An Integrated Model of U.S. State Attorney General Behavior in Multi-State Litigation

Multi-state lawsuits, filed by U.S. state attorneys general (AGs), have become an important method by which state consumer protection laws are enforced. Patterns of participation in these lawsuits vary tremendously across the states, yet little is known about the factors driving this variation. I argue that state AGs are primarily concerned with achieving electoral and policymaking goals. Consequently, I expect AGs to be responsive to strong consumer interests and to participate in cases with severe infractions. Additionally, whether it is for public interest or electoral goals, AGs should be attracted to cases that promise bigger settlements. My analysis of each state’s decision to join each of 172 multi-state lawsuits filed and settled between 1989 and 2002 provides support for each of these hypotheses. the dynamics of consumer protection regulation have changed considerably in the past 30 years. While the Federal Trade Commission (FTC) was once the primary enforcer of deceptive advertising laws and, to a lesser extent, antitrust laws, the U.S. states have assumed much of the responsibility for ensuring fair competition and truthful advertising in the marketplace. Specifically, state attorneys general (state AGs) bring joint lawsuits against companies that violate state laws, thereby regulating at the national level. Multi-state lawsuits were first employed in the mid-1980s as a response to a perceived weakening of federal regulatory enforcement under the Reagan Administration (Clayton 1994; Lynch 2001; Ross 1990; Zimmerman 1998), but during the 1990s, they grew beyond advertising cases to include cases involving price-fixing, mergers, product safety, and privacy issues.1 As a result, multi-state litigation has come to symbolize the increasingly visible and powerful role of state AGs in state, national, and even international, policymaking. While the overall number of lawsuits as well as the number of states participating has increased over time (Provost 2006), some states are much more active participants than others, but few studies have examined the factors behind this variation. Multi-state litigation deserves scholarly attention because its dynamics are still poorly understood, yet the key players involved believe that it has had profound effects on regulatory governance. While state AGs regard their role in consumer protection as a vital complement to the role played by the FTC, business leaders and conservative policymakers and commentators perceive multi-state litigation as an excessively punitive regulatory tool that violates the separation of powers and creates multiple, confusing standards for business to follow (Debow 2002; Pryor 2000; Zimmerman 1999). Consequently, in 1999, several Republican state AGs created the Republican Attorneys General Association (RAGA) as a balance to what they perceived as the excessive regulatory meddling of Democratic state AGs. Additionally, in 2002, the U.S. Chamber of Commerce started to campaign for Republican AG candidates as they ran television commercials against “aggressive Democratic candidates for attorney general” (Greenblatt 2003, 54). Regardless of the commercials’ effects, Republicans won 10 of the 15 open seats for AG. Given the significant potential effects of multi-state litigation as a regulatory tool, it is important to understand how such lawsuits proceed. Despite these new developments, relatively little is known about the patterns of state participation in multi-state lawsuits. Much of the current research on multi-state litigation has been done by lawyers and economists whose primary aim is to evaluate the legal and economic merits of multi-state litigation (see e.g., Calkins 2003; Greve 2005; Lynch 2001; Meyer 2007; Posner 2004; Pryor 2000; Zimmerman 1999). To be sure, these studies are useful, but they do not help us understand the circumstances under which different states participate in multi-state litigation. Studies of the politics of consumer protection regulation at the state level have focused mostly on the legislative side of the regulation (Bernacchi 1976; Ford 1977; Meier 1987; Sigelman and Smith 1980) and, as a result, they have little to say about the enforcement of these regulations. Other scholars have analyzed state AG participation in the tobacco litigation of the 1990s (Derthick 2002; Schmeling 2003; Spill, Licari, and Ray 2001), yet this is a unique case that is not representative of most multi-state lawsuits. Finally, some scholars have attempted to understand the inter-state differences in multi-state litigation systematically (Bowman 2004; Provost 2003, 2006), yet these studies all omitted case-level variables crucial to understanding state participation, and hence, they fail to develop a theory of state AG motivation and behavior in such cases. In this article, I fill in the gaps of the literature by explaining what motivates state AGs in their decisionmaking. As elected politicians in the majority 2 provost spring 2010 / state politics and policy quarterly 3 of the states, state AGs are driven by the desire to satisfy policy objectives, but also by political ambition. Thus, in states where consumer concerns are strong, we ought to see greater AG responsiveness, conditioned on case factors. However, each AG’s desire to achieve his/her policy goals should dictate that if the severity of the infraction is high and extensive damage to consumers can be proven, then a larger proportion of states will unite to file suit against the offending business. Additionally, if the target of a particular lawsuit is a large company with