"Customization and the Diffusion"
Customization and the Diffusion of Policy Innovations Andrew Karch University of Texas at Austin firstname.lastname@example.org Paper prepared for presentation at the Annual Meeting of the Southern Political Science Association, January 2005, New Orleans, LA. According to legend, Willie Sutton once said that he robbed banks because that was where the money was. For equally pragmatic reasons, observers of American public policy have recently turned their attention to the fifty states. For the past two decades, the states have served as the main locus of policymaking across a wide range of policy arenas. From abortion and capital punishment to education and the environment, many important decisions are being made in state houses across the country rather than in the nation’s capital. When state lawmakers view existing conditions as problematic, they can change the existing policy repertoire modestly and at the margins. For instance, they can provide a program with increased funding or slightly alter the conditions of program eligibility. Under some circumstances, however, officials will not be satisfied with these incremental adjustments. Instead, they will propose and adopt new and relatively untested solutions to perceived problems, establishing innovative programs that have not been implemented previously. These new policies are termed “policy innovations” and mark an alternative to the incremental changes that typify the making of public policy.1 The fifty states have long been a source of important policy innovations. In fact, one of the most famous metaphors in American history speaks to the significance of the states. In an oft-cited 1932 dissent, Supreme Court Justice Louis Brandeis wrote, “It is one of the happy incidents of the federal system that a state may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the 1 Jack L. Walker, Jr., “The Diffusion of Innovations among the American States,” American Political Science Review, Volume 63, Number 3 (1969): 880-899. In this seminal work, Walker defines a policy innovation as “a program or policy which is new to the states adopting it, no matter how old the program may be or how many other states may have adopted it” (881). 1 rest of the country.”2 Since this landmark dissent, the fifty states have been referred to as “laboratories of democracy” with great regularity. Actors from across the political spectrum invoke the laboratories of democracy metaphor to describe the states’ innovative potential. In the foreword to a book on statelevel economic policy during the 1980s, Arkansas Governor Bill Clinton used the laboratories of democracy metaphor to describe how the “states learn from one another, borrowing, adapting, and improving on each other’s best efforts.”3 A few years later, President George Herbert Walker Bush used the laboratories of democracy metaphor in his 1991 State of the Union Address. As he advocated an increased role for the states in managing public policy, Bush praised the “innovative power” of states as laboratories.4 Similarly, liberal and conservative judges have cited Brandeis’s dissent more than three dozen times. It possesses clear appeal to a nation of “compulsive tinkerers.”5 The laboratories metaphor captures two major arguments in favor of allowing the states to set public policy. The first argument stresses their innovative potential. States can serve as laboratories, with their policies serving as experiments. The states can serve as testing grounds for new policy ideas. These innovations can be transferred elsewhere if they prove successful. In other words, state-level programs can reveal which policy innovations achieve their objectives and which do not. Successes can be emulated, and the damage inflicted by unsuccessful innovations can be limited to the state in which the experiment is conducted. Allowing the states to formulate and evaluate policy facilitates 2 New State Ice Co. v. Liebmann 285 U.S. 262 (1932). In this case, the Supreme Court ruled on the constitutionality of an Oklahoma statute that characterized the manufacture, sale, and distribution of ice as a public business and regulated it accordingly. By a vote of six to two, the Supreme Court ruled the statute unconstitutional. 3 David Osborne, Laboratories of Democracy (Boston, MA: Harvard Business School Press, 1990), p. xii. 4 George Herbert Walker Bush, “State of the Union Address 1991.” 5 Michael S. Greve, “Laboratories of Democracy: Anatomy of a Metaphor,” AEI Federalist Outlook No. 6 (2001). This piece also includes a critical assessment of Brandeis’s dissent. 2 feedback and learning, promoting a process of trial and error that allows successful new programs to flourish.6 The second argument for increased state authority relates to the second half of the laboratories of democracy metaphor. Some argue that the states are a more democratic arena in which to make policy decisions. President Bush advanced this argument in his 1991 State of the Union Address by claiming that the devolution of authority to state officials “moves power and decision-making closer to the people.”7 Since officials must respond to intrastate constituencies if they wish to remain in office, they will be inclined to adjust policy templates to social and political conditions within their own jurisdictions. Thus state-based policymaking, it is argued, facilitates customization to meet local needs. This dynamic may not only be democratic. It may also be more efficient. Sometimes “one size fits all” policy solutions will not be appropriate, and programs will function more effectively if they are adapted to particular localities. The remainder of this paper focuses on the relationship between local conditions and policy content. The preceding discussion implies that policy innovations that purport to be the same might actually vary significantly across states. In other words, lawmakers might adjust an existing policy template to “fit” their state. Most policy diffusion studies, however, focus exclusively on the question of adoption.8 They ask whether policymakers Some analysts contend that federated organizations possess inherent advantages during the search for policy solutions because they can empower multiple subunits to search for a solution and then combine the information discovered by the subunits to come up with better solutions. See Ken Kollman, John H. Miller, and Scott E. Page, “Decentralization and the Search for Policy Solutions,” Journal of Law, Economics, & Organization 16 (2000): 102-128. 7 George Herbert Walker Bush, “State of the Union Address 1991.” 8 This emphasis on program adoption results from the use of event history analysis (EHA). The dependent variable of interest is the hazard rate, defined as the probability that a state will adopt a policy innovation in a specific year. Hazard rates cannot be directly observed because they are probabilities. Instead, EHA uses a dichotomous dependent variable to indicate whether the policy innovation was adopted that year (coded 1 6 3 enacted a policy innovation but rarely ask about its content. This emphasis places too much stock in the existence of a particular program. The same policy innovation may be adapted to certain states, taking account of particularities within a state. Alternatively, lawmakers might adjust the content of an innovation based on early adopters’ experiences with it. In either case, focusing exclusively on program enactment overlooks how state officials may have altered the existing template. The importance of program content holds in any study of public policy. Whether and when a country enacted a public pension program is informative, for example, but it is only the tip of the iceberg. Which individuals are eligible to receive a pension? How are benefits calculated, and when are citizens able to receive these benefits? How is the state pension system funded? How much do employers or individuals contribute? These details represent political decisions that probably provoked controversy.9 Policy content is especially important in the context of policy diffusion, a process through which elected officials learn about developments elsewhere and draw lessons from them. These lessons often involve subtle adjustments to an existing policy idea. These adjustments constitute a constitutive element of policy diffusion and merit careful examination. Reinvention, Customization, and Politics Most research on the spread of innovations emphasizes sameness across adopters. In organizational sociology this phenomenon is known as “institutional isomorphism.” if yes and 0 if no). While this strategy is useful for analyzing the question of program enactment, it is not well suited for distinguishing among the programs that have been enacted. 9 Scholars of comparative public policy often focus on these distinctions. See Gosta Esping-Andersen, The Three Worlds of Welfare Capitalism (Princeton, NJ: Princeton University Press, 1990); Giuliano Bonoli, “Classifying Welfare States: A Two-Dimension Approach,” Journal of Social Policy 26 (1997): 351-372. 4 Scholars in this field attempt to explain homogeneity in organizational forms, and they concentrate on the mechanisms that encourage organizations to resemble one another. At a certain point, the adoption of an innovation provides an imprimatur of legitimacy even if its effect on organizational performance is uncertain.10 A similar dynamic can emerge in social policy. During the welfare reform debate of the mid-1990s, for example, work requirements and time limits took on this level of importance. State lawmakers needed to grapple with these policy innovations if they wanted to be considered leading innovators. Applying the notion of isomorphism to social policy, however, fails to appreciate how policy innovations take on varying forms in the states where they are enacted. These variations in program content are not unique to the diffusion of public policy innovations. Differences across adopters also characterize technological and other innovations. These innovations are rarely monolithic and prepackaged, so adopters are able to modify an innovation or to incorporate some of its components while rejecting others.11 When state policymakers consider a policy innovation, their choices are not limited to adoption or rejection. They also customize a policy template to fit their particular state. For instance, the states coalesced around a general model of using hate crime law to address increases in intergroup violence, but these laws took on a variety of legal forms.12 When officials Paul J. DiMaggio and Walter G. Powell, “The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields,” American Sociological Review, Volume 48, Number 2 (1983): 147-160. Other scholars characterize these institutional rules as “myths” that organizations incorporate to enhance their legitimacy, resources, stability, and survival prospects. See John W. Meyer and Brian Rowan, “Institutionalized Organizations: Formal Structure as Myth and Ceremony,” American Journal of Sociology, Volume 83, Number 2 (1977): 340-363. These scholars also claim that there is a harmonization of organizational forms but a divergence in how these organizations actually function. 