery. A new despotism arises in the form of an executive with absolute power. Therefore, "it is not the head of an oligarchy of alleged special privilege, but the people’s avowed deliverer from such an oligarchy, who, when opportunity is ripe, is most ready and apt to put aside all legal restraints and exercise arbitrary personal rule"(p.43). It appears that the goals of structural reform were disputed even at the time of their invention and initial adoption. Hofstadter (1955) developed a more formal thesis of dual political ethics that greatly influenced theoretical development used to explain municipal structural change. In an often cited passage, he argued that these ethics emerged in the midst of the Progessive movement: ...out of the clash between the needs of the immigrants and the sentiments of the native there emerged two thoroughly different systems of political ethics... One, founded upon the indigenous Yankee-Protestant political traditions, and upon middle-class life, assumed and demanded the constant, disinterested activity of the citizen in public affairs, argued that political life ought to be run, to a greater degree than it was, in accordance with general principles and abstract laws apart from and superior to personal needs, and expressed a common feeling that government should be in good part an effort to moralize the lives of individuals while economic life should be intimately related to the stimulation and development of individual character. The other system, founded upon the European backgrounds of the immigrants, upon their unfamiliarity with independent political action, their familiarity with hierarchy and authority, and upon the urgent needs that so often grew out of their migration, took for granted that the political life of the individual would arise out of family needs, interpreted political and civic relations chiefly in terms of personal obligations, and placed strong personal loyalties above allegiance to abstract codes of law or morals. It was chiefly upon this system of values that the political life of the immigrant, the boss, and the urban machine was based, (p.8-9) This thesis led to further works along similar theoretical lines. The class conflict model, for example, was used to explain structural reform campaigns in St. Louis in the 1950’s by Salisbury (1961). He contended that there are two basic disputes: one over 28 policy as limited objectives and the other over fundamental structures affecting the status quo. Salisbury argued that "disputes over broad reform of the local governmental structure uncover certain durable antagonisms in the community" which inevitably divide along class lines and thus get resolved "according to the remarkably unchanging strength of the two groupings"(p.260). In fact, while most cities have deep cleavages, there are "few issues other than those which evoke the symbols of structural reforms, broadly conceived and widely articulated...sufficiently potent to divide the community in this way"(p.261). The two sides in the St. Louis reform battles were formed according to divisions posited by reform theorists. One group of upper class interests included the major, especially downtown, businessmen, middle and upper class residents and newspapers. Opposing this group stood a body of lower class interests including neighborhood business, blacks, local unions, and lower and lowermiddle class residents. The issue of reform, its symbolism and perceived meanings tend to spark conflicts between these classes which otherwise pursue different interests and political battles. Expressed goals of efficiency and economy, broader taxing authority, and anti-machine government were embodied in the St. Louis reform effort. Reform opponents focused on perceptions of higher taxes and loss of representation in municipal government. Salisbury claimed that substantive issues were put down and "on both sides the concern for concrete objectives was gradually superceded by concern for the slogan of a quasi-class struggle"(p.273). Salisbury also pointed out a difference between early Progressive 29 and modem day reformers. Contemporary reformers "are more self­ consciously concerned with such problems as economic growth and the revival of the core city...and less interested in continuing the social groups which threaten whatever hegemony middle class interests may still have”(p.273). The emphasis on establishing reforms that Hofstadter attributed to the middle class may have diminished in its intensity. In his concluding remarks, Salisbury argued that a quasi­ class conflict between middle and lower class will arise over reform in communities with larger and more diverse population(p. 275). Once activated, class lines will ultimately determine vote totals in referendums on structural change. Banfield and Wilson (1963) also distilled Hofstadter’s work to develop a theory seeking to explain municipal structure. Their political culture or "ethos" theory posited two general sets of cultural values in American cities. value weis synonymous with municipal reformers. An indiginous public-regarding Progressive middle-class values guiding Opposed to this native "ethos" was an ethnic- based culture of "private-regarding" value. It was derived from immigrants flooding American cities beginning in the middle 1800’s. Two different local government structures emerged out of these cultures. The first structure was "Unreformed". It represented the carry-over of a fragmented, English system of local government modified by Jacksonian democracy. It had many elected officers effectively decentralizing administrative power. A "private- regarding" ethos was retained under this structure. The second structure was "reformed". It was the result of the Progressive institution building period of 1880-1920. It is characterized by the 30 National Municipal League's Model Charter. now undergoing its seventh revision. It included a small elected council and appointed executive with centralized administrative powers. This structure would guarantee the dominance of a "public-regarding" ethos. Wolfinger and Field (1966) attempted to operationalize this "ethos" theory and examine its relationship to municipal structure. They correlated socio-economic variables with the following structural features:[7] FEATURE UNREPORMED Form of Government Type of Ballot Constituency Size of District Civil Service Coverage City Plan Expenditures Urban Renewal Mayor Partisan Wards Small Less Low Low REFORMED Manager Non-partisan At-Large Large More High High Wolfinger and Field found little relationship between socio-economic characteristics and structure once controlling for region and so concluded that the "ethos" theory is not empirically valid. Lineberry and Fowler (1967) conducted further tests of the "ethos" theory. Their major research concern was structural effects on city policy-making. They hypothesized that, to the extent that city populations made demands for both structure and policy, policy outputs would be related to types of structure. Following the earlier works of Hofstadter and Banfield and Wilson, Lineberry and Fowler theorized that municipal reformers implemented structural changes to protect against particularistic interests. The reformers’ principal tools were non-partisan elections, the commission and later the manager form of government, and at-large constituencies. The same city populations that wanted certain policies were thought to 31 also bring about certain reforms. thesis: Lineberry and Fowler extend the governments which are products of the Progressive reform movement behave differently from those which have unreformed institutions, even if the socio-economic composition of their populations may be similar (p.701). The middle class "ethos" that produced reforms would also induce different city policies by virtue of structural reform. In testing their thesis, Lineberry and Fowler treat two city policy outputs as dependent variables - taxation and expenditure levels. They relate these outputs to both socio-economic and political structure characteristics. Their first step compared reformed and unreformed cities in terms of their socio-economic positions. They found differences in the composition of cities’ populations; unreformed have larger and stable or declining populations. In examining class, they point out that "what is striking is not the differences between the cities but the similarities of their class composition"(p.704). And except in the racial composition, reformed cities do appear to be more homogeneous. On the basis of these initial results, Lineberry and Fowler criticized Wolfinger and Field on two major points. First, to control for region removes much of the variation in population composition and so unfairly test the ethos theory. Controlling for region "is much more than controlling for historical experiences, because region as a variable is an undifferentiated potpourri of socio-economic, attitudinal, historical, and cultural variations"(p.707). Region is a grossly unrefined variable needing much more definition with demographic data. 32 Second, the authors take Wolfinger and Field to task for mistakenly applying contemporary socio-economic characteristics to historical decisions, and for inferring that middle class attitudes and political power are proportional to their urban populations. This highlights the problem of inferring political activity or its potential for action on the basis of ethnicity or class populations. The tests of their main thesis indicate that outputs, and more importantly, the relative "responsiveness" of cities to their social cleavages differ. The "reformed cities are less responsive to cleavages in their population than unreformed cities"(p.710). They extend this analysis and conclusion to the "ethos" theory: If one of the premises of the ’political ethos’ argument is that reformed institutions give less weight to the 'private regarding’ and ’artificial’ cleavages of the population, that premise receives striking support from our analysis. Our data suggest that when a city adopts reformed structures, it comes to be governed less on the basis of conflict and more on the basis of the rationalistic theory of administration. The making of public policy takes less count of the enduring differences between White and Negro, business and labor, Pole and WASP. The logic of the bureaucratic ethic demands an impersonal, apolitical settlement of issues, rather than the settlement of conflict in the arena of political battle.(p.710) Lineberry and Fowler also stated that the major difference is not so much demography but behavior among cities. free of cleavages. Reformed cities are not They treated "reformism" as a continuous variable by summing the number of Index of Reform features, suggesting "the more reformed the city, the less responsive it is to socio-economic cleavages in its political decision-making"(p.714). Finally, for future work, they note that "it is important in any analysis of reformism to distinguish between the factors which produce the adoption of reformed institutions and the impact of the new political forms once they have been established"(p.716). 33 The Lineberry and Fowler work had important effects on the agenda of urban political studies. directions. Its findings sent scholars in several First, "structure" came to be represented by a three-part Index of Reform. Change in this structure was implicitly related to change in "ethos" and hence related to change in populations. Other researchers followed Banfield and Wilson's lead and sought to predict city structure. Dye and MacManus (1976), vising discriminant function analysis and the three-part Reformism Index, found socio-economic variables do well as predictors within regions but are inconsistent among regions. They also note that "it is unlikely that state-imposed legal requirements also contribute to our (prediction) 'error'"(p.270). Diffusion of Reformed Charters Several scholars argued that municipal adoption of reform structures, particularly commission and managerial forms, was a diffusion of innovative city government ideas. Two traditional generalizations drawn from this literature suggest that cities respond to innovative reform if they are closer to the innovator (a neighborhood effect) or possess a higher ranking on some scaled hierarchy such as size or social status (a hierarchial effect). The rate or pace of diffusion depends on environmental factors such as state law, proximity to the innovator, or social interaction and intergovernmental communications (Scott, 1968; Knoke, 1982). Scott’s model (1968) focused on a four-stage process of diffusion within a relatively stable political system (p.1091). Fran the first Pre-Choice stage where all cities have similar governmental structure, 34 a single state’s set of cities change structures through Innovative and Emulation stages to reach the final Institutionalization stage. At this point, the set of cities accept and reinforce the original innovation thereby going back full circle to the first stage. Scott’s findings suggest that innovation and diffusion of governmental structures is most likely to occur in cities with stable, homogenous populations with "high social rank" as defined by occupation and education. These populations, "whose educational and occupational experiences and exposures have encouraged the values of efficiency, economy, and administrative responsibility are more likely to embrace the manager model for their local political system "(p.1104). Cities without these characteristics are less conditioned for structural change. A similar spatial diffusion model across time, according to Knoke (1982), held up under his analysis much better than either a "culture clash" or "class conflict" model. His diffusion model is one based on considerations of a technical nature. Knoke claims that "community growth and increasing organizational complexity generated social pressures compelling the substitution of ’mo d e m ’ for traditional forms of governmental administration and service delivery" (p. 1321). Once introduced, innovative modem forms spread by "imitation or contagion" according to proximity and ranking. Knoke’s findings differ from Scott’s in that they suggest "the least affluent communities, where working-class interests were presumably stronger, were the most rapid acceptors of both commission and managerial government" (p. 1336). His analysis implies that these technical-based modernizations "spread most rapidly among communities 35 whose limited economic resources probably made such structural innovations attractive as a possible solution to urban ills"(p.1336). Both Scott and Knoke use a dependent variable consisting of the Index of Reform’s administrative component. They neglect to include measures for "political innovation" even though the commission structure, for example, later became unpopular for its unification of policy-making and administrative functions. Reliance on a strictly administrative-efficiency dimension to define propensity to change leaves out other values like representation that may induce innovation, change, and acceptance of other structural forms. The Urban Policy School The Lineberry and Fowler work sent urban policy scholars in two directions. One group studied the effects of entire "reformed" structures on policy provisions. These policy scholars generally found urban service delivery to be unrelated to municipal government structure. The second group closely examined the effects of specific structural features on policy. Reformed Structure Effects On Policy In one of the first studies to use time-series analysis, Morgan and Pelissero (1981) sought to test the hypothesis that structural differences among cities cause differences in both their taxing and spending policies. Their basic question was: "do politics and governmental structure have any independent influence on policies" at the local level (p.999). They operationalized the Index of Reform as an independent rather than a dependent variable and examined changes 36 in policy among a set of cities whose structures had changed. By using a control group of cities whose structures did not change, the researchers hoped to minimize the effects of differing socio-economic characteristics on policy expenditures over time. In doing this, they in effect argue that structural form is a proxy for municipal population characteristics. Political structure thus loses its value as an intervening component between government constituency and policy. Morgan and Pelissero conclude their analysis by noting that "changing structure does not seem to matter much in allocating funds across different functural areas"(p.1005). They claim that their findings are indeed "unambiguous - urban reforms have few policy consequences"(p. 1005). Alternative municipal structures matter little in tax and spending policy. Their model, however, does not include explanatory variables for socio- economic character of municipal population, a vital factor in the theory they seek to test. The one independent variable most frequently found to be significant was time, a counter variable difficult to justify theoretically or conceptually. Their conclusions seem to be largely unsupported by their results. The Mladenka (1980) study of Chicago finds direct political activity to be unrelated to distribuion of service benefits. Mladenka finds little direct connection between, and mixed attitudes toward, activity by ward alderman and subsequent municipal service delivery. Some aldermen "are confident of their ability to manipulate the administrative apparatus to secure superior services. Their power stems from their ability to deliver the vote and they can count on the 37 vote because they provide their constituents with tangible, material rewards"(p.992). Other aldermen "believe that it is the bureaucracy that defines the rules that govern distribution choices. They lack the resources necessary to effectively challenge administrative outcomes. They play little role in sampling and monitoring city policy that is left to bureaucratic technical-rational rather than political considerations"(p .992). His study examined parks, fire, sanitation, and education services and in every policy area finds decisions made more on the basis of bureaucratic technical-rational criteria. Mladenka claims the political structure has minimal effects of service delivery. He notes that "although the determinant of distributional outcomes vary across types of bureaucracies, it is significant that distributional patterns in both reformed and machine cities are little affected by voter choices or elected officials. Organization rules and unanticipated events and develpment provide a better explanation of who gets what than any combination of distinctly political and electoral variables"(p.996). Alternative municipal structures appear to matter little in policy delivery decisions. Boyle and Jacobs (1982) sought to develop a more complete model of how city governments distribute services across different parts of the city in response to a series of studies characterizing an "unpattemed inequality" in service delivery (p.372). They looked for some process of distribution that may or may not be equitable yet is in fact conscious and determinate. To investigate their thesis, Boyle and Jacobs designed a crosssectional analysis using sixty-two community planning districts within 38 New York City and concentrated on the inter-neighborhood allocation of service expenditures. The authors found a patterned service delivery response attributable to the nature of services and bureaucratic response to city conditions. The bureaucracy responds to external pressures and their service provisions are made accordingly. They note that "even though the behavior of these agencies is an unmeasured intervening variable in their investigation! our results do not support the idea that their decisions are unresponsive to changed political and social conditions in at least one urban environment"{p.378). While direct political connections do not seem apparent, more subtle bureaucratic responses to political leadership may occur. Municipal organization may filter political changes according to internalized professionsal codes, values and norms. Other scholars tended to support claims of structural neutrality in policy and decision making (Dye, 1979, Gray, 1979). Stonecash (1979) responded by noting structure's intervening role. Its effects on policy could therefore best be seen over time with proper methods. Lyons (1978) used expenditure analysis to find political structure does matter between "reformed" and "unreformed" cities. Clark’s (1968) analysis of fifty-one American cities finds reformed governments to be strongly related to centralized community decision making. This study operationalizes the traditional Index of Reform and finds it to be an overriding integrative device in otherwise decentralized decision making communities. His basic conclusion states that greater social system differentiation leads to greater political elite differentiation. This in turn stimulates decentralization of decision making which, without an intergrative 39 mechanism to coordinate sectors, leads to lower levels of policy outputs (p. 591). Clark's sunmary relies on characterization of a political structure to overcome sectoral-based decisions. Its effect on policy is moderated to the that extent issues are more established. It also implies that structural effects are somehow independent of some, if not all, of the decision-making bases affecting municipal policies. Structure Effects on Mobility Another body of literature developed around the subject of city policies and resident mobility. Tiebout (1956), in his classic work "A Pure Theory of Local Expenditures," outlines a core problem in public finance theory: "the mechanism by which consumer-voters register their preferences for public goods" (p.417). Mobility offers the consumer-voter an opportunity to reveal his preferences for local government services. This person "may be viewed as picking that community which best satisfies his preference pattern for public goods." Unlike a central government with flexible revenue and expenditure levels and more or less fixed constituent service demands, a local government has the opposite limitations. expediture levels are relatively fixed. Its revenue and The adjustment between revenue/expenditure patterns and constituent demands is settled by mobile consumer-voters who move to and from communities according to their preferences. The burden is thus placed on cities to try and adjust their services and revenue demands to attract and retain desired residents. To accomplish this, cities "are forced to keep production costs at a 40 minimum either through the efficiency of city managers or through competition from other communities" (p.422). Political structure and jurisdictional competition operate in such a way as to help streamline city policy to attain allocation efficiency and attract residents. The proponents of changing structure are likely to respond to residential mobility. Quigley (1980) separates local citizen mobility into two categories: interurban and intraurban. Several basic analytic problems encompass these two forms of mobility and how mobility affects municipal structure as it produces policy. One important problem pertains to temporal ordering of mobilty as a response to some stimuli. How long must we wait to see mobility actually occur given some cause? The second problem is the differences both theoretically and empirically between long and short distance movements. Long distance or interurban mobility is theorized as a response to jobs and changing income, which make movement more an "investment". Income represents an endogenous variable in the calculus of mobility. Short distance or intraurban mobility occurring within cities is determined more on the basis of consumption costs and benefits and less on income. Analysis becomes much more complicated as neighborhood characteristics and aggregrated individual mobility decisions affect determinations of "cause". Quigley notes that the most difficult set of policy concerns at the local level affect patterns of residential mobility. Local policies, including structural reforms, affect and sure influenced by mobility. In discussing Toronto and Jacksonville metropolitan governments, Quigley notes that "the relationships among public 41 service characteristics, taxes, and intraurban residential mobility suggest that the redistributional implication of reforms in financing or distributing local services are far more complex than was recognized initially by the proponents of reform". The "principal considerations...are the incentives for mobility of such rearrangements in finance or service provision"(p.50). A second order change in redistribution occurs as "the mobilty of households in response to these incentives has differential effects upon property values in various parts of the metropolitan area"(p.51). These two works seem to suggest simultaneity among political structure, policy, and mobility. Intraurban differences in providing municipal services will subsequently affect the calculus of residential mobility. Liebert (1974) responded to these comparative urban policy studies by pointing out methodological and theoretical problems in using municipal expenditures and functions as dependent variables. The core of Liebert's critique rests with his case for functional "inclusiveness." Municipalities in different regions and states are responsible for different types of services. Larger cities, for example, are more likely to be responsible for some welfare expeditures whereas smaller cities, even within the same state, may not. Education and highway expenditures may or may not be a municipal responsibility. This kind of functional variance makes it difficult to assess empirical results inasmuch as inclusiveness is associated with structure. Liebert attempted to replicate both Clark’s study of decision making and Lineberry and Fowler’s study of reformed cities. When he considered the effects of functional "inclusiveness" on both municipal 42 expenditures and decentralized decision making, Liebert found that "the importance of both municipal reformism and decentralization among reputed elites dropped to a minimum" (p. 783). They were found to have little independent effect on expenditures and appeared to be related more to functional responsibilities and degree of inclusiveness. Liebert concluded his analysis by suggesting that the comparative analysis of municipalities "would benefit from greater attention to differences in the allocation of formal powers"(p.783). Different states not only provide different city structures but also mandate different city services. The relationship between city structural properties and policy outputs measured by expenditures is confounded by the intervening variable of "function" which is often affected by city size.