Andrea Prat1 Tilburg University and London School of Economics
28 October 1999
Campaign finance is an important, and much debated, phenomenon in democracies throughout the world. This article discusses a possible economic model of campaign finance, which could be used for policy evaluation. At the core of the model lies an asymmetry of information between lobbies and voters. Lobbies know more than voters about the quality of candidates. Campaign contributions constitute an indirect way to reveal lobbies’ information to voters. However, this informational benefit comes at the cost of candidates deviating from the median voter’s preferred policy in order to attract higher contributions.
Electoral campaigns can be very expensive. In several countries including the US, candidates buy TV time to broadcast advertising. What many voters are uncomfortable with is that often the money to pay for campaigns comes from lobbies whose goal is to influence policy-making. As a result, several countries have imposed or are considering various forms of campaign finance regulations. Economics may be useful in modeling this phenomenon and trying to structure the debate on regulation.
In this article, I review a model of campaign finance that has recently been proposed by economists working on political institutions (Potters, Sloof, and Van Winden (1997), Prat (1998, 1999)). The core assumption is that there is an informational asymmetry between lobbies and voters. Lobbies have more time and resources to find out about the quality of candidates. By choosing which candidates they donate the bulk of their money to, lobbies may signal to voters what they have learnt about candidates. Hence, campaign finance has an informational benefit. However, there is also a negative side. Lobbies do not give money in exchange for nothing. There is an implicit political exchange between lobbies and candidates. A candidate who wants to attract contributions from a given lobby should be prepared to adopt a policy stance that goes toward the lobby’s preferred policy and away from the median voter’s preferred policy (as is well known, in absence of contributions, candidates who maximize their election chance tend toward the median voter’s preferred policy). I refer to the deviation from the median voter’s preferred policy as the policy cost of campaign contributions. The tradeoff between informational benefit and policy cost will be the main topic of this article.
The plan of the article is as follows. In the next section, I provide a brief overview of the practice of campaign finance throughout the world. I then propose an informational model that fits the main observed features of campaign finance. I discuss what happens in the equilibrium of this model and suggest two empirical tests of the model. Finally, I use the model to evaluate a number of possible regulatory regimes, including a complete ban on contributions, a ban on split contributions, blackout times, public finance, and disclosure rules.
Overview of the Phenomenon
Campaign finance is a well-known phenomenon in the US, less so in Europe and in the rest of the world. Consequently there is abundant information about the US but not much about other countries. I start by drawing a picture of the American case, and then I review what is known about the rest of the world.2
First, let us get an idea of the order of magnitude of US campaign finance. In the 1998 election cycle, the combined receipt of candidates to the Senate and the House of Representatives was $781million. In the Senate, the average winning candidate spent $5.2 million. The equivalent figure for the House is $650,000. The top spender is Alphonse D'Amato (Republican – New York), who got elected to the Senate with a $27 million campaign. No Senate winner spent less than a million dollars.
Where does campaign finance come from? A small proportion comes from the candidates' own pockets. Some wealthy candidates pay for their own campaigns. However, the vast majority comes from contributions from interest groups. Interest groups can be divided into business (such as individual firms or industry associations), labor (trade unions), and ideology (pursuing specific social issues). Three quarters of the campaign money from interest groups comes from business groups. Among the top-twenty donors, we find corporations such as Philip Morris, Citigroup, Bank of America, and Bell Atlantic. Another interesting finding is that, while labor and ideology groups tend to give either to Republicans or to Democrats, business groups typically give to both parties. For instance, the four firms cited above give more to Republicans but also give a considerable chunk to Democrats.
Once they have the money, what do candidates spend it on? A political campaign involves many types of expenditure – canvassing, leaflets, radio and newspaper advertising – but the largest single item is television advertising. Almost all candidates run ads to promote themselves and, sometimes, discredit their opponents. Unlike regular commercial advertising, campaign advertising is protected by the First Amendment as political speech, and therefore is not subject to a code of conduct (A 1976 ruling of the US Supreme Court). Indeed, some political ads contain misleading statements that would not be acceptable in regular ads.
What are the limits and the restrictions on campaign money? The Supreme Court also ruled that the expenditure side of campaign finance cannot be limited in any way. The receipt side, however, can be, and is, regulated. There are upper limits for individual donors and for interest groups. Nevertheless, it is general opinion that those limits are easily avoided through several loopholes. As this article is written, US legislators are debating a tightening of campaign finance law.
