Recent political alarm signals make Lobbying as a research field highly relevant. The main objective is to challenge common approaches to model the mechanisms of how lobbyists gain political influence. I will explain what parameters or factors makes lobbyists more successful in their aim to influence political decision makers and what “channels” they use to do so. This issue is closely related to the question of what makes incumbents prone to lobbyists’ activities.
This paper starts with a short theoretical introduction to rent seeking games, which are the theoretical underpinning of a school of thought which models lobbying as a kind of political investment through well-aimed campaign contributions. I question the conjecture that financial contributions are the main leverage for lobbyists to gain political influence and provide empirical data which support my claim. My analysis focuses on the total amount of money, which is transferred from U.S. interest groups to U.S. politics and could be understood as political investment through campaign financing. To explain my findings in this context, the question is addressed whether political donations do have a measurable effect on political decisions.
In the second part of the paper another popular interpretation of a lobbyist’s main instrument to influence is discussed: The provision of special knowledge. Empirical analysis shows that the data delivers strong evidence to reject this idea. Instead, I develop a new way of interpreting the meaning of “special knowledge” by asking what advantage a politician has to socialise with lobbyists. By focusing this question in more detail a theoretical model is introduced which shows how important the concerns of special interest groups could be for politicians in their aim to become reelected. This model leads to the final result of this paper:
What makes lobbyists useful to politicians and therewith powerful is the provision of private information about preferences of possibly pivotal constituents.
Contents
1 Introduction
2 The Political Investment Approach
2.1 Focusing the Effect of Financial Contributions
2.2 Theory: Rent-Seeking Games
2.3 Empirical Findings: Financial Contribution as Leverage
3 The Consulting Approach
3.1 The Role of Individual Relationships
3.2 Empirical Findings: Expert Knowledge as Leverage
4 Why Incumbents Benefit from Lobbyist’s Influence
4.1 A Theoretical Model by Lohmann
4.2 Empirical Findings: Electoral Cycles and Influence of Interest Groups
5 Conclusion
A Appendix
List of Tables
1 The Effect of Variables on Roll Call Voting Scores by the US Chamber of Commerce. Collected Data for the US House 1978 - 1994
3 Do Lobbyists Follow Politicians They Are Connected to as Those Politi- cians Switch Committee Assignments?
4 The Effects of Variables on Quarterly Veterans’ Payments (GLS Estimates)
2 Do Connections through Campaign Contributions Predict Overlapping Issues between Lobbyists and Congressmen? - 108th Congress
5 The Effects of Variables on Quarterly Veteran’s Payments 1961-78 (GLS Estimates)
6 The Effects of Variables on Unemployment Payments 1961-78 (GLS Es- timates)
1 Introduction
In June 2013 the European organization “ Group of States against Corruption ” (GRECO) published its 13 th Report on lobbying and corruption in Europe, presenting alarming results for Germany’s democratic system.1 According to LobbyControl (2013), a German organization tasked with watch- ing and controlling lobbying activities, Germany repeatedly ignored GRECO’s recommendations to organize legal framing of the lobbying business more transparently and more strictly. According to LobbyControl the German government not just ignored recommendations from international organizations like the United Nations2 or GRECO, but also prohibited also political attempts from German parties to regulate the lobbying business. This is why Germany became a focus of attention of supranational organizations like GRECO, which recently launched its second stage of dunning process against Germany.
Those recent political alarm signals make Lobbying as a research field highly relevant. Because of the intransparency of the German lobbying system there is nearly no data available for economic research. That is why this seminar paper focuses on U.S. lobbying datasets. The main objective is to challenge common approaches to model the mechanisms of how lobbyists gain political influence. I will explain what parameters or factors makes lobbyists more successful in their aim to influence political decision makers and what “channels” they use to do so. This issue is closely related to the question of what makes incumbents prone to lobbyists’ activities.
This paper starts with a short theoretical introduction to rent seeking games, which are the theoretical underpinning of a school of thought which models lobbying as a kind of political investment through well-aimed campaign contributions. I question the conjecture that financial contributions are the main leverage for lobbyists to gain political influence and provide empirical data which support my claim. My analysis focuses on the total amount of money, which is transferred from U.S. interest groups to U.S. politics and could be understood as political investment through campaign financing. To explain my findings in this context, the question is addressed whether political donations do have a measurable effect on political decisions.
