Galina Zudenkova

Lobbying as a Guard against Extremism

De Gruyter | 2017

Abstract

This paper analyzes endogenous lobbying over a unidimensional policy issue. Individuals differ in policy preferences and decide either to take part in lobbying activities or not. They are assumed to be group rule-utilitarian such that they follow the rule that, if followed by everyone else in their special interest group, would maximize their group’s aggregate welfare. Once formed, lobbies make contributions to the incumbent government in exchange for a policy favor as in a common-agency model. I show the existence of equilibrium with two organized lobbies. Individuals with more extreme preferences are more likely to join lobbying activities. Therefore, the lobbyists are rather extremists than moderates. However, the competition between those extreme lobbies results in a more moderate policy outcome relative to that initially preferred by the biased government. Lobbies therefore guard against extremism, while acting as moderators of the government’s preferences.

1 Introduction

Lobbying is a highly controversial phenomenon, often seen in a negative light by the public at large. There has been, however, an argument put forward in support of lobbying. According to this view, lobbying permits opposing interests to battle preventing extremism (see DeKieffer 2007 ). In other words, the Madisonian view of politics – in which factions compete with each other precluding tyranny – works for lobbies too. The only difference is that disputes are settled by interests groups fighting each other in the post-election stage rather than by elections. The present paper aims to contribute to this line of thinking.

I analyze a model of endogenous lobbying over a unidimensional policy issue. Individuals differ in policy preferences and decide either to join a lobby that represents their special interest group (a pro-policy lobby or an anti-policy lobby) or not to take part in lobbying activities. The incumbent government cares about policy outcomes and lobbies’ contribution payments. Once formed, lobbies make contributions to the government in exchange for a policy favor as in a common-agency model of Bernheim and Whinston (1986) , adapted to lobbying by Grossman and Helpman (1994) .

I follow a group-rule utilitarian approach to the individuals’ coordination problem, inspired by Harsanyi (1980) . According to this approach, individuals are assumed to be group rule-utilitarian such that they follow the rule that, if followed by everyone else in their special interest group, would maximize their group’s aggregate welfare from the policy outcome. I show the existence of equilibrium with two organized lobbies. Individuals with more extreme preferences are more likely to be involved in lobbying activities. In equilibrium, each lobby is characterized by a threshold level of preferences such that all individuals with more pro-policy views (for the pro-policy lobby) or more anti-policy views (for the anti-policy lobby) participate in lobbying activities. This is in line with the results of Glazer and Gradstein (2005) and McCarty, Poole, and Rosenthal (2006) that extremists want to contribute the most.

While the lobby members are rather extremists than moderates, the competition between the lobbies moderates the government’s preferences. In particular, it shifts the final policy outcome in favor of individuals who are initially disadvantaged by the government. Intuitively, consider the case of a government biased in favor of high policy levels. Individuals with preferences similar to the government’s have less stake in the policy, since they are initially favored by the government’s preferred policy. Anti-policy individuals, however, are disadvantaged by the government’s preferences. Owing to the concavity of preferences, anti-policy individuals gain more than pro-policy individuals from the same (in absolute value) policy change.[1] Therefore, they are willing to contribute more to the government for a policy change. As a result, under a government biased in favor of high policy levels, the anti-policy lobby is more numerous and contributes more than the pro-policy lobby. The equilibrium policy level is more moderate than the one preferred by the biased government prior to lobbying, and thus favors anti-policy individuals. A similar argument, in reverse, applies to the case of a government biased in favor of low policy levels under which the equilibrium policy level favors pro-policy individuals. I show therefore that under lobbying, the final policy outcome favors individuals who are initially disadvantaged by the biased government. In other words, lobbies act as moderators of the government’s preferences preventing extremism. These findings contribute to the literature on biased experts and advocates, which also reports the moderating effects of competing lobbies on policy outcomes (see Dewatripont & Tirole, 1999 ; Krishna & Morgan, 2001 ). The difference is that in my setting, the government is assumed to be biased and so the lobbies make efforts to offset the government’s bias by means of contributions.

Under an unbiased government, lobbying does not affect the policy outcome. In this case the lobbies “neutralize” one another, so that the pro-policy lobby’s bids are matched in the equilibrium by the anti-policy lobby’s bids, and political competition results in an unbiased policy outcome. However, each lobby makes a positive contribution to the government to avoid an undesired policy promoted by a competitor. This is in line with the findings of Persson and Tabellini (2002) that diverging lobbies may neutralize each other’s influence.

This paper follows the most prevalent approach in the formal literature, based on the assumption that lobbies influence political decisions through contributions (Baron, 1989 ; Becker, 1983 ; Becker, 1985 ; Snyder, 1990 ). The recent literature follows this approach to address the questions of commercial lobbying (Groll and Ellis 2014 ), lobbying organization structure (Bombardini and Trebbi 2012 ), barriers to entry (Kerr, Lincoln, and Mishra 2014 ), and others. Reviews of this and alternative approaches can be found in Austen-Smith (1997) , Grossman and Helpman (2001) , and Persson and Tabellini (2002) . I use the common-agency model of Bernheim and Whinston (1986) applied to lobbying by Grossman and Helpman (1994) . Lobbying is modeled as a “menu auction”, where lobbies confront government with contribution schedules that map any possible policy into a contribution payment. Several authors have applied the common-agency model of lobbying to study trade policy, commodity taxation, the provision of local public goods, and other policies (Dixit, Grossman & Helpman, 1997 ; Grossman & Helpman, 1996 ; Helpman & Persson, 2001 ; Persson, 1998 ). To address the individuals’ coordination problem, I follow the group-rule utilitarian approach proposed by Harsanyi (1980) . This approach has been applied to political economy questions (in particular, to turnout) by Coate and Conlin (2004) and Feddersen and Sandroni (2006) . My setting is similar to theirs such that each group’s rule is characterized by a critical threshold. An individual with more extreme policy preferences in my case (with lower voting costs in their case) than a corresponding threshold should take part in lobbying (voting) while an individual with more moderate policy preferences (with higher voting costs) should not.

This paper complements the literature on collective action and lobby organization which dates back to the seminal work of Olson (1965) . Recent contributions have addressed the question of lobby formation in several different contexts. Some authors have focused mainly on the formation of lobbies from exogenously given special interest groups with a fixed cost (Drazen, Limão & Stratmann, 2007 ; Felli & Merlo, 2006 ; Felli & Merlo, 2007 ; Laussel 2006 ; Leaver and Makris 2006 ; Mitra 1999 ; Redoano 2010). Others have in some way addressed the problem of individuals’ decisions to lobby. For example, Damania and Fredriksson (2000 , 2003 ) , Magee (2002) analyzed incentives for two firms and for n identical firms, respectively, to organize into a single industry lobby to affect policy outcomes. In turn, Bombardini (2008) proposed an “optimal lobby criterion” that reads as follows: it is optimal for a firm to “join the lobby” if the joint surplus of a prospective member firm and the lobby is higher under participation of the firm. Anesi (2009) analyzed the impact of moral hazard in teams on collective action. Furusawa and Konishi (2011) suggested a “free-riding-proof core” solution concept for the problem of the provision of public goods, which determines the formation of a contribution group, the level of provision of public goods, and the allocation of payoffs within the group. The present paper contributes to this literature and provides a novel model of endogenous lobby formation.