Outside the US, there is very little systematic data collection. In Europe, the situation is quite heterogeneous. Typically, campaign finance plays little or no role in small Northern Democracies such as the Netherlands, Denmark, and Sweden. On the other hand, it is a well recognized issue in Britain and in Italy. In the 1997 UK national elections, the two main parties spent a combined sum of 45 million pounds, doubling the amount spent in 1992.3
Although it is hard to generalize across European countries, three patterns set the old continent apart from the US. First, parties play a much more important role in deciding how money is used. Second, the state often provides generous public funding of electoral campaigns (this is not true in the US, with the exception of Presidential campaigns). Third, campaign spending is not protected as free speech, and the expenditure side of campaign advertising can be regulated. Indeed, several countries have in place draconian limits, such as the prohibition of all TV advertising.
An Informational Model of Campaign Finance
Campaign finance is a complex phenomenon. Three types of political actors are involved: voters, lobbies, and candidates. Voters monitor and evaluate candidates. Candidates choose their policy positions and their campaign expenditures in order to influence voters. Lobbies select their campaign giving strategy to maximize their influence on political outcomes. Candidates know that their policy choices affect the decisions of lobbies to contribute to their campaign.
I now present a model of campaign finance introduced by Potters, Sloof, and Van Winden (1997) and developed by Prat (1998, 1999). It can be labelled an informational model because it puts the emphasis on the idea that lobbies are better informed than voters and that campaign advertising is used to convey some of the lobbies’ knowledge to voters. Despite relying on several simplifying assumptions, the model is still quite complex, due to the number of players involved and the intricate connections among them. The building blocks of the model are:
Candidates. I consider a first-past-the-post electoral race with two candidates (the model aims to describe the US case). Regarding the candidates’ objectives, I make the simplest assumption: they both maximize the probability of being elected. Hence, they view policy in an instrumental way. Their stance is purely motivated by the desire to be elected. As is well known, in such circumstances – without the effect of lobbying – both candidates choose the median voter’s preferred policy.
A candidate is characterized on two dimensions: his policy stance and his valence. The former is simply his position on current political issues, such as the strictness of gun control or the level of tax on tobacco products, and is chosen by the candidate before the election. The latter is meant to capture all the personal characteristics of the candidate that may be of interest to voters and that are not modifiable by the candidate. Valence may include capability, leadership, or a “clean” past.4
Voters. There are a large number of voters who care about both policy and valence. On policy, voters have heterogeneous preferences. Some are in favor of high taxes on cigarettes, some are against. Some want to ban guns, some believe in the right to bear arms. Instead, on valence, voters have more homogeneous tastes: they all prefer a capable leader with an exemplary record to an inept politician with a shameful past.
I make an extreme assumption on voters. I postulate that they are rational, that is, they choose in their best interest given the information they have. There are three reasons to follow this course. First, it provides a useful benchmark because there are many ways to be irrational but only one to be rational. Second, single voters may make mistakes but in aggregate these mistakes should cancel out by the law of large numbers. Third, democracy is based on the principle that no authority can second-guess the popular opinion as expressed in free elections. Hence, if we are to change electoral rules, we should do so under the assumption that voters are rational and not with the goal of correcting a distortion that stems from an alleged irrationality of voters.
Lobbies. In the US political arena, there are a large number of interest groups – 5000 with offices in Washington (Lehman and Shlozman-Tierney 1986). Typically, a lobby does not care about the big picture but only about the position of candidates on a narrow set of policy issues. For instance, the National Rifle Association cares about right to bear arms but not tobacco taxes, while Philip Morris worries about the policy on tobacco but not gun control.
In the model, I assume that there are a large number of policy dimensions and that on each dimension there are two lobbies with opposite interests. However, the two lobbies that operate on a given dimension need not have the same resources. In the case of tobacco, for instance, it is quite clear that the pro-tobacco side is much better organized and funded.
The utility function of a lobby includes money and policy. Utility depends positively on how close the policy implemented by the candidate is to the desired policy of the lobby and depends negatively on the amount contributed to candidates.
