Paper 061 - Provost 2012 - Antitrust Law and Distributive Politics in the American States

This builds on previous work which demonstrates that in consumer protection cases, AGs only exhibit responsiveness to the electorate when defendants are large and wealthy Provost 2010b..

Antitrust Law and Distributive Politics in the American

States

Colin ProvostDepartment of Political Science/School of Public Policy

University College London29/30 Tavistock SquareLondon, WC1H 9QUUnited KingdomEmail: c.provost@ucl.ac.uk

Phone: +44 (0)20 7679 4903

Abstract: State enforcement by state attorneys general has become a major component of American antitrust law Much has been written about state antitrust enforcement, but existing accounts of AG incentives and behavior are incomplete As elected officials in

43 states, AGs must represent their constituents and therefore will be drawn to cases that maximize the level of settlement reward These will be cases with large, wealthy

defendants and where clear wrongdoing has been established Settlement rewards are likely to be higher when there are clear violations of the law, as in price-fixing cases, rather than in merger cases, where no wrongdoing has necessarily been established Finally, in an important area such as antitrust, state AGs also represent their constituents along ideological lines, but this relationship is also conditioned by case characteristics that involve the potential settlement reward Thus, this study has implications for how distributive politics shapes political responsiveness to the electorate

Keywords: antitrust law, competition law, state attorneys general, multi-state litigation, regulation, distributive politics

State enforcement of antitrust law has long been an important part of antitrust law enforcement in the American federal system States, through their attorneys general, have the power to enforce both state and federal antitrust laws and a wealth of research has been published describing this role of state AGs in antitrust federalism (Calkins 2003;First 2001; Lemos 2011) Some scholars have been highly critical of state antitrust enforcement, arguing that state AGs are too driven by electoral concerns rather than by considerations of economic efficiency, and that their standards diverge from federal standards, creating a messy, overlapping process (Flexner and Racanelli 1994; Gelhorn 1989; Greve 2005; Posner 2004; Zimmerman 1999) Other research defends the role of state AGs, claiming that they supply additional regulation not provided by federal

enforcers and that they have better knowledge of local markets (Calkins 2003; First 2001;Hubbard and Yoon 2005)

Valuable as these studies are, little research has demonstrated what drives AG behavior in antitrust enforcement As elected officials in 43 states, state AGs know they must be responsive to their state constituents, as the large majority of them face reelectionand many later run for higher office (Mahtesian 1996; Provost 2010a) However, the level of responsiveness AGs display towards constituents will vary according to the characteristics of particular cases they pursue Previous research on state antitrust

enforcement has not systematically captured this dynamic of AG accountability towards the electorate (Feinberg and Reynolds 2010; Lemos 2011) In this paper, I argue that AG responsiveness to the electorate is conditional on case salience and salience in turn is represented by defendant characteristics in antitrust cases

AGs will strive to be ideologically responsive to their constituents, partly by bringing home pork in the form of lawsuit settlements In antitrust regulation, they will

do this in two primary ways First, they will target wealthier, larger companies, who havethe ability to pay larger settlements As defendants grow in wealth, differences in

ideological responsiveness to the electorate ought to gradually disappear, as a majority of the AGs rush to claim their share of the settlement This builds on previous work which demonstrates that in consumer protection cases, AGs only exhibit responsiveness to the electorate when defendants are large and wealthy (Provost 2010b) However, in this case,the high probability of a reward from an antitrust case, particularly against a large,

wealthy defendant, means that AGs from conservative states are more likely to

temporarily put aside conservative principles in order to claim a share of the reward Second, they will focus on cases in which significant wrongdoing has occurred, and therefore where there is also potential for a larger payoff For state AGs, these cases provide strong incentives to participate Examining the types of antitrust infractions AGs are most likely to pursue has been done in previous research, but also without systematic explanation (Feinberg and Reynolds 2010)

To test these propositions, I examine all multi-state antitrust cases between 1989 and 2008 I begin by briefly describing the arrangements of American antitrust

federalism, as well as describe precisely how multi-state lawsuits function I then discussthe incentives and motivations of state AGs, which is followed by the presentation of testable hypotheses Finally, I present an empirical model designed to evaluate the relationships between AG multi-state participation, responsiveness to the electorate and case characteristics