The rest of the paper is organized as follows. Section 2 lays out a model. Section 3 describes the common-agency model of lobbying. Section 4 develops the concept of lobby formation. Section 5 establishes the existence of equilibrium with two organized lobbies. Discussions are presented in Section 6. Finally, Section 7 concludes the paper.

2 Model

Suppose that a certain policy option y is available to a society. Think of this as a piece of legislation. The set of feasible policies is the closed interval 0 , 1 , where y = 0 stands for the lowest policy level and y = 1 stands for the highest policy level.

The society is inhabited by a large number (formally a continuum) of individuals, where the size (mass) of the population is normalized to unity. Individuals differ in their policy preferences. Denote by x 0 , 1 an individual’s preferred policy outcome. (I refer to an individual with an ideal policy x as “individual x ”.) I assume that x is distributed in the population according to a well-behaved continuous distribution. In what follows, F denotes x ’s cumulative distribution function. Individual x ’s utility from policy y is given by

u x y = x y 2 .

The incumbent government decides on a policy outcome y . The government is policy-motivated and so cares about the policy. Its preferred policy outcome is denoted by γ 0 , 1 .[2] The government also cares about the total level of political contributions denoted by C 0 .[3] Its preferences are represented by

U γ y = γ y 2 + C .

Note that in the absence of political contributions ( C = 0 ), the government chooses to implement its preferred policy γ . Then, individuals with lower ideal policies, x < γ , prefer a lower policy level to be implemented. I refer to those as anti-policy individuals. However, individuals with higher ideal policies, x > γ , prefer a higher policy level. I refer to those as pro-policy individuals.

Policy-making involves not only government decision-making but also special interest politics, or lobbying. Lobbying is modeled here as a two-stage game. The first stage of the game is a lobby formation stage, where individuals decide whether to participate in lobbying activities or not. The second stage of the game is a contribution game, where lobbies (organized in the first stage) offer the government contributions to affect the policy outcome. The game is described in detail in the following subsections.

2.1 First Stage: Lobby Formation

Given the policy setting, it is reasonable to focus on the case in which two lobbies can be formed: a lobby of anti-policy individuals, given by a set A 0 , γ , and a lobby of pro-policy individuals, given by a set P γ , 1 . If the individuals cared only about their net benefit from lobbying then in this setting with a continuum of individuals and costly lobbying, there would be no equilibrium in which a positive fraction of individuals would participate in lobbying activities. I assume however that the individuals are group rule-utilitarian in the following sense. Each individual has a rule that he understands he has to follow, and derives utility from acting according to this rule. This is the rule that, if followed by every other individual from the same special interest group, would maximize this group’s aggregate utility from the policy outcome y . The aggregate utility of anti-policy individuals from the policy outcome y is equal to

0 γ u x y d F x ,

while that of pro-policy individuals is equal to

γ 1 u x y d F x .

Then an individual faces the following trade-off when deciding on whether to participate in lobbying activities. On the one hand, he realizes that given a continuum of individuals, his participation would not affect the policy outcome but would give him utility from acting according to the rule. On the other hand, shirking would save him a contribution fee in case the rule specifies that he should participate in lobbying. In what follows, the utility from following the rule is assumed to be sufficiently high such that it exceeds the highest possible contribution fee in each lobby.[4] Then individuals have no incentive to shirk and so follow the rule that, if followed by every other individual from the same special interest group, would maximize the group’s aggregate utility from the policy outcome.[5]

Let each special interest group’s rule be a critical policy threshold such that an individual with more extreme policy preferences should take part in lobbying while an individual with more moderate policy preferences should not. Formally, I denote by α 0 , γ the critical policy threshold for anti-policy individuals with x < γ . Their rule specifies that anti-policy individuals with ideal policies lower than α should enter lobby A while those with ideal policies higher than α should not. Likewise, I denote by π γ , 1 the critical policy threshold for pro-policy individuals with x > γ . Their rule indicates that pro-policy individuals with preferred policies higher than π should enter lobby P while those with ideal policies lower than π should not. The rule α that maximizes the aggregate utility of anti-policy individuals from the policy outcome is the one that each anti-policy individual x < γ understands he should follow. Likewise, the rule π that maximizes the aggregate utility of pro-policy individuals from the policy outcome is the one that each pro-policy individual x > γ reasons he should follow.

2.2 Second Stage: Contribution Game

I focus on lobbying activities in the context of the common-agency model of Bernheim and Whinston (1986) , adapted to lobbying by Grossman and Helpman (1994) . In this approach, lobbying is modeled as a “menu auction”, where lobbies confront government with contribution schedules that map any possible policy into a contribution payment.

First, each lobby i A , P , noncooperatively and simultaneously, presents its common agent, the government, with a contribution schedule C i y , giving a binding promise of payment conditional on a chosen policy level y .[6] Following the literature cited in the previous paragraph, I concentrate on (globally) truthful contribution schedules that satisfy

C i y = max U i y b i , 0 ,

where U i denotes lobby i ’s gross objective function and b i is a constant chosen optimally by lobby i . So lobbies reveal their true preferences: they contribute to the government the maximum amount that they are willing to exchange for the government’s decision. I follow the existing literature (e. g., Drazen, Limão & Stratmann, 2007 ; Grossman & Helpman, 1994 ; Mitra, 1999 ) and assume that the objective of lobby i is to maximize the net aggregate utility of its members, namely

U i y C i y x i u x y d F x C i y .

Second, the government sets y to maximize its utility:

(1) γ y 2 + C A y + C P y .

Equilibrium of the game is a subgame perfect Nash equilibrium in the contribution schedules, the chosen policy and the lobbies’ compositions. In the following section, the game is analyzed backwards. First, I solve for the policy level y and lobbies contributions C A and C P . Second, I find the critical policy thresholds α and π that characterize the rules followed by the anti-policy and pro-policy individuals, respectively.

3 Common-Agency Lobbying

Suppose that two lobbies A and P have been formed. To derive equilibrium in truthful strategies, the following lemma is used. (See Bernheim & Whinston, 1986 ; Dixit, Grossman & Helpman, 1997 ; for a proof.)

Lemma 1

The equilibrium policy is Pareto optimal in the bilateral relation between the government and each lobby.

Therefore, the equilibrium policy maximizes the sum of the lobbies’ net objective functions

i A , P U i y C i y

and the government’s objective (1). This sum equals

(2) γ y 2 + U A y + U P y .

Maximizing eq. (2) yields the equilibrium policy level y :

y = arg max y 0 , 1 γ y 2 + U A y + U P y .

To find the contribution levels in the equilibrium, define y j to be the policy that would emerge if the contribution offered by lobby i were zero, i , j P , A , j i . So,

(3) y A = arg max y 0 , 1 γ y 2 + U A y ,

y P = arg max y 0 , 1 γ y 2 + U P y .