[8] The Policy Effects of Structural Components A second group of scholars turned their research efforts toward examining the effects of specific parts of municipal structure. These "structuralists" examined various singular components of structure to determine their effects on policies other than those measured by expenditures. Engstrom and McDonald (1981) examined the effects of at-large and district electoral formats on black representation on city councils. They note that although conventional wisdom once held at-large elections to be discriminatory against blacks given racially polarized elections, recent revisionist studies (MacManus, 1978) contradict conventional wisdom or conclude it to be "exaggerated"(p.344). Using a cross-sectional design including 239 cities, Engstrom and McDonald 43 find strong support for the conventional wisdom. In comparison to district-baaed electoral systems, at-large elections tend to proportionately underrepresent blacks. They estimate that the electoral format begins to exert this effect once the black population reaches 10 percent of the city’s total population. The at-large format tends to make under-representation more acute as the black population increases, and, once the black population reaches 15 percent, has even greater influence. The electoral format as a structural characteristic was found to have varied effects on representation. Stein (1986) found that city minority hiring and adherence to affirmative action plans is related to the degree of municipal "reformism." Stein surveyed 134 cities with over 100,000 people and over 10 percent minority population. Devising a model that includes economic, demographic and political and organization variables, the study sought to predict minority work force representation. The principal explanatory variable was the size of a city’s minority population. Economic well-being also acted as a strong conditional factor. More importantly for this study, organizational and political variables exerted significant effects. Bureaucratic autonomy in the form of an independent civil service commission tends to constrain minority hiring. A minority executive increases such hiring. The political and organizational characteristics have an interactive effect. Therefore, these factors can play mediating roles such that "certain configurations of local political and administrative circumstances clearly will be more likely to permit the translation of 44 demographics into representative bureaucracies"(p.708). Structure affects the direction and effort of otherwise strong demographic and economic factors. Another set of scholars have questioned the additive effects of the three part Reformism Index. Lyons (1977) considered the traditional use of reform in three ways: independently; more than one but without interactive combinations; and additively. additive index The use of an of reforms assumes, however, that each is equal in its policy impact and the presence of any two reforms is equal to the presence of any other two (p. 214). Combinations of two or of all three reforms presents a set of eight alternatives and Lyons develops hypotheses about their policy effects. Lyons’ findings suggest that structural characteristics exert different effects depending upon the "reform" combination and policy. In employment policy, elections and managerial forms, Lyons finds administrative and electoral reforms working against each other (p.223). In terms of public safety expenditures, these same factors enhance each others’ effects. For highway expenditures, an interactive model is best but the expected trends do not materialize — "political structure is not particularly salient for this policy"(p.225). Each reform affects a different dimension of policy and "has its unique effects in the policy arena in which its operational characteristics would most logically place it"(p.225). A "continuous reformism variable," analyzed by Lineberry and Fowler, does not seem useful or effective. The reform movement and its results were too complex to be captured by a simple additive model. Lyon suggests 45 using an interactive model, especially when studying structural effects on city policies. Maser (1985) followed this lead and developed a more sophisticated measure of reformed structure. Using fourteen charter features, Maser derived a cumulative index called a "Civil Rights Distribution". This weighted index purported to measure one underlying dimension of municipal structures called civil rights. He defined civil rights as having "one’s preferences included in making political decisions and to receive the entitlements and to incur the obligations that result"(p. 129). It is based on the economic contract view of charters. His work represents an advance in that, as Lyons suggests, different dimensions of reform can be tapped and measured. Although Maser looks at only one -representation- others like centralized administration exists. measure change. The scaled dimensions can also be used to Using an economic-charter contract theory, city populations change their structures as their preferences change. Class and demographic population variables are used as proxies for preferences. Maser finds charter change related closely to changes in income inequality, mobility, region and prior charter index scores. Cornwell (1985), as both a constitutional scholar and charter commission participant, raises serious points about Maser’s work. first concerns His constitutional decision making most frequently made by political elites. The highly abstract argument of change and consequence without more concrete policy referents "often leave the public indifferent, uncomprehending, or bemused"(p.220). Second, the degree of constitutional change ranges from most specific to most 46 abstract. Constitutional change "involves a series of qualitatively different political processes that make quite different demands on the electorate and its ability to cope with narrow concrete versus broad abstract issues"(p.222). Cornwell's third point focuses on news media coverage as an intervening variable. In constitutional revision, news focus on personalities and specifics tends to inform and sway voters one way or another on reform. Much success rests on the delegate composition and the inclusion of "disinterested reform zeal"(p.223). Finally, Cornwell notes the effects of state regulation and legislation on municipal charters. rests with the state. Meaningful change at the municipal level often The constitutional change process is complex and different along a series of dimensions (p.224). Any model of charter change, Cornwell suggests, must weigh these considerations. Conclusions Each school of urban theory developed some characterization of municipal political structure. The Community Power schools tend to downplay structure, although pluralists recognize some of its value. Bachrach and Baratz (1962) highlight the extraordinary position structure plays in both theories. The Reformism school focuses on structure as ways of achieving goals and reinforcing values. Finally, the Urban Policy school displays conflicting views of the value of structure. Measuring policy as expenditures in cross-sectional analysis seems to shield them from a conversion process for policy demands (Stonecash, 1979, Gray, 1979). Other scholars find differences and variation in structural characteristics. Others seek 47 new methods of characterizing structure. In this way, "structure" is operationalized in a much more sophisticated manner than the simple Index of Reform or dichotomous "reformed-unreformed" measures. The dissertation's next chapter will draw upon these works to develop a characterization of municipal charters and sets of concepts to help explain their change. NOTES 1. Ostrom explains that one important method of changing behavior is to change the rules defining incentives in varied situations. These rules can be changed while behavioral "lawB" cannot. This ability to change rules is what makes than interesting and one of their key characteristics. « 2. This drastic of an action by the state legislature is politically unrealistic. A city is likely to remain unless voters decide to incorporate. States may, however, intervene in city affairs in very dramatic ways. One recent example in Michigan is the State’s appointment of a receiver in the City of Ecorse with exceptional powers to raise taxes and make severe budget reductions to balance the city budget. In this case, the state appointed receiver is even recommending mergers with neighboring communities. 3. This consideration is modified in some cases that Peterson overlooks. Schools, for example, tend to redistribute wealth within cities. However, in many cases, cities use zoning and land use planning to influence other policy areas they might not otherwise control. 4. The state legislatures looked for cues from the city’s senate and house representatives. If these people agreed with the city’s proposed charter change, the full legislature often went along. Local bargaining, in this case included the local state legislative cadre. The state’s larger cities were often exceptions to this practice. 5. These authors base their critique on the works of Professor Schattschneider. These are further developed later in the dissertation. 6. The League publishes various pamphlets and booklets to help concerned citizens, charter commissions, and other activists to adopt its plan. For example, see National Municipal League (1960) A Guide for Charter Commissions. 4th ed. New York: National Municipal League. Its monthly publication also periodically includes stories about successful and not so successful charter campaigns. See, for example, Chute (1951, 1956) and Herbert (1971). 7. The first three features of this display, "form of government", "type of ballot", and "constituency" are commonly used in later works to create an index of "reformism". This condensed three part index purports to measure "unreformed" and "reformed" structures. 48 49 8. Dye and Garcia (1978) take up Liebert*s (1974) challenge. They find unreformed governments have more functional responsibilities than do reformed governments (p.111). Their study, however, implies that functional responsibility is a consequence of governmental form. It does not seen reasonable to conclude that form dictates function. At best, this is probably a simultaneous relationship defined over time. A reformed government, for example, may move to have its city hospital transferred to the county or regional authority. This would reduce it functional responsibilities. This indeed was the case for at least two cities in this dissertation. CHAPTER I I I THEORETICAL ORIENTATION The following chapter describing theoretical development consists of four parts. The first part presents and discusses a conceptual view of the charter that includes elements of both compacts and contracts. provisions. It also highlights limitations imposed under Home Rule The second part characterizes the charter compact/contract along a single dimension. This underlying dimension is an expanded version of "reformism" according to an ideal type promoted by the National Municipal League. The third part discusses those factors leading to change in the charter as characterized. The fourth part presents theoretical propositions or hypotheses that comprise a model of charter change. Definition of Municipal Charter This dissertation will define the municipal charter principally as a compact issued by the state to its cities. The compact includes carefully defined rules, procedures and offices comprising a municipal political structure. The charter as compact grants formal state powers to municipalities. This implictly represents an effort to divide governmental functions between state and city. The charter, for example, will grant to municipalities the power to provide or regulate utilities and to enforce the laws of the state. 50 Each of 51 these functions could be and often are performed by the state government. The underlying premise of this division is a belief that such powers may be divisible between purely state and purely local concerns. McGoldrick (1933) notes, however, this is rarely so clear in practice. Indeed, whenever legislatures attempt to create local governments in federated systems, functional allocation between jurisdictions becomes difficult to initiate and to ultimately maintain and enforce. In the case of Home Rule, McGoldrick argues that it has mostly befallen the courts to determine exact specifications of local versus state functionalization when the legislature cannot or will not do so (p.310). The sources of compact powers vary. There are five possible sources of municipal compact powers to be established via the state constitution (McGoldrick, p.312). First, constitutions may grant powers directly to cities thereby completely barring state legislatures from any involvement. Second, the legislature may be barred from the domain of municipal affairs except by its creation of general and uniform laws. Third, the legislature may be completely barred from a domain of municipal affairs yet may entrust the domain to municipalties. Fourth, the legislature may be barred from the domain of municipal affairs except where it may create uniform laws or entrust power to cities. Lastly, the legislature, without constitutional limits on its authority, may be permitted to grant cities some degree of power over their affairs.[1] In the first four of these cases, the state constitution limits the legislature’s involvement in some part of municipal affairs and 52 requires it to give municipalities seme measure of real control. This limitation is designed to force legislatures to grant and divide functions it otherwise would be reluctant to grant. The fifth version allows state legislatures to determine on their own any grant of power. This legislative action is unlikely to occur without constitutional compulsion. The compact, then, may be protected from the legislature as well as include substantial and independent powers in several functional areas. In many states, the legislature may try to influence compact provisions for some functions. The municipality may not have complete control over its affairs and could still encounter legislative interference. This degree of legislative involvement varies according to the state constitution and disposition of legislatures. The powers granted to cities via Home Rule usually include authority to amend their charter. This amendment power creates potential political conflict over charter change at the municipal level. Adopting a reformist view, we may see the charter as a set of rules governing local political conflict. Municipal rules dispense political power and privilege to citizens and sets of interests within the city. As a set of rules, the charter is biased against one segment of the citizenry while gaining advantage for another. citizens may get more rights and power than others. Some Conflicts may thus occur over the charter-defined balance of local political power. Any study examining municipal Home Rule charters across several states would necessarily deal with variability in their statutory and constitutional enactment. Maser (1985), for instance, uses constitutional provisions as an independent variable to help explain 53 charter change. This particular study is restricted to municipal home rule charters in the State of Michigan. This restriction controls for state constitutional and legislative definition of and involvement in municipal affairs. Such involvement is held constant throughout the study. Charter amendment power is one area of constitutional dictate. The amendment process is also one type of local political battle which, as Salisbury found in St. Louis, is often determined by existing cleavages within the electorate. Amending the charter involves issues of functional power as did New Haven’s failed charter proposal to create and insulate an urban renewal department. The compact granting local powers like amendment is restricted by constitutional and legislative means. A void occurs to the extent that a constitution establishes an area of municipal jurisdiction that the legislature fails to fill by enabling legislation. The enabling legislation ultimately adopted may also be deficient. In every state, the courts play important roles in interpreting and giving substance to the definition of functional distribution between municipality and state. In Michigan, for example, the functional area most susceptible to deficient legislative enactment in the early Home Rule years has been the municipal court system, schools and special assessments (McGoldrick, p. 199-209). A study restricted to Michigan Home Rule cities also controls for state court interpretation of the Home Rule constitutional provision and any state legislation pertaining to it. 54 An Element of Bargaining The self-amending compact powers allow charters to be thought of as a bargained contract. To borrow from Maser’s work (1985), the charter as a constitutional contract "produces a stream of benefits and costs associated with decisions made over time. An individual, preferring to maximize his expected utility, invests in a constitution. Making that investment is costly. some will prevail over the preferences of others. The preferences of By definition, the charter approved will tend to reflect the preferences of the dominant political coalition, those more willing and able to bear the costs of investing"(p.132). The bargaining process is undertaken among all citizens who, in view of their own costs and benefits, choose to promote change, resist it, or abstain from participation. The underlying motivation for contract investment rests on a concept of city structure as an instrumental goal. As an instrumental goal, structure is used to achieve intrinsic public policy goals (Miller, 1981). Just as people have preferences for policy, so too they have preferences for structure (Miller, 1985; Maser, 1985). Several urban policy scholars have tried to demonstrate certain configurations of city institutional structures will induce various kinds of policies. They argue that people may influence desired policy provision while excluding others’ articulation for policy by implementing various structural designs. In bargaining for local political advantage, rules themselves come up for change. The amendment process under Home Rule opens charters to occasional movements for change while ordinary policy battles may be fought within confines of existing municipal structures. In an 55 effort to change the type of policies coming from the political system, charter rules may be changed. Referenda on rule changes make charter change a political issue and battle. Effort to change the charter is a means to change the distribution of benefits and costs to the advantage of one group at the expense of another. For example, a charter change increasing the debt or property tax levy limit may favor those citizens benefitting from city spending while exposing those who pay taxes to greater financial costs. Change can enlarge or restrict the access of groups to city powers and ability to direct policy. It represents potential shifts in the status quo and a focus for election battles. The "bargaining" or electioneering over proposed rule changes would not involve every constituent. participants are not equal. The benefits and costs to Contractors and city employees may, for example, gain from certain tax rules while other citizens might see relatively few benefits and only marginal costs. The contractors and employees would be much more likely to get involved in these kinds of rule issues. While potential to participate is open to all voters, only those with substantial benefits outweighing any costs are likely to vote and change rules. Limitations on Bargaining The concept of "Home Rule" does not mean that all preferred changes may be made to a charter. Home Rule statutes in Michigan, for example, permit municipalities to change charters by adding and deleting various provisions(MCL 1970). But these statutes also mandate some charter provisions while offering options on others. Mandatory 56 provisions include creating a council or commission with ordinancemaking powers; power to levy and collect taxes; and establishing several city officers. Home Rule charter options deal with borrowing moneyi special assessment use, public improvements and choice mechanisms for selecting city officers. Limitations include number of possible elections and a property tax levy of no more than two percent of the assessed value of real and personal property in the city. Home Rule also permits the state legislature to mandate and change municipal charter provisions. Therefore, the charter as a bargained contract does not mean all citizen preferences may be included. The "contract" is still subject to state control. Limitations and enabling legislation at the state level restrict the types of policies a municipality may adopt. Large scale redistribution of wealth among populations within a city is limited by state statutes and constitutional provisions. Following Peterson’s argument, policy battles are mostly fought among citizens wishing to pursue land-based developmental policies. Even within the confines of state constitutional and legislative limits, marginal forms of redistribution can be proposed and accomplished via charter change. Retirement systems, for example, can redistribute wealth from citizens to retired city employees if the system is included in the charter. The structural changes engendered by charter amendment may not have dramatic policy change. However, any change in charters would have some consequences on city policy. To summarize this discussion, a charter is characterized as a state-granted compact of variable state interference and control. Some element of contractual bargaining over its content exists due to 57 either legislative or constitutional voids and action. While this differs across states, it as well as court interpretations may be controlled by limiting analysis to one state. The charter itself has at best an indirect effect on policies and the conduct of local politics. As such, at times it becomes a political issue as it is changed to facilitate the acquisition of other policy goals. We can now discuss a broad characteri zat ion of the charter general enough to encompass the most important structural components of municipal government. Charter Characterization Along A Single Dimension In this section, the charter is characterized by "reformism" as a single dimension. The term "reform" connotes different meanings to different persons in different contexts. For some, it means "good" government without the stain or burden of "politics." For others, it signifies a white, upper-middle class effort to impose their set of values and beliefs about municipal government on other racial and income groups. Several values promoted the practice and effects of "municipal reform". They were and still are proclaimed to be efficiency, economy, and honesty in public service.[2] These same values were borrowed from the developing private sector undergoing rapid industrialization in the late nineteenth century. Early Reform Strategies The earliest reformers were more likely to operate within existing political structures and seek changes in personnel. These people wanted to "toss the rascals out of office" and replace corrupt 58 officials with more socially-conscious and honest persons. Their problems of government were laid on the shoulders of a complacent or unknowing public opinion. Once reawakened, reformers supposed this same public opinion could be used to inject new civic blood into a defunct political system, be it state, county or city. In Detroit, reformers thrice elected Hazen Pingree, a social reformer who advanced Detroit's development in both physical and social reforms. He later was rewarded by election to the Michigan governor's office. [3] The reformers experienced mixed success in their method of personnel change. A reformist or good government administration could not always compete with entrenched administrations or legislative bodies. In these cases, reformers could not get to the first step of improving government from within. In some states and cities, reformist organizations could elect their candidate slates at some times but not others. A rotation of "good" and corrupt administrations created a condition Plunkitt referred to as "morning glories"(Riordon, 1963, p.17). Reformers would be elected in one year, make little gains, and soon fade into political history. These conditions demonstrated that while public support for reform could be evoked at various times over mixed issues, it could not be sustained. Reformers came to realize that an aware, excited public opinion always on watch for corruption was an unrealistic expectation. The National Municipal League and Structural Reform As the reform movement became more organized and national in scope, communications opened among good government groups across the 59 nation. Several reform organizations founded in cities eventually met in a national convention in January of 1894. The most significant and longlasting result of this meeting was formation of the National Municipal League later in the year. suggested municipal reforms: elections. The convention also endorsed two civil service and separate city The endorsement of structural reforms pointed a way for this new alliance to offer a reform agenda to all member organizations regardless of their electoral representation and political abilities. After the founding of the League by 16 affiliated local units, its number swelled to 180 branch units one year later. The following year, eighty more city organizations joined as public interest again arose. Most of these local civic organizations were from the middle- Atlantic states (especially New York and New Jersey) where municipal governments were most corrupt and problems most complex.(Patton, 1940, p. 34) The League offered an organizational base for local civic reform groups to encourage cooperation, ease communications among member units, and plan agendas for a national movement.[4] The National Municipal League became identified with structural changes as a common denominator of all local reform movements. While it could not offer much support for voter turnout or candidate support for local elections of "good governments", it could help in developing structural changes and promoting an agenda of change to complement election of new, honest officials. In 1897, the League appointed a committee of eminent reform activists, professionals and academics to draft a model city charter. The standard of "reform" is contained in the National Municipal League’s Model Charter. The model was first developed in 1898 and 60 revised five times since. Each edition was the product of outstanding reform practioners, intellectuals and theorists. The 1915 committee recommending the counci 1-manager plan was chaired by Richard S. Childs, the plan's "father". He sat on all later revision committees. The list of committee participants reads like a reformer’s list of Who’s Who: Luther Gulick, Arthur W. Bromage, Louis Brownlow, Charles E. Merriam, William B. Munro, Clarence E. Ridley, and Leonard D. White among others. If the municipal reform movement ever had an institution to protect and promote its values, it was the National Municipal League. The National Municipal League is currently revising its 6th edition of the Model Charter to incorporate new elements and update the 1964 version. Progress reports issued by the revision committee suggest much strife and controversy engulf the task.[5] In earlier versions, compromises were made by including possible alternatives to structural arrangements. For example, the legislative body could be selected by a mixed system of both wards and at-large districts rather than completely with at-large districts. The revision committee working on the seventh edition has similar differences requiring compromise. The major sticking points blocking final agreement are the role of the mayor in city government and offering alternatives with or without preferences. Given the plan’s long reliance on an apolitical city manager who would nonetheless be a unified leader of the administration, the mayor’s role is seen as somewhat threatening. Others argue, though, that while a manager may be the administrative leader, a city needs a political leader beyond that of a figurehead. This leader should also be directly chosen at large and not selected 61 from within the legislative body. rival the manager in authority. In such a case the mayor would A longstanding politics- administration dichotomy appears to have broken down to affect other values and cause friction. The values underlying the original and later charter versions are susceptible to challenge and competition from proponents of other values. Concepts of efficiency, for example, may be challenged by concepts of representation and effectiveness for emphasis in specific local environs. Furthermore, agreement over values does not prevent conflict over implementation of these same values. The charter, because it establishes a political structure, must by definition implement values. The rules become value-laden. This dissertation avoids the current debate over the 7th Model Charter by utilizing the 6th edition published in 1964 as the reform standard. The Ideal Charter The 6th model charter is one based on reformist principles of leadership, democratic participation and responsibility. Its goal is "efficient, economical and progressive municipal government"(1964, p.xvi). This is the benchmark of municipal reform. Any charter change may be gauged as moving it closer to or further from this Ideal Type. The League assumes its model will be adopted by cities enjoying the advantages of municipal home rule. If the state does not have home rule provisions in its constitution, the League offers model state constitutions with such features. Reliance on home rule gives cities the option of incorporating all or parts of the model into their own charter while leaving other parts out. It also allows for 62 the liberal interpretation of "bargained contracts" to be used in assessing and promoting change. The Model Charter can be conceptualized as a single dimension, one pole of which is defined as the "Ideal". The National Municipal League establishes the Model as an ideally constituted structure we call "reformed". Anything found in a charter contrary to the Ideal will be considered "unreformed." The standard of "reformism" is defined and incorporated in the Model Charter. This same standard of an Ideal Charter also contains or embodies other values or dimensions. Maser, for example, derives his "civil rights distribution"(CRD) from eleven charter characteristics contained within the League’s Model. His CRD could also characterize a degree of representativeness or the extent to which municipal government represents constituent interests in that it includes council size and mayoral selection methods as parts of the CRD measure. The same measure also includes administrative features like budgeting control powers and the executive’s departmental appointments. These could be considered part of "representativeness" and certainly affects the administrative responsiveness to citizen as well as governmental demands. Three Pillars of the Ideal Charter Three major values and dimensions may be distinguished in the original model. These are first, a strong executive with an administration placed beyond political interference; secondly, a strong council acting as a board of directors; and third, a civic goal of business-like efficiency and capital development. 63 The original model emphasized a strong executive branch. It called for a strong mayor form of government that was later changed to the commissioner plan after its introduction in Galveston, Texas in 1901. The commissioner plan had severe flaws which became manifest after several years of operation.[6] Within several years the city manager plan was invented and soon became the Model's preferred standard for the executive branch. experimentation: This was municipal what worked was retained and what failed was discarded. An appointed city manager was likened to a corporate manager responsible to an elected body of "city directors" for efficient city government. A clear separation of powers existed between the appointed executive and elected legislative body. With one executive responsible for overall city government, probability of department in­ fighting was reduced. The appointment proviso naturally worked to keep electoral politics more distant from the executive branch. The underlying value was a conceptual dichotomy between politics and administration. City administration in its pure form was believed to be free of political influences. In his classic work on the Study of Administration. Woodrow Wilson (1887) posits this principle of separation that is still with us today: Most important to be observed is the truth already so much and so fortunately insisted upon by our civil service reformers; namely, that administration lies outside the proper sphere of politics. Administrative questions are not political questions. Although politics sets the tasks of administration, it should not be suffered to manipulate its offices.