Antitrust Federalism in Context

State antitrust enforcement has become a prominent part of antitrust federalism for multiple key reasons.1 First, in 1976, Congress enabled states to collect treble

damages for antitrust violations through the Hart Scott Rodino Antitrust Improvements

Act of 1976 The Act allowed state AGs to collect damages to be distributed back to the

states, its consumers or injured businesses This gave state AGs a strong incentive to enforce the law, as monetary rewards could serve as a form of regulatory pork-barreling,

a dynamic I discuss in more detail in the next section Additionally, many state AGs serving in the 1980s and 90s believed that federal enforcement of both consumer

protection and antitrust laws weakened considerably during the 1980s (Clayton 1994; Lynch 2001; Ross 1990; Zimmerman 1998) In antitrust, this shift has been well

documented, as the Federal Trade Commission and Department of Justice moved from the “Structure Conduct Paradigm” (SCP School) to the Chicago School philosophy of antitrust The SCP School dictated that big was bad and that all mergers of large

companies ought to be viewed suspiciously, while the Chicago School condoned mergers and large market players, as long as they did not diminish consumer welfare (Eisner and Meier 1990) Many AGs believed the shift was an excuse to pursue a regime of weak enforcement and thus, they banded together to pursue multi-state cases (Greenblatt 2003; Lynch 2001). 2

Multi-state cases begin because one or more state AGs suspect that state or federalantitrust laws are being broken on a large enough scale to affect the public at large State 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 allow AGs to investigate and gather

information 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 also file, provide assistance on the case and share the settlement money (Lynch 2001; Tierney 2002).3 Most defendants prefer to settle cases quickly and it is not uncommon for some businesses to ask that all or most states be included in the settlement to preempt the possibility of future litigation

(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 from the defendant(s)

Antitrust Enforcement and the Motivations of State AGs

To explain AG behavior in multi-state litigation, it is necessary to evaluate the institutional arrangements, incentives and motivations of state AGs As chief state law enforcement officer, the office of state AG is a very powerful one State AGs are chargedwith representing the state’s legal interests, as well as the governor, state legislators and other members of state government, but they are also responsible for representing the public interest, a broad claim that gives them much discretion in their policy pursuits (Davids 2005; Ross 1990) Two institutional factors ensure that state AGs are able to pursue their own policy goals with a minimum of interference from other state actors First, they are elected in 43 states, which makes the large majority of them accountable tothe electorate first.4 Second, in all 50 states, AGs have either common law authority or

parens patriae authority to enforce the law in the public interest as they see fit.5 In the

case of antitrust law, states have parens patriae authority under Hart Scott Rodino to sue

companies in federal court to recover monetary damages Thus, state AGs are executive, elected officials who are responsible to a large extent for judicial policy-making, who are accountable to the electorate and simultaneously charged with representing the interests

of state government This lengthy list of responsibilities ultimately helps to make state

AG one of the most powerful offices in American state government

Because AGs are elected in 43 states, electoral factors are a major motive for stateAGs, as the office is frequently used as a springboard into governorships and U.S Senate seats (Clayton 1994; Mahtesian 1996; Provost 2010a) Thus, state AGs must be mindful

of the concerns of voters, businesses and interest groups when they pursue their policy goals Many scholars have tried to systematically demonstrate what drives AG antitrust enforcement, but much of this research lacks a systematic accounting of responsiveness

to AG constituents Lemos (2011) provides an excellent discussion of how state AGs caninfluence the law more broadly, but accountability to the electorate is more incidental to the discussion Feinberg and Reynolds (2010) tell us more about state economic factors and AG characteristics, and not as much about AG constituents and the responsiveness relationship Finally, other scholars have been critical of the role of state antitrust

enforcement and their goal is not necessarily to provide a thorough explanation of what drives AG behavior (Greve 2005; Posner 2004; Zimmerman 1999) Here, I attempt to present a model of AG responsiveness to the electorate in multi-state antitrust

enforcement

The degree to which AGs will be ideologically responsive to their constituents in their policy making depends to a large extent on the salience of the policy to those

constituents A great deal of research has found that state politicians are ideologically responsive to the electorate when making policy (Brace et al 2002; Erikson, Wright and McIver 1993; Gray et al 2004; Jacoby and Schneider 2001) However, this relationship

is conditional according to the level of salience placed on the issue by the electorate If a particular issue escapes public attention, then policy makers may be able to follow their own policy preferences (to the extent that they diverge from voter preferences) or be quietly responsive to elites or organized interests The conventional cases of regulatory capture in which regulators do the bidding of regulated firms occur, in part, because the issue in question is important only to the regulated interests, but not to consumers, voters

or watchdog organized interests (Bernstein 1955; Peltzman 1976; Stigler 1971)