In other words, y j is the policy that would emerge if lobby i were not formed.

Lobby i will raise the constant b i in its truthful contribution schedule to the point where the government is just indifferent between choosing policy y j and choosing the equilibrium policy y , i. e.,

γ y A 2 + C A y A = γ y 2 + C P y + C A y ,

γ y P 2 + C P y P = γ y 2 + C P y + C A y .

Now one can solve for the lobbies’ contributions in equilibrium. The following proposition summarizes the results of the lobbies’ common-agency contribution game. Proofs of this and other results are delegated to the Appendix.

Proposition 1

There exists an equilibrium in truthful strategies such that the equilibrium policy level is given by

y = arg max y 0 , 1 γ y 2 + U A y + U P y .

The lobbies’ equilibrium contributions are equal to

C A C A y = γ y P 2 + γ y 2 + U P y P U P y ,

C P C P y = γ y A 2 + γ y 2 + U A y A U A y .

If there is just one organized lobby, the government derives exactly the same utility as it would have achieved without any contribution. Thus, a lobby that faces no competition captures the entire surplus from lobbying activities. If two lobbies compete for the final policy, the government captures some surplus from lobbying activities, and each lobby pays according to the political strength of its rival.

The lobby formation stage of the game is analyzed in the following section.

4 Lobby Formation

The individuals follow the threshold rules α 0 , γ and π γ , 1 that specify that an anti-policy individual with bliss point x should enter lobby A if x < α and should not take part in lobbying otherwise, while a pro-policy individual should enter lobby P if x > π and should not take part in lobbying otherwise.

Once formed, the lobbies care about the aggregate utility of their members. Therefore, their gross objective functions read

U A y = 0 α u x y d F x ,

U P y = π 1 u x y d F x .

In what follows, for ease of exposition and analysis, x is assumed to be distributed uniformly with density 1 , x U 0 , 1 . Then the lobbies’ gross objective functions amount to

U A y = 0 α u x y d x = α y 2 + α 2 y 1 3 α 3 ,

U P y = π 1 u x y d x = 1 π y 2 + 1 π 2 y 1 3 1 π 3 .

The following lemma characterizes the equilibrium policy level y and the lobbies’ equilibrium contributions in the case of two organized lobbies A and P .

Lemma 2

Given threshold rules α and π , the equilibrium policy level with two organized lobbies A and P is equal to

y α , π = 1 + α 2 π 2 + 2 γ 2 2 + α π .

The lobbies’ equilibrium contributions in the case of two organized lobbies A and P are equal to

C A α , π = α 2 1 2 α + π α π 2 + 2 γ 2 4 2 π 2 + α π 2 ,

C P α , π = 1 π 2 1 + α + π + π α α 2 2 γ 2 4 1 + α 2 + α π 2 .

In what follows, I show the existence of nonempty organized lobbies A and P , i. e., I find α and π that maximize the aggregate utilities of the corresponding special interest groups from the policy outcome.

5 Equilibrium

The following proposition establishes the existence of equilibrium with two organized lobbies.

Proposition 2

Given the government’s preferred policy γ 0 , 1 , there exist the following equilibrium with two nonempty lobbies A and P , equilibrium policy level y and equilibrium contributions C A and C P :

– for γ 0 , 1 2 , A = 0 , γ , P = 2 + γ 3 + 2 γ , 1 , y = 2 + γ 3 + 2 γ ,

C A = γ 2 4 3 + 2 γ 6 + γ 3 + 2 γ 4 2 4 3 + 2 γ 3 + 2 γ γ , C P = 4 2 3 + 2 γ + γ 4 + γ 2 3 + 2 γ 2 4 1 + γ ;

– for γ 1 2 , 1 , A = 0 , 2 + γ + 5 2 γ , P = γ , 1 , y = 2 + γ + 5 2 γ ,

C A = 9 4 5 2 γ + γ 2 5 2 γ + γ 6 2 4 2 γ , C P = 1 γ 2 45 20 5 2 γ + γ 4 5 2 γ + γ 18 4 5 2 γ 1 + γ .

According to Proposition 2, for the government’s preferred policy γ 0 , 1 2 , all anti-policy individuals join the anti-policy lobby A in equilibrium but only a fraction of pro-policy individuals join the pro-policy lobby P . In particular, only individuals with preferred policies higher than 2 + γ 3 + 2 γ enter lobby P . For the government’s preferred policy γ 1 2 , 1 , all pro-policy individuals join the pro-policy lobby P in equilibrium but only a fraction of anti-policy individuals participate in lobbying. Specifically, only individuals with preferred policies lower than 2 + γ + 5 2 γ join lobby A . These results suggest that lobbyists are rather extremists than moderates. In other words, some individuals with moderate preferences do not lobby in equilibrium. Those are members of a larger special interest group (recall that all members of a smaller special interest group lobby in equilibrium). Intuitively, consider members of a smaller special interest group. They follow the rule that maximizes the group’s aggregate utility from the policy outcome and so end up including everyone in the lobby in order to represent the interests of every single member. In turn, members of the larger special interest group realize that in order to maximize their group’s aggregate utility given lobbying efforts of the smaller group and the government’s biased position, their lobby should advocate quite an extreme policy and so should include extremists. Including moderates into the lobby would moderate its objective, which would result in a more moderate policy outcome and so in a lower aggregate utility of the group’s members. Therefore, the moderate members of the larger special interest group do not join lobbying activities.

The result that some moderates do not lobby in equilibrium arises due to heterogeneity of preferences. The moderates’ interests are represented to some extent by the government that advocates a relatively moderate (though possibly biased) policy. Representing their interests also through lobbying would moderate the corresponding lobby. This would lead to a more moderate outcome, which would definitely benefit those moderates but not their special interest group. Since the individuals are group-rule utilitarian, they follow the rule that maximizes the group’s aggregate utility from the policy outcome and so leave the moderates aside. It is important to point out that this intuition applies to a larger group whose members have more heterogeneous preferences which range from quite extreme to rather moderate. A smaller group is formed by more homogeneous individuals with rather extreme preferences and so is better off including every member in lobbying activities.

It is important to emphasize that the rule each special interest group follows is assumed to take the form of a policy threshold. Consider the anti-policy group, for example. An anti-policy individual joins the corresponding lobbying group if his ideal policy lies in the interval 0 , α where α is determined by the rule. Alternatively, a rule might determine an interval α _ , α such that α _ 0 , α . In Appendix D, I show that given the rules α _ , α and π , 1 of the anti-policy and pro-policy groups, the anti-policy group’s aggregate utility from the policy outcome is maximized at α _ = 0 . Consider next the pro-policy group. Its rule determines the interval π , 1 such that a pro-policy individual participates in lobbying if his ideal policy is in this interval. A rule might instead determine an interval π _ , π such that π π _ , 1 . I prove in Appendix D that given the rules 0 , α and π , π , the pro-policy group’s aggregate utility from the policy outcome is maximized at π = 1 . Therefore, even if the rules took form of intervals α _ , α and π _ , π , the lobby composition, policy level and contributions described in Proposition 2 would still form equilibrium of the game.