(p.10) This first tenet of an administrative doctrine has become part of basic public administration that endures among many practioners today 64 because of its appealing simplicity and forthrightness. Wilson, once setting the basic dichotomy, prescribed vast powers for the governmental administration executive without fear or anxiety over misuse or subversion by "politics". Wilson wrote: There is no danger in power, if only it be not irresponsible. If it be divided, dealt out in shares to many, it is obscured; and if it be obscured, it is made irresponsible. But if it be centered in heads of the service and in heads of branches of the service, it is easily watched and brought to brook, (p. 12) In this way, the public servant entrusted with great powers is "less likely ...to abuse it, the more he is nerved and sobered and elevated by it"(p.12). The single check on the aggregration of administrative power is arguably public opinion as an authoritative critic. A problem arises over how to introduce the powers of public opinion without making it "meddlesome" in administrative affairs.[7] Wilson argued that public administration will find the best means of giving public opinion and criticism the "indispensable" role of "superintending the greater forces of formative policy alike in politics and administration"(p.13). To insure that this kind of administration remains, Wilson offered a civil service ideal that is "cultured and self-sufficient enough to act with sense and rigor, and yet so intimately connected with the popular thought, by means of election and constant public counsel, as to find arbitrariness or class spirit quite out of the question"(p.14). In this way, Wilson set the outline of a politically neutral administration of civil service experts with great powers accountable to public opinion or public interest. Its goal is greater efficiency in serving that public interest while undergoing close, guiding scrutiny. Politics, according to Wilson, 65 can be fully separated from administration and make other goals achievable. Childs' city manager plan, invented nearly twenty years later, fully incorporates Wilson's tenets. The legislative body, the city council, retains some image of or connection to "politics” but only in benign form compared to previous "boss politics". In early American history, the council was the premier branch of city government. With Jacksonian reforms, however, council power was diminished and fragmented. The election of lesser administrative officials sapped some council powers and created independent power bases beyond council control. The introduction of special select committees and their later evolution into advisory, then permanent, and then independent boards, also reduced council power. The council was in decline and, when reformers looked to introduce a stronger executive, a multi-membered body was not the best model to follow or pursue. The weakened council also made it less attractive for elected public service and leadership which further alarmed reformers.[8] The revised municipal framework introduced by the National Municipal League in 1898 advocated among other items the "Federal Plan" charter, separate city and state elections, a reduced council size, and the election of council members at large with the principle of continuity (Patton, 1940, p.64). The "Federal Plan" proposed separation of powers described above, and also gave more power to the mayor. At the urging of the Pennsylvania Governor, Philadelphia’s Bullit Charter of 1887 attempted to closely follow the national plan on a city level. following figure: The similarity of the plan may be illustrated by the 66 City of Philadelphia United States President for four years Two houses of Congress Senate from each state House of Representatives according to population Presidential message President’s veto passed over by twothirds vote President’s appointments confirmed by Senate President impeached by House of Representatives before Senate Penalty— dismissal Mayor for four years Two chambers Select Council from each ward Common Council according to population Mayor’s message Mayor’s veto passed over by two-thirds vote Mayor appointments confirmed by Select Council Mayor impeached by Common before Select Councils Penalty— dismissal Figure 1 The Federal Plan of Municipal Government (Patton, 1940, p.65) Separate elections for councillors and other elected city officials would help insulate them from state and national party politics. A reduced council size, later established to be seven members, was expected to be elected at large. The separated and well defined powers entrusted to the council would consist of those essential to determining the public interests of the city as a whole and criticising the city administration. Councillors would adopt policy and set administrative goals and future direction. The popularity of bicameral councils with an upper "select" council and lower "common" council grew in the late 1800’s. This plan more fully replicated the national plan although it fragmented legislative powers even further. The League’s Model Charter did not go so far as to propose a bicameral council, and its proposed powers were still less than that of pre-Jacksonian reforms. the council would be set at seven members. Its version of Some cities had chambers 67 with several dozen legislators. None of the League's seven councillors would be tied to any single ward or section of the city but would owe their allegience to the entire city. In a business metaphor, they would act more as a board of directors and less like an executive. Their principal task was to make policy and pass a deciphered public interest on to the manager. Salaries would be diminished so monetary rewards and compensation would not accrue to the office of councillor. The job would pay a very modest per diem amounting to a couple hundred dollars per year. Adrian and Press (1977) note that such a council structure leads to election of eminent local political figures with some independent income and name recognition (p. 212-213). It also excludes the professional politician hoping to make some "honest graft" in administrative affairs. Council elections were set two years apart even though continuity was a strong League theme. It did not recommend any limit on consecutive councillor terms. A third value underlying the Model Charter is one of business efficiency with development orientations. The model has had lengthy sections pertaining to granting of franchises, acquiring, owning and operating municipal utilities, and special improvements and assessments at one time or another. The Great Depression of the 1930’s induced the 1934 edition to include lengthy passages on taxes, budgeting and bonds. Each edition has an implicit time-related bias that affects a particular city policy area. The new edition now in the formative stage, for example, has reflections of the emerging mayoral role and will undoubtedly affect both administrative and political leadership. The theme, however, is most frequently cast in 68 terms of development and efficiency. The Charter’s Relationship to Preferences One might hypothesize that ideal-type charters are adopted by constituent populations having assumed preferences for them. Some majority of the municipal population must vote to create and/or amend the charter, and citizens presumably vote according to underlying preferences for charters and expectations of achieving policy objectives through a particular configuration of charter features. set of charter interests are found within a city. A Some voting segment of that set involves itself in charter votes. Population changes will lead to preference changes which ultimately induce charter changes. Lord Bryce (1906) noted that constitutional revision responded to societal change. "In the older and more conservative States," he wrote, "there have been but few revisions. In the newer States the more rapidly moving population, and swifter economic and social changes tend more to induce changes in the fundamental law"(p.185). Ingress and egress of persons adds to, diminishes, and changes the set of population preferences for charters. Sub-populations and hence new majorities may be created and dissappear over time. This mobility creates stimuli for alternate charters more or less following the Model. For example, the ethnic makeup of municipalities may become more or less homogeneous over time. Such change may form a new majority preference for charter change. Other scholars are quick to note second-order effects of population changes. While not explicitly examining change in 69 diversity or homogeneity, Hannah (1986) examines Michigan municipalities defined by their population losses and perceived "fiscal stress". She found that these cities tended to alter their charters by strengthening and balancing both executive and legislative branches. The composition of the population in her study, except for the change in black populations, does not matter as much as strains created by population loss. Change in charters were made to alleviate the consequences of that loss. Maser (1985) claims that population demography guides municipal demography. Preferences for government are related to population changes. This proposition assumes preferences are stable among demographic groups over time. Demographically-measured preferences are part of most studies of structure in each of the three schools of thought reviewed in Chapter II (Dye and MacManus, Banfield and Wilson, Lineberry and Fowler, Dahl, Wolfinger, Maser, Miller). The sets of charter preferences are inferred from the composition of city populations. A more middle-class city would have leanings toward a city management form. The measured population characteristics are assumed to be proxies for political preferences. Some scholars point out flaws and dangers of using this approach. Miller (1985) for example, points to persons with preferences for lower levels of services and are therefore unlikely to go along with a charter promising more expensive city government, or in this case the reformed Model. Lyons (1977) argues that political intentions are extremely difficult to discern from measures of population composition. The history of reform battles also reveals problems in making 70 this assumption. Salisbury's examination of St. Louis charter battles describes how the outcome is determined by cleavages. The voting population, however, is some segment of the city's population and they may not be willing to change charters. fate of amendments. Voter turnout may decide the Patton's (1940) discussion of the reform movement suggests a variable public opinion willing to promote some reforms while not coming out to support other similar ones because it was not a "hot" issue. unknown. The fickleness of voting populations is certainly not Further, as Lineberry and Fowler point out, it may be mistaken to assume that political power capable of changing the charter can be guaged by its relative size of the city's total population. Small minorities may be capable of making radical change to the charter. Two rejoinders can be offered to this problem of measuring preferences in the aggregrate. First, as Wildavsky (1987) notes, the ultimate source of preferences is often assumed to be "given" by most rational choice theorists. unsatisfying. However, this is theoretically The preferences, Wildavsky writes, are endogenous to political systems and are a function of culture, in this case at the municipal level. Cultural theory... is based on the premise that preferences are endogenous internal to organizations - so that they emerge from social interaction in defending or opposing different ways of life. When individuals make important decisions, these choices are simultaneously choices of culture - shared values legitimating different patterns of social practices. Always, in cultural theory, shared values and social relations go together: there are no disembodied values apart from the social relations they rationalize, and there are no social relations in which people do not give reasons for or otherwise attempt to justify their behavior (p.5). The preferences are thus constructed "in the process of decision- 71 making. Their continuing reinforcement, modification, and rejection of existing power relationships teaches them what to prefer"(p.5). This explanation suggests that measures of culture broadly interpreted may be desirable for measuring charter preferences. The level of experience with political decisions made under a particular charter create the basis for preferences. Therefore, measures of culture and population change also capture some of the endogenous preferences. A second condition supporting the use of population measures to tap charter preferences is inherent in the nature of charter adoption and amendment. urban scholars. The stability of such charters is well known among Simply put, change does not occur so frequently among state and sub-state jurisdictions. While lending credence to the decision-making experience for preferences noted above, this also suggests the voting majorities reflecting assumed preferences will not be as fickle as other scholars might imply. Amendments are often piecemeal and introduced sporadically across Michigan cities. Significant change does not occur frequently for any single city. Preferences affecting charter support appear more stable across time even as populations change. The majority appears stable. A test of this is whether or not amendments follow on the heels of close charter adoptions. In summary, demographically-measured preferences, or as Wildavsky notes, cultures, are justifiable as a focus of inquiry for charter adoption and change. What we can measure is aggregate political demography and charter preferences for the League's Model Charter. 72 Traditions of Charter Change The traditions found in municipal reform literature provide concepts useful in understanding and explaining change in municipal political structures. One tradition specifies as a cause change in population preference due to altered economic conditions. Another specifies change in overtly political cleavages that result in charter change. A third tradition specifies the status of or experience with a charter can help explain the liklihood of its amendment. The Rational-Economic Tradition Maser (1985), Dye (1979), and Gray (1979) posit economic measures affecting policy preferences. There are seven concepts in this tradition related to charter change. will change municipal preferences. First, change in city wealth Wealth is related to level of and capability for civic participation and ultimately expressions of policy preferences. It is also indicative of the strength of propertied interests needing charter protection. Second and closely related to wealth is income. Populations with higher incomes exhibit higher levels of political participation. Charter preferences would differ between higher and lower income populations. Income and wealth would both be related to citizen interest in taxing and spending policies. At the higher income levels, population participation levels would decline.[8] Third, change in levels of homeownership within a city would increase charter activity. Homeownership reflects "stakes" in city government issues not found with renters. Homeowner interest is largely directed toward protecting and promoting homes and 73 neighborhoods. The fourth conceptual change would be in the levels of education of the municipal population. Education reflects an ability to get politically involved as well as general knowledge helpful in comprehending issues like charter change. Levels of education relate to interest and participation in municipal affairs. They also reflect citizen comprehension of structure and goals of city government. Fifth, change in the age of city populations would influence the likelihood of change. Age has a bearing on service and policy demands placed on city government. An older population, for example, would be less interested in school and education issues than a younger population. Age would also affect the probability of political involvement. Middle-aged citizens would be more likely to participate than either younger or older citizens. Sixth, change in length of residence would affect the probability of charter activity. A population that on the whole lived in the city longer would be less likely to change its charter. A city of highly mobile citizens would shift preferences as it undergoes constituency change. A seventh concept relates municipal housing development to charter change. As new residential growth increases, the city charter will be more likely to change in compliance with the ideal. This type of growth differs from population increases because it is more inclusive of municipal development as a whole. It suggests as cities increase their supply of new housing, the need for city managers and related efficiency and economy charter provisions increase. This would be the case especially for those suburbs undergoing rapid growth immediately 74 after World War Two. Changes in these factors reflect changes in citizens’ charter preferences. The direction of change is toward a stronger or weaker version of the Ideal Charter Type. The Political Cleavage Tradition • Charter battles are but one field in a larger battle between the "ins" and "outs" of city government. As the political environment changes, cleavages shift and the balance of local power changes. A disadvantaged, politically weaker side tries to gain strength and alter power balances. Charter rules will become issues. The weaker side will try to change them while the stronger side seeks to maintain them. As Salisbury (1961) noted in the St. Louis Charter battles of the 50’s and 60’s, those proposed charter changes leading to changes in advantage followed deep cleavages in the city political environment. Charter change is one effort to increase relative political advantage.[10] This tradition provides a set of "political" concepts to help explain charter change. They refer to several kinds of cleavages appearing within urban settings. The most common division may be one of political partisanship. This is most apparent if cities host partisan elections. Yet many cities hold non-partisan elections. In these cases, partisanship may still elicit some bias and affect local government structure. State and national political party influences inevitably filter down to municipal levels (Hawley, 1973). The partisan influence would be greater as the balance of forces approached a threshold level. If both parties reached a parity in 75 voter strength, reform impulses would tend to fade. At-this point, other non-political party issues would come to dominate charter politics. Population growth and decline stimulates change. Local immigrants or emigrees change the size and nature of municipal voting populations. Old power holders may have to give way to newcomers. This most clearly illustrates Schattschneider’s concept of "scope of conflict." The local political climate may be such that two or more groups fight election battles over local offices commanding power. In this case, rule change may gain political stability. Adrain and Williams note that "the loci of leadership are directly related to the form of political structure"(1963, p.312). In the absence of a strong council, either an executive or interest groups generally assume leadership roles (Adrian, 1959). heard. Constituencies may or may not be Factional electoral fighting can therefore spill into charter change to increase or restrict representation. It may also serve to inhibit or depress the chances of gaining consensus for electoral victory. Racial cleavages may divide city electorates. Physical segregration of racial populations suggest this is more likely to be neighborhood-based conflict. As racial divisions grow, conflict may occur over rule changes for greater governmental access and representation. This conflict is more likely to manifest itself as the racial division increases above a threshold. Smaller minorities would be less likely to exert strong effects on electoral politics (Engstrom and McDonald, 1981). 76 Ethnic cleavages work in the same direction, although they may weaken over time as ethnics are assimilated. In some cities, new Middle Eastern or Asian ethnic cleavages are beginning to appear as older European immigrant divisions diminish. Closely connected to race and ethnicity are religious cleavages. This is principally a Catholic-Protestant division. influence family and political values. Religions These religions view government institutions and individual roles within society differently. As the religious mix within a city changes, the majority’s vision of local government would change. Changes in the workforce can also create change in the charter. A large and growing blue collar population is less likely to succumb to vote appeals for League ideals. It also would be more likely that blue collar voters would maintain stronger traditional ties to present city government forms rather than change to something new and different. Blue collar voters, for example, would be less trusting of professional, white collar city managers if they presently had a mayor or commission. Change would be more difficult to produce given such a population. Two additional factors tend to modify political battle lines and voting forces. First, annexation of significant territory and population might change the existing order of battle. That is, new residents will be brought into the city voting population. This is similar to the effects of population change although the annexation is a more distinct change and may be forced upon some residents. The second factor is the presence of a major college or university. an institution will serve to alter the political climate within Such 77 cities. The university often employs the kind of persons closely identified with the municipal reform movement. These people are also more likely to participate as local government activists. This special presence would not be as strong in communities with a major college or university. Charter Status Influence Several concepts relate a charter’s "status" to its likelihood for change. This status is derived from the prior effort of cities to acquire and amend a compact to suit their particular needs and satisfy their preferences. As such, the charter at any point in time reflects the shared experience and opinion of its constituent population. It is the summation of prior Progressive reform efforts to derive an ideal-type charter as well as their opposing forces hoping to retain or remove selected features. The charter document at any period is a summary statement of reform success. The status of a charter also recognizes that over time, change is incremental. The reformed charter of one period becomes the base from which subsequent change is guaged. incremental means small change. This is not to suggest that The charter experiences of Michigan Home Rule cities reveals this incremental change can span from full scale revision to minor amendment. In either case, shared, working experience under an actual charter will help explain the likelihood of future change. First, the charter’s relative standing in terms of present "reformism" or closeness to the ideal affects the liklihood of change. If a municipal charter is very close to the ideal promoted by the 78 League, then it would be less likely to undergo further change toward it. If, on the other hand, the charter is far from the ideal, then it has more potential for change toward it. There is "space" for bargained amendment and it is likely to be an incremental adjustment to the present charter summary of Progressive reform. Second, most of the substantial changes made to municipal charters are contained in wholesale revisions rather than piecemeal amendments. The revision and adoption of entire charters would create a condition requiring less change in the years immediately following it. As the number of years since the last adoption increase, it is more likely the charter becomes out of date in terms of the current model promoted by the League. Recall the League’s first proposal called for commission forms of city government. A charter adopted in the early 1900’s would be more likely to have this form than a charter revised in the 1950’s or 1960’s. By that time, the now out-of-date charter would be more likely to undergo change than one adopted in 1940. Third, initial support of the charter upon its adoption will affect the chances of subsequent change. If the revised or new charter is adopted by a wide majority, then it would appear unlikely that subsequent changes would be proposed and adopted by charter opponents. The voting population would be less likely to alter a charter they wholeheartedly supported. If, however, the city voters only narrowly passed a charter referendum, it would be much more likely to amend the offending characteristics. A faction may see less support for various features and seek change. A fourth consideration relates the combined effects of need for 79 periodic review and charter age with likelihood for amendment. Older charters are thought of as being more likely to change based solely on their age. However, if they also adhered closely to the ideal, change was not expected. Yet it can be the case that while there would be a sense of tradition to maintain a charter that is tried and true, it will also come up for periodic review and "modernization". In many cases, charter commissions seek mandates to "update" an otherwise acceptable document. Thus the reformism base and age work to cancel each other out while in practice modernization occurs. The final concept relating charter status to liklihood of change affects university town charters. The presence of very old charters may present opportunity to change them in experimental ways, expeically in university cities with populations of intellectuals and professionals. The positive effect of a university on a charter would be moderated by the effect of its age. Newer charters would not undergo the constant change of a university influence while older charters would experience accelerated changes. Presence of a university would promote charter change and, in this case, accelerate it if the charter was of some age. The Structural Model These sets of economic, political, and status concepts can be combined to form an explanation of charter change. arranged in a structural model shown in Figure 2. change would be toward or away from the Ideal Type. They can be The direction of Specific hypotheses about directions of change are discussed in the following section. 80 Pr(CHAN) = BO + B1X1 + B2X2 + B3X3 + B4X4 + B5X5 + B6X6 + B7X7 + B8X8 + B9X9 + BIOXIO + B11X11 + B12X12 + B13X13 + B14X14 + B15X15 + B16X16 + B17X17 + B18X18 + B19X19 + B20X20 + B21X21 + B22X22 + B23X23 + et where: Pr(CHAN) = city i’s likelihood of charter change XI change in WEALTH X2 — change in INCOME X3 = change in HOMEOWNERSHIP X4 change in EDUCATION X5 — change in POPULATION AGE X6 — change in LENGTH OF RESIDENCE X7 - change in HOUSING GROWTH X8 change in PARTISANSHIP X9 - change in PARITY X10 change in POPULATION XI1 change in COUNCIL REPRESENTATION X12 change in RACE XI3 — change in RACE THRESHOLD X14 change in ETHNICITY X15 - change in RELIGION X16 change in WORKFORCE XI7 ANNEXATION X18 UNIVERSITY TOWN X19 = CHARTER BASE X20 “ CHARTER AGE X21 - CHARTER SUPPORT X22 OLD CHARTER X23 NEW CHARTER et - disturbance term — — — — — — - — Figure 2 Structural Model 81 Theoretical Propositions This dissertation attempts to provide a rigorous test of the relative strength the sets of theoretical explanations of charter change. Following the distinct traditions found in urban politics literature, two sets of hypotheses can be offered. First Set of Inferences — Conditions for Change The first set of inferences define conditions promoting structural change of city government. It also tests the relative strengths of political and economic concepts found in policy debates (Dye, 1979; Gray, 1979; Stonecash, 1979). One explanation put forward by those scholars viewing charters strictly as contracts is that structural change results from change in resident preferences. If the preferences of a city's residents for a particular structural form changed, then the form would change to conform to them. These forms or preferred structures are a function of economic stakes in the city and therefore largely class-based. We can use this dissertation format to derive two theoretical propositions and test parts of this claim. PROPOSITION ONE: The probability of structural change increases as a city’s wealth, homeownership, income, education, age, residency and housing growth changes. These demographic indicators are used as proxies for unobserved structural preferences. The propositions assume preferences for structure are associated with economic class at the aggregrate level. It is an untested although often made assumption (Maser, 1985; Miller, 1981; Banfield and Wilson, 1963; Lineberry and Fowler, 1967). The second proposition is a corralary of the first. It specifies 82 the relative impacts of demographic changes on the probability of charter change. PROPOSITION TWO: Of the economic concepts related to population preferences, changes in wealth, homeownership and housing growth would exert a greater influence on the probability of structural change than income, education, age and residency. Wealth, homeownership, and housing growth are more property-based in nature and therefore more directly affected by municipal government. Many of city government’s greatest powers are those related to property control (Peterson, 1981). The other concepts, while also related to preferences, are more transferable and mobile, and therefore less closely connected to charter change. The political battles of charter reform suggest that socio­ economic preferences by themselves exert relatively little pressure on the probability of structural change. An alternative proposition is that change in a city’s political environment would determine structural change. environment. conflict. SES is but one factor contributing to a political Political structure as a set of rules governs local A sudden or great rise in political conflict may place different sets of demands on the political structure. The political preferences of the population would change with change in the population. Therefore, changes in the political environment lead to altered demands placed on the structure. The city’s political structure would be altered to regain political stability. PROPOSITION THREE: The probability of structural change increases as a city’s partisanship, representation, population, race, ethnicity, religion, and workforce change. These population characteristics exert varied political demands on the city government. As in the economic set of preferences, this set 83 includes elements with varying effects on structural change. One subset is more closely linked to structural change because it is more overtly "political." This can be formally stated in a fourth proposition. PROPOSITION POUR: The probability of structural change would be more sensitive to change in partisanship, official representation, and population. These concepts would be more directly linked to structural change than those of religious or social natures. A separate case for conditions that stimulate structural change can be made for territorial annexation. Annexation brings new residents and property into the city and helps to redefine conceptions of the "city," its limits, and expectations for local politics. Therefore annexation of properties that require a vote for approval would contribute to probability of structural change. PROPOSITION FIVE: The probability of structural change increases as the city annexes land and people requiring voter approval. A separate case can also be made for those cities with a large college or university. College towns would have a substantial population of intellectuals and local activists closely attuned to the League’s reformism and its associated values. would promote change toward the Ideal. This population subset A 1972 amendment to the Michigan Constitution also gave college students the right to vote. This added significant new voting populations to college towns and would have changed structural preferences. This condition is formally stated in Proposition Six. PROPOSITION SIX: The probability of structural change increases if the city contains a large college or university. 