In the second part of the paper another popular interpretation of a lobbyist’s main instrument to influence is discussed: The provision of special knowledge. Empirical analysis shows that the data delivers strong evidence to reject this idea. Instead, I develop a new way of interpreting the meaning of “special knowledge” by asking what advantage a politician has to socialise with lobbyists. By focusing this question in more detail a theoretical model is introduced which shows how important the concerns of special interest groups could be for politicians in their aim to become reelected. This model leads to the final result of this paper: What makes lobbyists useful to politicians and therewith powerful is the provision of private information about preferences of possibly pivotal constituents.
To underpin the intuition of this theory, I discuss empirical evidence which shows electoral cycles in budgetary growth of veteran’s benefit programs in the U.S, confirming my suggestion that influential power of lobbyists is greatest in times of elections. These results strengthen my hypothesis that the main leverage for lobbyists to influence political decision makers is provision of private information about a certain group of voter’s preferences, which enable incumbents to maximize their reelection probability.
2 The Political Investment Approach
2.1 Focusing the Effect of Financial Contributions
In most of the literature, lobbying is often understood as a market where different parties are competing for political influence. The market actors are different interest groups, which favor different outcomes of a certain political decision in progress. Therefore they are spending money on lobbying activities. Those are often quantified through campaign contributions of a certain lobbyist on a politician. This behavior causes inefficiencies, because the politician’s decision is not welfare maximizing anymore, but becomes distorted by the influence of a lobbyist representing a special interest group. In consequence deadweight loss is created and damages welfare of all citizens. In a worst case scenario a certain industry succeeds through lobbying in establishing monopoly rents on a certain marked, because politics decides to protect the market by raising tariffs for example. In this case one can use standard microeconomic theory to describe how deadweight loss accrues.
Starting with Harberger (1954) economists tried to estimate the deadweight loss empirically, but it turned out that the welfare losses are relatively small.3 Tullock (1967) criticized this approach of measuring the damaging effect of the lobbying industry by arguing that producers are likely to compete against each other for monopoly rents and waste resources doing so. These resources are also social costs and have to be taken into consideration by analysing welfare effects of the lobbying industry.
2.2 Theory: Rent-Seeking Games
Using Tullock’s idea rent-seeking games became a basic approach to model properties of compet- itive behavior of two or more lobbyists working on the same topic for competing interest groups. The model assumes the political decision maker to be able to allocate a rent, V, to the players (lobbyists). A player can win the competition by simply bidding money (B). All competitors are simultaneously placing their bids and the player with the highest bid wins the price (winner takes it all).
The Baseline Model
The following model is based on Myles (2013). In this version of the game attention is restricted to symmetric equilibria. Suppose a politician offers a price. The player with the highest bid wins the competition and gets the price for sure. It can be shown that there is no pure strategy equilibrium:
Suppose both players bid B *, the expected payoff will than be[illustration not visible in this excerpt]it would. have been best for a player to bid B ′ = 0, so we know that V
illustration not visible in this excerpt
If that is the case either player can deviate to[illustration not visible in this excerpt]), win for sure, and earn[illustration not visible in this excerpt] small enough. So there is no pure strategy equilibrium.
Let’s now consider mixed equilibrium strategies. Let equilibrium bids be (uniformly) distributed as F (B). Therefore each of the n players assigns a probability of f (B) to play each strategy B ∈ [0 ,V ]. Let the probability of beating one of the competing players be F (B). In the case of n − 1 other players there is a probability of F (B) n − 1 to win the game. In equilibrium the expected payoff of each player has to be zero, because otherwise there would be an incentive for one of the others to deviate by small ϵ. So,
illustration not visible in this excerpt
has to hold for any value of B ∈ [0 , V ]. Rearranging gives us the winning probability of:
illustration not visible in this excerpt
This equation describes the distribution of equilibrium bids and has the property that the probability applied to higher levels of B falls relative to that for lower levels as n increases. Given the assumption that expected utility has to be zero in equilibrium and that each of the competitors has the same chance of winning, denoted by1 n, the expected amount bid by each competitor will be V N. Given there are n players this will mean that the total amount of money spent to gain a price of V is also equal to V.
The game can be played with different setups. One can allow for free market entry and exit for lobbyists or decide whether the game should be deterministic, like in the example above, or stochastic, meaning the higher the bid the higher the probability of winning. Independent of the actual setup, the intuitive outcome is the same.4
The outcome of the baseline rent-seeking game shows, that the lobbying industry wastes enormous resources to achieve their goal to realize rents for their interest groups. The amount of resources wasted could be up to the size of the achievable rent itself. In consequence one can say that society is suffering from lobbying acivity not only because of deadweight loss, but also because resources are wasted in competing for rents.