The following corollary provides comparative statics results.

Corollary 1

The more biased the government in favor of high policy levels, the more numerous the anti-policy lobby A and the less numerous the pro-policy lobby P :

d α d γ > 0 , d 1 π d γ < 0.

The more biased the government in favor of high policy levels, the higher the policy level y implemented in equilibrium, the larger the anti-policy lobby contributions C A and the lower the pro-policy lobby contributions C P :

d y d γ > 0 , d C A d γ > 0 , d C P d γ < 0.

When the government is biased in favor of low policy levels, γ 0 , 1 2 , the pro-policy lobby P is more numerous and contributes more than the anti-policy lobby A . The equilibrium policy level favors pro-policy individuals, i. e., it is higher than the government’s preferred level (the policy level that would emerge without lobbying).

When the government is biased in favor of high policy levels, γ 1 2 , 1 , the pro-policy lobby P is less numerous and contributes less than the anti-policy lobby A . The equilibrium policy level favors anti-policy individuals, i. e., it is lower than that preferred by the government.

In case of the unbiased government, γ = 1 2 , the lobbies are of the same size and contribute the same amount to the government. Lobbying does not affect the policy outcome, i. e., the equilibrium policy level is equal to that preferred by the government.

Thus, lobbying somewhat moderates the government’s preferences. The implemented policy favors those individuals who are initially disadvantaged by the government’s preferred policy. Consider a biased government (which is in favor of either low or high policy levels). Then individuals whose preferences differ considerably from the government’s have more stake in the policy than individuals with preferences similar to the government’s. In other words, owing to the concavity of preferences, “disadvantaged” individuals are willing to pay more than “favored” individuals for the same (in absolute value) policy change. As a result, the lobby of “disadvantaged” individuals is more numerous and contributes more than the lobby of initially “favored” individuals. And the final policy becomes less extreme and goes in favor of the lobby with higher relative political strength, i. e., the lobby of individuals who are initially disadvantaged by the government’s preferred policy. It is important to emphasize that even though those individuals influence the final policy more than the advantaged group, they have to pay for it. Indeed, in equilibrium, the disadvantaged group’s contributions are higher than those of the advantaged group.

When the government is unbiased, lobbying does not affect the policy outcome. In this case the lobbies “neutralize” one another, so that in equilibrium P ’s bids for a higher policy level are matched by A ’s bids for a lower policy level, and political competition results in an unbiased policy outcome. Nonetheless, each lobby has to make a positive contribution in order to induce the government to choose this outcome rather than one that would be worse from that lobby’s perspective.

I show in Appendix F that the government utility in equilibrium is non-negative, increases in γ for γ 0 , 1 2 and decreases in γ for γ 1 2 , 1 . If there were no lobbies formed, then the government would implement its preferred policy γ and get zero utility. It implies that in equilibrium, the government extracts rents from competing lobbies. Moreover, the less biased the government, the more rents it extracts. The unbiased government, γ = 1 2 , extracts the most rents. Intuitively, in this case, the lobbies neutralize each other, the government implements its preferred policy and so faces no utility loss. Moreover, none of the special interest groups has a policy advantage and so both lobbies have to contribute considerably to match each other’s contributions.

Note that the equilibrium policy level under a government biased in favor of high policy levels is higher than that under a government biased in favor of low policy levels, with or without lobbying. Therefore, the pro-policy individuals would prefer a government biased in favor of high policy levels to that biased in favor of low policy levels in spite of the fact that under the latter they could influence the final policy in their favor by lobbying. In turn, the anti-policy individuals would prefer a government biased in favor of low policy levels, even though they could lobby more successfully under a government biased in favor of high policy levels.

6 Discussion

My model of lobby formation predicts that lobbying over a particular policy issue moderates policy outcomes. This is in line with various facts about lobby impacts on policymaking. For example, in 2000, President Clinton signed the China Trade Bill which extended permanent trade relations with mainland China. Supporters of the bill (such as the US Chamber of Commerce and the Business Roundtable) argued that the bill would provide ample business opportunities and so took a lot of lobbying efforts to get it signed. Opponents of the bill (such as labor unions and environmental groups) opposed the bill because of China’s human rights and labor standards issues. They also believed that it would threaten the U.S. labor market because of job losses. The opponents’ concerns were addressed by a number of provisions, which led to a compromise towards a more moderate policy outcome relative to the initial pro-bill position of the Clinton administration.[7] Another example is the California Clean Energy and Pollution Reduction Act signed into law in 2015. The bill supporters (such as climate change groups and some Silicon Valley companies) advocated establishing new clean energy and greenhouse gas reduction goals. The initial proposal called for a 50 % reduction in gasoline use by 2030. The opponents (petroleum companies and drivers associations) argued that the bill offered no specific policy plans and could lead to gas rationing, driving restricting and fuel prices rising. They launched a massive lobbying campaign which resulted in the bill being amended and the petroleum provisions being dropped. Therefore, their lobbying efforts moderated the policy outcome to some extent.[8]

It is important to emphasize, however, that in the present model the individuals differ only in their policy preferences but not in their financial endowments. In other words, all individuals can pay their corresponding contribution fees. This assumption is crucial for my results. Intuitively, in the case of financial constraints, the individuals would not be able to make the contribution payments and so would not be able to offset the government’s bias. Then the moderating effect of lobbying would not arise. One can think of my setting as an “equal opportunity” environment in which the special interest groups face no or insignificant financial constraints.

The other important assumption of my model is that of no contribution caps (for models of contribution caps, see Cotton, 2012 ; Dahm & Porteiro, 2008 ; Drazen, Limão & Stratmann, 2007 ). Contribution limits would somewhat offset the moderating effect of lobbying in my model. Intuitively, consider a government biased in favor of low policy levels. In case of no contribution caps, the pro-policy lobby contributes more than the anti-policy lobby and so is able to “convince” the government to implement a less biased policy, which gives rise to the moderating effect. Suppose next that contribution limits apply and that the contribution cap falls between the equilibrium contributions of two lobbies. In this case, the pro-policy lobby would contribute the amount equal to the contribution cap while the anti-policy lobby would somewhat adjust its contribution downwards. Still, due to the concave utilities, the pro-policy lobby adjustment would exceed that of the anti-policy lobby. Therefore, the final policy would not be as moderate as in the case of no contribution limits. It implies that in my model, the contribution limits lead to less moderate outcomes and so partially neutralize the guarding effect of lobbying. It is important to stress that this is due to the “equal opportunity” nature of my framework in which individuals have the same financial endowments. In the case of endowment heterogeneity, the introduction of contribution caps might amplify the guarding effect of lobbying.