84 Structural change is more likely to occur if prior charter adoption and amendments do not follow progressive reform efforts. charter acts in part as a summary statement of prior reformism. The If it reflects minor change toward the Model, it is much more likely to be amended or revised in the future. If it conforms closely to the Model, there is less "space" available for progressive reform to occur. This is also dependent upon relative support for the current charter. These conditions are formally stated in Proposition Seven: PROPOSITION SEVEN: Structural change is more likely to occur if the charter is old, its base and original support are low, the need for modernization is high, and a university is present. Second Set of Inferences — Directions of Change This set of hypotheses specify the directions of structural change in city government. As the probability of structural change increases, the changes that do occur will move the structure into greater or lesser conformance with the Ideal Type. The economic tradition considers structural change a product of the "dominant political coalition" (Maser, 1985). This coalition adheres to common policy preferences and presumably has greater political resources necessary for charter investment. and its resources grow as the upper class grows. The coalition This class-based coalition prefers structural forms closer to the Ideal because it limits lower-class access. This is formally stated as Proposition Eight: PROPOSITION EIGHT: Structures will conform more to the Ideal Type as a city’s wealth, income, age, education, residency, and housing growth increases. Preferred structure is derived from conflict among political 85 rivals and coalitions. These rivals and coalitions form on the basis of class, partisanship, neighborhood, and interests. Structure as a mediating effect between these interests and city government changes as rivals change. The balance of political forces would lead to a reformed structure closer to the Ideal Type. An unbalanced system would be more likely to be dominated by one like-minded set of forces. This is formally stated as Proposition Nine: PROPOSITION NINE: Structure will conform to the Ideal Type as a city’s Republican partisanship and population increase and council turnover, race, ethnicity, Catholicism and manu­ facturing workforce decreases. The charter’s status conditions the potential for future structural change in the direction of the Ideal type. The Ideal charter has an explicit minimum and maximum level of reformism. If the charter is positioned at one end of the reform dimension, then it is more likely to move in the opposite direction. There exists more "space" for reform or unreform depending on its prior status. This is presented as Proposition Ten: PROPOSITION TEN: Structures will conform more to the Ideal type if their prior status is near the unreformed extreme. A low charter base, old age, weak support, need for modernization, and presence of a university would increase the probability of change toward the Ideal type. We have now developed theoretical explanations of municipal charters, why they change, and the direction such change takes in relation to the Ideal type. Several propositions were carefully deduced from these explanations. The next chapter of the dissertation outlines an empirical process used for testing the theory and propositions in the context of the State of Michigan and sixty of its Home Rule cities. NOTES 1. McGoldrick states that Michigan follows the fourth case and has legislative Home Rule. With Washington State, Michigan has "not only given to cities a generous degree of discretion in local matters but ... have refrained from impairing this grant by local and special legislation"(p.315). 11118 was his assessment of the 1908 Michigan Constitution and its enabling legislation. Michigan adopted a new Constitution in 1963 and it keeps the Home Rule provision largely intact or strengthens it. 2. These values are most clearly emphasized in the documents published by the National Municipal League. The most recent Model Charter adopted in 1964 is designed to help achieve and sustain these values. 3. Judd (1984) presents a fine description of Pingree's work in the context of the municipal reform movement. It stands as a contrast to the business influenced structural reform method. For a fine study of Pingree, see Holli, M. (1969) Reform in Detroit: Hazen S. Pingree and Urban Politics. New York: Oxford University Press. 4. Concerned municipal officials in Michigan formed the League of Michigan Municipalities at about this time and quickly affiliated with the National Municipal League. Several members played prominent roles in each organization. For a good discussion of this organization and its influence on reform in Michigan, see Cummings, E.B. (1938) The First Thirty Years: History of Michigan Municipal League (1898 - 1928). Ann Arbor, MI: Michigan Municipal League. 5. These reports are contained in the National Civic Review. The September-October 1986 edition notes disagreement over "how the models should present alternatives" and cites the council’s electoral Systran and the manner of selecting the mayor within the councilmanager plan as "issues [about] which there is serious differences of opinion"(p.264). The final version is due to be released in 1987. 6. Its worst drawback was lack of separation of powers, followed by a necessity for electing leadership. Since each elected commissioner was also an executive, nothing separated the two governmental functions within one city structure. A group of commissioners acting as a cohesive legislative body was unlikely or at best a tenuous relationship. Logrolling and protecting one’s department at the expense of city-wide interests made the commission system much less attractive. 7. Patton (1940) notes that many reformers placed faith in an attentive public opinion which was later downplayed when structural changes seemed more obvious and permanent a solution. Public opinion, however, still remains a central tenet of reformism. 86 87 8. For an excellent history of early city governments, see Griffith, E.S. (1927) The Modern Development of City Government in the United Kingdom and the United States. New York: Oxford University Press. 9. The relationship is curvilinear. According to Crain, R.L. and Rosenthal, D.B., "Community Status As A Dimension of Local Decision-Making" American Sociological Review, higher status as measured by income and education leads to "a more tightly organized and more potent decision-making structure"(p.970). This is similar to their findings of lower status cities. This suggests higher status cities may be more homogeneous in their values and attitudes toward municipal structure. A lower participation in "community decisionmaking" results. 10. The division between "political" and "economic" concepts is recognized as fairly arbitrary. The general rule of thumb used here is a recognition of others' use of them in their theoretical models, especially the positive choice scholars. These economic concepts try to distinguish the substantive economic interests of the population. The political concepts, on the other hand, try to gain some indication of "typical" cleavages other than economic occuring in voting populations. CHAPTER IV THE EMPIRICAL TEST This chapter of the dissertation discusses methodological concerns related to testing the theory presented in the previous chapter. It is divided into several sections. The first section describes a process of operationalizing and measuring the dependent variable, which is based on "reformism." This is followed by a section describing the operationalization and measurement of the independent variables. The third section discusses the data set compiled for model estimation. the statistical model itself. This is followed by a presentation of The fifth section describes a set of estimation methods and statistical techniques used to test the model. To conclude this chapter, we examine the statistical tests and relevant criteria used to evaluate the model’s performance. The Dependent Variable This section discusses the construction, operationalization and measurement of the dependent variable: charter change. It describes the construction of an initial index, an a priori weighting system, and a final ordinal variable. This final variable is derived from the index scores and is the one ultimately used in testing the model. The dependent variable’s foundation consists of an index constructed to tap and measure "reformism." 88 This concept of 89 "reformism" is embodied in the Charter (1964). National Municipal League’s Model The Model is predicated on a liberal state home rule constitutional provision that permits cities to frame and adopt the entire Model or any part of it.[1] The 1964 edition of the Charter includes preferred features or characteristics that the League considers important to proper municipal governance. A total of eighty-three preferred features were extracted from the Model Charter to form the basis of an index. Appendix A. These features are presented in Each characteristic of a city’s charter can be coded according to its proximity to the League-established ideal [see Appendix B]. The Model Charter, for example, prefers an appointed city manager as chief administrative officer. Actual charters contain provisions for chief administrative officers, and these may be elected or appointed. Elected officers would not conform to the Model and the code (feature #30, Appendix B) reflects this type of charter provision agreement. The coding format in Appendix B provides us with index scores ranging from a low of zero to a high of one hundred thirty-eight. Each provision is equally weighted by this method as it stands here. However, the League and others would place more weight on certain provisions they see as primary or "necessary" and less on others viewed as secondary or "recommended". This is a relative degree of importance requiring the establishment of some weighting system to assign weights to the index coding scores. Two basic methods were available to devise these weights. The first would be to utilize a statistical program such as factor analysis to mechanically produce loadings which could then be used as 90 index weights. This method was used by Maser (1985) in his work on charter civil rights. reasons. It was rejected for this dissertation for two First, it may compel elimination of important characteristics necessary to complete an accurate portrayal of the League’s ideal. As Maser noted in his work, some charter features were dropped because their loadings were too low given standard procedural guidelines for factor analysis. Besides losing a full identification of the ideal, corresponding observations of actual charter changes would also be dropped. The infrequent charter changes across time and location made it necessary to maintain a broadlyconstructed index. Second, using a technique like factor analysis may substitute mechanical computation for critical thinking about relative importance of features. The rich and exhaustive body of literature and research on reform contributes a profound depth of knowledge about the charter and what is considered most important. The League’s work on the Model Charter itself is such that it provides substantial assistance in determining the relative importance of different features. On a more practical note, actual change experience supported an a priori approach to guaging relative weight.[2] For these reasons, a second approach was used to devise index weights. The weighting method used for this dissertation relied on rankings provided by a five member panel of experts. Each panel member independently ranked each charter characteristic listed in Appendix B on a scale of High, Medium and Low importance relative to the League’s model of reform. The panel members’ individual rankings are included as Appendix C. The panel method utilized here represents a more theoretically- 91 based a priori approach to weighting the charter.[3] The expertise represented by the panel included three well-known scholars in the urban politics discipline and two practitioners in local government. One member was a long-time city manager of East Lansing and was later made director of the Michigan Municipal League, a state affiliate of the National Municipal League. The panel’s knowledge of city charters and model reform permitted it to establish the relative importance of charter features without any of the drawbacks encountered with first approach. the All features were included as relevant to the ideal by using this approach. One problem, however, was the panel’s indecisive rankings on several features. A summary of the panel’s rankings are displayed in Figure 3. panel ranked ten features consistently high, medium or low. The The decision rule used to judge other features was first, majority rule, and secondly, a medium rank for split decisions. The final ranks listed in Figure 3 were made on the basis of this decision rule and then assigned values. The High-ranked features were weighted with five points, the Medium-ranked features were weighted by three points, and the Low-ranked features were given one point. This point ranking, while somewhat arbitrary, operates to distinguish among several kinds of features. A weighting formula for non-majority panel decisions presented difficulty. was assigned. In every case of this type, the middle weight In the case of presence of a city manager, this exceptional and central feature of the Model Charter received an extraordinary weight of seven points. Even so, this extra weight does not make manager cities stand out from others in terms of scores. 92 ALL MEMBERS AGHRR (5 of 5) High Rankings Medium Rankings Low Rankings 1. Council Constituen. 61. Brd.Rev.Compensate 13. Election Type 62. Brd.Rev. Assessor 14. Mayor selection 17. Non-interfer; adm 24. Assessor appoint 30. City manager form 33. Mgr appoint heads 35. Mgr directs admin POUR OF FIVE MEMBERS AGREE Medium Rankings Low High Rankings 7. Council terms 3. Council elig 4. 15. Mayor terms 5. Mayor elig 60. 18. Non-interfer; appt 11. Council compens 63. 19. Anti-nepotism 12. Mayor compens 68. 73. 22. Clerk appt 39. Attomy appoint 23. Treasurer appt 47. Circulate comm 75. 82. 26. Audit require 83. 34. Executive budget 37. Civil service 45. Initiative limit 46. Referendum limit 66. Property tax rate THREE OF FIVE MEMBERS AGREE High Rankings Medium Rankings Low 9. Election timing 2. Council size 8. 10. 16. Elected officials 21. Council vacancy 41. 20. Conflict clause 25. Assessor Indiv. 43. 54. Mayor item veto 27. Council meeting 44. 55. Mayor ordin veto 28. Ordin effective 56. Mayor vote 29. Emergency ordin 69. 71. 58. Number of boards 31. Asst executive 72. 79. Judicial terms 32. Mgr appoints 74. 36. Dept creation 40. Plan commission 76. 48. Init signatures 77. 49. Ref signatures 78. 65. Contract bid 80. Rankings Candidate filing Brd.Rev. size Brd.Rev. executive Fiscal year date Spec.Asses, pay Spec.Asses, create Constable present Constable terms Rankings Councillor resign Term begin date Budget capital pgm Primary signatures Petition signature Tax days Property powers Spec.Asses. plan Spec.Asses, limit Judicial residency Judicial req't Judicial comp Court jurisdiction Figure 3 Summary of Charter Characteristic Ranking 93 TWO OF FIVE MEMBERS AfiftKE Medium Rankings Low Rankings 6. Elect majorite 38. Employee comp 42. Nomination meth 50. Refer time 51. Refer election 52. Init time 53. Init election 57. Number of depts 59. Brd.Rev. comp 64. Retire system 67. City income tax 81. Court systems High Rankings Figure 3 (cont’d). Point rankings were applied to the city charter characteristics’ coding formula to derive index scores. Total possible charter index scores ranged from a low of zero to a high of four hundred thirty eight. A low score represents less "reformism" and correspondence to the League’s Ideal. A high score indicates closer resemblance to the Ideal and more "reformism." The index was calculated for the charters of sixty Michigan Home Rule cities for 1950, 1960, 1970, and 1980. The distribution of charter index base scores is shown in Figure 4. Periods of charter change were defined by decade and the several charter changes made between 1950 and 1980 were also ranked by the weighted index scores. These changes in index scores were computed by decade to give charter change scores ranging from -12 to +126. distribution of the charter changes is shown in Figure 5. The A high score represents a move to closer agreement between the Ideal and actual city charter within a particular decade. The a priori weights and subsequent scores reveal change and its direction relative to the Model Charter. However, it is somewhat more 94 Frequency Distribution 326 322 318 314 310 306 302 298 294 290 286 282 278 274 270 266 262 258 254 250 246 242 238 234 230 226 222 218 214 210 206 202 198 194 190 186 182 178 174 170 166 162 x XX XXX XXX xxxxxxxx xxxxxxxxxx xxxxxxx xxxxxxxxx xxxxxxxxxxx xxxxxxx xxxxxxxxxxxx xxxx xxxxxxx XX xxxxxxxxxx XXX xxxxxxx xxxxx XXX X xxxxxxxx xxxxxxxxxx xxxxx XXX XXX XXX XXX xxxxxxx XXX xxxxxx xxxxx XXX X X XXX X Figure 4 Distribution of Charter Base Scores 95 Index Score Frequency Distribution 126 123 120 117 114 111 108 105 102 99 96 93 90 87 84 81 78 75 72 69 66 63 60 57 54 51 48 45 42 39 36 33 30 27 24 21 18 15 12 9 6 3 0 -3 -6 -9 -12 XX X X X XXX XX X XX X X X XX X XX XXX XXX XXXXX XXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXX XXXXX xxxxxx XXX Figure 5 Distribution of Charter Change Scores 96 tenuous in a theoretical sense to argue that an index score is double, triple or even more the weight of other scores. The index is not well tested enough to undergo these kinds of theoretical and practical strains. This condition is relieved to a great extent by further grouping index scores. In this way the dependent variable can adequately capture both change and the magnitude of its direction. To accomplish this task, weighted index scores were placed on a frequency distribution to locate any natural breaks. One break did occur in the positive direction between nineteen and twenty-six. On the basis of this information, the charter scores were placed into one of four categories of change. These categories were: negative (coded -1); none (coded 0); low positive (coded 1); and high positive (coded 2). This categorization of an interval level into an ordinal level variable was made as a safeguard against possible problems with index rankings and weights. For example, a higher positive index score well separated from lower positive or even negative scores can be categorically defined and better defended as representing high positive change in the direction of the Model Charter. A continuous level variable implies that smaller differences in index scores are meaningful. While defense of the index as a continuous, interval measure is avoided, catergorization tosses out data and knowledge about charter change. Yet given the circumstances surrounding the index construction and coding, it was determined that catergorization was a theoretically safer alternative. 97 The Independent Variables We now present an operationalization and measure for each of the independent variables. The first set of concepts to be operationalized relate to the city’s economic conditions and demographic attributes. These include taxable wealth, income, homeownership, education, age, residence, and housing growth. Wealth To operationalize taxable wealth, a measure was constructed to gauge the median housing value per city. This base figure was taken from the U.S. Census Bureau decennial census. however, were not adjusted for inflation. The raw figures, The 1967 Consumer Price Index (CPI) for Detroit was used to adjust each of the housing values thereby equalizing them across decades. In this way, change in this variable was likewise adjusted for inflation. The State of Michigan Tax Commission’s State Equalized Valuation (SEV) is an alternative to using the Census Bureau’s housing values. SEV is available for each year by city and includes all taxable property values. This alternative was not selected for two reasons. First, SEV figures include industrial and commercial as well as housing values. Industrial and commercial valuations would screen the effects of personal taxable wealth on a voting population’s charter choices. For example, a city reporting a low median housing value would, with all other things being equal, prefer a charter that conforms less to the Model Charter. However, if this same city had a SEV boosted by large industrial plants, it’s "wealth" would suggest that it would prefer a more conforming charter. measure the voter’s taxable wealth. The goal is to Second and more importantly for 98 measurement purposes, the SEV formula and property tax laws were changed in the early 1970's to exclude personal property and business inventories. This change makes comparisons of SEV across time more difficult and would provide misleading results when calculating changes in wealth. It is hypothesized that as the median housing value of the city increases, the charter will move toward the Ideal. No or negative movement in housing values reflect cities’ losses of wealth. This variable by itself will help to determine if indeed the poorer cities are more likely to go with the Model. Some scholars argue that since the Model represents values of efficiency and economy in municipal government, relatively poorer cities might be more willing to invest in charter reform. Income The income of a city is operationalized by utilizing the U.S. Census Bureau’s measure of median family income for families and unrelated individuals. As with the housing value above, the Bureau’s income measure was unadjusted for inflation. It was therefore adjusted for inflationary effects by equalizing values according to the Detroit CPI. It is hypothesized that the effect of income on charter change will resemble that of housing values. That is, the greater the increase in median family income, the greater the charter changes toward the Ideal form. Homeownership Homeownership is operationalized and measured by utilizing the U.S. Census Bureau’s percentage of owner-occupied housing statistic. This percentage in the 1950, 1960 and 1970 censuses includes single 99 family dwellings. In the 1980 census, this figure also included multiple family dwellings. condominium living. It This change includes the growing trend in does not substantially alter the measurement of homeownership in this model. The percentage would still be a fraction of total eligible housing units. Owner-occupation effects would still be reflected in the charter whether or not the owner lived in a condominium. The theoretical model hypothesizes homeownership to have a positive effect on charter change. Therefore, as the percentage of owner-occupied dwellings increases, the charter would be changed to more closely approximate the League's ideal. Education To operationalize and measure education, we turn again to the U.S. Census Bureau’s decenniel census. The census includes a count of the median number of years of formal schooling for each city’s residents. This includes both male and female residents. The model holds that as the city’s median number of years of schooling increases, the charter will be more likely to change to conform to the Model. Age The age of a city’s population is operationalized and measured in a manner similar to education. The U.S. Census Bureau’s statistic of median number of years in age of the city’s residents provides an indication of population age. median number of years. An older population would have a higher Growing cities would tend to have younger populations and hence a lower median number of years. It is hypothesized that a city will change its charter to conform to the 100 model as its median age decreases. A city with an increasing median number of years would conversely be associated with less probability of change. Length of Residence Length of residence may be operationalized by utilizing the U.S. Census Bureau’s measure of residence in the year immediately prior to the census. In 1950, for example, the census respondent was asked his place of residence in 1949. The percentage of those in the same city is recorded for 1950, 1960 and 1970. In 1980, the question was altered to provide for place of residence five years prior to the census. To get a measure similar to earlier censuses, the 1980 percentage of same-city respondents was divided by five. This would annualize the five year percentage to estimate a one year percentage. The percent of same place residency indicates cities with more or less mobile populations. It is hypothesized that this percentage would be negatively related to the probability of charter change. As the percentage increases, the likelihood of change toward the model charter would decrease. Housing Growth The growth of new residential areas in the cities was operationalized by taking the percent of residential units added to the total housing stock within the most recent decade. The data for this variable was contained in the U.S. Census for 1950, 1960, 1970, and 1980. It is hypothesized the residential housing growth separate from that captured in the population growth statistic contributes a positive effect to the probability of charter change. As housing growth increases, the city charter will be more likely to change in 101 compliance with the Ideal. A second set of concepts to be operationalized and measured are loosely defined here as more politically related to charter change and city politics in general. These concepts include partisanship, population, representation, race, ethnicity, and religion. Several of these variables could arguably have been included in the first set. The utility of this division becomes clearer as these combinations of variables are tested for their relative influence within the context of the statistical model. Partisanship Partisanship was operationalized by using the city’s vote for the Michigan State University’s Board of Trustees in the state-wide elections of April 1951, April 1961, November 1970 and November 1980. In each election, voters selected two of four candidates running under party labels to fill two vacant Board positions. A vote for these lesser partisan offices is considered to be cast mostly on grounds of partisanship. The city’s two- candidate vote for each party was totaled and the percentage of Democratic vote calculated. This percentage measures an estimated base percentage of Democrats among the city’s voter age population. Provided that partisanship changes very little or slowly for individuals, this measure is relatively stable among a city’s voting population. Change in partisanship occurs as the composition of the population itself changes. In some cases the vote for the MSU Board of Trustees was unavailable and the Michigan Secreatary of State votes were used as a proxy measure. In these instances, the vote is the percent Democratic for this office. Correlation analysis shows that these measures are 102 approximately the same for the 1950 and 1960 elections, the two most frequently substituted elections. [4] It is hypothesized that the percent Democratic vote is negatively related to charter change. The issues of charter reform are more likely to be Republican issues. Partisan Threshold The generally negative relationship between Democratic partisanship and charter change, however, should not necessarily hold as the change in Democratic voting population approaches a threshold. As the Democratic population grows to approach the threshold of from forty to sixty percent of the total population, the presumed Republican majority would diminish. Both political partisan forces would be in positions to attempt changes in the city charter to attain structurally-induced political advantage. This condition can be measured by a dummy variable that takes on the value of one if the percent Democratic vote for the MSU Board of Trustees surpasses the forty percent mark or falls below the sixty percent mark. variable would take the value of zero otherwise. The It is hypothesized that this dummy variable would be positively related to municipal charter change. As the threshold is passed, the city would increase the probability of having a charter change. Population The city population is operationalized and measured by utilizing the U.S. Census Bureau's statistic on total city population. It is hypothesized that as population increases, the city will change its charter to conform more closely to the Model Charter. Population decreases are hypothesized to lead to movement away from the Ideal charter. 