11 Ronald E. Rice and Everett M. Rogers, “Reinvention in the Innovation Process,” Knowledge, Volume 1, Number 4 (1980): 499-514. The authors examine the diffusion of “Dial-A-Ride” across U.S. cities, noting that the many programs falling under this heading “actually represent a variety of different versions of this innovation.” 12 Ryken Grattet, Valerie Jenness, and Thedore R. Curry, “The Homogenization and Differentiation of Hate Crime Law in the United States, 1978 to 1995: Innovation and Diffusion in the Criminalization of Bigotry,” American Sociological Review, Volume 63 (1998): 286-307. 10 5 altered state welfare programs in the 1990s, a similar dynamic emerged. The new work requirements and time limits varied along many dimensions such as the individuals to whom they were applied, the length of time before they were imposed, and provisions for extensions and exemptions.13 This substantive variation had important consequences for the individuals affected by welfare reform legislation. Thus policymakers in two or more states can enact the “same” policy innovation, but its specific provisions might differ so substantially that it is difficult to believe that these different versions fall under the same heading. The concept of reinvention recognizes that state policymakers create substantively different programs even though the policies can be grouped into a broad general category. Most scholarly examinations of reinvention imply, however, that these changes are due to the order in which the programs were enacted. Do later adopters establish less ambitious policies than leaders, do late adopters enact more extensive programs, or do late adopters practice wholesale borrowing?14 Most analyses of reinvention assume that late adopters will enact innovations that are more expansive than the tentative approaches that leaders adopt. Late adopters might possess more reliable information about the political and the administrative feasibility of a policy innovation. A new program that was controversial when leaders adopted it can become less contentious by the time that late adopters enact the same program. This increased acceptability can eliminate the need to take a minimal approach, enabling late adopters to enact more expansive programs.15 Various analyses 13 Matthew C. Fellowes and Gretchen Rowe, “Politics and the New American Welfare States,” American Journal of Political Science, Volume 48, Number 2 (2004): 362-373. 14 This formulation comes from Jill Clark and J. Lawrence French, “Innovation and Program Content in State Tax Policies,” State and Local Government Review, Volume 16 (1984): 11-16. 15 Jill Clark, “Policy Diffusion and Program Scope: Research Directions,” Publius: The Journal of Federalism, Volume 15, Number 4 (1985): 61-70. 6 provide empirical support for this assumption. The earliest versions of living will laws, for instance, were less facilitative than the statutes that followed them.16 Late adopters might not “lead the pack,” but their programs can nonetheless contain new approaches for dealing with social problems. Reinvention is a useful idea because it draws attention to differences in program content, but it is also problematic. There is no a priori reason to expect that the content of a policy innovation will shift in a more expansive direction. The passage of time might also encourage lawmakers to take a more restrictive approach. The experiences of other states with a program, for example, might generate a backlash that limits its acceptability or convinces state officials that the policy is not feasible administratively. Late adopters might move in either a more expansive or a more restrictive direction. More fundamentally, the concept of reinvention describes a mechanical process in which the order of policy adoption overrides political factors. It attributes variation in policy content to a passive process that leaves little room for state politics. According to this notion, policy content is largely determined by the order of policy adoption. Scholars have different expectations for early adopters, middle-range adopters, and late adopters. This explanation of program content is unsatisfactory because it underestimates the role of politics, an inherently dynamic process in which groups frame issues, shape agendas, create coalitions, and try to affect public policy.17 The order of adoption cannot explain program content in a theoretic sense. 16 Henry R. Glick and Scott P. Hays, “Innovation and Reinvention in State Policymaking: Theory and the Evolution of Living Will Laws,” Journal of Politics, Volume 53, Number 3 (1991): 835-850. For a similar analysis of state abortion regulations see Christopher Z. Mooney and Mei-Hsien Lee, “Legislating Morality in the American States: The Case of Pre-Roe Abortion Regulation Reform,” American Journal of Political Science, Volume 39, Number 3 (1995): 599-627. 17 Jeffrey M. Stonecash, “The State Politics Literature: Moving Beyond Covariation Studies and Pursuing Politics,” Polity, Volume 28, Number 4 (1996): 559-579. 7 Because of the mechanical connotations of the term “reinvention,” this paper refers to the process that affects policy content as “customization.” Lawmakers tailor a policy innovation to fit their state just as individuals have a suit tailored after they buy it off the rack. There is ample reason to believe that politics, and not the order of program enactment, affects the specific provisions of a policy innovation. Once a policy template is imported into a state, it undergoes a process of customization that alters its content, and many political forces can influence this process. Lawmakers might attempt to persuade reluctant groups to endorse a policy innovation by amending the original legislation. For example, they might try to convince women’s groups to endorse welfare time limits by exempting victims of domestic violence. Customization also includes technical changes. Executive branch officials might promote amendments that make a policy innovation easier to implement, and objective conditions might push lawmakers in a certain direction. Some officials in Oregon, for instance, argued that high enrollment in health maintenance organizations (HMOs) in the state would limit the efficacy of medical savings accounts. Differences in program content, whether they result from political maneuvering or technical changes, seem likely to reflect intrastate developments. Elected officials must respond to their constituents if they wish to remain in office and might therefore amend a policy template when intrastate interest groups voice objections to its provisions. These changes will not always convince an opponent to endorse the final product, but outright hostility toward opponents can have electoral ramifications. Ignoring the sentiments of the “labor lobby,” the “Catholic lobby,” or another constituency might mobilize a vocal opposition campaign. When interest groups representing intrastate constituencies support or oppose specific provisions, their stances are likely to be taken into account. Elected 8 officials must face intrastate constituencies at the ballot box, and they may therefore be willing to address their concerns about policy innovations. Even when the opponents of a policy innovation are not able to block its enactment, their vocal opposition can alter the content of a proposal. For example, advocates for low-income individuals in many states fought welfare legislation in the early 1990s and sometimes succeeded in moderating its strictest provisions.18 The electoral connection between policymakers and intrastate constituencies is extremely important. Professional associations such as the Council of State Governments (CSG) sometimes recommend specific legislative provisions. These organizations do not represent intrastate constituencies, however, implying that state officials can ignore their recommendations with virtual impunity. State lawmakers are more likely to respond to interest groups located within their state. Ignoring the views of a powerful constituency carries electoral risk. The remaining sections of this paper examine the politics of customization. They proceed in two steps. First, they classify the witnesses who testified before legislative committees during public hearings on four innovations in health care and welfare policy: medical savings accounts (MSAs), individual development accounts (IDAs), time limits, and family caps.19 This section compares the relative visibility of intrastate and external organizations, and it reveals that witnesses represent groups with intrastate constituencies. Second, they examine the programmatic content of welfare and health care policy in 18 Pamela Winston, Welfare Policymaking in the States: The Devil in Devolution (Washington, DC: Georgetown University Press, 2002). 19 These four policy innovations vary along a number of important dimensions. First, they represent different conceptualizations of the proper role and reach of government. While MSAs and IDAs combine governmental efforts with those of the private sector, time limits and family caps retrench existing policies. Second, the five policy innovations vary in the extent and speed with which they diffused across the states. Examining the varied paths of these diverse policies enhances our confidence in the conclusions drawn from this analysis. 