Note also that the quantitative results of my model depend on the way the anti- and pro-policy special interest groups are defined (with respect to the government’s preferred policy γ ). Intuitively, this threshold divides individuals in two groups: those who prefer a policy lower than the government’s and those who prefer a policy higher than the government’s. It is therefore a natural critical point to define the special interest groups with divergent preferences. Alternatively, the interest groups could be defined by some other threshold (e. g., a median policy 1 2 ) or by a pair of thresholds x _ and x ( x _ < x ) such that individuals with ideal policies lower than x _ belong to the anti-policy special interest group while individuals with ideal policies higher than x belong to the pro-policy special interest group. Under those alternative definitions of interest groups, the main intuition for my results would still prevail. So the qualitative results of my model would hold unless the thresholds x _ and x were so extreme that the corner solutions arose in equilibrium, i. e., α = x _ and π = x . In this case, the results about moderating effects of lobbying would depend on the size of special interest groups, x _ and 1 x .

7 Conclusion

The present paper studies the impact of lobbying on government decision-making. I develop a model of endogenous lobby formation, which predicts that individuals with more extreme preferences are more likely to participate in lobbying. It follows that the lobbyists are rather extremists than moderates. However, the competition between the lobbies benefits the public at large resulting in a less extreme policy outcome relative to that initially preferred by the government. I show that lobbying somewhat moderates the government’s preferences. The final policy under lobbying favors individuals who are initially disadvantaged by the government’s preferred policy. Indeed, under a government biased in favor of high policy levels, the final policy outcome is somewhat more moderate than that initially preferred by the biased policymakers. Under a government biased in favor of low policy levels, lobbying moderates the final policy in favor of pro-policy individuals. In the case of an unbiased government, lobbying does not affect the final policy, and political competition results in an unbiased outcome. However, each lobby has to contribute the same amount to the government to maintain this policy level.

My paper therefore presents a formal argument in support of lobbying as a guard against extremism. According to my results, the competition between extreme special interests results in a quite moderate policy outcome, which would not be achieved with no lobbies around. Nevertheless, this paper constitutes a partial attempt to study individual lobbying behavior and focuses on a specific form of coordination among citizens. It might be interesting to consider different coordination mechanisms that citizens use to overcome the free-riding problem.[9] It might also be interesting to consider a nonsymmetric distribution of preferences. In this case, I expect that the qualitative results presented above will still hold (except, probably, those for the case of an unbiased government).

Acknowledgments

The author is grateful to Daron Acemoglu, Antonio Cabrales, Filipe Campante, Luis Corchón, Jacques Crémer, M. A. de Frutos, Raquel Fernández, Rebecca Morton, Ignacio Ortuño Ortn, Albert Solé-Ollé, Allard Van Der Made, Bengt-Arne Wickström, two anonymous referees and Johann Brunner, the editor, for helpful comments, suggestions and encouragement; and to Cung Truong Hoang for research assistance. The grant from Karin-Islinger-Stiftung foundation is gratefully acknowledged. The usual disclaimer applies.

Appendix

Proof of Proposition 1

The lobbies’ truthful contribution schedules are

(4) C P y = max U P y b P , 0 , C A y = max U A y b A , 0 .

The constants b P and b A in the lobbies’ truthful contribution schedules satisfy

(5) γ y A 2 + C A y A = γ y 2 + C P y + C A y , γ y P 2 + C P y P = γ y 2 + C P y + C A y .

Plugging eq. (4) into eq. (5) yields

U P y b P = γ y A 2 + γ y 2 + U A y A U A y ,

U A y b A = γ y P 2 + γ y 2 + U P y P U P y ,

where the first line is lobby P ’s equilibrium contribution C P , and the second line is lobby A ’s equilibrium contribution C A . ■

Proof of Lemma 2

The equilibrium policy level is equal to

y = arg max y 0 , 1 γ y 2 + U A y + U P y = arg max y 0 , 1 γ y 2 α y 2 + α 2 y 1 3 α 3 1 π y 2 + 1 π 2 y 1 3 1 π 3 .

The first-order condition of this maximization problem amounts to

1 + α 2 π 2 + 2 γ 2 2 + α π y = 0.

The second-order condition is satisfied and so

y α , π = 1 + α 2 π 2 + 2 γ 2 2 + α π .

If lobby P did not contribute to the government, the following policy would emerge:

y A = arg max y 0 , 1 γ y 2 + U A y = arg max y 0 , 1 γ y 2 α y 2 + α 2 y 1 3 α 3 .

The first-order condition is given by

α 2 + 2 γ 2 1 + α y = 0.

The second-order condition is satisfied and therefore

y A α = α 2 + 2 γ 2 1 + α .

Likewise, if lobby A did not contribute then the government would choose the following policy:

y P = arg max y 0 , 1 γ y 2 + U P y = arg max y 0 , 1 γ y 2 1 π y 2 + 1 π 2 y 1 3 1 π 3 .

The first-order condition amounts to

1 π 2 + 2 γ 2 2 π y = 0.

The second-order condition is satisfied and so

y P π = 1 π 2 + 2 γ 2 2 π .

According to Proposition 1, the lobbies’ equilibrium contributions are equal to

C A C A y = γ y P 2 + γ y 2 + U P y P U P y ,

C P C P y = γ y A 2 + γ y 2 + U A y A U A y .

Plugging y α , π , y A α and y P π yields

C A = γ 1 π 2 + 2 γ 2 2 π 2 + γ 1 + α 2 π 2 + 2 γ 2 2 + α π 2 + 1 π 1 π 2 + 2 γ 2 2 π 2 + 1 π 2 1 π 2 + 2 γ 2 2 π 1 3 1 π 3 1 π 1 + α 2 π 2 + 2 γ 2 2 + α π 2 + 1 π 2 1 + α 2 π 2 + 2 γ 2 2 + α π 1 3 1 π 3 = α 2 1 2 α + π α π 2 + 2 γ 2 4 2 π 2 + α π 2 ,

and

C P = γ α 2 + 2 γ 2 1 + α 2 + γ 1 + α 2 π 2 + 2 γ 2 2 + α π 2 + α α 2 + 2 γ 2 1 + α 2 + α 2 α 2 + 2 γ 2 1 + α 1 3 α 3 α 1 + α 2 π 2 + 2 γ 2 2 + α π 2 + α 2 1 + α 2 π 2 + 2 γ 2 2 + α π 1 3 α 3 = 1 π 2 1 + α + π + π α α 2 2 γ 2 4 1 + α 2 + α π 2

Proof of Proposition 2

The threshold rule α 0 , γ maximizes the aggregate utility of anti-policy individuals from the policy outcome y given by

0 γ u x y d x = γ y 2 + γ 2 y 1 3 γ 3 .

Plugging in the expression for y α , π from Lemma 2 yields

(6) γ 1 + α 2 π 2 + 2 γ 2 2 + α π 2 + γ 2 1 + α 2 π 2 + 2 γ 2 2 + α π 1 3 γ 3 .

The first derivative of eq. (6) with respect to α is equal to

(7) γ 1 α 4 + α + 2 γ + 2 α π π 2 1 + α π α γ + π 2 2 + α π 3 .