103 Council Turnover Council turnover was operationalized by calculating the percentage of council seat changes within each decade. The numerator was the total number of actual seat changes while the denominator was the total number of possible seat changes given an election cycle. The number of possible seat changes was determined by council election provisions detailed in each city charter. The actual number of council seat changes was measured by comparing the Michigan Municipal League’s Directory of Michigan Municipal Officers for each year from 1950 through 1980. A seat change was considered to occur whenever there was a change in the names listed between directories of two succeeding annual editions. This measure of seat change also includes those council members who resigned, died or otherwise left office and were replaced by appointments. As measured in this manner, council turnover is hypothesized to be negatively related to charter change: as the percentage of council seat change increases, the city charter would be less likely to change to conform to the Model Charter. Turmoil over city politics and local representation would distract the population from pursuing consensus for the purpose of municipal reform. A fight over representation would indicate factions too entrenched or volatile to fight for charter changes. This also reflects a charter-defined rule allowing for competition over office rather than a structure-induced political hierarchy. It may act to screen the effects of population changes by intervening between them and efforts to reform charters. 104 Race Race was operationalized and measured by using the U.S. Census Bureau's estimate of the percentage of each city’s non-white population. This gives a conception of race. population measure based on a dichotomous It is hypothesized that as the percentage of non­ white population increases, a city’s charter will be less likely to change to agree with the League’s Model. Race Threshold The likelihood of change, however, will increase as the percentage of non-white population approaches a threshold level of about thirty to forty percent. As the non-white population percentage crosses this threshold, a city would be more likely to change its charter toward the League’s ideal. It is operationalized by having a dummy variable take on the value of one if the non-white population peases and/or stays above the threshold of thirty percent of the total city population. It takes the value of zero otherwise. is hypothesized to have a negative sign. This variable As it takes on a value of one, the probability of charter change decreeises. Ethnicity To operationalize and measure ethnicity, we used the U.S. Census Bureau’s measure of the percent of the population that is foreignborn. This is based on another dichotomous population characteristic: birth place. It is an approximation of ethnicity, culture, language and social background based strictly on place of birth. It excludes the influence of second, third and later generation ethnicity that follows in later years. However, the Census Bureau’s record on ethnic heritage was not consistent across census periods making more refined 105 measures inaccurate. The use of percent foreign born can still approximate ethnic heritage and culture yet on a lower proportionate basis. This measure of the foreign b o m population would provide a lower estimate than other ethnic makeup measures. Using the percent foreign b o m also captures the new wave of non-European immigration to U.S. cities, especially among the Eastern and Middle-Eastern populations. It is hypothesized that the percent foreign-born is negatively related to the probability of charter change toward the Model Charter. Religion The concept of religious cleavages was operationalized by estimating the percentage of Catholics within the city population. This estimate was derived from the 1936 Census of Religions conducted in counties and cities over 25,000. For cities in this study under 25,000 population, the county percentage was used as a proxy. measurement does not include change across decades. This Other scholars have used private school enrollment as a proportion of total school age population to measure percent Catholic. available for the 1960, 1970 and 1980 census. 1950. This statistic is It is unavailable for Comparisons for several cities in the 1936 Religious Census and the 1960 census of private school attendance reveal problems in combining the two measures. Several cities with large percentage of Catholics such as Hamtramck show a substantially lower percentage of private school attendance, probably due in part to income characteristics. Also, in western Michigan with few Catholics, private school enrollments are high based on the number of private Protestant schools. The 1970 change in Catholic school policies also 106 closed many Catholic schools, lowering enrollments. In light of these conditions, the 1936 Census of Religions and percent Catholic was used as a static variable across all decades and the private school measure dropped altogether. that the percent Model Charter. It is hypothesized Catholic population is negatively related to the A higher percentage of Catholics would be associated with a charter less likely to conform to the ideal. Workforce The working class population was operationalized by using the Census Bureau’s measure of the percent of city workers employed in manufacturing. No data transformations were required. It is hypothesized that as the percent of the manufacturing workforce increases, the probability of charter change in the direction of the Ideal decreases. Annexation Territorial annexation was operationalized by the use of a dummy variable. The variable takes the value of one if a city annexed any territory or sum of territories that included more than one hundred residents within the decade. otherwise. The variable takes the value of zero The U.S. Census Bureau’s 1950, 1960, and 1970 census records include tables showing the number of city residents added to the population totals as a result of annexations. The 1980 census contained a listing of territories with base population figures of the prior census adjusted to reflect additions. The adjusted figures were subtracted from original population figures to derive 1971-1980 annexation populations. It is hypothesized that annexation is positively related to 107 charter change. If a city annexes new residents, it will be more likely to subsequently change its charter to accomodate the new residents. The direction would be more toward the Model Charter. University Town Finally, the concept of a University town was operationalized by a dummy variable that takes the value of one if the city includes one of fourteen state-supported universities or four-year colleges within its boundaries. The variable takes on a value of zero otherwise. It is hypothesized that the presence of a state-supported college or university would be positively related to charter change. The presence of a university would promote such change in the direction of the League’s model. Charter Base The variable of charter base seeks to capture the "space" available for charter change. It is operationalized and measured by taking the charter index weighted point values for the year prior to the decade of change. That is, it measures charters by the index scores for 1950, 1960, and 1970. It is hypothesized this variable has a negative relationship to the dependent variable. A high charter base score would mean less chance for a city to change its charter. Charter Age This variable accounts for the time since the charter adoption and is a simple counting measure. The number of years since the most recent charter adoption was measured to the mid-points of each decade. An adoption in 1915, for example, was given a value of forty in the first period to the mid-point year 1955. The bench marks were the years 1955, 1965, and 1975 for the three periods. It is hypothesized 108 this variable is positively related to the probability of charter change. As the number of years since the last major revision increase, the probability of significant charter change increases. Charter Support This variable seek to measure the initial satisfaction of the charter upon its adoption. It is operationalized and measured by calculating the percent of "yes" votes on the original charter adoption referendum. 50% plus one to 100%. This margin of yes votes range from a possible It is hypothesized this variable is negatively related to the dependent variable. As the support increases, the probability of later charter amendment decreases. Modernization This variable was created to pick up the interactive effects of the charter base score and charter age. The effort to "modernize" or update an otherwise acceptable charter opens the way for some movement toward the Ideal. It is hypothesized that this variable would be positively related to the probability of charter change. Experimentation The interactive effects of old charters in university towns allows cities to expand municipal reform in their cities. It is hypothesized this interactive variable constructed by multiplying the dummy variable university town by charter age will have a positive effect on the probability of charter change. The presence of a university accelerates reformism, especially if it is an older document. This concludes the operationalization and measurement section of Chapter Four. The next section summarizes operationalization and 109 measurement within the scope of the entire data set. The Data Set This section of the chapter discusses the data set as a whole. It includes descriptions of various problems encountered in collecting and coding the data. It also includes a section on various assumptions made in pooling observations from three time periods. The data collection and coding of charter characteristics formed the initial basis for construction of the dependent variable. These charters and their amendments or revisions were made available at the Michigan Secretary of State’s Great Seal Office at the Capitol Building. The coding itself was largely on the basis of the presence or absence each characteristic specified in the Model Charter. Compilation of these charters and their changes presented little problem outside of time consumption. A significant factor in weighting the ranked characteristics consisted of deciding the weights themselves. The choice of five for high, three for medium and one for low rankings is admittedly arbitrary. However, in most changes, the direction is unitary. is, change is mostly positive or negative in direction. That The weight spread is less troublesome than if most amendments crossed directions within decades. In mixed directional changes within a decade, the weights themselves may decide the overall outcome of measuring the dependent variable. It was thought that relatively low proportionate weighting would allow seme combination of low-ranked characteristics to outweigh other higher-ranked ones. In this way, one highly ranked change cannot skew the coding for possible combinations of lesser- 110 ranked changes. The independent variables based on Census Bureau data offered less of a problem to collect and code than the dependent variable. However, some cities almost consistently offered problems of missing data. These cities were mostly those that had populations less them 10,000 at some point between 1950 and 1980. A level of 10,000 residents was the usual cut-off for presenting census statistics. The city of Farmington, for example, had a population of 2,325 in 1950 and is therefore not reported in some categories of data. In these cases, the county-wide figure was used as a proxy measure. Several census statistics had to be interpolated, as with the five year residence figure in the 1980 census, to conform to prior years' reporting. It is assumed that this would be the best method of keeping data relatively consistent across censuses. Equalizing financial statistics based on the Detroit CPI also led to valid timeseries comparisons. The non-census based measures were in most cases more difficult to collect and code. The partisanship measure based on MSU Board of Trustee votes would have been the most straightforward had all of the data been located. The proxy vote consisting of the Secretary of State vote for 1950 and 1960 will be used until the other votes are collected. All council seat change data were made available at the Michigan Municipal League headquarters. The most important assumption underlying this variable concerns non-elected appointments. no way to control for this feature given the data source. There is Non-elected turnover is therefore included as part of council change in representation. Finally, the religion variable is perhaps the weakest Ill measure of all the independent variables. The percent Catholic from 1936 is a fair measure for those cities with populations over 25,000 at that time. However, to the extent that cities in the sample did not fit that criterion, this variable is significantly misspecified. It was decided to retain the variable as it is rather than leave it out of the model altogether. There were a total of sixty cities in the sample [see Table 1]. These cities had to meet two criteria for inclusion. First, they had to have a population of at least 10,000 at any decennial census between 1940 and 1980. Some cities had less than 10,000 at some point during this time which accounts for problems in data collection in some years. Second, these cities had to operate under Michigan Home Rule statutes since 1950. The cities had to have the ability provided under Home Rule to change their charters. This criterion excluded some large cities such as Ann Arbor that did not operate under Home Rule since 1950. The study period spans the years 1950 to 1980 for a total of thirty years. Data were collected for three separate time periods: 1951 - 1960, 1961 - 1970, and 1971 - 1980. The four censuses comprised base points needed to calculate change across decades. This provided 3 x 60 or 180 observations of change in the dependent and independent variables. The three sets of observations were "stacked" as one single data set for the purpose of increasing the total number of available observations. This "pooling" of data combines the cross- sectional and time series data and requires two assumptions (Pindyck and Rubinfeld, 1981, p. 252-261). The cross-sectional parameters are assumed to be stable across all time periods. It is further assumed 112 Table 1 List of Cities CITY Adrian Albion Alpena Battle Creek Bay City Benton Harbor Berkley Big Rapids Birmingham Cadillac Center Line Clawson Dearborn Detroit E Detroit E Grand Rapids E Lansing Ecorse Escanaba Farmington Femdale Flint Garden City Grand Haven Grand Rapids Grandville Grosse P Farms Grosse P Park Grosse P Woods Hamtramck Hazel Park Highland Park Holland Iron Mountain Ironwood Jackson Kalamazoo Lansing Lincoln Park Livonia Marquette Melvindale Midland Monroe Mt. Clemens Mt. Pleasant Muskegon INC POP 40 POP 50 POP 60 POP 70 POP 80 20 382 21 186 20 347 14 230 18 393 1853 12 112 11 059 12 749 8 345 10 406 1885 13 805 12 214 14 682 13 135 1871 12 808 38 931 35 724 48 774 43 453 48 666 1859 49 449 41 593 53 604 47 956 52 523 1865 16 481 14 707 18 769 19 136 16 668 1891 18 637 23 275 21 879 17 931 6 406 1932 14 361 8 686 11 995 6 736 4 987 1869 21 689 26 404 15 467 25 525 1933 11 196 10 199 9 990 10 425 10 112 1877 9 855 9 293 10 379 10 164 7 659 3 198 1936 17 617 15 103 5 196 14 795 4 006 1940 90 660 104 199 94 994 112 007 63 584 1927 1806 1,623 452 1,849 568 1,670 144 1,514 063 1,203 339 38 280 45 920 45 756 21 461 8 584 1929 10 914 12 565 10 924 6 403 4 899 1927 47 540 51 392 30 208 20 325 1907 5 839 17 515 14 447 17 328 17 948 1942 13 209 15 368 14 355 15 391 15 170 1883 14 830 10 329 11 022 6 881 2 325 1926 1 510 26 227 30 850 31 347 29 675 1927 22 523 193 317 159 611 196 940 163 143 151 543 1855 35 640 38 017 41 864 9 012 1933 4 096 11 763 11 844 11 066 9 536 1867 8 799 181 843 197 649 197 193 176 515 1850 164 292 12 412 10 764 7 975 4 348 1933 1 566 10 551 12 172 11 701 9 410 1949 7 217 13 639 15 641 15 457 13 075 1950 12 646 18 886 18 580 21 878 10 381 1950 2 805 27 783 21 300 34 137 43 355 1921 49 839 23 784 20 914 25 631 17 770 1942 15 380 27 909 35 444 38 063 46 393 1918 50 810 26 337 26 281 24 777 15 858 1867 14 616 8 341 8 702 9 299 9 679 1889 11 080 7 741 8 711 10 265 11 466 1889 13 369 39 739 45 484 51 088 50 720 1857 49 656 79 722 85 555 82 089 73 045 54 097 1883 130 414 131 403 113 058 92 129 1859 78 753 45 105 52 984 29 310 53 933 1925 15 236 104 814 110 109 17 534 66 702 1950 8 728 23 288 19 824 21 967 17 202 1871 15 928 12 322 13 862 9 483 13 089 4 764 1933 37 250 35 176 27 779 21 065 1887 10 329 23 531 23 894 22 968 21 467 1837 18 478 18 806 20 476 21 016 17 027 1879 14 389 23 746 20 524 14 875 11 393 1889 8 413 40 823 48 429 46 485 44 631 1869 47 697 113 Table 1 (cont’d). CITY INC Muskegon Hts. Oak Park Owosso Plymouth Pontiac Port Huron River Rouge Royal Oak Saginaw St. Joseph S. Ste. Marie Traverse City Wyandotte 1903 1945 1859 1932 1861 1857 1922 1921 1857 1891 1887 1895 1867 POP 40 16,047 1,169 14,424 5,360 66,626 32,759 17,008 25,087 82,794 8,963 15,847 14,455 30,618 POP 50 18,828 28,138 15,948 6,637 73,681 35,725 20,549 46,898 92,918 10,452 17,912 16,974 36,846 POP 60 19,552 36,632 17,006 8,766 82,233 36,084 18,147 80,612 98,265 11,755 18,722 18,432 43,519 POP 70 17,304 36,762 17,179 11,758 82,279 35,794 15,947 86,238 91,849 11,042 15,136 18,048 41,061 POP 80 14,611 31,537 16,455 9,986 76,715 33,981 12,912 70,893 77,508 9,622 14,448 15,516 34,006 that the periods are not related to prior periods across time, although several time series’ explanatory variables can be included in the model. This last point pertains to the model’s disturbance term. By pooling the data, we assume that the disturbance term contains the effects of cross-sectional disturbances, time series disturbances, and elements of both. The Statistical Model In devising a model for this theory, two considerations arose. First, the dependent variable is operationalized in an ordinal fashion. The model seeks to explain municipal charter change. This is often a simple yes or no choice made by the city voters who are offered a referendum on a charter amendment or revision. the model's dependent variable must be n-chotomous. Secondly, The reasons for this are largely theoretical. According to the theory, cities must first undergo some type of 114 population change preceeding actual charter amendment. They must incur some change in the population’s set of preferences or political cleavages. In both cases there is in effect a self-selection process made by those cities deciding to amend their charters. A purely random sample of all cities would be biased, compounded by the relatively low number of charter changes in an average city. Local special elections for charter change occur infrequently when sampled in this manner. The bias would be against those cities undergoing frequent charter change. A subset of cities undergoing change out of a sample set of all cities would consequently be biased by exclusion of cities not changing. Absence of a self-selection component to the model permits bias and constitutes a severe form of model misspecification error.[5] This kind of bias would be absent in a model utilizing a nchotomous dependent variable. City charters measured as a n-chotomous variable will exhibit one of four possible outcomes for each period of change. Charters would: 1)change or not change during the period, and 2) if changing, would be closer to or further from the Ideal Type. The dependent variable will therefore be operationalized as follows: Charter Charter Charter Charter amended to conform less to ideal unchanged amended to conform to ideal greatly amended to conform toideal -1 0 +1 +2 This type of dependent variable will account for those cities not changing their charters during the periods in question by including a separate category. A statistical form of the model can now be offered and it is shown in Figure 6. 115 Pr(SOOR) = BO + B1 B7 X7 + B13 X13 B18 X18 B23 X23 where: XI + B2 X2 + B3 X3 + B4 X4 + B5 X5 + B6 X6 + B8 X8 + B9 X9 + BIO XIO +Bll Xll + B12 X12 + + B14 X14 + B15 X15 + B16X16 + B17 X17 + + B19 X19 + B20 X20 + B21 X21 + B22 X22 + Pr(SOOR) = city i’s charter change score XI = change in MEDIAN HOUSING VALUE X2 = change in MEDIAN FAMILY INCOME X3 = change in PERCENT OWNER-OCCUPIED HOUSING X4 = change in MEDIAN YEARS OF SCHOOLING X5 = change in MEDIAN YEARS AGE X6 = change in PERCENT SAME PLACE RESIDENCE X7 = change in PERCENT NEW HOUSING UNITS X8 = PERCENT DEMOCRATIC VOTE X9 = PERCENT DEMOCRATIC THRESHOLD XIO = change in POPULATION XII = change in COUNCIL SEATS XI2 = change in PERCENT NON-WHITE POPULATION XI3 = PERCENT NON-WHITE THRESHOLD X14 = change in PERCENT FOREIGN BORN X15 = PERCENT CATHOLIC X16 = change in PERCENT MANUFACTURING XI7 = ANNEXATION X18 = UNVERSITY TOWN X19 = CHARTER INDEX BASE SCORE X20 = YEARS SINCE ADOPTION X21 = PERCENT YES VOTE X22 = MODERN [BASE*YEARS] X23 = EXPERIMENT [UNIVERSITY*YEARS] Figure 6 Statistical Model 116 This equation characterizes each city’s decision whether or not to change its charter and if it decided to change, whether or not (and to what extent) it will conform to the National Municipal League’s Model. The city has a n-chotomous choice from a four element set in each equation. It may choose to change or not change. If it decides to change, it may choose to support or oppose the Ideal type. The probabilities associated with each event may be inferred directly from this equation. A Set of Methods The structural form of the model suggests use of a n-chotomous probit model to estimate its parameters. There are several significant theoretical and methodological reasons for this selection. They relate to the dependent variable's n-chotomous nature and subsequent inadequacies inherent in use of ordinary least squares (OLS) estimation given this model structure. The dependent variable’s n-chotomous structure makes other assumptions beyond those used for OLS necessary. The unit of analysis is the city charter. The dependent variable is not levels of "reformism" but rather city’s expressed preference for reform as measured by charter amendment. What is observed is a city’s condition for charter change and elective choices. Four-choice categories and an absence of observed linear relationships prescribe the use of a n-chotomous probit model. The behavior being modeled is election choices. Interpretation of a model designed to replicate this behavior has to be made in terms of probability of making dichotomous choices. Estimation procedures have 117 to produce parameter estimates in the form of coefficients that can be used to construct Z-scores with some probability distribution to use in interpreting them. Traditional OLS does not provide coefficients with this capability whereas probit modeling will. The functional form of the model is assumed to be non-linear thus making ordinary least squares (OLS) inappropriate as a modeling technique. By using OLS and its underlying assumption of linearity with these cases, the models would be misspecified. Incorrect model specification undermines all technique assumptions and leads to mistaken inferences. The plausible substitutes, justified by prior theoretical consideration, are non-linear functional forms. The form selected for the model is the cumulative normal distribution function, although Aldrich and Nelson (1984) note that "the choice of (non­ linear) specification remains fairly arbitrary"(p. 35).(6] Model Tests The set of criteria to guide empirical model testing includes individual and grouped tests of estimated coefficients, three Goodness of Fit statistics, conditions under which cities decide to change and how they change, and the relative success of its predictions given model estimates and actual cases. Once estimated, the model will provide individual coefficients to be tested for statistical significance. Results will indicate the relative size of Maximum Likelihood Estimates, their standard errors and the Student’s t-distribution. By themselves, these estimates provide some indication of how well theoretical concepts were specified, measured and integrated into the model. Joint hypothesis 118 tests will be arranged for subsets of common independent variables. The Goodness of Fit statistics do not present as strong a measure of fit as do those used in models utilizing OLS estimation techniques (Aldrich and Nelson, 1984). Yet they do offer evidence which cannot be ignored or lightly dismissed. measures: This dissertation will use three the log likelihood ratio, estimated R-Square, and an alternative "pseudo" R-Square. The log likelihood ratio is comparable to using an F-statistic in regression analysis when one wishes to test a joint hypothesis that all coefficients except the intercept are zero. The estimated R-Square offered by McKelvey and Zavoina (1975) and pseudo R-Square measures should be used with caution because of the unknown distribution of the real R-Square. The model’s percentage of correctly predicted cases will comprise a very powerful summary statistic. A crosstabulation between actual and predicted cases will visually transmit an idea of strength and accuracy of each model. Finally, the model’s parameter estimates can be utilized to construct hypothetical city cases likely to be found in Michigan. These can help demonstrate probabilities associated with common cases fulfilling model expectations. This part of the empirical test is important in probit modeling because coefficients by themselves do not provide much information. The probit model’s strength is in its estimates being translated into probabilities. OLS will not be used principally because it cannot provide a means to determine probability without potentially severe difficulties. Hypothetical cases will assist in assessing the model’s overall plausibility. In the following chapter, this evidence will be summarized for the 119 model and reviewed given its theoretical foundation. The model should be judged partly by how well it serves its predictive task. This is a subjective assessment based on the weight of the model's statistical performance conditioned by how well it plausibly "explains" the process of charter change. NOTES 1. If a state does not have a liberal Home Rule clause in its Constitution, the National Municipal League offers a model state constitution with strong Home Rule provisions. The League hopes to have its members broaden their reform interests to include state and county governments. 2. Certain charter features tend to be changed more frequently than others. This would imply voters have decided to alter what they consider important and left those features intact that are either unimportant or in line with their preferences. In most cases, change in features is made in both directions: toward and away from the Model in different cities at different times. 3. The panel of experts were: Kenneth VerBurg, Professor, Michigan State University Alvin House, Professor, Michigan State University Charles Press, Professor, Michigan State University John Patriarche, Director (ret.), Michigan Municipal League Steven Thomas, County Commissioner, County of Ingham, Michigan A special thanks is extended to each for their assistance in participating in the panel of experts. 4. The Secretary of State vote in 1951 and 1961 were between individuals without long incumbency records. The experience with Secretary of State Richard Austin in 1970 and 1980 indicates the longer incumbency record leads to inflated percent Democratic vote totals. The correlation analysis was for those cities in Oakland and Ingham counties for 1951 and 1961 elections of MSU Board of Trustees and for Secretary of State. The correlation was .912 in these cases. 5. For a fuller discussion of sample selection bias, see Heckman (1976, 1979). The selection bias is diminished by including a component in the dependent variable to pick up those charters not being changed within a period. 6. Other sources for discussion of non-linear assumptions include Pindyck and Rubinfeld (1981) and Maddala (1983). 120 CHAPTER V EMPIRICAL RESULTS In this chapter, we will examine the results of model estimation and evaluate how well they support our theory of charter change presented earlier. The chapter is divided into several parts. First, we present a set of estimated coefficients for the initial model. We discuss various tests of significance for this model and its strengths and shortcomings. This includes presentation of a set of Goodness of Fit statistics including the likelihood ratio, R-square, and a pseudo R-square. We then present joint hypothesis tests for the subsets of independent variables. In the second part of this chapter, we present and discuss a more parsimonious version of the estimated model. Its estimated coefficients and their related tests of significance are reviewed and compared to the initial model. The third part of Chapter Five infers from the model's estimated coefficients the conditions under which cities will change their charters and to what degree. Recall that one advantage to using probit estimation is the derivation of coefficients best interpreted as probability scores. We are therefore able to construct various municipal scenarios and compute their associated probabilities of charter change. We present a substantive analysis of the model to assess the accuracy and utility of the theoretical development discussed in Chapter Three. 