However, for issues that are salient to voters, politicians exhibit higher levels of

responsiveness to the electorate (Burstein 2003; Gormley 1986; Jones 1994; Wlezien 2004) That is to say, politicians from liberal areas will produce liberal policies and politicians from conservative areas will produce conservative policies

Consumer protection has traditionally been classified as a salient issue (Gerber and Teske 2000; Gormley 1986; Ringquist, Worsham and Eisner 2003), as everybody in society is a consumer and can fall victim to market failures, such as poor information (fraudulent or deceptive advertising) or market power (price fixing, bid rigging or supply restrictions) This salience is apparent in previous studies which demonstrate political responsiveness to the electorate in issues of consumer protection, as liberal states are more likely to pass consumer protection statutes and spend money on consumer

protection programs (Bernacchi 1976; Ford 1977; Sigelman and Smith 1980)

Additionally, state AGs from liberal states are likely to participate more frequently in multi-state consumer protection cases than those from conservative states (Provost 2006)

Salience in consumer protection can vary, however, with defendant

characteristics Some defendants are small and will generate neither attention nor the probability of a high settlement reward However, a case against a large, wealthy

defendant increases the probability of a high reward, as well as case attention, thereby raising the overall level of salience Indeed, Provost (2010b) has found that in consumer protection cases against small defendants, the relationship between AG behavior and citizen ideology does not hold up, as too few AGs join these suits However, in salient cases, (those with Fortune 500 defendants), the relationship does bear out, as

participation increases across the spectrum of AGs, but particularly so for AGs from liberal states

We should expect a different dynamic from antitrust cases Antitrust policy has traditionally been considered a less salient issue than consumer protection, due to its greater complexity and its focus on assessing the merits of mergers (Gerber and Teske 2000; Gormley 1986; Ringquist, Worsham and Eisner 2003) However, this

characterization is based largely on federal antitrust enforcement Multi-state antitrust cases can bring monetary rewards to state government, to the electorate and to injured businesses, thereby raising its level of issue salience considerably Under federal law, state AGs can pursue violators of antitrust statutes and recoup treble damages, which can then be paid back to the states or to affected consumers Thus, the potential promise of large payouts make antitrust cases among the most salient ones that state AGs pursue and

thus, we ought to see responsiveness on the part of state AGs to their constituents However, as defendants become large and wealthy, the political cost of not participating increases for AGs in conservative states, as the potential rewards of the case become too great to ignore These AGs then become more likely to participate, and the responsive relationship actually breaks down as a majority of the AGs rush in to claim their share of the spoils Thus, as case salience increases in consumer protection cases, the

conditioning effect is one of strengthening the relationship between citizen ideology and

AG behavior (Provost 2010b), but in antitrust cases, increases in salience have the effect

of weakening the relationship, as a majority of AGs are lured in by the potential rewards

of the case

Finally, the motivation to bring home large rewards will also influence the type of cases in which state AGs participate Price fixing and supply restriction cases are cases inwhich market power has clearly been abused and consumers have been made worse off as

a result of firms artificially raising prices or decreasing supply For state AGs, the

opportunity is perceived not only as a chance to bring justice to those consumers, but simultaneously as a method to bring funds to the state Merger cases do not necessarily share this trait, as potential wrongdoing is under consideration, rather than actual

wrongdoing Thus, the larger potential payoffs associated with price fixing and supply restriction cases ought to bring greater AG participation in those cases

Hypotheses

The salient nature of antitrust cases should ensure that AGs are responsive to theirelectorates Therefore, we should expect those AGs from liberal states to participate in

cases more often, as voters in those states will be more likely to support lawsuits to curtail market power abuses Consumers do not typically bring the complaints that generate antitrust lawsuits, as they do not have access to the information required to know if the law is being broken Rather, most complaints are brought by rival

businesses, which I address shortly However, consumers and voters will still react to settlement outcomes, along ideological lines Voters in conservative states are likely to look at intervention in the markets a little more suspiciously than voters from liberal states, whereas voters in liberal states are more likely to view the efforts of AGs more favorably