abundant financial resources, AGs might perceive the case as important to constituents, as well as detect a chance to distribute pork throughout the state. To observe these motivations, I then examine the effect of various stateand case-level factors on the state AG decision to join or not join each of the 172 multi-state lawsuits filed between 1989 and 2002. First, I will describe briefly how multi-state litigation functions. Second, I list and explain the twin AG motivations of electoral and policymaking considerations. Third, I describe what these motivations mean for responsiveness to state constituents and for participation in particular types of cases. Fourth, I explain the data behind all the variables, and fifth and finally, I provide a test of the model, along with results and a discussion of those results. multi-state litigation Multi-state litigation employed by state AGs is an important method by which states deal with deceptive advertising, antitrust violations, and consumer fraud. The power to investigate has a tremendous impact on the decision to follow a case to the end, as most AGs have the authority to issue civil investigative demands (CIDs) “to obtain both documentary and testamentary evidence from anyone who may have information relevant to the investigation” (Ross 1990, 208). CIDs essentially amount to subpoenas and they allow AGs to gather all the information they need before deciding whether a lawsuit should be filed or not. In most multi-state cases, one state or a small group of states does much of the early investigative work and then once a lawsuit is filed, other states will file lawsuits in their own states,2 help with the remainder of the work to be done and share the settlement money (Lynch 2001; Tierney 2002). Rather than spend a protracted length of time in court, most defendants choose to settle cases quickly. In fact, some businesses will ask that all or most states be included in the settlement to avoid the specter of future litigation from other states (Greenblatt 2003, 56). Final settlements usually include a cash settlement for the states or for state consumers, along with mandates that the defendant(s) will not repeat the forbidden activity, yet they rarely include an admission of guilt on the part of the defendant(s). the motivations of state ags Understanding why consumer protection is more important to some state AGs than others requires an understanding of AG motivations. AGs represent the governor, legislators, and other members of state government as their clients, but they are also charged with representing the public interest (Davids 2005; Ross 1990). As such, they are granted a significant level of discretion to use the tools of their office to pursue policies that they believe will benefit the state. In all 50 states, AGs have either common law or parens patriae authority to enforce the law in the public interest.3 Given these powers, state AGs have strong motivations to formulate and implement legal policy as they see fit. However, the state AG is also elected in 43 states,4 which means that state AGs must face reelection. Additionally, the office has a reputation for being a springboard into higher office positions, such as governor and U.S. senator (Clayton 1994; Mahtesian 1996; Provost 2003). Most AGs are generally reelected if they do not commit serious blunders while in office. But some evidence explicitly links participation in multi-state litigation to the pursuit of higher office. For example, Provost (2008) examined a sample of 104 state AGs and found that among the 28 most active participants in multi-state litigation, 21 (75 percent) ran for governor or U.S. senator. Consequently, when pursuing their policy initiatives, AGs consider how they will garner sufficient public support to achieve their electoral goals. Thus, state AGs are driven by normative concerns of what policies are best for the state, but also by pragmatic concerns such as reelection and, for the more ambitious, reaching higher offices, such as governor. Previous research on state AGs supports the idea that they have the twin goals of shaping policy and accomplishing career/electoral objectives. In their study of the state tobacco litigation, Spill, Licari, and Ray (2001) find that AGs filed early if their state had a high level of Medicaid expenditures due to smoking-related disease, a finding that reveals a concern for how smoking was affecting state health care programs. On the other hand, their discovery that states with large tobacco-farming sectors were more likely to file late or not at all, indicates responsiveness to important constituents within states. Thus, Spill, Licari, and Ray (2001) find evidence not only that state AGs are responsive to their constituents, but that they also make policy according to their perception of what is in the publics’ best interest. 