121 122 Empirical Results of the Model The model was estimated using probit techniques and variables operationalized and specified in the theoretical development. from this estimation are presented in Table 2. Results This table shows the Maximum Likelihood Estimates) their standard errors) Student's tstatistics for related significance tests, and estimated probability values. In first looking at the t-statistics and levels of significance, we find most of the variables are insignificant. At the .05 level, only eight variables, housing value, education, housing growth, partisanship, population, turnover, charter base and charter age, prove significant using a one-tailed test. All of these estimates except education and housing growth are negative. All but housing value, population and charter age are in the predicted direction. Charter base and turnover are the most significant with p-values of .001 and .005 respectively. The constant is the most significant variable with a p>-value of .0002. The probit model’s two threshold estimates are also very significant with p-values in excess of .0002. Three additional variables are significant at the .10 level as the tstatistic cutoff falls to 1.282. These three variables, university town, charter support and modernization, are in the predicted direction. The least significant variable estimates are change in owner-occupied, change in income, and change in age. When looking at the relative magnitude of demographic variable estimated coefficients, change in schooling carries the most weight. With an estimate of 3.902, it outweighs the impact of any other percentaged demographic variable. The next two estimates with large 123 Table 2 Model Estimation Results Maximum Likelihood Estimate Standard Error MLE/SE P-valu( 7.4380 1.6936 4.3918 .0002 Economic Variables median housing -1.1925 median income -0.0561 owner-occupied 0.2731 years schooling 3.9026 median age -0.2401 pet residence -0.7322 housing growth 2.4817 0.6590 0.2274 1.4747 1.9848 1.1254 0.5754 1.3359 -1.8094 -0.2466 0.1852 1.9662 -0.2133 -1.2724 1.8577 .0359 .4052 .4286 .0250 .4168 .1020 .0322 Political Variables pet Democratic -1.2410 Dem threshold 0.0944 population -0.7730 council turnover -1.2556 non-white pop 4.4854 non-white thres -0.4365 foreign b o m 0.2226 pet Catholic -0.0021 pet manufacture -0.1931 annexation -0.2075 university town 0.8696 0.6245 0.1934 0.4289 0.4929 3.6338 0.3435 0.4997 0.0059 0.6858 0.2384 0.5379 -1.9870 0.4881 -1.8024 -2.5473 1.2343 -1.2707 0.4456 -0.3585 -0.2815 -0.8705 1.6165 .0239 .3156 .0359 .0055 .1093 .1020 .3300 .3632 .3897 .1922 .0537 Charter Variables charter base charter age charter support modemi zation experimental -0.0166 -0.0750 -1.1744 0.0002 -0.0197 0.0056 0.0456 0.8124 0.0001 0.0167 -2.9606 -1.6431 -1.4455 1.4046 -1.1773 .0015 .0500 .0749 .0808 .1210 1.0710 2.6418 0.1095 0.1661 9.7740 15.8991 .0002 .0002 Variable constant Threshold 1 Threshold 2 -2xLLR 163.2 Estimated R-square .275 Pseudo R-square .475 % Predicted Correctly 48.89 Joint Hypothesis Test: without SubsetChi-square Economic 39.896 Political 11.864 Charter 20.976 Significance .0005 .8000 .0010 124 impacts are change in residency and change in housing value. The residency measure, however, lags considerably behind the effect of housing, which is very close to schooling in terms of impact. The two dummy variables, annexation and university town, reveal some disparity. The presence of a state-supported college or university adds .869 to the dependent variable score. is also significant at the .10 level. This estimate The occurrence of a large territorial annexation, however, only decreases the dependent variable score by .207, or about one-fifth that of the other dummy variable. The annexation variable is also statistically insignificant. Perhaps the most surprising finding is the wide disparity between hypothesized and actual variable signs. Seven of twenty-three estimates have signs opposite the predicted one. The variables housing value, income, population, charter age, and experimentation were expected to be positive but instead exert negative influence on charter change. The non-white and foreign-born variables were hypothesized to exert a negative influence yet are estimated to be positive. However, the effective direction of percent foreign-born is like that hypothesized. negative on average. The change in foreign-born population is A positive coefficient estimate operating on negative observations would thus create a negative influence on the dependent variable score. This is consistent with our hypothesized effect. One possible explanation for the effects of housing value and income is that poorer communities may feel more pressure to acquire modem municipal charters along the line of the League's model. The Model holds promises of greater efficiency and economy, features more 125 appealing to poorer cities. It may also be that wealthier communities already had charters closer to the League’s Model prior to 1950. Wealthier communities may also feel more comfortable with popular representation in city government at the expense of efficiency and economy. It is often argued that increased representation would decrease charter-induced efficiency and economy. the effect of wealth and income on charter change. This is one test of It supports Knoke’s (1982) claim that perceived needs of efficiency in city government promote reformism. The effects of foreign-born and Catholic populations may be partially explained by the effect of educational levels. Years of schooling consistently increased over the course of thirty years across most of the sample cities. Change in other variables, especially foreign-born populations, was mixed and small for most cities. Education would cut across these populations and reduce their impact. The measurement of percent Catholic is probably badly misspecified as well and this accounts for its poor showing in the estimation. In a similar manner, council turnover may be concealing these effects by absorbing demographic pressures on the charter. But the model was estimated without the council turnover variable, and results showed negligible improvement for these two demographic variables. The political cleavage variables exert considerable impact on the dependent variable. The estimated coefficient for change in the non­ white population is the largest in size. The partisanship and turnover variables are the greatest in terms of magnitude. The relative magnitude of the charter status variables also exert a very 126 strong impact in the model. The charter base score and charter age are both statistically significant and have large effects on the dependent variable. The age variable has the highest estimated coefficient other than the constant. We can rigorously test our conclusions on the relative effects of the three variable subsets on the dependent variable by separating them and conducting joint hypothesis tests. To do this, we estimated the model three times leaving out one subset at a time and conducted a Chi-square test. The results are shown in Table 2. Here we find that as a group, the political variables exert the most effect. this subset, the Chi-square is 11.864. exerts the least effect. Without The economic variable subset When it is removed from the model, the Chi- square statistic drops only 2.395 from 42.291 to 39.896. The remaining variables are still highly significant as a group. The charter subset’s Chi-square significance falls between these two extremes although it is closer to the political subset in terms of effect. Goodness of Fit for the Model Three Goodness of Fit statistics summarize the model's performance and accuracy in predicting municipal charter change. First, the basic test of significance of -2.0 times the log likelihood ratio with 180 degrees of freedom provides a statistic of 163.02 which is significant at the .001 level. This log likelihood ratio is comparable to using an F-statistic in regression analysis to test the joint hypothesis that all coefficients except the intercept are zero. In this case, even though several of the individual coefficients are 127 far from significant, as a whole they are very significant. According to the log likelihood ratio, the model performs well. The second Goodness of Fit statistic is an estimated R-square defined by McKelvey and Zavoina (1975) and used in the SST computer software package.[1] As an estimate of R-square, this statistic represents the total variance "explained” by the model as if it were similar to a regression equation. It is another summary statistic used to describe the probit model estimate as a whole. R-square for this initial version is .275. The estimated However, the reader should note the cautions associated with using this overall measure of fit in probit models. It may be under- or over-stated. Third, an alternative pseudo R-square statistic, offered by Aldrich and Nelson (1984), is .475. the McKelvey and Zavoina estimate. This is substantially higher than The difference between these two estimates of R-square demonstrates difficulties encountered when using R-square estimates in probit modeling when the actual distribution of R-square is unknown. In summarizing the initial model's performance on the basis of these three statistics, the researcher finds mixed results. The log likelihood ratio and pseudo-R-square indicate that the model does reasonably well. As a whole the model is significant and explains an estimated .475 of the variance. However, the estimated R-square is .275 which is lower by comparison. By these statistics, then, the model's utility is undetermined. To add further evidence, we turn to the model's success in correctly predicting actual charter changes. This is a very powerful summary measure. The initial model predicted 48.89% of the cases 128 correctly. These results and a cross tabulation are presented in Figure 7. The crosstabulation shows that the model most accurately predicts low positive charter changes [coded 1]. The model does more poorly when attempting to predict no change [coded 0], where thirtysix of fifty-seven actual cases are mistakenly predicted to have changed. Undoubtedly the poorest performance of the model occurs when it should predict negative changes [coded -1]. In this category, the model only predicts one of twenty-six cases correctly. The best it can do with high positive change is predict four of twenty-one correctly. However, if one considers the model’s main task to be predicting change and only secondarily direction of change, its performance improves. It predicts fifteen of twenty-six high positives to change as low positives. Only two are not predicted to change. The model does well in the extremes. It does not predict any negative changes to occur as high positives or high positives to occur as negatives. Also, most of the failed predictions fall within one cell of their proper places. To gain a summary statistic of this crosstabulation, we look at the Chi-square and modal category. The Chi-square statistic of 42.291 is significant at about the .0001 level. The modal category, [coded 1] also reveals a relatively good performance. Here we see the model prediction is an 8.2 point increase over the modal or null category of 42.5. represents a 19.2% improvement over the null. The model thus 129 ********** Crosstabulation of change by predict ********** predict COUNT COL POT negative -1 none 0 low pos 1 high pos 2 ROW TOTAL -1 negative 1 14.3 12 26.7 13 10.6 0 0.0 26 14.4 0 none 3 42.9 21 46.7 33 26.8 0 0.0 57 31.7 1 low pos 3 42.9 10 22.2 62 50.4 1 20.0 76 42.2 2 high pos 0 0.0 2 4.4 15 12.2 4 80.0 21 11.7 COLUMN TOTAL 7 3.9 45 25.0 123 68.3 5 2.8 180 100.0 Chi-square(9 d.f.) = 42.291338 Figure 7 Model Summary Measures 130 Given these summary statistics and predictive results, the initial model can be judged to have fair results, especially when attempting to predict low positive changes to municipal charters. It appears to do less well in other change categories.[2] Testing the Theoretical Propositions Our substantive analyses of the initial model results begins by using the theoretical propositions stated in Chapter Three as a framework to guide discussion. We can explore and assess the results by how they support our propositions. The first proposition stated that charter change is a function of changes in wealth, income, owner-occupation, education, age, residency and housing growth. These concepts were used as proxies for population preferences. The results indicate that at the mean, the probability of change is above the mark for low positive change. The only significant variables are housing value, education and housing growth. The strongest effects are registered by education changes. Home value and residency changes act to depress the probability of charter change toward the Ideal. If income is also considered, the combined class variables tend to inhibit change. That is, as a city prospers, it is less inclined to change its political structure toward the reform Ideal. Three other class variables temper this effect. The changes in education, housing growth and owner-occupation work to counter negative effects of class to generally promote change. As city prosperity declines, a change toward the ideal seems to promise more efficiency and savings in local government. The proposition that 131 charter change is solely the function of class change, however, is not supported by these results. The second proposition concerns the relative strength of these class concepts in promoting structural change. It holds that wealth, homeownership, and housing growth as measured by median home value, percent owner-occupied, and percent housing growth are the strongest class measures. We find, however, that when measured at the mean value, only change in housing growth has the stronger effect on charter change. residency. This effect is closely followed by education and This suggests that citizenship values may be more important than property-related stakes in the community when deciding on government structure. Citizenship participation and activity is often dependent on education. More stable or permanent residency are visually associated with greater participation. As shown here, however, participation based on residency does not necessarily extend to charter reform. The third proposition examines the impact of political cleavages on political structure. These cleavages include partisanship, council turnover, population, race, ethnicity, religion, and workforce. Of these, council turnover has the strongest effect on charter change and in the negative direction. It appears that competition over political office and related administrative control signals challenges within the confines of current structural rules. Reform-oriented rule changes in the charter are not made if representation is unstable. Several reasons may cause such effects. The move to alter rules during competitive periods may be stopped because voters perceive it as unfair. The change may be such that it would give unfair advantage 132 . to one side or the other. This might cause adverse reaction to proposed charter amendments. Competitiveness over representation may also precede charter changes away from the ideal-type. That is, competition over seats may promote change toward more open, unreformed charters in a later period. A lagged reaction may occur. Of all cleavage variables, competition over council representation exerts the most influence on charter changes and depresses the probability of change toward the Ideal. When measured at the mean, these political cleavages explain little change. It is possible that the council turnover variable masks the effects of other cleavages. Once again, it was removed to examine the separate effects of other variables. There was a decrease in model predictive results as the other cleavage estimates responded in a mixed manner. Partisanship influence declined substantially only to be marginally offset by smaller increases in the effects of population, ethnicity and Catholicism. Therefore, we conclude that the council turnover effect seems to have a separate and distinct influence on charter change. The predicted effect of political changes alone (measured at the mean) is that city charters would not change. The fourth proposition specifies one subset of political cleavages as more influential than another subset. Measured at the mean, partisanship, council turnover, and population exert more influence than race, ethnicity, religion, and workforce on the probabiltiy of structural changes. strongest. Of these, turnover is the Partisanship has less effect. This can be due to a lack of strong partisan ties to charter change campaigns, especially if the 133 partisan balance is overwhelmingly one-sided. The partisan factions may also be unwilling to publicly back proposals for fear of voter reprisals in those cities that operate under non-partisan elections. The population changes are generally positive and promote change. However, it should be recognized that population decline also promotes change away from the charter ideal. In this case, population declines in combination with fierce council competition will provoke movement in a negative direction. The fifth proposition hypothesizing annexation effects is not supported by the results. The estimated coefficient for annexation has relatively little effect on charter change. Its level of statistical significance is also extremely low. This suggests that by itself, annexation is not a reason to change charters. The annexed territory is generally not a sufficient voting block to force change on its own accord. It may be brought into regular city politics as would any other new neighborhood or subdivision. It seems plausible, however, that local activitism within annexed areas would be a factor in some change if the area were one of importance to city prosperity and. politics. Any change would have to be sparked by a political entrepreneur within the annexed area. Annexation affects routine political activity more than it does charter activity. The sixth proposition about charter change pertains to the presence of a state-supported college or university. The probability is hypothesized to increase if the city contains a state-supported school within its boundaries. Results for the initial model indicate that the presence of such institutions increases the likelihood of changes in charters. As operationalized, its effect is very 134 substantial. In terms of magnitude, its impact is second only to council representation. This result suggests that university influence extends to charter operations and degree of charter reformism. The influence is primarily related to city culture and academic leadership in municipal government. Academic leaders formed a significant component of the National Municipal League and its work. The Model Charter itself was drafted with the help of many university-educated personnel. The university community, then, has an impact on promoting change. The seventh proposition specifies incremental charter change based on its status. The results suggest mixed effects. In general, charter base and age have the greatest effects on structural change. The base score works as hypothesized in that low scores promote positive change. Charter age, however, tends to reduce the probability of change. Of the interactive variables, modernization has positive effects on the probability of change toward the Model Charter. The next three propositions relate to the direction of charter change. Proposition eight specifies a positive influence for those variables measuring social-economic class. That is, reformed city government is a response to improved class welfare. However, the results presented in Table 2 indicate a mixed direction for class measures. The effects of changes in income and home value is toward unreformed charters. As wealth and income increase, the probability of greater Model compliance decreases. Indeed, if wealth and income increase dramatically, charter reformism is likely to be undone. As potential city resources decrease, the charter is more likely to be 135 changed to conform to the Ideal. Promises of greater economy would seem to have greater appeal to voters. efficiencyand Residency also contributes to this effect but in a negative direction. As residency stabilizes and increases, the charter is less likely to change. As population mobility increases, the charter becomes more likely to change. Mobility is likely to increase as society in general becomes more mobile over time. staying in place. Fewer people are Itwas hypothesized that stability would lead to increased charter reformism because residents would feel a stake in its government. The opposite seems to be suggested here. population will acquire a more reformed charter. offered here. A changing Two reasons may be As one moves across several cities in a lifetime, perhaps he or she would prefer standardized city charters and the Ideal is such a standard. The new population subset, then, would prefer reformed city government. It may also be the case that as residents leave, the remaining residents decide to reform the charter and the newer residents do not participate. Age and education also have positive influences on the charter’s reformism. As both increase, charter change is in the direction of the Ideal. These are constant influences on probability of change toward municipal reform. Demographically, city populations are becoming more educated and older as well as more mobile. All three work in favor of positive charter change toward reformism. Proposition nine represents a test of Schattschneider’s argument about proponents changing the rules to gain new policy advantage. specifies that els It the political variables approach an unbalanced condition, change will occur toward the Ideal Charter. To test this 136 proposition directly, we can review the signs and magnitude of the coefficients. The proposition is generally supported by the results. of Republicanism promotes reformism. The rise This is reinforced if a rough balance of partisanship is achieved as indicated by positive effects of the partisan threshold variable. Lower levels of change in council turnover, percent foreign-born, percent Catholic, and manufacturing workforce also support movement toward reformed charters. The differences lie in results for changes in population growth and non­ white constituencies. Population losses and growth in non-white populations are related to change toward reformed charters. These two differences tend to support the findings from the test of proposition eight whereby increased municipal "stress" coupled with lower income levels provoke changes toward reformism. The magnitude of these coefficients indicates partisanship and council turnover exert the strongest impact on the dependent variable. The turnover rate and percent Democratic population each have large negative effects on charter change toward the Ideal. The negative partisanship variable is moderated by the effects of the partisan threshold coefficient. This suggests that partisan parity promotes movement toward the Ideal type charter. The tenth and final proposition specifies the direction of charter change given the charter’s status. The proposition is generally supported by the coefficient estimates. The charter base score exerts the strongest effects on the likelihood for positive change. Extremely low base scores indicate greater probability of change toward the reform Ideal and higher base scores. Charter age, 137 however, tends to provoke change away from the Ideal. Older charters are less likely to be changed to conform with the Ideal. Weak voter support for adoption increases likelihood of later change and this is away from the Ideal. The modernization effect partly offsets the negative effect of charter age. That is, if a charter is old with a high base score, it is more likely to be updated and revised to conform more closely to the Ideal. The experimental effects suggest a slight decrease in this modernization as it concerns university cities. These cities are more likely to move away from the Ideal if they do change. The next section outlines a parsimonious model. We present this second version estimated to improve the initial version in both summary statistics and predictions. A Parsimonious Model The first part of the estimated final model to be discussed is shown in Table 3. This streamlined model drops the annexation, percent Catholic, change in median years, percent owner-occupied, and median income variables. The annexation variable was dropped from the model because its effect appears to be on the political environment and not the battles fought over rules. is thus minimized. Its effect on charter politics The religion variable was dropped because it is not measuring what it should due to poor operationalization. The population age variable is deleted from the model based on its poor showing and likelihood of its potential impact being captured by other variables in the model. Housing growth, to the extent it captures growth in numbers of young families, would acquire some of the age 138 effects. Housing growth likewise captures some of the change in owner-occupied housing given the likely ownership of single-family dwellings in suburban cities. Finally, the income variable was dropped after experimenting with its interactive effects with other variables. It would appear that wealth measured by home value would pick up some of the income effects. from the model. Therefore, it too was deleted Each of these variables has relatively small effect on the dependent variable and is statistically insignificant. The streamlined model contains a total of eighteen variables plus a constant term. We may now move to a discussion of this final model's estimated coefficients, their standard errors, Student’s t-statistics, and estimated p-values. The coefficients appearing in both models are affected in a minor way by the deletion of the five statistically insignificant variables. None of the variables changes direction. Nine of nineteen variables including the constant are significant at a minimum .05 level. of the others are close as indicated by their p-values. Most The largest estimates in terms of impact on the dependent variable belong to the charter-status subset. The demographic variables exert relatively less influence on probability of change. lies in the summary features. The substantial difference The next section will present the summary statistics for Goodness of Fit before moving on to a substantive discussion and interpretation of the model results. 139 Table 3 Final Version Estimation Results Maximum Likelihood Estimate Standard Error MLE/SE P-value 7.1729 1.6414 4.3700 .00001 Economic Variables median housing -1.1904 years schooling 3.8003 pet residence -0.7015 housing growth 0.0237 0.6460 1.9216 0.4736 0.0129 -1.8426 1.9776 -1.4810 1.8371 .0329 .0244 .0694 .0336 Political Variables pet Democratic -0.0116 Dan threshold 0.1038 population -0.7248 council turnover -1.2778 non-white pop 4.2090 non-white thres -0.3573 foreign b o m 0.2240 pet manufacture -0.2214 university town 0.7790 0.0060 0.1877 0.4153 0.4851 3.5292 0.3284 0.4718 0.6761 0.5004 -1.9423 0.5530 -1.7450 -2.6341 1.1926 -1.0880 0.4748 -0.3275 1.5566 .0262 .2912 .0409 .0043 .1170 .1401 .3192 .3745 .0606 Charter Variables charter base charter age charter support modernization experimental -0.0166 -0.0768 -0.9896 0.0002 -0.0186 0.0055 0.0444 0.7758 0.0001 0.0164 -3.0000 -1.7297 -1.2756 1.4734 -1.1301 .0013 .0427 .1020 .0708 .1292 1.0721 2.6328 0.1092 0.1646 9.8137 15.9929 .0002 .0002 Variable constant Threshold 1 Threshold 2 -2xLLR 162.09 Estimated R-square .270 Pseudo R-square .473 % Predicted Correctly 50.56 140 Comparing Goodness of Fit We can calculate the same three Goodness of Fit statistics used in the initial version for the final one and compare the two results. The first basic test, the log likelihood ratio with 180 degrees of freedom, provides a statistic of 162.090. From the Chi-square tables, we find this is significant at the .001 level. This is small decrease of .935 from the initial version’s ratio and indicates the relative statistical insignificance of the deleted variables. Again, the ratio is similar to using an F-statistic in regression analysis. The second summary statistic comparison is that of the estimated R-square provided by McKelvey and Zavoina (1975). It is calculated to be .270 for the final version and again, is close to that of the original version. The third summary statistic is Aldrich and Nelson’s (1984) pseudo R-square. There is almost no difference here as it goes from .475 to .473. These three Goodness of Fit statistics reveal the final version to be a marginally improved version of charter change. The final model, however, does much better at predicting individual categories of change. From the crosstabulation presented in Figure 8 we see that additional correct predictions are made in each category except the negative category. In the high positive category, the final version predicts five of twenty-one or 23.8% correctly. It also continues to predict one correctly in the lower category of negative change. When we look to the extremes, we see the final model again fails to make predictions far off target. negatives and vice versa. It does not predict high positives to be Of those incorrect predictions, seventy- three of eighty-nine, or 82.0% fall within one cell of their proper 141 ********** Crosstabulation of change by predict ********** predict COUNT COL PCT negative -1 none 0 low pos 1 high pos 2 ROW TOTAL -1 negative 1 16.7 13 27.7 12 9.9 0 0.0 26 14.4 0 none 3 50.0 22 46.8 32 26.4 0 0.0 57 31.7 1 low pos 2 33.3 10 21.3 63 52.1 1 16.7 76 42.2 2 high pos 0 0.0 2 4.3 14 11.6 5 83.3 21 11.7 COLUMN TOTAL 6 3.3 47 26.1 121 67.2 6 3.3 180 100.0 Chi-square(9 d.f.) = 53.296 Figure 8 Final Version Summary Measures 142 place. The total count of correct predictions improves to where the final model shows a majority of correct ones. A total of ninety-one of one hundred eighty, or 50.56% fall within their proper cells. represents a small 3.41% improvement over the initial version. This The best measure of the significance of this model version is its improvement in the Chi-square measure. The Chi-square statistic for this crosstabulation increases substantially to 53.296 which is significant at over the .005 level. This represents a 26.02% improvement in the model’s overall statistical significance. The null continues to be the modal category, low positive change [code 1]. The final version correctly predicts sixty-three of one hundred twenty one, or 52.1% in this cell. This is a 9.9% increase over the null of 42.2%, or a 23.45% improvement. Together with the Chi-square statistic and correct predictions in all four cells, this improvement reveals the final version to be an advance over the initial model. In the next section, we begin our discussion of conclusions by presenting analysis of conditions under which cities will be more likely to change their charters. This is performed by constructing Z- scores out of the estimated coefficients and comparing them to a probability distribution. Municipal Cases As Illustrations of Change An advantage to using a probit model estimation technique is its provision of coefficients best interpreted as additions to or subtractions from a final product. This product acts as a Z-score which, when compared to a standard normal probability distribution, 143 will give us an indication of probability of change. hypothetical municipal cases are presented in Table 4. Nine We can project the likelihood of change by varying the independent variables. Usually, the independent variables are measured at their means but we are not bound to this or confined by it when developing our cases. In fact, we deviate from it in a couple of cases to check on the possibility of radical charter change. We now move to a presentation of each of the nine cases. The first case in Table 4 is one in which all continuous level variables are at their mean values. This excludes the dunmy variables of partisan and non-white population thresholds, the university town effect, as well as the interactive variable experimental. The total Z-score is calculated to be 1.17, or just over the threshold of low positive change [1.072]. Using a standard normal probability distribution for calculations, results show a probability of .4638 for low positive change. The chance of no change at all is relatively high with a probability of .3431. The related probabilities for other changes are presented in the lower part of Table 4. These probabilities for Case 1 indicate that on average, cities will change their charters toward the ideal-type but only incrementally. This conclusion is tentative, however, given a strong alternative possibility of showing no change. Other political and demographic conditions may depress the chances that a charter will undergo change. The difference between change and no change in this average case is only one of .098 in the Z-score. With such a low margin, most of the other variables can shift to lower the prediction to one of no change [code 0]. 