Information. This is the crucial piece of the model. Lobbies are in a much better position than voters to collect information about candidates, not only public records but also hearsay and first-hand impressions. The typical lobby leader has a more precise opinion than the typical voter on the valence of the candidate. In the beginning of the electoral campaign, each lobby receives an insider signal on the valence of the candidate. For simplicity, the insider signal can only take two values: good or bad. Later in the campaign, voters and lobbies receive a public signal about the valence of the candidate. To fix ideas, one may think of the insider signal as the result of a face-to-face conversation between the lobbyist and the candidate and the public signal as the result of a televised debate between candidates.5
Campaign Contributions and Policy Deals. After observing the insider signal, the lobbyist offers campaign money to the candidate, conditional on the candidate adopting a particular policy stance. I assume that this implicit contract is established and well known by all parties involved, so that a lobby or a candidate that reneged on their promises would be excluded by all future deals. Each candidate may make deals with various lobbies. Voters are fully aware of this process. In this simple version of the model, I assume that these deals are done under everybody’s eyes.6
Campaign Spending. In line with the US reality, candidates cannot pocket campaign money. The only use they can make of the money they receive from lobbies is to finance their campaigns.7 Candidates can buy TV time to broadcast their political ads. At this point, it is necessary to spell out the role of advertising. In economic theory, there are three theories of advertising.
First, one can assume that advertising changes the preferences of viewers (Dixit and Norman 1978). This theory is just a formalization of the idea that voters are systematically irrational and, for the arguments provided above, I wish to model voters as rational agents. Moreover, such a theory is not suitable for welfare comparisons and, hence, policy implications.
Second, advertising could provide viewers with verifiable information.8 For instance, some commercial advertising informs viewers about the existence of a product, its technical specifications, and its price. While political advertising does provide some verifiable information, this appears to be a small part. This is because, in the US, political advertising is protected as free speech and, hence, is not subject to the codes of conduct that apply to commercial advertising. If I claim that my yogurt is fat free while in reality it has a high fat content, I can face legal consequences. If I promise that I put personal integrity before everything while in reality I have a history of corruption and embezzlement, I cannot be taken to court by deceived voters.
The third way to model advertising is to assume that advertising provides viewers with information in an indirect way (Milgrom-Roberts 1986). Suppose a new business is opening. Potential customers are not sure about the quality of the goods sold by this business. They may suspect that this is a fly-by-night operator, who will rip off consumers and disappear. How can the new business signal that it has a good product? Suppose it launches into an expensive advertising campaign that provides customers with no direct information about the product. Still, customers see that the new business is spending more money on the campaign than it would get if it were a fly-by-night operator and they must deduce that the business plans to be there for the long term and hence must have a quality product. The information is not in the message of the ads but rather in the money that is burnt to buy the ads.
This third theory does not apply directly to political advertising because candidates are not businesses and, usually, rely on external funds to finance their campaigns. However, what I am after here – which will become clear shortly – is a role for campaign spending as indirect certification of the valence of the candidate.
Timing. As this model has many features, it is worth giving a brief summary. It is a game of incomplete information (because there is asymmetric information on valence) with two candidates, a large number of voters, and several lobbies. The timing is as follows:
Lobbies receive a good/bad insider signal about the valence of the two candidates.
Each lobby offers to each candidate a monetary contribution conditional on the candidate adopting a certain policy.
Each candidate chooses a policy stance and, in so doing, determines the amount of money he receives from lobbies.
Each candidate spends the money he got from lobbies on TV ads.
Voters – who are fully rational – receive a public signal about the valence of the candidates. They observe the policy choices and they view TV ads.
Voters vote for one of the two candidates (no abstention). The elected candidate implements the announced policy.
What happens in equilibrium in the model of campaign finance just presented? An equilibrium of the game (formally, a perfect Bayesian equilibrium) must satisfy the following conditions: (a) in stage 2, each lobby selects the optimal contribution strategy, given the insider signal it has received and the way it expects voters and candidates to behave; (b) in stage 3, each candidate selects the policy stance that maximizes her election chance given the contribution offers of lobbies and the way he expects voters to behave; (c) in stage 5, voters form rational beliefs about the valence of candidates; and (d) in stage 6, each voter – on the basis of policy preferences and rational beliefs—votes for the candidate that maximizes her expected utility.
It can be shown that there exists an equilibrium with full revelation of the lobbies’ insider signal to voters through TV advertising (see Prat 1998 for details). This article presents an informal version of the argument.
The starting point is that each lobby views the contribution to a candidate as an investment with an uncertain return. By offering money, the lobby is able to move the candidate away from the median voter’s preferred policy toward the policy preferred by the lobby. However, this is useful only if the candidate is elected. Hence, the amount that a lobby is willing to contribute to a given candidate depends on the probability that that candidate is elected. In turn, this probability depends on the insider signal. If the insider signal is good, the candidate has a higher expected valence and is more likely to give a good public signal as well. Hence, for the same policy, a lobby is willing to give more money to a candidate with a good insider signal.