The first derivative (7) is strictly positive for all α 0 , γ only when 0 < γ 1 2 and γ π < 1 2 γ + γ . In this case, α = γ maximizes eq. (6). The first derivative (7) is never strictly negative for all α 0 , γ . The first derivative (7) equals zero in the following cases:

α = 0 when γ = 0 and 0 π 1 ,

α = γ when 0 < γ 1 2 and π = 1 2 γ + γ ,

α = 2 + 5 + 2 γ 4 π + π when 0 < γ 1 2 and 1 2 γ + γ < π 1 , or when 1 2 < γ < 1 and γ π 1 ,

α = 1 + 3 when γ = 1 and π = 1 .The second derivative of eq. (6) is given by 2 α γ γ 1 + α 4 + α 2 γ 2 α π + π 2 2 2 + α π 3 γ 4 + 2 α 2 π 1 + α π α γ + π 2 2 + α π 3 + 3 γ 1 + α 4 + α 2 γ 2 α π + π 2 1 + α π α γ + π 2 2 + α π 4 .

Evaluating it at the critical points yields that the second-order condition holds for α = γ , α = 2 + 5 + 2 γ 4 π + π and α = 1 + 3 . Therefore, the argmax of eq. (6) is as follows. For γ = 0 , α = 0 for all π γ , 1 . For γ 0 , 1 2 ,

α = { γ i f π γ , 1 2 γ + γ , 2 + 5 + 2 γ 4 π + π i f π 1 2 γ + γ , 1 .

For γ 1 2 , 1 , α = 2 + 5 + 2 γ 4 π + π for all π γ , 1 . For γ = 1 , α = 1 + 3 for π = 1 .

The threshold rule π γ , 1 maximizes the aggregate utility of pro-policy individuals from the policy outcome y given by

γ 1 u x y d x = 1 γ y 2 + 1 γ 2 y 1 3 1 γ 3 .

Plugging in the expression for y α , π from Lemma 2 yields

(8) 1 γ 1 + α 2 π 2 + 2 γ 2 2 + α π 2 + 1 γ 2 1 + α 2 π 2 + 2 γ 2 2 + α π 1 3 1 γ 3 .

The first derivative of eq. (8) with respect to π is equal to

(9) 1 γ 1 + 2 γ + α π 2 4 π 1 α 2 + α 1 + γ 1 + γ π π 2 2 + α π 3 .

The first derivative (9) is never strictly positive for all π γ , 1 . The first derivative (9) is strictly negative for all π γ , 1 only when 1 2 < γ < 1 and γ 1 + 2 γ < α γ . In this case, π = γ maximizes eq. (8). The first derivative (9) equals zero in the following cases:

π = 2 3 when γ = 0 and α = 0 ,

π = 2 + α 3 + 4 α 2 γ when 0 < γ 1 2 and 0 α γ , or when 1 2 < γ < 1 and 0 α γ 1 + 2 γ ,

π = 1 when γ = 1 and 0 α 1 .

The second derivative of eq. (8) is equal to

1 γ 1 + 2 γ + α π 2 4 π 1 + γ 2 π 2 2 + α π 3 + 3 1 γ 1 + 2 γ + α π 2 4 π 1 α 2 + α 1 + γ 1 + γ π π 2 2 + α π 4 1 γ 4 + 2 α π 1 α 2 + α 1 + γ 1 + γ π π 2 2 + α π 3 .

Evaluating it at the critical points yields that the second-order condition holds for π = 2 3 and π = 2 + α 3 + 4 α 2 γ . Thus, the argmax of eq. (8) is as follows. For γ = 0 , π = 2 3 for α = 0 . For γ 0 , 1 2 , π = 2 + α 3 + 4 α 2 γ for all α 0 , γ . For γ 1 2 , 1 ,

π = { 2 + α 3 + 4 α 2 γ i f α 0 , γ 1 + 2 γ , γ i f α γ 1 + 2 γ , γ .

For γ = 1 , π = 1 for all α 0 , γ .

I find next the equilibrium threshold rules α 0 , γ and π γ , 1 given the government’s preferred policy γ 0 , 1 . Given the argmax of eq. (6) and argmax of eq. (8), I find that

for γ = 0 , α = 0 and π = 2 3 ,

for γ 0 , 1 2 , α = γ and π = 2 + γ 3 + 2 γ ,

for γ 1 2 , 1 , α = 2 + γ + 5 2 γ and π = γ ,

for γ = 1 , α = 1 + 3 and π = 1 .

This can be summarized as follows. For γ 0 , 1 2 , α = γ and π = 2 + γ 3 + 2 γ . For γ 1 2 , 1 , α = 2 + γ + 5 2 γ and π = γ . Plugging α and π into y α , π , C A α , π and C P α , π in Lemma 2 yields the equilibrium policy level y and the equilibrium contributions C A and C P . In particular, for γ 0 , 1 2 , those are equal to

y = 2 + γ 3 + 2 γ ,

C A = γ 2 4 3 + 2 γ 6 + γ 3 + 2 γ 4 2 4 3 + 2 γ 3 + 2 γ γ ,

C P = 4 2 3 + 2 γ + γ 4 + γ 2 3 + 2 γ 2 4 1 + γ .

For γ 1 2 , 1 , they amount to

y = 2 + γ + 5 2 γ ,

C A = 9 4 5 2 γ + γ 2 5 2 γ + γ 6 2 4 2 γ ,

C P = 1 γ 2 45 20 5 2 γ + γ 4 5 2 γ + γ 18 4 5 2 γ 1 + γ .

Rules Characterized by Two Thresholds

I show first that given the rules α _ , α and π , 1 , the anti-policy group’s aggregate utility from the policy outcome is maximized at α _ = 0 . Consider the following rules: α _ , α and π , 1 . The lobbies’ gross objective functions are

U A y = α _ α u x y d x = α y 2 + α 2 y 1 3 α 3 + α _ y 2 α _ 2 y + 1 3 α _ 3 ,

U P y = π 1 u x y d x = 1 π y 2 + 1 π 2 y 1 3 1 π 3 .

The equilibrium policy level is equal to

y α _ , α , π = arg max y 0 , 1 ( γ y 2 α y 2 + α 2 y 1 3 α 3 + α _ y 2 α _ 2 y + 1 3 α _ 3

1 π y 2 + 1 π 2 y 1 3 1 π 3 ) = 1 + α 2 α _ 2 π 2 + 2 γ 2 2 + α α _ π .

The aggregate utility of anti-policy individuals from the policy outcome y α _ , α , π is equal to

0 γ u x y d x = γ 1 + α 2 α _ 2 π 2 + 2 γ 2 2 + α α _ π 2 + γ 2 1 + α 2 α _ 2 π 2 + 2 γ 2 2 + α α _ π 1 3 γ 3 .

Plugging in the expressions for α and π from Proposition 2 yields

0 γ u x y d x =

{ γ 1 + γ 2 α _ 2 2 + γ 3 + 2 γ 2 + 2 γ 2 3 + 2 γ α _ 2 + γ 2 1 + γ 2 α _ 2 2 + γ 3 + 2 γ 2 + 2 γ 2 3 + 2 γ α _ 1 3 γ 3 f o r γ 0 , 1 2 , γ 1 + 2 + γ + 5 2 γ 2 α _ 2 γ 2 + 2 γ 2 5 2 γ α _ 2 + γ 2 1 + 2 + γ + 5 2 γ 2 α _ 2 γ 2 + 2 γ 2 5 2 γ α _ 1 3 γ 3 f o r γ 1 2 , 1 ,

which is a decreasing function of α _ 0 , α . It follows that the anti-policy group’s aggregate utility from the policy outcome is maximized at α _ = 0 .