144 Table 4 Hypothetical Municipal Cases Variable #1 constant 7.172 #2 #3 #4 #5 #6 #7 #8 #9 Economic Variables heme val -0.108 school 0.250 residen 0.124 hous gwth 0.399 Political Variables pet Dan -0.567 Dem thrs — 0.103 populat 0.077 turnover -0.626 non-whit 0.052 n-w thrs — -0.357 for b o m -0.034 manufact 0.029 univ — Charter Status Variables ch base -4.385 age -2.147 support -0.685 modem 1.773 experim Z-score* Prob P(-l) P< 0) P( 1) P( 2) 1.177 .1210 .3431 .4638 .0721 .813 1.273 .2090 .3897 .3662 .0351 .1020 .3187 .4908 .0885 0.103 0.103 -0.357 -0.357 0.779 0.779 0.779 -3.104-5.66 -0.071 1.878 .0307 .1812 .5615 .2266 ♦thresholds are 0, 1.072 and 2.632 1.283 2.341 -0.071 -0.071 .916 1.521 1.981 3.734 2.236 .1814 .3782 .3968 .0436 .0643 .2657 .5365 .1335 .0239 .1602 .5581 .2578 .0001 .0038 .1318 .8643 .0129 .1101 .5287 .3483 145 The second case reveals just such an instance. This case shows the change in the Z-score when the non-white population increases beyond the threshold. When it crosses the threshold, it reduces both the Z-score and the probability of change in either negative or positive directions. The probability is not, however, clearly distinct from the low positive category. The probability of low positive change falls from .463 to .366. The probability of no change, the new prediction, increases from .343 in the first case to .389. A sizeable increase is also made in the probability of negative charter change as the associated measure moves from .1210 to .2090. Once again, we can imagine cases when changes related to increases in the non-white population would support movement further from the ideal-type charter. Blacks, for instance, are usually Democratic voters (Abramson, Aldrich, and Rohde, 1987). If the non-white population growth threshold were crossed, the increase would also be likely to increase the size of the Democratic voting block to further reduce the Z-score. The non-white population is also associated with relatively less formal schooling, lower home values, and "white" flight. Each of these changes, in combination with the variable threshold, would work to further reduce the Z-score and provoke change away from the Model. The third scenario shows that the partisanship threshold has an independent effect. We see the Z-score increase by .103 to 1.273. places the probability of change more squarely into the low positive category. This reveals a tendency for both partisan factions to settle on a pro-reform charter. points to partisan compromise. This particular effect of parity There is almost a fifty percent It 146 probability that the charter will change in a low positive fashion given these political and demographic conditions. The fourth hypothetical case is similar to the first case but includes the effects created by presence of a state-supported university. This also includes the effects of the interactive variable, experimental. Here we see the Z-score increase substantially from 1.17 to 1.878 for a net gain of .708. In probabilistic terms, the chances of low positive change increase to .5615. The probability of high positive change under these conditions increases to .2266. This suggests that the university has a great deal of impact on local charter politics. Its influence is in the direction of greater reform toward the ideal-type. This occurs even in the case of a charter with a relatively high initial base score. The next three examples show the effects of adding the dummy variables in combinations of two. The Z-scores fluctuate from .916 to 1.981 due to the conflicting signs of these variables. In the fifth case, it becomes difficult to distinguish a clear pattern of change. The probability between low positive and no change is less than two percent. The sixth and seventh cases show increasing probability of positive change as the dummy variables tend to reinforce each other. Note, however, that in the seventh case, the probability of high positive change is increased to .2578. The last two hypothetical scenarios illustrate the strong impact of the charter base index score on the probability of change. We vary these scores by two standard deviations from the mean to show the upper and lower limits of its effect alone. Certainly the supporting status measures as well as demographic and political variables can 147 help influence the final Z-score. But the variance of charter base has an independent effects on probability of charter change. In the eighth case, the base score was decreased by two standard deviations. Even with this reduction, the score was still twenty points higher than the lowest score actually observed. This reduction is partially offset by the interactive variable modernization which is calculated by multiplying the age with the base score. The Z-score increases to 3.734 and clearly drives up the probability of high positive change. The probability of change occuring in the top category increases to .864. Note that the probability of the incidence of no or negative charter change effectively drops to zero. When the base score is increased by two standard deviations in case nine, we see that the other variables still produce a probability of low positive change. From this we conclude that the base score seems to have the capacity to swing charter reform to great lengths if such "unreformed" preconditions are met. What is surprising in these last two cases is that even without a high index score, the probability of negative change never rises to great heights. This suggests that once reformed, a charter is less likely to be undone without very significant political and demographic changes in the population. This may be likened to a kind of ratchet effect as marginal positive reform continues throughout the study period. With these illustrations as a guide, we can now present a summary of the dissertation’s major findings and substantive conclusions in the next chapter. NOTES 1. The data analysis was conducted by using Dubin, J.A. and Rivers, R.D. Statistical Software Tools (SST) Version 1.0. Pasadena, CA: Dubin/Rivers Research. The probit calculation follows those provided by McKelvey and Zavoina (1975). 2. The distribution of the dependent variable was altered by lowering the index cut off value from seventeen to ten. A value of ten represents a change of two high-ranked charter characteristics. This alteration reduces the number of low positive observations while increasing the number of high positive observations. The initial model was estimated using this altered dependent variable. Results were mixed. The estimated R-square was .258, the pseudo R-square was .479, and the percentage of correct predictions was 40.56%. Compared to the final version reported in Figure 8, the Chi-square statistic of the altered model fell to 37.366. These results suggest that the index supporting the dependent variable is not a very accurate point estimator of charter change. Future index refinements may be able to increase its utility as a point estimator. 4 148 CHAPTER VI CONCLUSION This final chapter of the dissertation presents and discusses a substantive summary of the model estimates. several parts. It is divided into The first part presents four basic arguments and concludes with a fifth point in partial support of Cornwell’s observations regarding study of municipal constitutional change. A second part highlights three general observations as an overview of municipal structural change. The third part attempts to generalize our findings to other places and times. Finally, in judging our results, we can point to areas requiring further analysis. The theoretical implications of our model require additional testing within the context of a model of not only aggregrate but also individual level action. If rules are indeed important for shaping political outcomes, the individual voter’s intent can and should be examined. Summary of Model's Substantive Performance This section outlines four basic arguments and concludes with a fifth point regarding the study of constitutional change. The first substantive point concerns the direction of charter changes in relation to the National Municipal League’s Model Charter. Home rule cities in Michigan continue to bring their municipal 149 150 charters into greater conformance with the League’s ideal-type. Charter change is overwhelming in the direction toward "reformism." Maser asserts that municipal charters will undergo severe change away from reformism if population preferences so dictate. Hannah’s study of Michigan cities also follows this tack by arguing that cities under "stressful conditions" will undo reform. These conditions for negative change are included in the model and even so, these conditions do not provoke negative change. Very little change away from reformist direction takes place even as population preferences can be assumed to support it in several cities. This divergence may be partly explained by Maser and Hannah’s operationalization of their dependent variables as well as Hannah’s lack of a control group. The charter as operationalized in this dissertation includes many more features than does Maser’s CM) index. This increases the sensitivity of the index to change and adds more support to concluding its positive directions. Including a feature for measuring lack of change also improves the predictive performance of our model over Maser*s. Hannah’s conclusion regarding the balance of power is still basically supported here. The Model Charter includes features that describe a strong council as well as strong executive. Both are central to the League’s position on city structure and powers. When Hannah notes that some cities strengthen the council as well as the executive in the face of adversity, this permits more representativeness and responsiveness contained within the Ideal-type document. Some negative changes do occur, and these appear as mostly 151 "corrections" to compensate for initial charter revisions. None of the cities sampled here ever reach full compatibility with the Ideal. If a charter is changed to greatly approach the Model, it is very likely to undergo some negative change. It remains comparatively "reformed" and does not add significant change in the overall trend toward greater "reformism" across the board. The second finding relates to Maser’s argument that "political demography follows population demography." This was tested in the context of the dissertation by utilizing aggregrate population measures as proxies for population preferences. If we continue to assume that this kind of measure is accurate, then results show Maser's argument to be wanting. Political demography does not necessarily follow population demography. We can see this through an examination of demographic movement and corresponding charter reaction. By themselves, the population variables do not explain much charter change. This is due primarily to the mixed direction each separate variable change takes. Recall that our initial hypotheses proposed that these variables have similar effects on charter change. For example, the class concepts of home value and income would support and reinforce the effect of education and age. However, we find that the effect of these variables are in directions. Their combined effect on charter change is therefore minimized. As a whole, they do not promote significant change in municipal political structure even if all were consistent. There are two qualifications to this finding. First, as we see from the first scenario in Table 4, mean changes in demographic 152 variables promote minor change in charters. An aging and increasingly educated population helps to promote the general trend of movement toward the Model Charter. For all cities, then, population demography adds some support to political demography. Second, this support can be dramatically increased to help provoke major changes. That is, when population changes take paths that minimize negative effects and maximize positive effects, they may add such force to existing political factors that great changes to the charter may be made. As we see from the eighth scenerio, extreme charter-status scores alone can barely breech the upper threshold to gain greater changes in the municipal structure. Certainly the right mix of demographic factors may help promote greater change. Population demography may make a small difference. These scenarios are not difficult to imagine. For example, the educational levels of the population continue to increase as does its age. Age is perhaps more certain to increase as suburbanization slows in several cities. The residential longevity of some cities may also decrease as "gentrification" processes begin to occur. This is especially likely in older suburbs where original settlers are into retirement years. We may expect new preference replacements. factors would combine to promote change. These This kind of replacement would also tend to lower the combined economic status as people on the top end of the income level are replaced with presumably younger and less affluent residents. These kinds of changes with a relatively unreformed charter would promote change in the political structure. The conclusion we make here is that in general, charter change does not result solely or even largely from change in the population. 153 The preferences of such cities may change, but this would be over time and at the margins. According to Wildavsky’s explanation of preference origin, the newer segments of the populations would mix in with the old. If preferences are a function of shared experience, a sizeable change over time would have to take place and this "newly" composed population would have to share similar municipal-political events for new charter-specific preferences to emerge. This explanation helps justify the conclusion here that by itself, population demography will not cause charter change. Such demographic changes as may occur would be marginal and "shared" population experience would moderate its immediate effect. The third point we may conclude from these results is that knowing the "status" of a municipal charter will tell us more about the likelihood of its change that simply knowing city demographics. When the variables measuring such characteristics as the charters’s age, its margin of voter support, and most importantly its prior degree of reformism were added to the probit model, its prediction improved dramatically. The overall trend institigated by demographic change toward reformed charters was accelerated by its comparative relationship to the Model Charter. While it is difficult to imagine the status of a charter "causing" rapid change, it is easier to picture a process of "fits and starts" when a dormant, unreformed charter is suddenly altered to closely conform with the ideal-type. The gently upward trend of positive change includes many occasions of extreme change. If the direction of change is positive according to demographic movement, and if the charter is relatively unreformed, then those variables of 154 charter status become vital to predicting future change. One must recognize, however, that population changes in the right direction and maginitude are necessary for pushing change into the upper limits. Some "modernization" occurs in charters of cities like Detroit and Hamtramck but radical reformist change requires supporting demographic shifts. Charter "status" variables provide a more substantial base for predicting change than do population variables alone. A fourth point that is quite surprising is that a "non-political" atmosphere seems necessary for charter revision and change. That is, charter change in any direction must appear bipartisan or apolitical to be successful. Two findings support this conclusion. First, fighting over council seats indicates a highly politicized atmosphere. Competing sides may take their turns putting one another out of office. The political battle is over control over existing municipal structures. With more council seat fighting, the charter has less likelihood of being changed. This seems to suggest that rule change in mid-battle is unfair and will not be adopted. A calmer atmosphere where political lines are restrained would have to prevail before change is adopted. Partisanship activity also supports this conclusion. From the results, we see that a Democratic dominance of city partisanship dampens the possibility of change. The more Democratic a city voting block, the less likely it is able to adopt reformed charters. Conversely, the more Republican a city voting population, the greater its likelihood of change. issue. Reformism thus tends to be a Republican This observed pattern in our model supports other scholarly observations (Williams and Adrian, 1963) The reform agenda is one 155 more in time with upper class business interests (Judd, 1984). interests are also usually Republican. These A rough class division may underlie this partisan alignment if one assumes partisanship is based on class. Williams and Adrian make this assumption in their study of four Michigan cities. When measured by a base vote, Republicanism leads to reform while Democrats oppose it. Change in the partisan alignment, however, does lead to charter reform if parity between parties occurs. That is, as the voting population approaches a rough balance between partisan voting power, city charters will be changed toward the Ideal. A bi-partisan balance is struck which allows change to occur providing the electoral fighting ceases. If such balance had negative effects, one would expect that the rule change would be further from the Ideal-type. But the bipartisan effort provides a chance for either party to veto change. If rule change strengthened one or the other partisan sectors, its opponents could muster a vote to prevent its adoption. Schattschneider's claim regarding changing the rules is supported in the context of this model if such change is in effect "non-partisan". It is more difficult to explain the direction of change rather than its occurrence when parity exists. If Democrats oppose reforming charters, then what compels them to change once parity exists? Parity, in this case, occurs as either the Republican or Democratic voting block increases relative to the other. One clue may lie in the partisan position on reform when one or the other side is dominant. oppose it. economy. Republicans promote reform; Democrats Republicans promote it on the basis of efficiency and Democrats oppose it because they value responsiveness and 156 representativeness. When parity is achieved, however, the goals hoped to be achieved by reform may change. Each partisan faction may see reform in a different light given new alignments. They may view and emphasize reformed charters for different reasons and intent. Each side may see reformed charters more in terms of relative control and insulation from partisan fighting. Each side may be willing to compromise by removing the administration from politics and strengthening it. Given the characteristics comprising the index, reformism may also mean a corresponding strengthening of the electoral devices for government control — council powers and direct democracy techniques. The parties may see that it is to their advantage to promote a reformed city charter to protect whatever advantage they consider important. The Republicans may strengthen the executive branch while the Democrats may strengthen the council and electoral devices. If parity meant electoral conflict and changing rules for advantage, then a negative sign would have been associated with it. case. This is not the Parity results in compromise. The last set of findings we infer from our model estimation support Cornwell's argument on the uniqueness of constitutional change. Constitutional change, he writes, is an individualized contest that is best predicted on the basis of unique characteristics surrounding the proposed change (1985). Each charter contest is a unique event affected by media coverage, personalities, and differing levels of technical complexity. If this were absolutely correct, then no overall trend could be predicted. general trend. Our results do indicate such a The model's missed predictions within the overall 157 trend, however, lend credence to unique interpretations of charter change. The population characteristics do not explain general trends without some consideration of the charter status. greater effect on predicting change. The status has Yet it appears that an overall trend toward the Ideal Charter is stimulated by unique sparks in city politics. That is, demographic changes need some kind of political expression if charter status conditions are met. Once these preconditions of population change and status are met, charters can be changed and some even radically. However, as Cornwell notes, a change needs to be articulated by political groups willing to promote such change. These "sparks" for reform are not directly picked up by this model although demographic and political variables may indirectly measure them. The charter change needs supporting groups willing to champion reform in a non-political environment. To that extent, charter change can be cast as mostly non-partisan campaigns. It would be difficult to incorporate such individualized, unique conditions into such a model as estimated here. The use of aggregrate, cross-sectional data prevents this. Our analysis does work to identify those related factors most likely to act as "kindling" for the "sparks" of charter reform campaigns. General Conclusions There are three general statements we can make about municipal charter change after comparing our bodies of reviewed literature and dissertation results. First, in response to Maser and Heckathom’s (1987) argument, city charter change is not strictly a response to 158 population demographic changes. One finds that population preferences do not completely determine either the direction or magnitude of changes. They do, however, contribute some important effects at the margins of change. In this sense, Bryce (1906) is correct in claiming that constitutions respond to changed societal conditions. not, however, a deterministic relationship. This is Such a relationship would depend on much more than voter structural preferences neutrally and absolutely transformed into political reality. Second, while the status and condition of the charter document indicates something about the need for revision and updating, it does not solely determine a city’s likelihood of change. Examining an overtly political environment contributes to an understanding of a balance of forces supporting the rule structure. These forces are guided and influenced by their charter’s present status and what is largely accepted as an ideal. Revised theories and plans of local government continue to appear and add impetus for change. The current revision of the Model Charter, and the fact that it has undergone several changes since its inception, illustrate this concept of a moving target, the ideal-type. Even so, the Model and its supporters have consistently upheld several principles embodied within it. Non- partisanship, for example, is fairly well entrenched in the reform ideology and will not easily diminish in intensity. continue to uphold strong administrative leadership. Any Model will At the margin, however, theoretically specified and institutionally recognized changes occur in the ideal-type which bring actual charters nearer or further to compliance. Incremental changes in the ideal-type make actual charter status an important part of assessing the need for 159 revision. The balance of forces are presumably aware of both the ideal and present status. These political and theoretical implications for charter status, while appearing to exert great influence on the likelihood of change, do not contribute entirely to determination of its occurrence. Thirdly, we can successfully argue that in relation to am idealtype charter, several trends of change do emerge. Charters are changing over long periods of time and these changes are not random. Overall, the pattern of changes fall in the positive direction when contrasted with the Model Charter. Cities are becoming more uniform in their charters as they amend them to encompass ideal-type provisions. Within this general positive pattern emerges a negative movement by some cities. Some of these cities are very high in charter status rankings. They are in effect "over"reformed and choose to "undo" some charter provisions. Other cities are undergoing stress and also seek change in their present charter away from the Model. Yet on the whole, cities with moderate degrees of population change and current difficulty with generating revenues are seeking moderate change toward the Ideal-type. The promise of greater efficiency and economy seem appealing for those cities with relatively poorer conditions. These three general observations are made on the basis of the models’ results and success in predicting charter change on an aggregrate basis. Individual city analysis would undoubtedly raise different general observations given case-specific circumstances and voter alignments. 