2.3 Empirical Findings: Financial Contribution as Leverage
The Legal Situation of Campaign Contributions in the U.S.
The economic literature dealing with lobbying use published data5 concerning to campaign contributions in the US as a quantitative measure for lobbying activity. Even though financial contributions coming from the industry adressing politicians are strongly regulated and limited in the US,6 theory suggests that there is an incentive for interest groups to exhaust all legal opportunities to succeed in lobbying competition.
To start with the empirical investigation, the essentials of contribution spending regulation in the US have to be explained:
The FECA specifies two different types of campaign organizations: party committees and candidate campaign committees. Furthermore two different sources of financial contributions to politicians are defined: Individuals, who spend their private money, and interest groups. The latter could include all kinds of firms, associations or unions. Independent of the kind of contributor, all payments have to be reported to the Federal Elections Commission (FEC). Contributions underlie strong regulation, the most important ones for our analysis are summarized in the following:7
- On purposes of transparency in accounting, commercial organizations have to establish po- litical action committees (PACs) if they wish to support politicians financially. A company is not allowed to donate money directly to politics. It is even not allowed to give money directly from the organization’s treasury to their own PAC. Instead, the PAC has to raise money from their individual members on a voluntary basis. In practice that does mean com- pany’s PAC contributions usually come from the organization’s board of management. So, in the end also for interest group spending, individuals are the source of contributions.
- PACs are only allowed to give up to 10.000 Dollar in a two-year election cycle to a candidate committee and up to 17.500$/y to a party committee. Also an individual is allowed to donate a maximum amount of 5.000$/y to a PAC and 20.000$/y to a party committee.
- Moreover all kinds of financial contributions have to be reported to the FEC to ensure maximum transparency.
Furthermore Ansolabehere (2003) identifies two important loopholes in spending regulations:
- Independent Expenditures: Individuals are allowed to donate an unlimited amount to politicians if the contribution is not coordinated with a candidate’s campaign.
- Soft Money Regulation: “Soft Money” is defined as money raised by national parties for non- federal accounts. Individuals and interest groups are allowed to give an unlimited amount of “Soft Money”, which is used finance non-federal election activities (i.e. governor’s elections)
Do Contributions Influence Political Decisions? A Regression Model by Ansolabehere (2003)
After understanding the basics of contribution regulation in the US one would conclude that there is a strong need for regulating financial contributions, because of the described above theoretical characteristics of political investments. But the data reported to the FEC shows that only 4% of the corporate PACs contributed nearly the maximum amount of 10.000$ to a political candidate. The average amount donated by corporate PACs is, according to Ansolabehere (2003), just about 1.400$. In addition to that, only 60% of the Fortune 500 companies even have a PAC, whereas one could assume that all companies listed in the Fortune 500 are surely directly affected by political decisions, because of their size. The data shows that the most important source for political campaign funding are not PACs, but individuals spending independently from interest groups.
This is an empirical observation worth discussing, because it contradicts theoretical intuition fundamentally.8 There might be several explanations for this. Ansolabehere (2003) at first discusses whether the reason might be that campaign contributions do not have a significant influence on a politician’s decision process. If this would be the case there might be no incentive for the industry to donate at all. To measure whether contributions have a significant effect on a politician’s decisions in roll call votes is hard for two reasons (Ansolabehere (2003)):
Firstly, there is good reason to assume a simultaneity bias between contributions and votes. To give an intuition: On the one hand, one could think of an incumbent voting in favor of an interest group, because they spent a large amount of money. But on the other hand one could also easily imagine an interest group contributing a large amount of money to any politician, because he voted in their favor. Therefore all specifications of the model are tested with the OLS method and with a model using instrumental variables9 which are introduced to overcome the simultaneity problem. To measure important factors which make a politician more attractive for financial contributors, the degree of a senator’s electoral competition and the relative power of the politician are considered. Intuition suggests that a senator trying to win a close election is more open for political influence by interest groups, because he is more dependent from political support of potential votership. This makes him more likely to “sell” his services and adjust his roll call voting behavior after receiving campaign contributions. It is also quite plausible to assume that a politician who holds an office of relative importance to the industry is receiving more financial contribution.