Hypothesis 1: As the electorate of a state becomes more liberal, AGs ought to participate in multi-state antitrust cases more frequently

However, this relationship is conditional on the types of cases in which AGs participate When a case is against a large and wealthy defendant, an AG is likely to think two things First, the company’s size is likely to lead to a higher level of attention for the case and therefore to consumers Second, the company’s large net worth may lead

to a very large settlement, which might directly benefit citizens of the states AGs from liberal states are still just as likely, if not more likely, to participate in these cases, as an AG’s participation still comports with those voters’ policy preferences, as well as brings distributive benefits to the states However, AGs from conservative states have more of adilemma, as they must balance conservative principles against the benefits of a

potentially large settlement As the potential for a large settlement increases, AGs from conservative states are more likely to participate, as the benefits become too significant too ignore Failing to join such a case is “like turning down federal stimulus money”,

according to one former state AG At this point, the responsiveness relationship begins tobreak down, as all AGs join in order to claim their share of a potential settlement

Hypothesis 2A: As defendants become larger and wealthier, more AGs are likely

representing consumers will pay attention to AG behavior and whether they sincerely enforce antitrust laws However, some states have a multitude of consumer groups, whileother states have no organized groups at all Previous research has demonstrated that the number of consumer groups drives multi-state participation in all types of consumer protection cases, not just antitrust cases (Provost 2006, 2010b) Thus, AGs ought to participate more frequently in multi-state antitrust cases when their states have a large number of consumer groups

Hypotheses 3: State AGs ought to participate in more multi-state antitrust cases, the larger the number of consumer groups in that state

Businesses also care a great deal about how state AGs enforce antitrust laws, but they do not necessarily speak with one voice For large businesses, there may be a desire for AGs to be minimally involved, as they may want to pursue mergers with other

companies and expand their market share in a particular industry However, the majority

of businesses are more likely to favor a level playing field and therefore, stronger

antitrust protections Mergers can make market competition difficult for the remaining

competitors in an industry, or for producers of other products further down the supply chain For example, retail gasoline stations may object to the merger of two large oil companies, claiming that it will result in higher crude oil prices Additionally,

manufacturers who do not engage in price fixing with their retailers will surely resent those manufacturers that do, as the price fixers will earn higher, albeit illegal, profits Consequently, according to one assistant state AG, many antitrust cases originate from complaints from businesses about other businesses Thus, it is possible that a vibrant, competitive business atmosphere in a state will lead to greater transparency and more complaints when antitrust violations are present On the other hand, such complaints may be idiosyncratic depending on the industry and may not follow a pattern of how large a particular state’s business community is

Hypothesis 4A: AGs will participate more frequently in multi-state antitrust cases the larger the state’s business community

Hypothesis 4B: AG participation in multi-state antitrust cases will not be affected

by the size of the state business community.

Finally, the type of antitrust case being pursued also has an impact on the level of

AG participation In antitrust cases, cases where wrongdoing has been identified and damages from the offense can be calculated will appeal more to AGs because they are more likely to attract the attention of consumers For example, in price fixing or supply restriction cases, AGs will examine how much consumers were charged over the

prevailing market rate and such calculations will, in part, determine how much restitutionthe business(es) have to pay back These cases are substantively different from merger cases in which wrongdoing has not necessarily occurred, but the potential size of a company’s market share may arouse suspicion among AGs or other businesses Mergers

might, but do not necessarily lead to abuses of market power, which is why they are less likely to generate wide consensus on participation among AGs AGs are more likely to join price-fixing and supply restriction cases because of the potential settlement, and this potential settlement is large because it is easier to establish a violation of the law.