I show next that given the rules 0 , α and π , π , the pro-policy group’s aggregate utility from the policy outcome is maximized at π = 1 . Given the rules 0 , α and π , π , the lobbies’ gross objective functions are

U A y = 0 α u x y d x = α y 2 + α 2 y 1 3 α 3 ,

U P y = π π u x y d x = π y 2 + π 2 y 1 3 π 3 + π y 2 π 2 y + 1 3 π 3 .

The equilibrium policy level is equal to

y α , π , π = arg max y 0 , 1 ( γ y 2 α y 2 + α 2 y 1 3 α 3

π y 2 + π 2 y 1 3 π 3 + π y 2 π 2 y + 1 3 π 3 ) = α 2 + 2 γ π 2 + π 2 2 1 + α π + π .

The aggregate utility of pro-policy individuals from the policy outcome y α , π , π equals

γ 1 u x y d x = 1 γ α 2 + 2 γ π 2 + π 2 2 1 + α π + π 2 + 1 γ 2 α 2 + 2 γ π 2 + π 2 2 1 + α π + π 1 3 1 γ 3 .

Plugging in the expressions for α and π from Proposition 2 yields

γ 1 u x ( y ) d x = { ( 1 γ ) ( γ 2 + 2 γ ( 2 + γ 3 + 2 γ ) 2 + π ¯ 2 2 ( 3 + 2 γ + π ¯ 1 ) ) 2 + ( 1 γ 2 ) γ 2 + 2 γ ( 2 + γ 3 + 2 γ ) 2 + π ¯ 2 2 ( 3 + 2 γ + π ¯ 1 ) 1 3 ( 1 γ 3 ) for γ [ 0, 1 2 ] , ( 1 γ ) ( ( 2 + γ + 5 2 γ ) 2 + 2 γ γ 2 + π ¯ 2 2 ( 5 2 γ + π ¯ 1 ) ) 2 + ( 1 γ 2 ) ( 2 + γ + 5 2 γ ) 2 + 2 γ γ 2 + π ¯ 2 2 ( 5 2 γ + π ¯ 1 ) 1 3 ( 1 γ 3 ) for γ ( 1 2 ,1 ] ,

which is an increasing function of π π , 1 . It implies that the pro-policy group’s aggregate utility from the policy outcome is maximized at π = 1 .

Proof of Corollary 1

The first derivatives of α , π , y , C A and C P with respect to γ are equal to

d α d γ = { 1 f o r γ 0 , 1 2 , 1 1 5 2 γ f o r γ 1 2 , 1 , d π d γ = { 1 1 3 + 2 γ f o r γ 0 , 1 2 , 1 f o r γ 1 2 , 1 ,

d y d γ = { 1 1 3 + 2 γ f o r γ 0 , 1 2 , 1 1 5 2 γ f o r γ 1 2 , 1 ,

d C A d γ = { γ 168 + γ 276 + γ 164 + 27 γ 3 + 2 γ 96 + γ 16 + γ 8 + 3 γ 4 3 + 2 γ γ 3 + 2 γ 2 f o r γ 0 , 1 2 , 4 3 γ 2 γ 2 13 5 γ + 5 2 γ 279 + γ 424 γ 226 + 3 γ γ 16 4 5 2 γ 2 γ 2 f o r γ 1 2 , 1 ,

d C P d γ = { 2 + γ 2 2 1 + γ 3 + 2 γ 2 1 + γ 5 + 4 γ + 2 + γ 3 + 2 γ 2 + 3 γ 4 1 + γ 2 3 + 2 γ f o r γ 0 , 1 2 , 1 γ 5 2 γ 283 + γ γ 65 3 γ 249 635 + γ 685 + γ 27 γ 245 4 5 2 γ 5 2 γ 1 + γ 2 f o r γ 1 2 , 1 .

These derivatives are such that d α d γ > 0 , d π d γ > 0 , d y d γ > 0 , d C A d γ > 0 and d C P d γ < 0 .

When the government is biased in favor of low policy levels, i. e., γ 0 , 1 2 , then the lobby sizes are equal to

A = γ , P = 3 + 2 γ 1 γ .

It follows that P > A , i. e., the pro-policy lobby P is more numerous than the anti-policy lobby A . The equilibrium policy level is higher than that preferred by the government:

2 + γ 3 + 2 γ > γ f o r γ 0 , 1 2 .

Lobby P contributes more than lobby A :

C P C A = 4 2 3 + 2 γ + γ 4 + γ 2 3 + 2 γ 2 4 1 + γ γ 2 4 3 + 2 γ 6 + γ 3 + 2 γ 4 2 4 3 + 2 γ 3 + 2 γ γ > 0 f o r γ 0 , 1 2 .

When the government is biased in favor of high policy levels, γ 1 2 , 1 , then the lobby sizes are equal to

A = 2 + γ + 5 2 γ , P = 1 γ .

It follows that A > P , i. e., the anti-policy lobby A is more numerous than the pro-policy lobby P . The equilibrium policy level is lower than that preferred by the government:

2 + γ + 5 2 γ < γ f o r γ 1 2 , 1 .

Lobby A contributes more than lobby P :

C A C P = 9 4 5 2 γ + γ 2 5 2 γ + γ 6 2 4 2 γ 1 γ 2 45 20 5 2 γ + γ 4 5 2 γ + γ 18 4 5 2 γ 1 + γ > 0 f o r γ 1 2 , 1 .

When the government in unbiased, γ = 1 2 , the lobby sizes are equal to A = P = 1 2 , i. e., the lobbies are of the same size. They contribute the same amount to the government: C A = C P = 1 96 . The equilibrium policy level is equal to that preferred by the government: y = 1 2 .

Government Utility in Equilibrium

The government utility in equilibrium is equal to

U γ y = γ y 2 + C A + C P =

{ γ 3 + 2 γ 28 + γ 44 + γ 16 + γ 48 γ 92 + γ 52 + 7 γ 4 1 + γ 3 + 2 γ γ f o r γ 0 , 1 2 , 1 γ 5 2 γ 89 γ 79 + γ γ 19 199 + γ 217 + γ 7 γ 73 4 2 γ 5 2 γ 1 + γ f o r γ 1 2 , 1 .

It is non-negative, increases in γ for γ 0 , 1 2 and decreases in γ for γ 1 2 , 1 .