4 provost spring 2010 / state politics and policy quarterly 5 Research has also shown that other public prosecutors are strongly motivated by the public interest and by career/electoral goals. For example, Gordon and Huber (2002) find that voters will reelect local prosecutors for obtaining convictions at trial, thus giving prosecutors the incentive to maximize their conviction rate. Even for appointed prosecutors, these motives appear to be paramount. Whitford (2002) finds that U.S. attorneys will pursue more regulation cases in liberal districts, thus showing responsiveness to their local constituents. Moreover, in what might be seen as pursuing the public interest in order to advance one’s career, Boylan (2005) finds that U.S. attorneys who secure longer sentences for guilty defendants are more likely to become federal judges, an occupation to which many U.S. attorneys aspire (Eisenstein 1978). Employing this previous research as a guide, I argue similarly that in multistate litigation, the decisions of state AGs are driven by desires for reelection, to reach higher offices, and to shape policy in their vision of the public interest. In accordance with achieving their electoral goals, we should expect that AGs will participate in multi-state lawsuits more frequently when they expect their constituents to reward them for such participation. State AGs’ desire to shape policy will be borne out by the seriousness of the crime.5 The more overall damage done by a business to consumers or to the economy in a particular case, the more likely it is that AGs will overcome their partisan differences and rally together to prosecute collectively the business in question. Finally, multi-state litigation is an ideal context for studying these twin motivations of state AGs because it provides a policy making measure comparable across all the states. This is the case for three major reasons. First, as already indicated, every state gives its AG primary responsibility for enforcing consumer protection and antitrust laws and state AGs have authority to enforce such laws as they see fit. This means that in no state does another agency have gate-keeping power over the decision to join a multi-state consumer protection lawsuit. Second, and related to the first point, every state gives its AG the power to issue CIDs to investigate potential wrongdoing, with the exception of Connecticut, Nevada, and Utah (Ross 1990). However, these states can still join lawsuits that have already been initiated, as well as help states that are initiating investigations. Third and finally, multi-state lawsuits deal with businesses that operate throughout the nation, and thus, every state has the same potential incentive to join a lawsuit, once one has been initiated. Even if a company’s wrongdoing does not affect citizens from every state, an AG might join a suit to prevent that business from committing future infractions in his/her state. Additionally, Bowman (2004) has found evidence that multi-state lawsuits do not appear to follow any sort of regional pattern. Taken together, these factors reveal that a state AG’s decision to initiate or join a lawsuit depends primarily on the AG’s own motivations and political considerations. electoral and policy making goals In order to fulfill their goals, state AGs must decide to whom in the state polity they will be most responsive, as well as how best to achieve their policy objectives. In this section, I evaluate the possibilities of AGs being responsive to interest groups, specifically consumer and business groups, voters at large, and actors in state government. I then explain how state AGs pursue policies in accordance with their vision of the public interest. Responsiveness to Organized Interests Responsiveness will be determined largely by the salience and complexity of the issue (Gerber and Teske 2000; Gormley 1986; Ringquist, Worsham, and Eisner 2003). Salience refers to the overall importance of an issue to the electorate at large, while complexity refers to the technical knowledge necessary to administer a particular policy issue. Gormley states that “salience and complexity shape the contours of regulatory politics. They affect incentives to participate, the choice of tactics, the selection of a forum, and the kinds of criteria that are invoked” (1986, 603). Consumer protection can be considered a salient issue because virtually everybody in society is a consumer and can therefore potentially fall victim to market failures, such as poor information (deceptive advertising), abuse of market power (product-tying or price fixing), or fraud (phony telemarketers or pyramid schemes). Moreover, consumer protection has been classified as a salient issue in previous studies (Gerber and Teske 2000; Gormley 1986; Ringquist, Worsham, and Eisner 2003). When an issue is salient, but also low in technical complexity, it is easier for regular voters to understand, and as a result, “regulators and business groups cannot hide behind a cloak of technical expertise” (Gormley 1986, 608). Because the cost of acquiring information is low, collective action problems are fewer and consumer groups can more easily form. Research has shown that citizen groups exert more influence over utility issues when they are not complex (Berry 1984; Gormley 1983). Moreover, Provost (2006) finds that AG participation in multi-state litigation is higher in states with a larger consumer group presence, and I expect to find the same result in this analysis. If we expect consumer interests to influence the level of multi-state liti6 provost spring 2010 / state politics and policy quarterly 7 gation, we might suspect that business interests will have a role in deterring multi-state lawsuit participation in conservative states. The ability of business interests to influence regulators when the issue is complex or when the benefits of regulation are narrowly concentrated on particular industries has been well documented (Gormley 1986; Meier 1987; Peltzman 1976; Posner 1975; Stigler 1971; Wilson 1980). However, in consumer protection regulation, makers of all types of products are potentially affected, so companies from multiple industries have potential collective action problems in organizing against strong regulation. Recent findings are mixed on the power of big business to influence policymaking in a broad fashion. Some authors have found a modest impact of business lobbying in the states on overall policy liberalism (Gray et al. 2004), but others have found that business lobbyists affect neither the state’s policy liberalism nor the business policy climate (Witko and Newmark 2005). Additionally, and as suggested by the 2002 election evidence previously cited, business groups might target AG elections more so than actual AG activities. Thus, when taken as a whole, the evidence does not suggest that AGs will be responsive to business groups when deciding whether to participate in multi-state litigation.6 Responsiveness to the Electorate State AGs might also be responsive to the electorate in general when enforcing consumer protection regulations. Much research at the U.S. state level has found that policymakers are responsive to the wishes of the electorate (Brace et al. 2002; Erikson, Wright, and McIver 1993; Gray et al. 2004; Jacoby and Schneider 2001). Moreover, and more specifically, evidence exists that politicians are responsive to the electorate in the area of consumer protection. For example, in liberal states, state legislatures are more likely to pass consumer protection statutes and to spend money on consumer protection programs (Bernacchi 1976; Ford 1977; Sigelman and Smith 1980). These findings suggest that enforcement of consumer protection regulations will also be influenced by the electorate’s ideology. However, while consumer protection is a salient issue, not all multi-state lawsuits are equally salient. Some lawsuits target relatively small companies while others focus on Fortune 500 companies, such as Reebok or Walgreens, and are thus more likely to garner more media attention. The question then becomes, how does the salience of each case affect the behavior of state AGs in multi-state litigation? In addition to the evidence showing government policy responsiveness to the electorate, many scholars have found that representation is often stronger when the policy issues in question are salient to voters (Burstein 2003; Gormley 1986; Jones 1994; Wlezien 2004). When voters pay closer attention to the issues, politicians work harder to make sure that policies match the preferences of voters. Consequently, I expect citizen ideology to have an impact on AG participation, but only when the case in question is salient. In this article, I use the size of a company as a proxy for case salience for two main reasons. First, as already suggested, when the subject of the case is a large, well-known company, the case is more likely to capture the attention of average people, partly through increased media attention. Second, cases that deal with large companies might be perceived by state AGs as salient to the public because of their wealth. Indeed, many conservative commentators and policymakers have complained that state AGs are primarily interested in prosecuting big companies, in order to extract large settlements, which are then delivered to either state consumers or to state government budgets (see e.g., Debow 2002; Pryor 1997, 2000). Moreover, in their study of the tobacco litigation, Spill, Licari, and Ray (2001) find that Republican AGs were more likely to file suit once the Liggett Company had settled, a move that demonstrated vulnerability and division on the part of the tobacco companies. This indicates that although Republicans did file later overall, they were still willing to file when it was clear that there was a higher probability of victory and larger cash settlements. Based on this research, I can formulate two expectations. First, because of the potential for larger settlements from large companies, state AGs might be more likely to join lawsuits that involve such companies. However, because there should also be greater responsiveness on the part of AGs in these salient cases, I expect state AGs from liberal states to join more salient lawsuits than AGs from conservative states would. In other words, a salient case should result in heightened participation for all state AGs, but particularly for those from liberal states. In sum, we should expect salient cases (cases with large, wealthy companies) to attract a larger number of AGs overall, but they will attract AGs from liberal states at a faster rate. Responsiveness to State Government Finally, will state AGs be responsive to other actors in state government, such as the governor or state legislature? The constitutional and statutory authority of state AGs to pursue policy initiatives in the public interest ensures that they can act as independent state executives, whether the governor or state legislators agree with AG policymaking or not. State legislatures do have the power to reduce AG budgets as well as the statutory or constitutional power of the AG, but “legislatures typically have not changed 8 provost spring 2010 / state politics and policy quarterly 9 the common law authority of State Attorneys General” (Ross 1990, 35). Moreover, Spill, Licari, and Ray (2001) found no gubernatorial influence over the timing of state lawsuits in the tobacco litigation, and when Mississippi Governor Kirk Fordice sued his AG, Democrat Mike Moore, in 1994 to prevent Moore from suing the tobacco companies, the Mississippi Supreme Court ruled in favor of Moore (Zimmerman 1998). Similarly, state courts also have little influence on AG behavior, as judges cannot constrain an AG’s common law or parens patriae authority to investigate or prosecute cases (Ross 1990). Finally, Provost (2003, 2006) finds virtually no state government influence on AG multi-state litigation decisions while Bowman (2004) finds only modest government influence. Thus, I expect that other actors in state government will have little to no influence on AG multi-state participation.