9 Massachusetts, Oregon, and Virginia, focusing on the divergent priorities, strategies, and strength of intrastate interest groups. This intrastate variation begins to explain why the same policy innovation typically takes on a number of different forms. Legislative Testimony and Customization In his seminal work on the diffusion of policy innovations across the states, Jack Walker noted the increasing availability of national information but expressed skepticism that it affected program enactment. He explained, “Officials in Illinois may know of the procedures and performance levels in New York and California, but they are unlikely to think of events in these states as legitimate guides to action.”20 As state officials tailor a policy innovation to fit their states, they are likely to rely on intrastate forces to serve as “legitimate guides to action.” Interest groups representing both intrastate and national constituencies possess a number of tools to make state lawmakers aware of their positions on legislative proposals. They utilize a wide variety of tactics, including testimony at committee hearings. During these hearings, officials gauge support and opposition to a proposal among the constituencies that it would affect. They frequently respond to this testimony by amending proposals in an effort to ameliorate opposition and to generate additional support. These amendments are an important component of the customization process, and this section categorizes witnesses based on the type of constituency that they represent. This section uses testimony at legislative hearings as a proxy for influence during the customization process. Although this approach cannot establish definitively that the 20 Walker, “The Diffusion of Innovations among the American States,” 892. 10 testimony altered the content of a proposal, it provides a good first cut at answering this question. A recent survey of state lobbyists and state organizations suggests that virtually all interest groups engage in this activity. It found no significant variation across group types in the use of this tactic. “Insider” and “outsider” groups are equally likely to count legislative testimony among their lobbying techniques.21 As a result, we can be confident that this measurement strategy is not biased in terms of finding specific types of groups to be especially influential. This section uses appearances before legislative committees to assess the balance between witnesses representing intrastate and national constituencies. My analysis utilizes a five-fold classification scheme to categorize the witnesses who appear before state legislative committees. “Individual citizens” do not represent an organization, although interest groups sometimes ask these constituents to testify. Their testimony often represents an attempt to personalize an issue. During a hearing on a bill to establish a senior pharmaceutical assistance program, for instance, an elderly citizen of the state might describe her struggles to pay for the medicine she needs. At a hearing on welfare reform, a witness might testify about her struggles or improved self-esteem after entering the workforce. Organizational representatives fall into one of four different categories. First, “national organizations” incorporates the representatives of groups with no presence in a state. This category includes witnesses who represent professional associations and think tanks headquartered in Washington, DC. Second, “state chapters” includes witnesses who represent a group located within a state that is affiliated with a national organization. The Oregon Medical Association is a state affiliate of the American Medical Association 21 Anthony J. Nownes and Patricia Freeman, “Interest Group Activity in the States,” Journal of Politics, Volume 60, Number 1 (February 1998): 86-112. 11 (AMA) and belongs in this category. State chapters represent federations, organizations that possess both a national headquarters and an organizational presence in each of the states. Connections between state chapters and their national leadership vary in their intensity. These bonds can be strong or weak depending on the organization. In their day-to-day operations, state chapters usually set their own agendas and determine their own strategies. Third, “intrastate groups” include witnesses from groups or coalitions that operate exclusively within a state. These organizations are not isolated from national forces, but their national connections are less formal than the links that characterize state chapters.22 Fourth, “state and local government” includes testimony by elected officials, members of the state executive branch, and individuals who work for local government agencies. Sometimes these witnesses voice their political opinions. At other times, they provide a professional assessment of how a policy innovation will work. Dividing committee witnesses into five categories allows me to assess the relative prominence of intrastate and national organizations. Based on the previous discussion, I expect witnesses representing intrastate constituencies (state chapters, intrastate groups, and state and local government) to testify with greater frequency. My analysis examines patterns in Oregon. Legislative committees in Oregon, particularly in the state house, are extremely powerful. The rules carve out a central role for committees in the legislative process, giving them significant influence over policy content. Committees in the house can introduce bills with the committee listed as the author, offer substitute bills in place 22 For example, the Oregon Health Action Campaign is part of the Universal Health Care Action Network (UHCAN). UHCAN was formed in 1992 to bring together diverse state groups and activists working for comprehensive health care. It differs from groups categorized as state chapters because this network exists to connect state groups that were already in place. It is a political campaign rather than an organization per se, and it therefore differs from an interest group or professional association that establishes state chapters in order to create a presence in all fifty states. 12 of the original, have committee amendments automatically incorporated into a proposal rather than have each amendment accepted or rejected by the floor, and make it difficult to amend legislation on the floor by requiring unanimous consent.23 Committees do not have such favorable rules in every state. The Oregon rules provide interested parties with a powerful incentive to testify before legislative committees on issues about which they feel strongly. In addition, the superior archives maintained in Oregon make an analysis of committee witnesses relatively straightforward. My analysis assesses the balance between witnesses representing intrastate and national constituencies along three different dimensions. First, I simply count the number of witnesses in each category. This approach weighs each appearance equally. Second, I examine multiple appearances as a proposal moves through the legislative process. A bill typically receives multiple hearings. When witnesses appear more than once, it might indicate the importance of the group they represent. Sometimes policymakers amend the original legislation after the initial testimony and then ask whether the amended bill earns the group’s endorsement during a subsequent appearance. This dynamic is particularly common during the consideration of omnibus legislation as provisions are dropped and added in an attempt to build a broad coalition. I also compare the balance of national and intrastate constituencies along a third dimension that grows out of the particularities of the legislative process in Oregon. The work of legislative committees in Oregon takes place in two stages: public hearings and work sessions. During public hearings, witnesses generally describe their positions on a policy proposal. They explain why they hold these views and speculate on the effects of 23 Keith E. Hamm and Gary F. Moncrief, “Legislative Politics in the States,” in Virginia Gray and Russell L. Hanson, eds., Politics in the American States: A Comparative Analysis, 8th ed. (Washington, DC: CQ Press, 2004), 180. 13 a new program. These public hearings give Oregon legislators a sense of where specific constituencies stand and what it would take to earn their support. Witnesses sometimes discuss their positions on specific provisions within the legislation, but more often they provide a general evaluation of the overall proposal. During work sessions, which generally follow public hearings, legislators focus intently on legislative language. They examine specific provisions and discuss both their meaning and their administrative feasibility. Work sessions generally involve a careful reading of the bill. A work session on medical savings accounts illustrates this point. In 1995, the House State and School Finance Committee discussed MSA legislation during a contentious work session. Members questioned how specific sections of the bill would be implemented. The chair of the committee eventually said, “This is obviously a pretty complicated thing. I am beginning to get the feeling right now that we are not going to be able to do this or have it in any form that we are going to be ready to move on. I think that we have too many questions and too much to adjust.”24 This example illustrates how work sessions shape program content. Appearances at work sessions therefore represent the third dimension along which I compare the influence of intrastate and national forces. Legislative Testimony in Oregon My analysis of committee appearances in Oregon reveals three notable patterns. First, the representatives of national organizations rarely testify. Second, state chapters and intrastate groups often provide committee witnesses. Third, officials from state and local governments officials frequently appear before legislative committees. Tables 1, 2, 24 Chair John Schoon, House State and School Finance Committee, Work Session on House Bill 2865, 5 April 1995, 68th Oregon Legislative Assembly, 1995 Regular Session. Oregon Archives. 14 3, and 4 illustrate these patterns. Tables 1 and 2 classify witness testimony on omnibus welfare legislation in 1993 and 1995, a period during which state legislators considered time limits and family caps. Tables 3 and 4 categorize the witnesses who testified on medical savings accounts and individual development accounts, respectively. These four tables count the number of witnesses who fall into each of the five categories described earlier. The limited number of appearances by witnesses from national organizations is the first noteworthy pattern in these tables. Despite the national prominence of welfare reform in the early and mid-1990s, Oregon policymakers generally did not hear from the representatives of national organizations. The only “national” witness who testified on either of the two welfare reform bills was a Virginia-based consultant who specialized in welfare-to-work programs. None of the witnesses represented a professional association or think tank based in Washington, DC. Representatives of national organizations were more prominent during hearings on medical savings accounts. Witnesses testified on behalf of Kaiser Permanente,25 the National Federation of Independent Businesses, and Pharmaceutical Research and Manufacturers of America. Overall, the limited number of appearances by the representatives of national organizations suggests that these groups have a relatively minor impact on the customization process. Combined, state chapters of national organizations and intrastate interest groups typically provided more than half of committee witnesses. This is the second noteworthy pattern in Tables 1 through 4. These witnesses testified on behalf of groups that represent intrastate constituencies. When they testified, these witnesses could credibly claim that 25 The Government Relations Manager of Kaiser Foundation Health Plan of the Northwest submitted written testimony but did not testify before the House Committee on State and School Finance. In addition, the organization lacked truly “national” reach. It operated in only nine states and the District of Columbia. 