Secondly, there is a strong need for control variables (i.e. the interest of a incumbent in a certain district). Those control variables are important, because of the empirically and theoretically well- grounded tendency of interest groups to donate to pivotal legislators or to politicians who share their political bent.10 If that is the case and control variables are not good enough, estimates are biased, because omitted variables tend to be correlated with a group’s contributions. So, by examining this problem empirically one have to care for several fixed effects: One fixed effect could arise if it is assumed that a senator is tending to vote in favor of his home district if a certain political issue effects the industrial structure of his district for example. (District fixed effect). Another problem of this type could arise if one wants to capture the fact that a senator’s voting decisions are not independent from each other, but express the legislator’s own political preferences (legislator’s fixed effects). Ansolabehere (2003) therefore runs several specifications caring about party affiliation combined with the described fixed effects.
The regression model is set up in two steps: In the first step contributions from corporate and labor PACs are defined as dependent variables and the mentioned instrumental variables serve as regressors. The coefficients of this model can be used to determine the predicted contributions from corporate and labor PACs to a certain individual. In the second step the predicted donations of the first regression serve as independent variables explaining a voting score produced by the US Chamber of Commerce, which is collected for the US House from 1978 - 1994.11
Table 1: The Effect of Variables on Roll Call Voting Scores by the US Chamber of Commerce. Collected Data for the US House 1978 - 1994
illustration not visible in this excerpt
All specifications include year fixed-effects. (2) includes district fixed-effects. (3) includes member fixed-effects. Furthermore it is defined: *= significant at the .05 level; **= significant at the .01 level
It can now be analyzed whether contributions do have an significant effect on a incumbent’s voting behavior. The results of the second regression are listed in the table. The first specification is mostly used in literature and covers party affiliation and a measure of district preferences based on voting patterns. In this setting contributions do have an significant effect on the voting behavior of the incumbent. Although the influence is significant, it is relatively small compared to other factors. Hence, being a member of the Republican Party does have a tenfold higher effect on the voting score than corporate contributions. Interestingly, if we look at the same specification in the setting where instrumental variables are used, corporate contributions stay significant, but change its sign meaning that more contributions of corporate PACs lead to a worse voting behavior in the eyes of the interest group. This result is counterintuitive. In addition to this, if we consider a more realistic case and control for the politician’s political or district preferences the influence becomes non significant. In contrast to this, the effect of being a member of the Republican Party stays significant and increases.
We can conclude that empirical evidence of the influence of corporate PAC contributions is, if significant at all, not very large. Especially in comparison with the influence of party affiliation. If this fact influences the voting behavior of the politician stronger than monetary incentives, than it is more likely that a politician is influenced by party line than by donations of a certain interest group. This makes sense, because of the fact that a politician is mostly interested in becoming reelected. Therefore his political standpoint, which he expresses inter alia with his roll call votes, has to be optimally compatible with the opinion of his party, which nominates him as a candidate.12 The same argument could explain the strong influence of the parameter “District Partisanship”. Once a politician is nominated by his party as a candidate, he needs to become reelected by the people in his district. So, he tries to vote in favor of his home region. The positive correlated influence on the voting score of the CCUS could be explained by the following intuition: If a politician tries to vote in favor of a district, he would be attempted to support local businesses to ensure high employment rates. This voting behavior surely lies in the interest of the CCUS. So, if the politician’s voting patterns indicate a strong district partisanship it is more likely that the legislator achieves a higher voting score in the eyes of the CCUS.
These empirical findings could explain why there is so little money spent by interest groups to participate in lobbying competition: The expected return of a political investment through contributions is negligibly small. This observation raises doubts whether the main tool to influence political decision makers through lobbying activities is really financial contribution, like the rentseeking game suggests. Can we interpret campaign contributions really as political investment? Is there really lobbying competition based on financial support like it is described in the theoretical model? Because of the empirical results we need to search for another approach to model the mechanisms of the lobbying industry seeking political influence.
3 The Consulting Approach
Another popular approach to model the way of how lobbyists influence legislators is to assume that lobbyists mainly come from certain industries and posses expert knowledge in special topics, which they deliver to politicians and enable them to make informed decisions.13 Despite the discussion whether this mechanism could have welfare enhancing effects the focus of this paper addresses the question whether this informational behalf could be a leverage to influence uninformed politicians in the sense of the interest group.
In the following I seek for empirical hints to find some information about what characterizes the relationship between a lobbyist and a politician: Is the lobbyist’s main asset special knowledge which he shares with incumbents? Or lies a lobbyist’s main interest not in subject-specific consulting, but in establishing a (maybe personal?) relationship to the politician to gain the legislator’s trust and thereby the ability to influence him?