Hypothesis 5: AGs are more likely to join price-fixing and supply restriction cases than they are to join merger cases

Finally, it is less clear how state AGs will approach monopoly cases Monopoly cases are similar to merger cases in that wrongdoing has not been proven, and there is only suspicion that wrongdoing might occur as a result of concentrated market share for afirm or small group of firms On the other hand, many multi-state monopoly cases tend

to involve pharmaceutical companies that have paid other companies not to release a generic version of a drug In these cases, state AGs do believe that wrongdoing has occurred, just as it would in a price-fixing case Thus, given the conflicting expectations,

I refrain from posing a testable hypothesis about state AGs and monopoly cases

Data and Methods

This analysis covers all multi-state antitrust cases initiated between 1989 and

2008 Data were gathered primarily from two sources: from the National Association of

Attorney General (NAAG) publication, Antitrust Bulletin and from NAAG’s Multi-State

Antitrust Litigation Database In this analysis, I only examine multi-state lawsuits rather than lawsuits that originate from single states This is not because single state actions are not worthy of study, but rather multi-state actions allow us to observe the decision of eachstate to join or not join a particular type of case.6 I perform statistical analyses to see precisely which state and case-level factors drive participation in multi-state lawsuits

Cases were divided into five categories according to information provided in each of the databases Merger cases are those in which a proposed merger of two or more companies

is challenged by multiple AGs Monopoly cases are those in which a company illegally conspires to ensure it has a monopoly over the production of a particular product These are common in the pharmaceutical industry when companies sometimes attempt to prevent the entry of generic versions of a drug onto the market Price-fixing cases are those cases in which a small number of companies have coordinated their actions and agreed on what prices to set, in effect rendering the market uncompetitive This can happen “horizontally” where retailers agree to fix prices, but it is also common in the

“vertical” sense of the infraction, in which manufacturers with a strong position in the market coerce retailers to set prices at certain levels Such cases are known as “resale price maintenance” cases Supply restriction cases, another category, have similar anti-competitive effects as price-fixing cases, but in these cases, companies conspire to limit the availability of a certain product, rather than conspire to fix prices.7 AG participation ought be greater in price-fixing and supply restriction cases than in merger cases, as consumers have lost money and business competitors have lost market position Finally, bid rigging cases are those cases in which two or more companies have conspired to fix a process of competitive tendering, in order to win, typically, a contract.8

Hypothesis 1 states that as state electorates become more liberal, AG multi-state participation is likely to increase In order to measure liberalism, I utilize the Berry et al (1998) citizen ideology measures It is necessary to have a measure that captures inter-year variation, as well as differences within parties across states and the Berry et al measures does both those things.9 In hypothesis 2, I argue that AGs are more likely to

join cases against large and wealthy defendants Large and wealthy defendants are defined as being companies that made the American or Global Fortune 500 in the year that the lawsuit was initiated Hypothesis 2A is tested by including an interaction term that combines citizen ideology and Fortune 500 status In hypothesis 3, I argue that as the number of consumer groups in a given state increases, AG participation in multi-state lawsuits also ought to increase To measure the presence of consumer groups, I take the number of member groups from the Consumer Federation of America Hypothesis 4 examines the relationship between a state’s business community and the AG’s

participation in multi-state lawsuits In order to gauge the strength of state business communities, I take the proportion of each state’s for profit lobby registrants as a

proportion of the overall lobbying community.10 Finally, hypothesis 5 looks at the effect

on participation of the type of case initiated The sources of these data were described earlier in this section, and they are used to generate four dichotomous variables for each type of case initiated: price-fixing, supply restriction, monopoly and bid rigging Merger cases do not appear as a variable in the model and thus, represent the baseline against which the other types of cases are compared

In addition to these variables that account for the tests in the five main

hypotheses, there are state-level variables which must be present in order for the model to

be properly specified First, strong consumer protection policy is typically a more

significant priority for wealthier states Sigelman and Smith argue that “to the extent thateducational and income characteristics of a state’s population are indicative of the

presence of middle-class values, we would expect these characteristics to be related to theadoption of consumer-oriented legislation” (1980: 61) Ford (1976) and Bernacchi

(1976) both find that states with high income levels are likely to pass consumer

protection laws Thus, I include annual median income to account for level of wealth in each state Second,, the presence of urban population should also have an impact on AG participation, as it is another important socioeconomic reason why states adopt consumer protection laws (Meier 1987) I generate annual population density figures by dividing each state’s population by its land area Third, there is evidence that antitrust

enforcement is counter-cyclical, meaning that enforcement tends to rise during periods of high unemployment (Feinberg and Reynolds 2010; Ghosal and Gallo 2001) Thus, I include the lagged annual average unemployment rate for each state Finally, to gauge office resources, I include a dummy variable for whether the state AG office includes a separate antitrust bureau Complete variable summaries and descriptive statistics are to