15 their testimony reflected the interest and position of voters within Oregon. State chapters usually provided between one-fourth and two-fifths of the witnesses who testified. Many state chapters, including Planned Parenthood Advocates of Oregon and Oregon Right to Life, are a part of national interest groups. A second category of state chapters includes labor organizations such as the Oregon Public Employees Union and Local Union 290. Both of these groups testified during welfare reform hearings in 1993. The state affiliates of national professional associations form a third category of state chapters. The Oregon Society of Certified Public Accountants (OSCPA) and the Oregon Association of Health Underwriters both appeared at hearings on medical savings accounts. Like union groups, these associations are organized on an occupational basis. These different types of state chapters provided a much larger percentage of witnesses than national organizations did. Intrastate interest groups based exclusively within the state usually provided about one-fourth of the committee witnesses in Oregon. These organizations include advocacy groups like Oregon Student Lobby and Oregonians for Medical Savings Accounts and officials from local academic institutions such as Portland State University. Even local businesses can make an electoral impact. In 1990, Jeld-Wen, Inc. was the primary force behind the passage of a welfare reform ballot initiative. The prominence of state and local government officials is the third noteworthy pattern in Tables 1 through 4. These officials usually represent between one-third and two-fifths of all committee witnesses. Sometimes state and local government officials appear as policy advocates. Legislators, for example, can testify in support of proposals that they themselves have written. These officials can also voice their opposition to a new program. In Oregon, a member of the Commission for Women objected to welfare 16 reform legislation at a hearing in 1995.26 State and local government witnesses can also provide technical or programmatic expertise about how current programs function and how a policy innovation might operate if it is implemented. An official from the Oregon State Treasury provided technical expertise during discussions of individual development accounts, and a representative of the Department of Revenue played a similar role during a hearing on medical savings accounts. Legislative testimony is one of many ways that executive branch officials keep lawmakers apprised of how programs are working. State policymakers often rely on the technical expertise of program administrators to customize policy innovations. Organizations representing intrastate constituencies and witnesses from state and local government offices provide most committee witnesses in Oregon. Similar patterns prevail if we focus only on witnesses who make multiple appearances before legislative committees. Sometimes state lawmakers invite witnesses to return and give their opinion on an amended piece of legislation. Table 5 lists all of the witnesses who made multiple appearances. During welfare reform hearings in 1995, witnesses from the Oregon Public Employees Union and Multnomah County Legal Aid Service made two appearances.27 The union also made three appearances as lawmakers debated MSA legislation in 1997. The OSCPA made six distinct appearances during the MSA debate in 1997, testifying to the centrality of this professional association. In 1995, a representative of Oregonians for 26 The Commission for Women was created by executive order in 1964 and received its first budget in 1977. The governor appoints nine of its eleven members, while legislative leaders choose the remaining two. The Commission is “directed to work for the implementation and establishment of economic, social, legal, and political equality for women and to maintain a continuing assessment of the issues and needs confronting women in Oregon.” For more information on the history and activities of the Commission for Women see http://bluebook.state.or.us/state/executive/commis_for_women/comm_women_history.hmtl. 27 The witness from Multnomah County Legal Aid Service represented an umbrella organization called the Oregon Human Rights Coalition, a membership organization that represents individuals in the Oregon state welfare system. 17 Medical Savings Accounts appeared twice. All of these witnesses represented intrastate groups or state chapters of national organizations. No one representing a purely national organization testified multiple times. Some state and local government officials also made multiple appearances before Oregon committees. When lawmakers considered welfare reform measures in 1993 and 1995, the administrator and assistant administrator of Adult and Family Services testified seven times. The director of the Department of Human Resources testified twice, as did a representative of Clackamas County Social Services and the human services advisor to Governor John Kitzhaber. An official from the Department of Revenue appeared at three hearings on MSA legislation in 1997. In sum, executive branch officials were prominent at committee hearings. The division between public hearings and work sessions in Oregon provides a third way to assess the impact of groups representing intrastate constituencies. Statutory language is a primary topic of conversation during work sessions, so witnesses at these sessions are especially likely to have an effect on program content. Table 6 provides this information.28 It illustrates how program administrators and state and local government officials are particularly visible. For example, witnesses testified on behalf of Adult and Family Services, the Division of Child Support, and the governor during welfare reform work sessions in 1993 and 1995. In addition, lawmakers invite the representatives of influential intrastate interest groups to these work sessions. In 1995, representatives of Jeld-Wen, Inc. and Multnomah County Legal Aid Service participated in welfare reform work sessions. Committee members heard almost exclusively from the representatives of 28 Table 6 excludes House Bill 2865 (1995) because its work session did not feature any witnesses. It also excludes officials from the Legislative Revenue Office and the Office of Legislative Counsel because these offices are housed within the legislature and employ individuals who assist legislators in bill drafting. 18 intrastate constituencies as they focused on the provisions of welfare reform legislation. A Virginia-based consultant on welfare-to-work programs represented the lone exception. Relatively few witnesses appeared during work sessions on medical savings accounts and individual development accounts, but all of them represented intrastate groups or a state executive agency. Work sessions often precipitate important changes to the legislation being considered, so this pattern is significant. Thus witnesses from intrastate groups, state and local governments, and the state chapters of national organizations dominate legislative hearings in Oregon. Compared to representatives from purely national organizations, they testify more frequently, they are more likely to appear at multiple hearings on the same proposal, and they are more likely to participate in work sessions. These patterns suggest that organizations representing intrastate constituencies have a strong influence during the customization process. The next section builds on this suggestive evidence by taking a closer look at differences in program content across three states. Differences across States in the Politics of Customization: Welfare Reform While intrastate forces seem likely to affect the customization process in every state, these forces can push state officials in all sorts of directions. Differences in the size, strength, and strategies pursued by intrastate groups might lead to varied program content. Policy content is more likely to reflect these constituencies’ preferences in states where they are large and strong. Diverse strategies might also affect program content. In one state a group might compromise to remove a particularly objectionable provision from a 19 bill, but a similar group in another state might devote all of its energy to stopping the bill from gaining enactment at all. This section examines the politics of welfare reform in Massachusetts, Oregon, and Virginia. Officials in all three states considered and enacted omnibus welfare legislation in the mid-1990s, but they endorsed very different bills. The welfare advocacy community objected strenuously to policy innovations like family caps and time limits, but welfare reform legislation still gained enactment. State policymakers frequently responded to the concerns of the welfare advocacy community, however, by amending proposals during the customization process. Welfare Reform in Massachusetts Even though Massachusetts maintains a liberal reputation, in 1995 lawmakers in the state adopted one of the most stringent welfare reform programs in the country. The Massachusetts welfare reform legislation imposed many strict requirements on welfare recipients, including time limit and family cap provisions that are often associated with a conservative political stance. The welfare program imposed a “work requirement time limit.”29 It required non-exempt adults to work at least 20 hours each week after sixty days of independent job search. Recipients who could not find jobs would be placed in community service positions. In the event of noncompliance, welfare grants would be reduced by the proportion attributed to the individual required to work, while continued noncompliance could result in a full family sanction. The family cap provisions of the Massachusetts program were similarly stringent. The program denied families an increase in welfare benefits when a child was born more 29 Setting the Baseline: A Report on State Welfare Waivers (Washington, DC: Office of the Assistant Secretary for Planning and Evaluation, Department of Health and Human Services, June 1997). This format is qualitatively different from a time limit that eliminates benefits after a certain number of months. 