Thus, candidates with a good insider signal can collect more money than candidates with a bad insider signal. If voters observe a large number of ads from one candidate, they must conclude that that candidate collected large contributions from lobbies and hence that lobbies received a good insider signal about him. In this way, the insider signal of lobbies is revealed to voters through the number of ads.
If this discussion seems abstract, one can think about the following case. Suppose there are two candidates. By talking to insiders, lobbies learn that candidate 1 has never been suspected of any illegal action (good insider signal) while candidate two is the object of a rumor about a possible corruption case that occurred ten years ago (bad insider signal). Lobbyists know that the rumor is founded but there is no hard evidence for it. However, there is a chance that some journalist will come up with incriminating evidence before the election (a bad public signal). If evidence turns up, candidate 1 is less likely to win the election. Clearly, Candidate 1 is a better investment from the point of view of lobbies and, hence, he receives and spends more money. In turn, voters must ask themselves why candidate 2 received less money and they will suspect that there is something wrong with him, although they cannot exactly point out what. This suspicion is indeed correct and voters are rational in using the extent of campaign spending (or more simply the number of TV ads) to judge the valence of the candidate. Note that voters do not need to make this reasoning in a conscious way. All they need to do is to respond positively to TV ads in equilibrium. That is an optimal response on their part and it does not matter why they do it.9
The other side of the equilibrium is of course the deal between lobbies and candidates. Lobbies offer money in exchange of political favors. Formally, a political favor is a deviation from the median voter’s preferred policy towards the lobby’s preferred policy. Keeping all the rest constant, a political favor is costly for the candidate because the majority of voters will like him less. Keeping all the rest constant, money is useful for the candidate because it buys campaign ads and reveals him as a candidate with a good insider signal. Hence, the candidate faces a tradeoff. Lobbies realize this tradeoff and they do not ask from the candidate more political favors than he is willing to give.
Is this model testable? The main assumption of the model, namely voters’ aggregate (implicit) rationality, cannot be tested directly. However, there are two indirect implications that can in principle be checked against existing data.
First, the model predicts that the wider the informational asymmetry between lobbies and voters, the higher the importance of campaign contributions. In the limit, if voters know exactly what lobbies know, there is no reason why candidates should seek contributions. This prediction can be tested with cross-constituency data. In constituencies in which voters are more involved and informed, the amount of campaign contributions should be lower. A cross-country test may work as well. Indeed, a comparison of European countries suggests that campaign finance is almost absent from countries that are smaller and have higher levels of education (such as Scandinavian countries and the Netherlands as opposed to Italy and Britain).
Second, a slight modification of the model yields another potential test. Suppose that candidates can access both lobbies’ money and personal funds. Also suppose that voters (or at least some of them) are aware of which candidates use personal money to finance their campaign. As a candidate will always prefer to finance his campaign with lobbies’ money, only candidates who cannot get enough from lobbies will resort to personal wealth. Hence, campaign spending from personal funds should be less effective in attracting votes than spending from lobbies’ contributions. As the effectiveness of spending is observable, this prediction is testable. Indeed, Palda and Palda (1998) finds that in French national elections a franc spent on a self-financed campaign is less effective than a franc spent on an externally financed campaign.10
One can think of many possible forms of regulation. They range from an outright ban, to disclosure requirements, to blackout times. Indeed, different countries have implemented or are debating different regulatory regimes. With our model in mind, we can attempt an evaluation of the possible alternatives.
Complete Ban. A simple form of regulation is total prohibition.11 One may conjecture that, in a model with full rationality, banning contributions cannot be optimal because rational voters can already, if they find it in their interest, ignore advertising and hence render contributions useless. This conjecture is incorrect. A ban on contributions may be optimal. Let us adopt the point of view of the median voter. The presence of contributions brings an informational benefit (she can make better informed choices because she learns the lobbies insider signal) and a policy cost (by offering contributions, lobbies are able to sway candidates from the median voter’s ideal policy). There are situations in which the policy cost is higher than the informational benefit. This occurs because, in equilibrium, the lobbies have a powerful threat against a candidate with a good insider signal. If the candidate rejects the lobbies’ offers, he does not spend money and he is perceived by voters as a candidate with a bad insider signal. Lobbies have the ability to extract, in the form of deviations from the median voter’s preferred policy, the whole difference between a candidate with a good insider signal and a candidate with a bad insider signal. Prat (1998, 1999) identifies several factors under which a ban on advertising is more likely to make the median voter better off: (a) When lobbies are more extreme (they ask for policies that are far away from the median voter; (b) When some lobbies have more resources than other (if resources are evenly distributed, lobbying efforts from opposite directions cancel out); and (c) When the insider signal is more precise than the public signal (this is nontrivial because a higher insider signal also brings a higher informational benefit.