References

Anesi, V. 2009. “Moral Hazard and Free Riding in Collective Action.” Social Choice and Welfare 32 : 197–219. Search in Google Scholar

Austen-Smith, D. 1997. “Interest Groups: Money, Information, and Influence.” In Perspectives on Public Choice. A Handbook, edited by D. C. Mueller. Cambridge University Press. Search in Google Scholar

Baron, D. 1989. “Service-Induced Campaign Contributions and the Electoral Equilibrium.” Quarterly Journal of Economics 104 : 45–72. Search in Google Scholar

Becker, G. S. 1983. “A Theory of Competition among Pressure Groups for Political Influence.” Quarterly Journal of Economics 98 : 371–400. Search in Google Scholar

Becker, G. S. 1985. “Public Policies, Pressure Groups and Deadweight Costs.” Journal of Public Economics 28 : 330–347. Search in Google Scholar

Bernheim, B. D., and M. D. Whinston. 1986. “Menu Auctions, Resource Allocation, and Economic Influence.” Quarterly Journal of Economics 101 : 1–31. Search in Google Scholar

Besley, T., and S. Coate. 1997. “An Economic Model of Representative Democracy.” Quarterly Journal of Economics 112 : 85–114. Search in Google Scholar

Bombardini, M. 2008. “Firm Heterogeneity and Lobby Participation.” Journal of International Economics 75 : 329–348. Search in Google Scholar

Bombardini, M., and F. Trebbi. 2012. “Competition and Political Organization: Together or Alone in Lobbying for Trade Policy?” Journal of International Economics 87 : 18–26. Search in Google Scholar

Coate, S., and M. Conlin. 2004. “A Group Rule-Utilitarian Approach to Voter Turnout: Theory and Evidence.” American Economic Review 94 : 1476–1504. Search in Google Scholar

Cotton, C. 2012. “Pay-To-Play Politics: Informational Lobbying and Contribution Limits When Money Buys Access.” Journal of Public Economics 96 : 369–386. Search in Google Scholar

Dahm, M., and N. Porteiro. 2008. “Side Effects of Campaign Finance Reform.” Journal of the European Economic Association 6 : 1057–1077. Search in Google Scholar

Damania, R., and P. G. Fredriksson. 2000. “On the Formation of Industry Lobby Groups.” Journal of Economic Behavior & Organization 41 : 315–335. Search in Google Scholar

Damania, R., and P. G. Fredriksson. 2003. “Trade Policy Reform, Endogenous Lobby Group Formation, and Environmental Policy.” Journal of Economic Behavior & Organization 52 : 47–69. Search in Google Scholar

DeKieffer, D. E. 2007. The Citizen’s Guide to Lobbying Congress: Revised and Updated. Chicago, IL: Chicago Review Press. Search in Google Scholar

Dewatripont, M., and J. Tirole. 1999. “Advocates.” Journal of Political Economy 107 : 1–39. Search in Google Scholar

Dixit, A., G. M. Grossman, and E. Helpman. 1997. “Common Agency and Coordination: General Theory and Application to Government Policy Making.” Journal of Political Economy 105 : 752–769. Search in Google Scholar

Drazen, A., N. Limão, and T. Stratmann. 2007. “Political Contribution Caps and Lobby Formation: Theory and Evidence.” Journal of Public Economics 91 : 723–754. Search in Google Scholar

Feddersen, T., and A. Sandroni. 2006. “A Theory of Participation in Elections.” American Economic Review 96 : 1271–1282. Search in Google Scholar

Felli, L., and A. Merlo. 2006. “Endogenous Lobbying.” Journal of the European Economic Association 4 : 180–215. Search in Google Scholar

Felli, L., and A. Merlo. 2007. “If You Cannot Get Your Friends Elected Lobby Your Enemies.” Journal of the European Economic Association 5 : 624–635. Search in Google Scholar

Furusawa, T., and H. Konishi. 2011. “Contributing or Free-Riding? Voluntary Participation in a Public Good Economy.” Theoretical Economics 6 : 219–256. Search in Google Scholar

Glazer, A., and M. Gradstein. 2005. “Elections with Contribution-Maximizing Candidates.” Public Choice 122 : 467–482. Search in Google Scholar

Groll, T., and C. J. Ellis. 2014. “A Simple Model of the Commercial Lobbying Industry.” European Economic Review 70 : 299–316. Search in Google Scholar

Grossman, G. M., and E. Helpman. 1994. “Protection for Sale.” American Economic Review 84 : 833–850. Search in Google Scholar

Grossman, G. M., and E. Helpman. 1996. “Electoral Competition and Special Interest Politics.” Review of Economic Studies 63 : 265–286. Search in Google Scholar

Grossman, G. M., and E. Helpman. 2001. Special Interest Politics. MIT Press. Search in Google Scholar

Harsanyi, J. C. 1980. “Rule Utilitarianism, Rights, Obligations and the Theory of Rational Behavior.” Theory and Decision 12 : 115–133. Search in Google Scholar

Helpman, E., and T. Persson. 2001. “Lobbying and Legislative Bargaining.” Advances in Economic Analysis & Policy 1 :. Article 3. Search in Google Scholar

Kerr, W. R., W. F. Lincoln, and P. Mishra. 2014. “The Dynamics of Firm Lobbying.” American Economic Journal: Economic Policy 6 : 343–379. Search in Google Scholar

Krishna, V., and J. Morgan. 2001. “A Model of Expertise.” Quarterly Journal of Economics 116 : 747–775. Search in Google Scholar

Laussel, D. G. 2006. “Special Interest Politics and Endogenous Lobby Formation.” Topics in Theoretical Economics 6 (Article 1). Search in Google Scholar

Leaver, C., and M. Makris. 2006. “Passive Industry Interests in a Large Polity.” Journal of Public Economic Theory 8 : 571–602. Search in Google Scholar

Magee, C. 2002. “Endogenous Trade Policy and Lobby Formation: An Application to the Free-Rider Problem.” Journal of International Economics 57 : 449–471. Search in Google Scholar

McCarty, N., K. Poole, and H. Rosenthal. 2006. Polarized America. MIT Press. Search in Google Scholar

Mitra, D. 1999. “Endogenous Lobby Formation and Endogenous Protection: A Long-Run Model of Trade Policy Determination.” American Economic Review 89 : 1116–1134. Search in Google Scholar

Olson, M. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Harvard University Press. Search in Google Scholar

Osborne, M. J., and A. Slivinski. 1996. “A Model of Political Competition with Citizen-Candidates.” Quarterly Journal of Economics 111 : 65–96. Search in Google Scholar

Persson, T. 1998. “Economic Policy and Special Interest Politics.” Economic Journal 108 : 310–327. Search in Google Scholar

Persson, T., and G. Tabellini. 2002. Political Economics: Explaining Economic Policy. MIT Press. Search in Google Scholar

Redoano, M. 2010. “Does Centralization Affect the Number and Size of Lobbies?” Journal of Public Economic Theory 12 : 407–435. Search in Google Scholar

Smith, M. A. 2000. American Business and Political Power: Public Opinion, Elections, and Democracy. Chicago: University of Chicago Press. Search in Google Scholar

Snyder, J. 1990. “Campaign Contributions as Investments: The U.S. House of Representatives, 1980–86.” Journal of Political Economy 98 : 1195–1227. Search in Google Scholar