20 than ten months after the family applied for AFDC. In addition, state policymakers also incorporated two especially strict requirements. The Massachusetts family cap required parents to participate in the Job Opportunities and Basic Skills Training (JOBS) Program when the affected child turned three months of age, a provision known as the “reduced JOBS exemption.” It also applied the family cap to children conceived within a year of the family leaving AFDC.30 In addition, the Massachusetts family cap did not include certain provisions thought to be favorable to affected individuals. Lawmakers chose not to provide a partial increase in welfare benefits or vouchers for goods and services in lieu of increased benefits. Nor did the Massachusetts policy allow families to keep a larger percentage of their earnings in order to compensate for the benefits they were denied due to the family cap. The Massachusetts family cap incorporated one generous provision, a child support “pass through” allowing families to keep all child support money collected on behalf of the child who was excluded from benefits.31 The Massachusetts family cap was most notable, however, for its stringency. Why did strict time limit and family cap policies gain enactment in such a liberal state? These outcomes can be traced to two factors. First, Republican Governor William Weld provided strong executive leadership. In 1990, Weld was elected on a platform that included calls for a crackdown on welfare. In 1993, he introduced a reform proposal that required work after two years of benefit receipt. That same year, the Joint Committee on Human Services and Elderly Affairs debated a family cap bill. Neither of these proposals 30 Family caps in California, Connecticut, Indiana, Nebraska, Virginia, and Wisconsin incorporated the “reduced JOBS exemption.” Family caps in Arizona, Illinois, Mississippi, Nebraska, and Virginia included an extension after families left AFDC. 31 Family caps in Connecticut and Florida provided a partial increase, while those in Maryland, South Carolina, and Indiana provided vouchers in lieu of increased benefits. Family caps in Arizona, Georgia, Illinois, and New Jersey incorporated an increased earnings disregard. Family caps in California, Delaware, Maryland, Mississippi, Nebraska, and Virginia incorporated a child support “pass through.” 21 gained enactment in 1993, but they represented the opening salvo in what became a twoyear struggle over welfare reform. Second, the welfare advocacy community pursued a confrontational strategy that backfired. It alienated the very policymakers that it intended to influence. Excluded from the customization process, these interest groups were unable to modify the stringent proposals that the governor and other state officials advanced. In 1994, family cap proposals played a central role during the debate over welfare reform in Massachusetts. In early June 1994, the State Senate rejected a family cap bill by a 21-18 vote. In addition to this stand-alone bill, however, family cap proponents had also embedded the policy within a larger budget bill. It was one of the issues that stalled the budget negotiations. Shortly after the Senate vote, however, a Boston Globe editorial argued that negotiators should accept the family cap provisions to send a “powerful, if largely symbolic, message that the state advocated personal responsibility.”32 This strong endorsement from a relatively liberal media outlet indicated that a bipartisan consensus had emerged behind the controversial policy innovation. Even though the editorial did not have a causal effect, the conference committee reached a budget deal less than a week later. This deal included a family cap. The debate over time limits in Massachusetts proved less contentious than did the debate over family caps. In February 1994, the State Senate overwhelmingly endorsed a time limit bill proposed by Democrat John O’Brien. This bill provided a middle ground between the conservative proposal of Governor Weld and a more liberal House proposal. When the Joint Committee on Human Services and Elderly Affairs dropped the time limit provision in March, its move enraged Republicans and moderate Democrats. The House responded by passing a more stringent bill that included a time limit and a family cap. A 32 “Time to Bridge the Budget,” The Boston Globe, 18 June 1994, 18. 22 poll conducted by KRC Communications Research of Newton suggested that lawmakers were responding to the positions of their constituents. Seventy percent of the respondents approved of the time limit and family cap provisions contained in the House bill, and less than one-fourth of the respondents voiced opposition.33 This poll reflected the increasing popularity of tough, punitive welfare reform proposals. Legislators and the public agreed that the existing welfare system was flawed and in need of repair. Allegations of rampant fraud and child abuse led many to conclude that a complete overhaul was necessary. The welfare advocacy community took a radical stance and vehemently rejected the emerging consensus on welfare reform. Rather than pursuing a consensual lobbying strategy or seeking to modify the strictest provisions of these proposals by participating in legislative discussions, the advocacy community took a confrontational posture and staged a number of colorful protests.34 Activists left a bag of manure outside the office of the Speaker of the House, disrupted floor proceedings by throwing fake money from the House balcony, and covered State House statues with black cloth. They staged a morality play on the State House steps that included a “legislator” with bloodied hands and a child in a body bag. They also initiated a disorderly protest at the State House that resulted in eight arrests. By pursuing this adversarial strategy, the advocacy community alienated the very state officials that it wanted to influence. The activists refused to acknowledge the palpable shift in public opinion that favored strict welfare reform proposals. With the electorate strongly in favor of reform, legislators knew that the advocacy community did not pose much of a threat at the ballot box. The virtually complete disengagement of the 33 Anthony Flint, “Welfare Reform Wins Big Support: Residents Polled Back Benefits Cuts,” The Boston Globe, 15 May 1994, 29. 34 Michelle Buis Michaux, “Taking States Seriously: Welfare Reform Implementation in Massachusetts and Oregon” (Ph.D. dissertation, Brandeis University, 2002), Chapter 4. 23 advocacy community did not guarantee that a stringent reform package would pass, but this strategic choice proved critical. In Oregon and Virginia, the advocacy community voiced its objections through conventional political channels and had a more substantial impact on program content. In Massachusetts, welfare reform took many more twists and turns before finally gaining enactment. In 1994 Governor Weld vetoed reform legislation because he felt it was not sufficiently strict and liberal Democratic Senators Dianne Wilkerson and Lois Pines surprised their colleagues by voting to uphold his veto. Many Democrats believed that the two senators were “suckered by a Republican governor who has no intention of accommodating their concerns.”35 When the two senators decided to switch their votes and override the veto, another senator changed his vote and upheld it.36 In the meantime, the governor won reelection by a convincing margin in November 1994. The debate then resumed in 1995. Governor Weld was in a strong position and the advocacy community was on the sidelines. The two-year struggle culminated that year in a reform package that incorporated strict time limit and family cap provisions. Welfare Reform in Oregon The welfare advocacy community in Oregon pursued a more consensual strategy than its counterparts in Massachusetts did. These activists affected the content of welfare reform legislation through their thorough engagement in the political process. Activists in Oregon worked through conventional political channels such as task force participation 35 Dan Aucoin, “Welfare Reform Dash Slows to a Crawl: Roosevelt, Weld Bicker over Lack of Action,” The Boston Globe, 27 September 1994, 25. 36 Senator William Keating, involved in a leadership dispute with Senate President William Bulger, claimed that the heavy-handed tactics of his main rival justified the vote switch. See Doris Sue Wong, “Bulger Foe Stymies Veto Override: Keating Switches, Decries Maneuvers,” The Boston Globe, 20 December 1994, 21. 24 and legislative testimony. This consensual strategy was one reason why welfare reform legislation in the state incorporated a relatively lax time limit and did not include a family cap. Oregon is a less liberal state than Massachusetts is. Opponents of welfare reform in Oregon nonetheless made a large impact on the content of welfare legislation by adopting consensual tactics. Welfare reform surfaced in Oregon in 1990 and was a prominent issue throughout the early 1990s. The Oregon Full Employment Program, a ballot initiative that replaced welfare benefits with wages for transitional jobs leading to regular employment, received 58 percent of the vote in 1990. Its implementation proved immensely controversial, and its supporters filed a lawsuit against the administration of Democratic Governor Barbara Roberts. They argued that the administration had not taken sufficient steps to obtain the necessary waivers from the U.S. Department of Health and Human Services. Both sides eventually went to the bargaining table. Negotiations included the sponsors of the ballot measure, members of the legislature, gubernatorial representatives, and top officials in the Department of Human Resources.37 In 1993, these discussions produced House Bill 2459, known as the JOBS Plus Program. This welfare reform legislation focused on onthe-job training and skill enhancement and was established as a three-year pilot program. After a contentious debate in 1995, JOBS Plus was expanded statewide. Welfare reform in Oregon grew out of a complex political process. State officials adjusted the framework provided by the original ballot measure over a number of years. The welfare advocacy community played an active role in these discussions. In contrast to their counterparts in Massachusetts, activists in Oregon participated in hearings as well 37 Representative Gene Derfler, Exhibit O to House Bill 2459, dated 21 July 1993, 67th Oregon Legislative Assembly, 1993 Regular Session. Oregon Archives. 