3.1 The Role of Individual Relationships
In the paper by M. Bertrand (2011), campaign contributions serve as a proxy for the “connections” a lobbyist has. Because personal ties and relationships to politicians are not observable, there is the need for a quantitative measure to capture this variable. So, M. Bertrand (2011) assumes that if a lobbyist is contributing money to a politician, he has some kind of personal or at least professional relationship to this legislator. This assumption has to be verified, because a lobbyist who gets a certain budget from his client could also distribute contributions freely and support politicians who share the client’s political opinion independently of an existing relationship between the lobbyist and the politician.
To test this assumption, M. Bertrand (2011) tries to examine whether a significant relationship between contributions and politicians, who work on the same topic as the lobbyist, do exist. If contributions have a significant influence on the probability that a politician and a lobbyist share topics they are working on, one could assume that a lobbyist tends to have a personal or at least professional relationship to politicians he is contributing to. If that is the case we could justify our assumption to proxy relationships by using contributions as quantitative measure.
M. Bertrand (2011) creates a dataset which contains all possible lobbyist-legislator pairs by crossing the pool of active lobbyists with the pool of lawmakers. By doing so we get about 4.6m observations which serve as a sound basis for the analysis. M. Bertrand (2011) runs two different specifications of his regression models measuring the “issue overlap” within lobbyists-legislators pairs. An issue is “overlapping” if it is associated with the politician and coincidentally appears in at least one of the lobbying records associated with the corresponding lobbyist. In the first specification the dependent variable is defined by the number of issues overlapping. In the second specification a dummy variable that equals one if the observation is perfectly overlapping serves as dependent variable. I will focus my attention on the second specification here, because interpreta- tion seems easier and results are similar. The author argues that the overlap measure could vary across legislators, because some committees politicians are assigned to are more narrowly-focused than others. Therefore M. Bertrand (2011) considers legislator-fixed effects and lobbyist-fixed effects. Table 2 in the appendix shows the regression results.
The data shows that there seems to be a significant effect of campaign contributions on issue overlap. If a lobbyist at least contributed once in a two year Congress election cycle to a politician, the probability that their issues overlap perfectly increases by 2.9 percentage points compared to a random draw (11.16%). Similar results hold for other Congress election cycles. This result seems to justify the mentioned proxy measure.
If one now interprets campaign contributions as a proxy variable for personal relationship of any kind, one can conclude that there is a relationship between whom lobbyists know and what they work on. Of course the proxy does not convey any information about the quality of the assumed relationship, so one cannot say whether there are strong or loose ties between them. Nevertheless the described results give reason to assume, that what lobbyists are working on is somehow connected to who they know.
[...]
1 for further reading, see GRECO (2012)
2 see Brühl (2013)
3 For US manufacturing 1929, estimates were around 1% of GNP; Harberger (1954)
4 Myles (2013) summarize the effect of the game becoming stochastic in the Partial Dissipation Theorem: If there are two or more competitors in a probabilistic rent-seeking game, the total expected value of resources expended by the competitors in seeking a prize of V is a fraction n − 1 n ofthepricevalue V,andisincreasingwith the number of competitors.
5 United States: http://www.opensecrets.org and http://www.lobbyists.info
6 regulated by the Federal Election Campaign Act of 1974 (FECA)
7 Ansolabehere (2003)
8 In literature this phenomenon is often called “Tullock’s puzzle”.
9 Ansolabehere (2003) follows Chappell (1982) in choosing the IVs. I do not describe those in further detail, because attention of my argument lies on the regression shown in table 1, which does not use IVs as independent variable.
10 Ansolabehere (2003) gives here several references such as Rosenthal (1987) for example.
11 To give a short explanation of what this score is about: Several well-organized interest groups prepare annual surveys for their members in which voting behavior of roll call votes of all senators concerning to certain topics of interest is listed. From this data the interest group calculates a score reaching from 0 to 100 for each incumbent. This score serves as an indicator how well the voting behavior of the politician fitted to the political standpoint of the interest group.
12 Theoretically, one could think of an application of the median voters theorem here: The best a politician, who wants to be renominated as a candidate from his own party, can do is to represent the median political opinion within the party. If he deviates from this, because of interest group’s influence this could be harmful for his renomination probability.
13 This view is also represented by the American League of Lobbyists, which states on its website: ” Lob- bying is a legitimate and necessary part of our democratic political process. Government decisions affect both people and organizations, and information must be provided in order to produce informed decisions ” (http://www.alldc.org/publicresources/lobbying.cfm)