160 Major Conclusions The results of testing the explanations of change in municipal charters derived from fairly distinct theoretical traditions leads us to present five major conclusions. These relate to the reviewed literature and theoretical orientation. The first conclusion is that while it is not a sole determinant of predicting structural change, change in population demography is a necessary precondition for it. If a city has a stable population without any change of influx on new residents, we can expect very little change. On the other hand, if a city is following an average trend with respect to change in its demographic character, then the likelihood of change increases. If there is a lot of change, the probability of change increases as well. This signals an effect in the shared preference for city structure that underlies the support for a particular set of rules for local government. With Wildavsky’s explanation, shared experience and a stable population can promote change. However, the influx of new residents is almost a given in today's society. Such change works to promote efforts at obtaining structural alternatives. Secondly, we find the trend of change is toward the 1964 edition of the National Municipal League’s Model Charter. It remains to be seen, however, if the new seventh edition of the Model will retreat from its present level of reformism. This raises the question that the Model itself is too "reformed." Like cities in Michigan over the last several decades, the Model too may retreat on some of its more extreme positions, especially on preferred representation and the Mayor’s role. 161 The third conclusion is that population changes need political expression. The demographic change in a city will not in and of itself "cause" charter amendment. The amendment process demands a petition for change or referral action by the city council. In either proposal method, the amendment must be ratified by a majority of the residents voting on the question. The very fact that an amendment requires a referenda election should illustrate the "political" ramifications of charter change. The political structure does not automatically react to any and all population change. Preferences need to be formed as well as articulated in these kinds of political situations. This is best done by a political entrepreneur, a person or group of persons willing to initiate the process of formal charter change. Once we open the door to "political" initiation, one is struck by Cornwell's valid objection to aggregrate measure of change. Each election result can be explained in terms of the special conditions surrounding it. Cornwell suggests that style and symbolic issues activate voters to participate and make choices. The "contagiousness" of an election follow's Schattscheneider’s description of a political conflict (1960, p.2). In such cases, the outcome is determined less by analyzing the entire population preferences, but rather only some segment involving itself with voting — preferences via the ballot box. that is revealing their The outcomes of electoral reform battles can be treated as Salisbury did that of St. Louis in the 1950's and 1960's. Political entries and their relative voting strengths thus determine an electoral outcome and in this case, structural change. This is not to say, however, that trends cannot be 162 found, measured and defended as accurate. Among the overtly political contributions to change measured on an aggregrate level, we found charter status to be most influential in projecting the likelihood of change. These indicators reveal the amount of "slack" necessary for structural modification. This slack is based upon an idea of an upward or positive trend toward achieving the ideal-type. Several cities will be below or above this trend, a moving average of reformism. Thus, given certain population changes and the status of a charter, in fits and starts a city will move to amend its structure. The extent of change becomes in part a function of how close the city is to the ideal-type charter. A highly reformed charter will not be as likely to undergo positive change as would the city with a highly unreformed charter. Charter status incorporates this concept of slack and relative proximity to the ideal-type. Finally, the political environment of a city is important to judging the potential for structural change. The two most significant and overtly "political" factors in our model were partisanship alignments and council representation. From our results, we conclude that partisan alignments help determine the odds of reforming the charter. When one partisan side or another dominates the city voting population, it is generally likely to either promote reform if it is Republican or oppose reform if it is Democratic. When partisanship change swings the opposing sides closer to parity, the balance generates some change toward reform. While difficult to say exactly why this occurs, it appears that bi-partisan coalitions may form to insure some mutual advantages in city structure by altering parts of the rules, all of which may conform to the Model. 163 This bi-partisan effort at positive reform is conditioned by the requirement for some stability in council representation. A high degree of turnover in local political office may substitute for issues of "proper" local structure. amendment. charter. It decreases the likelihood of A majority promoting change cannot form to amend the This representation effect is related to partisan balances if such balance determines the outcomes of competition over council seats. In this way, each conditions the probabilties and direction of change. To summarize our generalizations, municipal structural change occurs in a positive direction toward the Model Charter. It is promoted partly by change in constituent demography and partly by a population’s political character. to accomodate change. The charter itself must have slack The change that does occur is the result of political entrepreneurs articulating or giving expression to some but not necessarily all underlying population conditions. Future Research The findings of a trend or pattern of change and political and demographic conditions underlying it are linked to the results of electoral battles over charter referenda. It can therefore be difficult to analyze aggregrate level data to judge the relative effects of socio-economic and political preferences. Indeed, change in a charter may not be made toward an optimum structure based on population preferences but may only be a compromise by some faction of a concerned voting population. This type of theoretical perspective flows naturally from the dissertation results presented here. It 164 requires a different level of data with the individual as the unit of analysis. The assumption in this dissertation is that population preferences would dictate subsequent structure and its change. Further research to test the relationship between preferences and charters should be conducted within the context of a charter revision. The individual voters and not city would become the decision unit. The methodology contained in this dissertation would be readily adaptable to such a test. The analysis would be designed according to a two-stage model. The first equation would specify the conditions under which an individual would participate in changing a charter. The second would then specify a supporting or unsupporting vote cast on the basis of preferences. Within such a model, one could refine the conclusion made in this dissertation about relative strength of preference influence and define more exact demographic, socio-economic, and political preferences. In the absence of such analysis, aggregrate analysis like that conducted here could be performed to check on the relationship between newer and older charters. growing. The pool of newly incorporated cities is As their charters are formed and adopted, they too can be assessed in relationshp to the Model Charter. It may be that older cities may be following the lead of these new cities, many of whom are wealthier and more economically viable. Given the results of our model, such newer cities would be further from the ideal-type. Perhaps these are the examples older cities try to emmulate with their charters, and not the Model itself. The change in the 1964 Model Charter edition may establish an 165 ideal-type with closer proximity to most of the existing charters. The effects in direction and magnitude would change if this occurred. Revision of the Model would promote an analysis of particular new features contained in it as well as their relationship to actual city charters. In a broad sense, rules were assumed to be important in their effects on local politics. Alternative sets of rules would mean alternative political outcomes. This relationship can be tested by linking the municipal charters studied here to their policy outputs contained in the U.S. Census of Governments. These censuses exist for 1958, 1962, 1967, 1972, 1977, 1982, and 1987. The estimation of decennial census information for these years can be provided by various statistical techniques and assumptions. Using these sources of data, one could vary the charters to discover their effects on policy output. expenditures. measure, its Outputs in this case would require measurement using While this is not the best kind of policy output use is more palatable when utilizing aggregrate level data for several dozen cities. These sources and analysis of charter effects would permit the introduction of a model based on simultaneity. The model results in this dissertation do not control for simultaneity and yet it is theoretically of interest. Wildavsky’s discussion implies that preferences would not only result in charter change but that such change would also affect preferences. As a population changes its charter and presumably its policy output, does the charter and city policy then react on the population? The charter and policy may work to promote exodus of certain constitutents while attracting others. A 166 simultaneous equation model can estimate this new level of complexity, opening up the possiblity of testing a host of additional propositions on the causes and consequences of charter change. In this dissertation, the dependent variable consisted of the categorized index scores for the entire charter. However, just as certain population changes seem to gamer political expression, various dimensions of the charter may be more responsive to particular changes. That is, the Model Charter as a whole includes alternative dimensions. The administrative and representative strengths were two highlighted in the literature on change. Other segments of the model may prove sensitive to the demographic and political changes. Further testing of the dimensions of this index could more accurately measure influence on structural change. The theoretical development posed here is not refuted by the dissertation’s results. Neither tradition is judged as correct. do refute several propositions. Some also prove inconclusive. We Here we suggest a refinement of the theoretical connections between city populations and their structure of government. This will lead to new, more rigorous propositions hopefully incorporating simultaneity and individual expression of preferences. These propositions with more refined data will reveal a fuller picture of municipal political structures and its complementary influence on local politics. APPENDICES APPENDIX A Charter Characteristics # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Charter Characteristic Council constituency (% at-large) Council size (7 members at-large) Eligibility (qualified voter) council Filing date for candidacy Eligibility (qualified voter) mayor Staggered terms Council four year terms Councillor resignation required before campaigning Odd year election Term beginning date Compensation for councillors Compensation for mayor Non-partisan election Mayor selected from council Mayor four year term No other elected offices Non-interference clause - administration Non-interference clause - appointments Anti-nepotism clause Conflict of interest clause Filling vacancy, special election w/i 120 days Appoint clerk by council Appoint treasurer by council Appoint assessor by council Assessor is individual Independent Audit requirements Meeting once month Ordinance effective date w/i 30 days (for Ref.) Emergency ordinance requires 5 votes City manager Assistant city manager appointed by city manager Manager makes all city employee appointments Manager makes all department head appointments Manager (executive) makes budget Manager directs all administrative activity Council creates departments Merit system by ordinance Employee compensation by ordinance Attorney appointed by manager Comprehensive plan and advisory planning commission Figure 9 National Municipal League’s Model Charter (6th ed.) 167 168 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 Budget Includes capital program Nomination methods Primary nomination # signatures Petition nomination # signatures Initiative (except budget, cap pgm, approp/tax/sal ord) Referendum (except budget, cap pgm, approp/tax/emer ord) Circulate by committee, affidavit Initiative # signatures = 15% register to vote Referendum # signatures Referendum within 30 days Referendum election timing Initiative unlimited time Initiative election timing Mayor item veto Mayor ordinance veto Mayor vote Number of departments Number of boards Board of Review composition Board of Review size Board of Review compensation Board of Review assessor Board of Review executive Retirement system Contract bid limitation Property tax rate City income tax Fiscal year date Number of tax collection days Incidence of dedicated revenues Council property powers Special assessment annual work plan Special assessment installments 25% limit on special assessments Council create special assessment districts Judicial residency Judicial office requirements Judicial compensation Judicial terms Court jurisdictions Court systems Constable Constable terms Figure 9 (cont’d.). APPENDIX B Characteristic Coding Coding Characteristic 1. Council constituency {%) 2 if 100 % at large 1 if >0 and <100 % at large 0 if 0 % at large 2. Council size 0 1 2 3 if if if if 3 members 5 members 7 members >7 members 3. Council residency 0 1 2 3 if if if if four year or more three year two year one or no years 4. Filing date 0 if 0 - 1 4 days prior 1 if 15 - 28 days prior 2 if > 28 days prior 5. Mayor residency 0 1 2 3 6. Elect majority 1 if yes 0 if no 7. Council term 0 1 2 3 8. Council resignation 0 if not required before running 1 if required before running 9. Election timing 0 if fall/even year 1 if fall/odd year 0 if otherwise 10. Term beginning date if if if if if if if if four or more years three years two years one or none years one year two year three year four year 0 if within 30 days of election 1 if after 30 days of election Figure 10 Charter Characteristics and Coding 169 170 11. Council compensation 2 if part-time $0 - $1040 year 1 if part-time > $1040 year 0 if full-time 12. Mayor compensation 2 if part-time $0 - $1040 year 1 if part-time > $1040 year 0 if full-time 13. Election type 1 if non-partisan 0 if partisan 14. Mayor selection 1 if appointed 0 if elected 15. Mayor term 0 1 2 3 if if if if one year two year three year four year 16. Elected officials (# other than the mayor and council) 3 2 1 0 if if if if 0 elected officials 1 - 3 elected officials 4 - 6 elected officials > 7 elected officials 17. Non-interference admin 1 if yes 0 if no 18. Non-interference appt 1 if yes 0 if no 19. Anti-nepotism clause 1 if yes 0 if no 20. Conflict of interest 1 if yes 0 if no 21. Council vacancies 2 if by election 1 if by limited appointment 0 if by appointment 22. Clerk selection 1 if appointed 0 if elected 23. Treasurer selection 1 if appointed 0 if elected 24. Assessor selection 1 if appointed 0 if elected Figure 10 (cont’d. ). 171 25. Assessor 1 if individual 0 if board/comnission/council 26. Audit requirements 0 if no requirement 1 if yes, by city dept/officer 2 if yes, independent of city 27. Number of council meetings 3 2 1 0 28. Ordin. effect w/i 30 days 1 if yes 0 if no 29. Emer. ordin. req 5 votes 1 if yes 0 if no 30. Chief Executive officer 1 if appointed 0 if elective 31. Asst exec appt by exec 1 if yes 0 if no 32. Employee appointment 2 if by chief exec,brd or comm 1 if by executive with council 0 if by council 33. Department Head appointment 3 2 1 0 34. Who budgets? 2 if appointed 1 if mixed appoint and elect 0 if elective 35. Mgr directs all admin 1 if yes 0 if no 36. Department creation 1 if by council 0 if by charter 37. Civil service 0 if no provision 1 if set by charter 2 if set by ordinance 38. Employee compensation 2 if set by council 1 if set by civil service 0 if set by executive if if if if if if if if set by ordinance 1 - 2 per month 3 - 4 per month > 4 per month 100 percent appointed 100 - 66 percent appointd 65 - 33 percent appointed 3 2 - 0 percent appointed Figure 10 (cont’d. ). 39. Attorney selection 0 if appoint by council 1 if appoint by executive 40. Plan & Advis Plan Conm? 1 if yes 0 if no 41. Budget include Cap Pgm? 1 if yes 0 if no 42. Nomination 1 if petition 0 if election (primary) 43. Primary nomination 0 if 1 to 50 signatures 1 if 50 to 100 signatures 2 if >100 signatures 44. Petition nomination 0 if 1 to 100 signatures 1 if 101 to 250 signatures 2 if >250 signatures 45. Initiative subjects 0 if unlimited 1 if limited 46. Referendum subjects 0 if unlimited 1 if limited 47. Circulate petition req? 0 if no 1 if yes 48. Initiative signatures 0 1 2 3 if if if if no initiative > 25% of voters(last elect) 15% - 24% of voters 0% - 14% of vtrs 49. Referendum signatures 0 1 2 3 if if if if no referendum > 25% of voters 15% - 24% of voters 0% - 14% of voters 50. Referendum time limit 0 1 2 3 if if if if 0 - 10 days 11 - 21 days 21 - 30 days > 30 days 51. Referendum election 0 if within 10 days of certification 1 if within 1 1 - 2 1 days of certific 2 if > 21 days of certification Figure (cont'd.). 173 if if if if 0 - 30 days 31 - 60 days 61 - 90 days > 90 days 52. Initiative time limit 0 1 2 3 53. Initiative election 0 if within 30 days of certification 1 if within 31 - 60 days of certific 2 if > 60 dayB of certification 54. Mayor item veto 0 if yes 1 if no 55. Mayor ordinance veto 0 if yes 1 if no 56. Mayor votes 2 if always 1 if tie 0 if never 57. Number of departments 0 if 4 departments 1 if 5 to 7 departments 0 if > 7 departments 58. Number of boards 2 if board of review 1 if 2 - 3 boards 0 if > 3 boards 59. Board of Review composition 2 if 100 percent citizens 1 if mixed 0 if 0 percent citizens 60. Board of Review size 3 2 1 0 61. Board of Review compensation 0 if compensation 1 if no compensation 62. Board of Review assessor 0 if assessor sits of Board 1 if otherwise 63. Board of Review executive 0 if city executive sits on Board 1 if otherwise 64. Retirement system 0 if no provision 1 if set by ordinance 0 if set by charter if if if if 3 5 7 > members members members 7 members Figure 10 {cont'd.). 174 65. Contract bid limitation 0 1 2 3 if if if if 66. Tax rate 0 if 0 - 1 0 mills 1 if 1 1 - 1 5 mills 2 if 16 - 20 mills 67. Income tax 0 if yes, city has income tax 1 if no income tax 68. Fiscal year 0 if with state/national 1 if independent of state/national 69. Tax days 0 if once a year 1 if twice a year 70. Dedicated revenues? 0 yes 1 otherwise 71. Property buy,sell 0 if no power 1 if yes, limited 2 if yes, unlimited 72. Special assessments plan 0 if no annual plan 1 if annual plan 73. Special assessments payment 0 if no installment payments 1 if installment payments 74. Special assessments limit 0 if 25% limit on spec assess 1 if no limit on spec assess 75. Special assessments create 0 if citizen petition 1 if council may create dists 76. Judicial residency 0 1 2 3 77. Judicial office requirements 1 if no requirements 0 if requirements 78. Judicial compensation 0 if fees (part-time) 1 if salary (full-time) if if if if $0 to $500 $501 to $1000 > $1000 amount set by ordinance four or more years three years two years one or none years Figure 10 (cont'd.). 175 79. Judicial terms 0 if two years 1 if four years 2 if six years 80. Court Jurisdictions 0 if $0 - $500 1 if $501 - $1000 2 if > $1000 81. Court system 0 if Justice of Peace 1 if Municipal Court 82. Constable 1 if no constable 0 if yes 83. Constable term 0 1 2 3 if if if if one year two years three years four or more years Figure 10 (cont'd.). APPENDIX C Characteristic Ranking 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 H H M L M H H L L L H H H H H H H H H M M H H H H M M M M H M M H H H M H H M M L M L M L M L M M H H M H H M H H M M M H M H M M M H M M H M H M H H M L L L M H H H L M M H H H L H H M M M H H H H H L L L H H H H H H M H H M M L M H H L H M M M H H H H H H H H L H H H M H M M M H H M H H H H H H H M H M L H H H H M M H H H M H H H H H H H H M H H H H H M H H H H H L Figure 11 Characteristic Ranking and Weights 176 H M M L M M H L H L M M H H H H H H H H M H H H M H M M M H M M H H H M H Weight OlUOlOlOlW W -JW UUW W UIW OlW OlW OlOlW OIUlOlUW I-iW HOltdOSHW W W ----------- Panel of Experts -----------Char. VerBurg Press House Thomas Patriarche Final Rank # 177 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 H M M L L L L H H M M M M M M M H H L M H M L L L L M M L M L L M L M M M L M M H H M M L L Panel of Experts -----------Press House Thomas Patriarche Final Rank M M M L L L L M M L L L L L L L H L M H H L L L L L L M H M L L M L L L L L L L L L L M L L L H L L H M M H H M M M H H H L H H H L L L L L L L L L H H L L L L L L L H L L L H L L L L H M M H M L L H H M M M M M M M M H H M H M L L L L M H H H L H M H L L L L M H H H M H H M Figure 11 (cont’d.). L M H H H H H H H M H H H H H H M M H L L H H L L H H M H L H H L H H L H L L L L L L L L L M M M L M L L H H M M M M M M M H H H M H M L L L L M M H M L L M L L L L L L L L H L M L L Weight ucnuuMHPKUcnuoioioiuuuucocducnoiMHUMuuu ------Char. VerBurg # BIBLIOGRAPHY BIBLIOGRAPHY Abramson, P.R., Aldrich, J.H., and Rohde, D.W. (1987) Change and Continuity in the 1984 Elections, rev.ed. Washington, DC: Congressional Quarterly. Adrian, C.R., and Press, C. (1977) Governing Urban America. 5th New York: McGraw-Hill Book Company. ed. Aldrich, J.H., and Nelson, F.D. (1984) Linear Probability. Logit, and Probit Models. Beverly Hills, CA: Sage Publications. Babcock, R.S. (1957) State and Local Government and Politics. New York: Random House. Bachrach, P. and Baratz, M.S. (1962) "Two Faces of Power" Political Science Review. 56(December): 947-952. Banfield, E., and Wilson, J.Q. (1963) Mass.: Harvard University Press. American City Politics. Cambridge, Boyle, J. and Jacobs, D. (1982) "The Intracity Distribution of Services: A Multivariate Analysis" American Political Science Review. 76(June): 371-379. Bryce, J. (1906) The American Commonwealth. Philadelphia: Morris and Co. Chute, C.F. (1951) "How to Get a New City Charter" Municipal Review. 40(September): 403-410. _____ (1956) "Charter Campaigning" 45(December): 537-545. 178 John D. National National Municipal Review. 179 Clark, T.N. (1968) "Community Structure, Decision-Making, Budget Expenditures, and Urban Renewal in 51 American Communities" American Sociological Review. 33(August): 576-593. Cornwell, E.E. (1985) "Comments on the Denzau and Maser Papers" Public Choice. 47: 219-229. Crain, R.L., and Rosenthal, D.B. "Community Status As A Dimension of Local Decision-Making" American Sociological Review. 33(August): 576-593. Dahl, R.A. (1961) Who Governs? Democracy and Power in an American City. New Haven: Yale University Press. Dye, T.R. (1979) "Politics Versus Economics: The Development of the Literature on Policy Determination" Policy Studies Journal. 7(Summer): 652-662. Dye, T.R., and Garcia, J.A. (1978) "Structure, Function, and Policy in American Cities" Urban Affairs Quarterly. 14(September): 103-122. Dye, T.R., and Macmanus, S.A. (1976) "Predicting City Government Structure" American Journal of Political Science. 20(May): 257-271. Engstrom, R. and McDonald, M. 1981. "The Election of Blacks to City Councils: Clarifying the Impact of Electoral Arrangements on the Seats/Population Relationship" American Political Science Review. 75(June): 344-354. Gamm, L. (1976) "Voter Education and Participation: Pennsylvania Local Referenda" National Civic Review. 65(February): 75-82. Gilbertson, H.S. (1912) "The Recall, Its Provisions and Significance" The Annals of the American Academy. XLIII (September): 216-226. Gosnell, C.B. and Holland, L.M. (1951) State and Local Government in the United States. New York: Prentice-Hall. Gray, V.H. (1979) "Conclusions By the Symposium Co-Editor" Policy Studies Journal. 7(Sumner): 799-803. 180 Hannah, S.B. (1986) "Checks and Balances in Local Government: City Charter Amendments and Revisions in Michigan 1960-1985" Paper presented at 1986 Annual Meeting of Midwest Political Science Association. Hawley, A.H. and Ziraner, B.G. (1970) The Metropolitan Community. Its People and Government. Beverly Hills, Calif.: Sage Publications Hawley, W.D. (1973) Nonpartisan Elections and the Case for Party Politics. New York: John Wiley and Sons, Inc. Hays, S.P. (1984) "The Politics of Reform in Municipal Government in the Progressive Era" in Readings in Urban Politics: Past. Present and Future. 2nd ed. (ed.) Hahn, H. and Levine, C.H. New York: Longman, Inc. Heckathorn, D.D. and Maser, S.M. (1987) "Bargaining and Constitutional Contracts" American Journal of Political Science. 31(February): 142-168. Heckman, J.J. (1976) "The Common Structure of Statistical Models of Truncation, Sample Selection and Limited Dependent Variables and A Simple Estimator For Such Models" Annals of Economic and Social Measurement. 5/4: 475-492. Heckman, J.J. (1979) "Sample Selection Bias As a Specification Error" Econometrics, 47(January): 153-161. Herbert, A.W. (1971) "The Los Angeles Charter: The Lessons of Defeat" National Civic Review. 60(December): 603-607, 636. Hofstadter, R. Knopf. (1955) The Age of Reform. New York: Alfred A. Hollingsworth, C.M. (1912) "The So-Called Progressive Movement: Real Nature, Causes and Significance" The Annals of the American Academy. 43(September): 32-48. Hunter, F. (1953) Community Power Structure: A Study of Decision Makers. The University of North Carolina Press. Its 181 Judd, D.R. (1984) The Politics of American Cities: Private Power and Public Policy. 2nd ed. Boston: Little, Brown and Company. Knoke, D. (1982) "The Spread of Municipal Reform: Temporal, Spatial, and Social Dynamics" American Journal of Sociology. 87(May): 1314-1339. Knott, J.H., and Miller, G.J. (1987) Reforming Bureaucracy: The Politics of Institutional Choice. Englewood Cliffs, NJ: Prentice-Hall. Inc. Liebert, R. (1974) "Municipal Functions, Structures, and Expenditures" Social Science Quarterly. 55: 765-783. Lineberry, R.L., and Fowler, E.P. (1967) "Reformism and Public Policy in American Cities" American Political Science Review. 61(September): 701-716. Lyons, W. (1977) "Urban Structures and Policy: Reassessing Additive Assumptions of Reform" Political Methodology. 4: 213-226. (1978) "Reform and Response in American Cities: Structure and Policy Reconsidered" Social Science Quarterly. 59(June): 118-132. MacManus, S.A. (1978) "City Council Election Procedures and Minority Representation: Are They Related?" Social Science Quarterly. 59: 153-161. Maddala, G.S. (1983) Limited-dependent and Qualitative Variables in Econometrics. New York: Cambridge University Press. March, J.G., and Olsen, J.P. (1984) "The New Institutionalism: Organizational Factors in Political Life" American Political Science Review, 78(September): 734-749. Marra, R.F., and Ostrom, C.W. Jr. (1985) "Model Building Under Uncertainity" Paper presented at 1985 Annual Meeting of American Political Science Association. 182 Maser, S.M. (1985) "Demographic Factors Affecting Constitutional Decisions: The Case of Municipal Charters" Public Choice. 47: 121-162. McGoldrick, J. (1933) Law and Practice of Municipal Home Rule: 1930. New York: Columbia University Press. 1916- McKelvey, R.D. and Zavoina, W. (1975) "A Statistical Model for the Analysis of Ordinal Level Dependent Variables" Journal of Mathematical Sociology. 4: 103-120. Michigan. (1908) Constitution. Michigan. (1963) Constitution. Miller, G.J. Press. (1981) Cities by Contract. Cambridge, Mass.: The MIT _____ (1985) "Progressive Reform as Induced Institutional Preferences" Public Choice. 47: 163-181. Mills, C.W. Press. (1956) The Power Elite. New York: Oxford University Mladenka, K.R. "The Urban Bureaucracy and the Chicago Political Machine: Who Gets What and the Limits to Political Control" American Political Science Review. 74(December): 991-998. Morgan, D.R. and Pelissero, J.P. "Urban Policy: Does Political Structures Matter?" American Political Science Review. 74(December): 999-1006. National Municipal League. (1960) A Guide for Charter Commissions. 4th ed. New York: National Municipal League. National Municipal League. (1964) National Municipal League. Model Charter. 6th ed. New York: Nelson, D.C. (1979) "Ethnicity and Socioeconomic Status as Sources of Participation: The Case for Ethnic Political Culture" American Political Science Review. 73(December): 1024-1038. 183 Ostrom, E. (1986) "An Agenda for the Study of Institutions" Choice. 48: 3-25. Public Patton, C.W. (1940) The Battle for Municipal Reform: Mobilization and Attack. 1875 to 1900. Washington, DC: American Council on Public Affairs. Peterson, P.E. Press. (1981) City Limits. Chicago: Pindyck, R.S. and Rubinfeld, D.L. Economic Forecasts. 2nd ed. Company. University of Chicago (1981) Econometric Models and New York: McGraw-Hill Book Polsby, N.W. (1963) Community Power and Political Theory. New Haven: Yale University Press. Quigley, John M. (1980) "Local Residential Mobility and Local Government Policy" in Residential Mobility and Public Policy, (ed) Clark, W.A.V. and Moore, E.G. Beverly Hills, Calif.: Sage Publications. Riker, W.H. (1980) "Implications from the Disequilibrium of Majority Rule for the Study of Institutions" American Political Science Review. 74(June): 432-447. Riordon, W.L. Dutton. Salisbury, R.H. St. Louis" 260-275. (1963) Plunkitt of Tammany Hall. New York: E.P. (1961) "The Dynamics of Reform: Charter Politics in Midwest Journal of Political Science. 5(August): Schattschneider, E.E. (1960) The Semi-Sovereign People: A Realist*s View of Democracy in America. New York: Holt, Rinehart and Winston. Schneider, M. (1985) "Suburban Fiscal Disparities and the Location Decisions of Firms" American Journal of Political Science. 29(August): 587-605. 184 Schneider, M. (1987) "Income Homogeniety and the Size of Suburban Government" The Journal of Politics. 49(February): 36-53. Scott, T.M. (1968) "The Diffusion of Urban Governmental Forms As A Case of Social Learning" The Journal of Politics. 30(November): 1091-1108. Shepard, W.B. (1975) "Participation in Local Policy Making: The Case of Referenda" Social Science Quarterly. 56(June): 55-70. Stein, L. (1986) "Representative Local Government: Minorities in the Municipal Work Force" The Journal of Politics. 48(August): 694-713. Stone, C.N. (1980) "Systematic Power in Community Decision-Making: A Restatement of Stratification Theory" American Political Science Review. 74(December): 978-990. Stonecash, J. "Politics, Wealth and Public Policy: The Significance of Political Systems" Policy Studies Journal. 7(Summer): 670675. Tiebout, C. (1956) "A Pure Theory of Local Expenditures" Political Economy. 64 (October): 416-424. Journal of White, A.G. (1973) Charters as Municipal Constitutions. New Council of Planning Librarians. York: Williams, O. and Adrain, C. (1963) Four Cities. Philadelphia: University of Pennsylvania Press. Wilson, W. (1887) "The Study of Administration" in Classics of Public Administration. (ed.) Shafritz, J.M. and Hyde, A.C. Oak Park, Illinois: Moore Publishing Company, Inc. Wolfinger, R.E. (1974) The Politics of Progress. Englewood New Jersey: Prentice-Hall, Inc. Cliffs, Wolfinger, R.E., and Field, J.O. (1966) "Political Ethos and the Structure of City Government" American Political Science Review. 60(June): 306-326. 185 Zimmer, B.G. (1976) "Suburbanization and Changing Political Structures" in The Changing Face of the Suburbs, (ed.) B. Schwartz. Chicago: University of Chicago Press. Zimmerman, J.F. (1972) The Federated City: Community Control in Large Cities. New York: St. Martin’s Press.