25 as other conventional settings rather than relying on protest tactics. Two advocates in Oregon joined a thirteen-member welfare reform task force created by Representatives Mary Ann Ford, a Republican, and Beverly Stein, a Democrat. A representative of the Multnomah County Legal Aid Service participated in public hearings and work sessions. During his testimony he suggested numerous amendments to the JOBS Plus legislation. Some of them were incorporated into the final bill. In 1993, the Assistant Administrator for Adult and Family Services explained that new language had been incorporated into welfare reform legislation based on the objections voiced by this witness during an earlier hearing.38 This example illustrates the willingness of lawmakers to customize a policy innovation in response to objections from the advocacy community when these activists work through conventional channels. By pursuing a consensual lobbying strategy, the welfare advocacy community in Oregon was able to eliminate certain provisions of JOBS Plus and to modify others. This impact did not mean that these lobbyists were satisfied with or enthusiastic about the final legislation. Nor did it imply that lawmakers embraced every amendment they suggested. It was, however, an important reason why JOBS Plus was significantly less stringent than the reform package that was adopted in Massachusetts. In part, their consensual strategy reflected a realistic recognition of the political situation. The momentum behind welfare reform in Oregon meant that its opponents stood little chance of blocking the legislation. They were better positioned to modify specific provisions that they found objectionable. Republican Stan Bunn, Chair of the Senate Committee on Health and Human Services and a leading architect of welfare reform legislation in 1995, indicated his willingness to 38 Jim Neeley, Assistant Administrator of Adult and Family Services, House Legislative Rules and Reorganization Committee, Work Session on House Bill 2459, 26 May 1993, 67th Oregon Legislative Assembly, 1993 Regular Session. Oregon Archives. 26 make changes to his original proposal. When he described an iteration of this legislation, he noted that it included amendments proposed by the governor, legislators, and “parties in the audience.” Bunn explained, “I think that almost everybody gets something, and nobody gets as much as they want.”39 Members of the advocacy community did not get everything they wanted. Their primary objective was to prevent the statewide expansion of JOBS Plus. Even though this effort failed, the relatively moderate content of the final legislation reflected their active participation during the customization process. By acting through conventional political channels, they affected program content. JOBS Plus incorporated neither strict time limits nor a family cap. Lawmakers in Oregon proposed both policy innovations, but they received only lukewarm support. In 1995, Republican Representatives Charles Starr and Jane Lokan submitted separate bills calling for a strict two-year time limit. Starr also introduced family cap legislation. The House Committee on Children Families held a public hearing on all three bills in March, but none of them gained enactment. JOBS Plus legislation overshadowed these narrow bills and served as the focal point for the welfare debate. This omnibus welfare package, in keeping with the original intent of the 1990 ballot initiative, emphasized employment and training. JOBS Plus incorporated a modified time limit. The time limit provisions were a compromise between the welfare advocacy community and conservative advocates of a more rigid policy. JOBS Plus limited receipt of AFDC benefits to no more than 24 out of 84 months for Oregon families with employable parents. This time limit did not apply to caretakers other than parents or to households where the parent was disabled or needed to 39 Chair Stan Bunn, Senate Committee on Health and Human Services, Work Session on Senate Bill 1117, 10 May 1995, 68th Oregon Legislative Assembly, 1995 Regular Session. Oregon Archives. 27 care for a disabled person. It also exempted individuals who were enrolled in JOBS and welfare recipients who were required to participate in JOBS but had not been offered the opportunity to enroll or participate in an education, employment, or job training program. This wide range of exemptions, combined with relatively lax sanctions policies, led one analyst to conclude that “welfare recipients [in Oregon] are not subject to the two-year limit at all.”40 Opponents of welfare reform “did not get everything they wanted, but many of their specific concerns with the legislation were addressed.”41 In addition, JOBS Plus did not include a family cap. The lenient time limit and the absence of a family cap were due in part to the efforts of the welfare advocacy community. The policymaking environment in Oregon aided the welfare advocacy community in two important ways. First, welfare reform supporters were willing to compromise. The coalition that sponsored the initiative cared deeply about its focus on employment and training. It was not enthusiastic about time limits and family caps. Republicans in the state legislature added a time limit provision to the original ballot initiative but were willing to moderate its provisions. Second, the opposition of Governor Roberts aided the welfare advocacy community. Her skepticism validated the efforts to amend the original ballot initiative. When Roberts endorsed a compromise under immense political pressure, however, her agreement made the adoption of omnibus legislation a foregone conclusion. In 1995, Democratic Governor John Kitzhaber played a similar role, participating in the negotiations that amended the content of JOBS Plus legislation. The willingness of welfare reform proponents to compromise and gubernatorial support aided the efforts of the advocacy community. Activists took advantage of this 40 41 Michaux, “Taking States Seriously,” 132. Michaux, “Taking States Seriously,” 134. 28 opportunity by adopting a consensual stance and working through conventional political channels. They participated on welfare reform task forces and testified during committee hearings. Although lawmakers enacted JOBS Plus over the objections of the advocacy community, they took these objections into account as they formulated and reformulated this landmark legislation. Thus the efforts of the advocacy community contributed to the removal of fairly strict provisions from JOBS Plus. Policymakers in Oregon established a welfare reform program with a lax time limit and no family cap. Welfare Reform in Virginia Like its counterpart in Oregon, the welfare advocacy community in Virginia pursued a consensual lobbying strategy. Its utilization of conventional political channels produced important amendments to the original welfare proposal, but the scope of these changes was relatively limited. The political terrain in Virginia placed liberal advocacy groups in a difficult position. The inclusion of strict time limit and family cap provisions in the Virginia Independence Program (VIP) seems to match the conservative political environment in the southern state. A strict time limit was a key component of welfare reform in Virginia. The VIP limited welfare benefit receipt to 24 cumulative months for cases headed by non-exempt caretakers.42 It provided hardship exemptions for recipients residing in localities where the unemployment rate exceeded ten percent. It also granted exemptions if they would 42 The program exempted: individuals under age 16; individuals between 16 and 19 enrolled full-time in elementary or secondary schools; individuals incapacitated by a temporary medical condition; individuals receiving SSI or S.D. benefits; individuals over the age of 60; the sole caregiver of an incapacitated household member; a parent/caretaker of a child under 18 months; children receiving AFDC-Foster Care; families where the primary caretakers are not the adoptive or biological parents of the child; and females in their fourth through ninth month of pregnancy. 29 allow an individual to complete a program of employment-related education or training, and 90-day extensions were available to individuals who were actively seeking a job but who could not find work that would equal the AFDC cash benefit. While important, the exemptions did not alter the fundamentally strict nature of the Virginia program. Other states provided transitional benefits or limited assistance to a specified number of months within a given period, but the Virginia plan imposed a stringent “termination time limit” at the 24-month mark. The Virginia Independence Program also incorporated a strict family cap. This provision required parents to participate in JOBS after the affected child turned six weeks of age, a particularly stringent provision. Virginia policymakers also extended the family cap to children conceived within six months of the family leaving AFDC. The family cap did not include an increased earnings disregard, and it provided neither a partial increase nor vouchers for goods and services in lieu of increased benefits. Like their colleagues in Massachusetts, lawmakers in Virginia adopted especially strict family cap provisions and rejected policies that would have softened their impact. Welfare reform figured prominently in Virginia state politics in the early 1990s. It was a prominent issue during the 1993 gubernatorial campaign. Democratic candidate Mary Sue Terry endorsed time limits, and Republican candidate George F. Allen outlined an especially restrictive time limit proposal. In 1994, omnibus welfare reform legislation faced “little opposition” as it moved through the Virginia General Assembly.43 This bill incorporated the recommendations of the Commission to Stimulate Personal Initiative to Overcome Poverty, headed by Democratic Lieutenant Governor Don Beyer. It reflected the bipartisan consensus on the necessity of overhauling the existing welfare program. 43 Leslie Taylor, “Welfare Reform Arrives,” The Roanoke Times, 16 March 1994, C1. 