Note that in the cases in which a ban on campaign money is optimal, voters would be willing to commit “not to listen to political advertising.” However, this commitment is not credible because, once the ads are paid for, voters find it optimal to get information from them – not to speak about the fact that they cannot avoid seeing them. This type of result is by no means new in economics. In his celebrated discussion of the Rat Race, Akerlof (1974) shows that an employer could benefit from committing not to use certain signals about her employers’ performance when she takes promotion decisions. The logic of the present argument is similar.
Split Contributions. As we saw in the Overview, US business lobbies tend to donate to both parties. In fact, they sometimes donate to more than one candidate in the same race. This phenomenon – known as split contributions – is certainly bad for the median voter. Both candidates receiving money has the same informational value of neither candidate receiving money. However, when both candidates receive money they must both promise favors to the lobby. The informational benefit is zero and the policy cost is positive. In this model, a ban on split contributions is always beneficial to the median voter. Each lobby should be allowed to give to at most one candidate in any given race.
Transparency. An obvious, but important, policy that is certainly beneficial to voters is a disclosure requirement on the identities of contributors and the amounts involved. This point is made formally in a model similar to the one presented here by Sloof (1999). To see why transparency is beneficial, we need to assume that voters are not completely sure about the candidates’ policy stance. In that case, knowing where the money comes from matters. If it comes from an extremist lobby, voters should infer that the candidate has accepted to support a rather extremist policy. A transparency requirement helps achieve the maximal informational benefit with the minimum policy cost, because it is in the interest of candidates not to get involved with extremist lobbies. The US has strict disclosure laws and makes updated information publicly available, including names of donors and amounts donated. Instead, most other countries have more lax requirements or none at all.
Blackout Periods. A proposal that has been put forward in some countries (Italy and India) is to prohibit advertising late in the campaign. For instance, in Italy there were talks to ban TV ads in the month before the election. This policy is usually motivated with the argument that voters are mystified by last minute messages. In this model, it can have a different motivation. Later in the campaign, lobbies base their contribution strategies not only on the insider signal but also on the public signal that was revealed during the campaign. This makes the communication of the insider signal more costly (recall that the informational value of a contribution lies in the fact that it is a bet on the candidate’s public signal). Hence, a policy that forces lobbies to make their choice early in the game increases the informational benefit.
Contributions from Foreign Sources. The US and the UK prohibit political donations from foreign persons or corporations. Does this prohibition make sense? Let us abstract from enforcement issues (as the Huang case in the US in 1996 and the Ashcroft case in the UK in 1999 show, the definition of foreign source is a difficult one). As we saw before, the policy cost of allowing contributions increases when lobbies have extreme policy preferences. One may argue that foreign lobbies are on average farther away from the median voter than national lobbies. National lobbies are composed of national citizens who are “in the same boat” with the median voter. If this argument is correct, then a prohibition on foreign contributions may reduce the policy cost of campaign finance.
Public Campaign Finance. In the present model, public campaign finance is at best useless. If contributions signal the insider information of lobbies, the presence of state money provides no useful information. In the worst scenario, public campaign finance garbles the signal coming from private contributions. This begs the question: why do most EU countries offer generous public campaign finance?
Quality of Politicians. So far we have assumed that politicians are not intrinsically reluctant to accept lobbies’ money. They do so if and only if it increases their chance of being elected. However, it may be realistic to assume that there are two classes of politicians: the honest ones and the dishonest ones. The honest ones find some actions repugnant, like embezzling public money, favoring their friends, or receiving money from lobbies. In the equilibrium of our model, voters cannot tell if a candidate has not received money because he is honest and has rejected offers or because he is inept and has received no offer. Capable honest politicians would find it harder to get elected than capable dishonest politicians and may decide not to get into politics altogether. In this case, banning contributions may be a good idea because it encourages honest people to take part in politics and thus it increases the level of honesty of the political class.
Lobby Concentration Power. We saw before that the policy cost of campaign contributions is higher the more concentrated lobbies are. If every citizen were equally represented in the lobbying arena, lobbies’ attempt to sway politicians would offset each other.12 The full informational benefit would be created but there would be no policy cost. Hence, it is desirable to have a lobbying system as representative as possible of the electorate. In reality, some sectors (like agriculture in the US, the EU, and Japan) are over-represented in the lobbying arena. However, it is not clear what policies can correct this over-representation.