30 The welfare advocacy community in Virginia advanced a vision of welfare reform that did not include time limits or family caps. As the omnibus proposal moved through the General Assembly in 1994, the director of Total Action against Poverty contended, “Simply cutting people off will result also in an increased cost in child neglect, foster care, delinquency and crime, and incarceration.”44 Similarly, the director of the Roanoke City Department of Social Services claimed that universal day care and universal health care would be more effective than any of the strict proposals favored by members of both parties.45 Given the bipartisan consensus behind time limits in Virginia, these arguments stood little chance of winning the day. The welfare advocacy community nonetheless earned a small victory by working through traditional lobbying channels. The 1994 reform legislation was a three-year pilot program, and its incorporation of a family cap proved particularly controversial. Some legislators opposed the reform proposal based solely on the inclusion of the family cap. The Black Caucus, women’s groups, and the American Civil Liberties Union (ACLU) of Virginia all voiced their opposition to the plan. Attempts to remove this provision failed repeatedly, but policymakers eventually brokered a compromise. The final legislation included a family cap but subjected it to a two-year trial period. Limiting the application of a family cap to the first two years of the pilot program was not the optimal outcome for either side. Yet it was necessary to move the legislation forward. Lieutenant Governor Beyer argued that this change would “provide objective data on the impact of the [family 44 Ted Edlich, “Welfare Reform: A Prescription for Real Welfare Reform,” The Roanoke Times, 18 March 1994, A11. Total Action against Poverty is an welfare advocacy group located in Roanoke, Virginia. 45 Corinne Gott, “Blaming the Poor Is No Way to Reform Welfare,” The Roanoke Times, 17 July 1994, E3. 31 cap] limitations.”46 In reality, this concession was made for political reasons. Given the unfavorable position of the welfare advocacy community in Virginia, this alteration was small but significant. Its victory proved fleeting, however. Welfare reform moved back on to the political agenda in 1995. Governor Allen greeted the General Assembly with a bold agenda that included deep budget cuts and an even more stringent welfare reform proposal. The governor suffered a “huge defeat” on the budget but his welfare plan was one of the few bills that was not undone by partisan sniping during the 1995 session.47 This bill incorporated a time limit, work requirements, and a strict family cap. The new legislation revoked the uneasy compromise of 1994 by making the family cap provisions permanent. This episode illustrates one of the perils of a consensual lobbying strategy. The welfare advocacy community earned a compromise initially on the family cap, but it was unable to preserve this small victory. Explaining Variation in Welfare Reform Developments in Massachusetts, Oregon, and Virginia illustrate the importance of lobbying strategies and underlying political alignments. Table 7 categorizes these three episodes along two dimensions, “terrain” and “tactics.” Massachusetts and Oregon are “favorable” terrains because of the relatively liberal policy histories of these two states. Virginia is “unfavorable” because of its conservative political environment. Oregon and Virginia qualify as “consensual” in terms of tactics because the advocacy community in 46 Don Beyer, “Virginia Independence Program,” The Virginian-Pilot, 3 May 1994, A22. For more on the family cap debate see Leslie Taylor, “Welfare Plan’s ‘Family Cap’ Is Controversial,” The Roanoke Times, 26 March 1994, C1. 47 “Out of the Rubble in Richmond,” The Washington Post, 2 March 1995, A26. 32 these states operated through conventional channels. Massachusetts is an “adversarial” state because the opponents of welfare reform pursued a protest-based strategy. Their tactics backfired and alienated the lawmakers that the groups hoped to influence, and it effectively excluded the welfare advocacy community from the political process. As a result, state policymakers adopted relatively stringent time limit and family cap policies. A consensual strategy might not have guaranteed a different outcome in Massachusetts, but developments in Oregon and in Virginia suggest that participating through standard political channels will affect program content. In Oregon, representatives of the welfare advocacy community submitted amendments, and state legislators followed some of their recommendations. These advocates supported additional amendments and the rejection of the original proposal, but they did not get everything that they wanted. However, their participation resulted in a moderate bill. Their consensual tactics were one reason why the Oregon welfare reform program incorporated a lax time limit and no family cap. Differences in program content, however, were not due exclusively to advocates’ tactical choices. Critics of welfare reform in Oregon claimed some advantages that their counterparts in Massachusetts did not possess. For example, Massachusetts Governor William Weld was a leading welfare reform advocate in his state while Oregon Governor Barbara Roberts objected to the original proposal in her state. In Virginia, the advocacy community worked on less favorable terrain. Governor Allen galvanized the momentum behind welfare reform, and the political environment was more conservative than that of either Massachusetts or Oregon. Even so, critics of the Virginia Independence Program nonetheless convinced lawmakers to modify the original reform legislation. Proponents and opponents settled on temporary family cap provisions. This policy innovation was 33 made permanent a year later. Even though this compromise proved fleeting, it illustrates the potential efficacy of pursuing a consensual strategy rather than an adversarial one. In both Oregon and Virginia, working through conventional political channels facilitated the modification of welfare reform proposals. Differences across States in the Politics of Customization: Medical Savings Accounts Program content also varies when the coalition supporting a policy innovation varies across states. In some cases, differences in legislative content can be attributed to disparities in the type of groups involved in formulation of the program. Policy proposals reflect the intentions of their primary supporters. When the supporters vary, innovations that purport to fall under the same heading can differ significantly. Differences in MSA legislation in Oregon and Virginia illustrate the impact of diverse constituencies. The nature of the Virginia MSA program reflected the ambitious political objectives of its primary supporter, while the Oregon legislation grew out of the technical objectives of its main advocate.48 The expansive provisions of the Virginia Medical Savings Account Plan reflect the political ambitions of the Jeffersonian Health Policy Foundation. Established by a group of Virginia physicians, the foundation claimed that MSAs would reestablish the “proper and beneficial” relationship between patients and the physician of their choice. Inspired by the federal debate over the Health Security Act, the foundation developed model MSA legislation called the “American Health Care Plan.” In 1995, the Virginia 48 This section does not describe developments in Massachusetts because no MSA legislation was enacted in that state. 34 General Assembly unanimously endorsed an MSA bill “based on the model developed by the Jeffersonian Health Policy Foundation.”49 The federal government had not yet acted on MSAs, and the Virginia statute was contingent on congressional authorization of this policy innovation. In other words, the Virginia program would not be operational until Congress passed legislation that defined eligible participants and prescribed criteria for the accounts. Since no federal legislation had gained enactment, the Virginia statute contains some unusual provisions. For instance, it authorizes the use of “direct debit cards” to access the accounts and incorporates “programs to educate recipients in handling health care services in a cost-effective manner while ensuring that necessary care is obtained.” The federal legislation that passed in 1996 did not address these topics. The Virginia plan also mandates the “integration of existing coverage,” “a system of refundable tax credits,” and “a system for calculating individual need for health care services in order to ensure that adequate sums are calculated for the care of individuals with great need.”50 These features also distinguish the Virginia MSA Plan from the federal legislation that established an MSA pilot program. The statute is a very clear statement of the ambitious policy objectives possessed by the foundation and other MSA supporters. In Oregon, the provisions of the state MSA program reflect the technical goals of its primary sponsor, the Oregon Society of Certified Public Accountants (OSCPA). The Oregon statute overlaps significantly with a federal statute. MSAs had achieved agenda status in Oregon in the early 1990s, but no legislation passed until 1997. The creation of 49 Joint Commission on Health Care, Virginia’s Medical Savings Account Program Study (Richmond, VA: Joint Commission on Health Care, 2002). See also Matt Murray, “Area Doctors Suggest Savings Account as Alternative Approach,” Daily Press, 6 March 1994, B5. 50 Virginia, Code of Virginia, Section 38.2-5603. 35 a pilot program by the federal government in 1996 provided the impetus for the Oregon bill.51 When the state legislature reconvened in 1997, the OSCPA became involved in the MSA debate for the first time. Its representatives made multiple appearances at public hearings and work sessions. They argued for the enactment of MSAs as part of a larger “tax reconnection” bill that aligned the Oregon tax code with the federal tax code. The Oregon MSA program differed significantly from the bill that was adopted in Virginia. Only three paragraphs of the lengthy reconnection bill mentioned MSAs, and all of them described the Internal Revenue Code as the current and future basis of the law in Oregon. In contrast to the more expansive Virginia legislation, the Oregon bill did not mention debit cards, educational programs, or the integration of existing coverage. The Oregon bill was based almost entirely on federal provisions.52 The adoption of MSAs in Oregon therefore might be described as a “technical” shift. One opponent of the statute, in fact, took umbrage at the political process that produced this bill. He argued that any recognition of the policy innovation “requires a separate policy decision [and] ought to be addressed through separate legislation.”53 The tax reconnection bill sailed to enactment as this argument fell on deaf ears. Thus the MSA provisions that gained enactment in Oregon differed significantly from those adopted in Virginia. In Oregon, OSCPA dominated the customization process and included MSAs within a larger bill linking the state tax code to the Internal Revenue The Health Insurance Portability and Accountability Act (HIPAA) permitted only the self-employed, businesses with fifty or fewer employees, and persons enrolled in high-deductible health insurance plans to purchase MSAs. It capped the total number of tax-advantaged MSAs at 750,000 for the four-year period between January 1, 1997 and December 31, 2000. 52 The House Committee on Revenue adopted one “technical amendment” addressing the “deductibility limit for part-year and non-residents’ contributions to medical savings accounts.” See House Committee on Revenue, “Staff Measure Summary: Senate Bill 347,” dated 19-21 March 1997, 69th Oregon Legislative Assembly, 1997 Regular Session. Oregon Archives. 53 Tim Nesbitt, Oregon Public Employees Union, House Revenue Committee, Public Hearing on Senate Bill 347, 21 March 1997, 69th Oregon Legislative Assembly, 1997 Regular Session. Oregon Archives. 51 36 Code. Motivated by administrative concerns rather than policy goals, the organization advocated limited statutory language. By contrast, the Virginia MSA program was standalone legislation that reflected the ambitious political objectives of its primary advocate. The Jeffersonian Health Policy Foundation lobbied successfully for a more complex bill. Although both statutes established state-level MSA programs, their specifics diverged substantially. Conclusion: Intrastate Forces and Customization The fifty states are frequently described as laboratories of democracy, underlining the notion that state lawmakers derive lessons from developments in other states. At the same time, however, advocates of granting increased policymaking prerogatives to statelevel officials commonly argue that devolution permits these officials to adjust programs to social and political circumstances in their state. They argue that it is more democratic to permit state officials to customize programs to fit their state than to impose a common program for all fifty states. Given the centrality of this political argument, it is surprising that few studies of policy diffusion systematically examine differences across states in terms of policy content. While studies of policy reinvention draw attention draw to these differences, they generally attribute this variation to the timing of policy adoption rather than to internal state politics. This paper focused specifically on customization, the political process through which lawmakers amend a policy template after importing it. It found that most of these changes come at the behest of intrastate actors who represent constituencies that can hold 37 elected officials accountable for their actions. It relied on two types of evidence. First, it examined the committee witnesses who testify before state legislative committees since their testimony is an important source of programmatic adjustments. Most committee witnesses in Oregon represented intrastate groups, state and local government agencies, and the state chapters of national organizations. Individuals representing purely national groups rarely testified. Intrastate witnesses appeared more often, testified multiple times on policy innovations, and dominated the work sessions at which legislators focused on statutory language. The second type of evidence presented in this paper was a more intensive look at the relationship between internal political forces and program content. By comparing the content of policy innovations in Massachusetts, Oregon, and Virginia, this paper stressed the importance of the coalitions supporting these programs. The different provisions of MSA statutes in Massachusetts and Virginia reflect the divergent objectives of the major forces behind this legislation in the two states. Even when a similar coalition exists, as in the case of welfare reform, tactical decisions can affect policy content. Intrastate groups utilize diverse tactics to achieve their objectives. Program content reflects this diversity. The strictness of welfare reform in Massachusetts was due, in part, to the adversarial pose of the welfare advocacy community, whereas advocates in Oregon pursued a consensual strategy and moderated many provisions of that state’s welfare program. In combination, these two sources of evidence begin to explain why policy innovations that purport to be the same actually vary substantially across the fifty states. 38 Table 1: Witnesses before Oregon Committees Topic/Bill Number/Year Welfare Reform House Bill 2459 (1993) Intrastate Groups Witnesses: 3 (11%) Multnomah County Legal Aid Service Portland State University Oregon Student Lobby State and Local Government Witnesses: 10 (36%) Administrator, Oregon Employment Department Assistant Administrator, Adult and Family Services Director, Department of Human Resources Clackamas County Social Services Representative Gene Derfler Representative Bob Shipreck Representative Sam Domini Representative Cynthia Wooten Representative Avel Gordley Clackamas County Regional Workforce Quality Committee State Chapters of National Groups Witnesses: 7 (25%) Northwest Seasonal Workers Association Oregon Head Start Association Oregon AFL-CIO Oregon Public Employees Union AFSCME Council 75 Local Union 290 (Plumbers, Steamfitters, and Marine Fitters) United Way of the ColumbiaWillamette National Organizations Witnesses: 1 (4%) Virginia-Based Consultant on Welfare-to-Work Programs Note: The chart does not include the seven individual citizens who testified. 39 Table 2: Witnesses before Oregon Committees Topic/Bill Number/Year Welfare Reform Senate Bill 1117 (1995) Intrastate Groups Witnesses: 5 (25%) Society of St. Vincent de Paul, Portland Council Oregon Legal Services Western Oregon State College Multnomah County Legal Aid Service Jeld-Wen, Inc. State and Local Government Witnesses: 8 (40%) Administrator, Adult and Family Services Workforce Quality Council Corvallis Adult and Family Services Representative Dennis R. Luke Policy Advisor to Governor Kitzhaber on Health and Human Resources Assistant Administrator, Adult and Family Services Commission for Women Support Enforcement Division State Chapters of National Groups Witnesses: 7 (35%) Catholic Charities in Oregon Elizabeth House Maternity Residence Oregon Right to Life Oregon Public Employees Union Planned Parenthood Advocates of Oregon Oregon Building and Construction Trades Council Oregon Catholic Conference National Organizations Witnesses: 0 (0%) 40 Table 3: Witnesses before Oregon Committees Topic/Bill Number/Year Medical Savings Accounts Senate Bill 347 (1997) House Bill 2488 (1997) Intrastate Groups Witnesses: 2 (22%) Oregonians for Medical Savings Accounts Oregon Health Action Campaign State and Local Government Witnesses: 2 (22%) Department of Revenue Department of Human Services, Oregon Health Plan Administrator State Chapters of National Groups Witnesses: 3 (33%) Oregon Society of Certified Public Accountants Oregon Public Employees Union Oregon Medical Association National Organizations Witnesses: 2 (22%) National Federation of Independent Businesses Pharmaceutical Research and Manufacturers of America Medical Savings Accounts House Bill 2865 (1995) Witnesses: 1 (20%) Oregonians for Medical Savings Accounts Witnesses: 1 (20%) Representative Patricia Milne Witnesses: 1 (20%) Oregon Association of Health Underwriters Witnesses: 1 (20%) Kaiser Permanente Note: One individual citizen, a local insurance agent who accompanied the representative of the Oregon Association of Health Underwriters, testified on House Bill 2865. 41 Table 4: Witnesses before Oregon Committees Topic/Bill Number/Year Individual Development Accounts House Bill 3600 (1999) Intrastate Groups Witnesses: 4 (31%) Human Solutions Farmworker Housing and Development Corporation Housing Development Center Association of Oregon Housing Authorities State and Local Government Witnesses: 4 (31%) Oregon State Treasury Multnomah County Oregon Housing and Community Services City of Portland Bureau of Housing and Community Development State Chapters of National Groups Witnesses: 3 (23%) Oregon Building Industry Association Oregon Association of Realtors Association of Oregon Community Development Organizations National Organizations Witnesses: 0 (0%) Note: The chart does not include the two individual citizens who testified. 42 Table 5: Multiple Appearances before Oregon Committees Topic/Bill Number/Year Welfare Reform House Bill 2459 (1993) Intrastate Groups None State and Local Government Assistant Administrator, Adult and Family Services (4) Director, Department of Human Resources (2) Clackamas County Social Services (2) State Chapters of National Groups None National Organizations None Welfare Reform Senate Bill 1117 (1995) Multnomah County Legal Aid Service (2) Administrator, Adult and Family Services (3) Policy Advisor to Governor Kitzhaber on Health and Human Resources (2) Oregon Public Employees Union (2) None Medical Savings Accounts Senate Bill 347 (1997) House Bill 2488 (1997) None Department of Revenue (3) Oregon Society of Certified Public Accountants (6) Oregon Public Employees Union (2) None Medical Savings Accounts House Bill 2865 (1995) Individual Development Accounts House Bill 3600 (1999) Oregonians for Medical Savings Accounts (2) None None None None None None None 43 Table 6: Witnesses at Work Sessions in Oregon Topic/Bill Number/Year Intrastate Groups State and Local Government State Chapters of National Groups None National Organizations Welfare Reform House Bill 2459 (1993) Multnomah County Legal Aid Service Administrator, Adult and Family Services Assistant Administrator, Adult and Family Services Director, Department of Human Resources Clackamas County Social Services Virginia-Based Consultant on Welfare-to-Work Programs Welfare Reform Senate Bill 1117 (1995) Multnomah County Legal Aid Service Jeld-Wen, Inc. Administrator, Adult and Family Services Policy Advisor to Governor Kitzhaber on Health and Human Resources Assistant Administrator, Adult and Family Services Commission for Women Division of Child Support Oregon Public Employees Union Oregon Building and Construction Trades Council Oregon Catholic Conference None Medical Savings Accounts Senate Bill 347 (1997) House Bill 2488 (1997) Individual Development Accounts House Bill 3600 (1997) None Department of Revenue Oregon Society of Certified Public Accountants None Human Solutions None None None 44 Table 7: Explaining Variation in Welfare Reform Terrain Favorable Unfavorable Consensual Oregon Modified Time Limit No Family Cap Massachusetts Strict Time Limit Strict Family Cap Virginia Strict Time Limit Modified Family Cap Tactics Adversarial 45