I have presented a simple informational model of campaign finance. Using this model, it is possible to provide some – tentative – evaluations of possible regulatory regimes. As campaign finance is a complex phenomenon, the model needs to be enriched in order to have a more realistic picture and more robust conclusions. Yet, its main contribution is to show that economics can provide a way to talk about campaign finance in a structured and rigorous way.
Admittedly, I have developed this model with the US in mind. As we saw in the Overview, American politics is peculiar in many respects. In most democracies outside the US, parties play a substantial control role in the allocation of campaign money and the state provides considerable public finance. Existing models on campaign finance focus on the US because little systematic data is available in other countries. It would be extremely useful to collect comparable information about a representative sample of countries. This will provide a testing ground for theories about the effects of alternative policy regimes.
George Akerlof, "The Economics of Caste and the Rat Race and Other Woeful Tales," Quarterly Journal of Economics, 90: 599-617, 1976
David Austen-Smith, "Interest Groups, Campaign Contributions, and Probabilistic Voting," Public Choice, 54: 123-139, 1987
Avinash Dixit and Victor Norman, "Advertising and Welfare," Bell Journal of Economics, 9(1): 1-17, 1978
Kay Lehman Schlozman and John T. Tierney, Organized Interests and American Democracy, Harper and Row, 1986
Gene M. Grossman and Elhanan Helpman, "Competing for Endorsements," Center for Economic Policy Research, Discussion Paper, 1996
Gene M. Grossman and Elhanan Helpman, "Electoral Competition and Special Interest Politics, " Review of Economic Studies, 63: 265-286, 1996
Paul Milgrom and John Roberts, "Price and Advertising Signals of Product Quality," Journal of Political Economy, 94: 796-821, 1986
Filip Palda and Kristian Palda, "The Impact of Campaign Expenditures on Political Competition in the French Legislative Elections of 1993," Public Choice, 94: 157-174, 1998
Jan Potters, Randolph Sloof, and Frans van Winden, "Campaign Expenditures, Contributions, and Direct Endorsements: The Strategic Use of Information and Money to Influence Voter Behavior," European Journal of Political Economy, 13: 1-31, 1997
Andrea Prat, "Campaign Spending with Office-Seeking Politicians, Rational Voters, and Multiple Lobbies," CentER Discussion Paper, Tilburg University, n. 98123, 1998
Andrea Prat, "Campaign Advertising and Voter Welfare," Center for Economic Policy and Research, Discussion Paper No. 2152, May 1999
Randolph Sloof, "Campaign Contributions and the desirability of Full Disclosure Laws," Economics and Politics, 11(1): 83-107, 1999
1 Address: STICERD, London School of Economics, Houghton Street, London WC2A 2AE, UK. Email: firstname.lastname@example.org. Homepage: center.kub.nl/staff/prat. I am grateful to Ellynne Dec and Patrick Francois for useful discussions.
2 Official up-to-date information on US campaign finance is available on the web site of the Federal Election Commission: www.fec.org. Another excellent source of data is the Center for Responsive Politics: www.opensecrets.com. For a cross-country examination of political advertising, see Kaid and Holtz-Bacha (1995).
3 Outside Europe and North America, campaign finance is an important phenomenon in several large democracies, from Brazil to Japan to India.
4 With a small modification of the model, valence can also be seen as the ideological stance of the candidate on moral issues on which the candidate is not willing to compromise. The crucial difference between valence and policy is that only the latter can be chosen by the candidate.
5 Of course, lobbies also have better information on the candidate’s policy stance. However, this informational asymmetry is less important (See Prat 1998 for details).
6 A more sophisticated version, in which lobby-candidate deals are secret, can be found in Prat (1998).
7 In line with the reality of US politics, candidates cannot pocket the money they receive from lobbies. This assumption may not be appropriate for other countries.
8 This is the line of research pursued by Austen-Smith (1987).
9 This does not mean that candidate 1 is elected for sure because there may be other factors (he may make some mistake during the campaign – bad public signal). It just makes the election of Candidate 1 more likely.
10 I am not aware of equivalent studies for the US or other countries.
11 In this model it is immaterial whether the ban is on giving (contributions are prohibited) or on spending (all forms of advertising are prohibited).
12 See Grossman and Helpman (1996) for a formal proof.