How a firm can induce legislators to adopt a bad policy

Transcript, Lawrence F. O’Brien Oral History Interview XVII, 12/17/86, by Michael L. Gillette, Internet Copy, LBJ Library.

Danielle Kurtzleben, $4 billion in election spending a drop in the bucket, http://​www.​usnews.​com/​news/​articles/​2010/​11/​09/​4-billion-in-election-spending-a-drop-in-the-bucket.​html, downloaded December 29, 2010.

Nor is the size of illegal contributions large. Investigations following the Watergate scandal found that 21 companies made illegal contributions in 1972, totaling only $968,000; the largest one was made by Northrop for a mere $150,000 (Alexander 1980).

Models that combine both influence instruments are Bennedsen and Feldmann (2006), and Dahm and Porteiro (2008a, 2008b). In these models special interests choose between providing policy-relevant information or engaging in non-informational influence activities (such as by making campaign contributions).

Analyzing a game with two rent-seekers, Corchón and Dahm (2010) show that some contest success functions can be derived as a utility-maximizing choice from a setting in which rent-seekers are uncertain about the type of politician. Corchón and Dahm (2011) derive contest success functions without the commitment assumption. See also the literature reviewed in these papers.

For works arguing that special interests inform legislators about political consequences of legislative votes, see Smith (1984), Hansen (1991), Austen-Smith (1993), Austen-Smith and Wright (1992, 1994), Rasmussen (1993) and Lohmann (1995, 1998).

See Uhlaner (1989) and Schwartz (1987), who discuss a “local public benefit” to a group that increases with the number of votes the winner received from that group. Lapp (1999), however, testing the model, finds little empirical evidence that ethnic leaders can increase turnout among their followers.

The firm may invest with probability less than one. The values of g and b can then be taken as expected values.

In Sect. 6.1 we relax the assumptions that m≥1+(n+1)/2 and that the proportion of districts favoring the investment is common knowledge. We could also describe our setting as follows. First, nature assigns types to legislators (so that with probability γ a legislator favors the investment) and each legislator knows his type. This implies that m could be very small and our assumption on m just says that we focus on realizations in which m is larger. Then, each legislator gets a perfect signal concerning the other legislators’ types. The firm does not receive this signal. Under this interpretation Sect. 6.1 investigates robustness when neither legislators nor the firm receive a signal and m≥1+(n+1)/2 need not hold.

Faccio and Parsley (2009) show that the sudden death of a politician causes a 1.7 % decline in the value of companies headquartered in the politician’s home town. This decline is followed by a decline in the rate of growth in sales and a decline in access to credit.

Obviously, we do not require a legislator favoring the investment to use the same strategy as a legislator opposing the investment, with the labels ‘voting for’ and ‘voting against’ swapped.

Strictly speaking, we assume that a district that does not want the investment cannot prevent the firm from locating in the district. Incorporating such a veto power, Proposition 1 below still holds, although the strategy of the legislators from districts that oppose the investment is then weakly dominant.

Note that, by definition of n and k, n−k−1 other legislators vote against the tariff, who constitute a majority if n−k−1≥(n+1)/2, implying k≤(n−3)/2.

Besides the two voting profiles described above, two other symmetric strategy profiles are candidates for an equilibrium. Each, however, requires legislators who do not favor investment to use a strictly dominated strategy.

To see that an equilibrium with a dominant strategy is compatible with an inefficiency, consider the example in which g=F=1. It must hold that F/n<C and C/g<2/(n+1). These expressions define a non-empty interval for C described by 1/n<C<2/(n+1). But, of course, as the inefficiency becomes increasingly severe, the equilibrium is no longer in dominant strategies.

This argument can be made precise. Denote by F ^m the firm’s benefits from a successful match. In the unique equilibrium, social welfare is F−nC+F ^m +g. Suppose the alternative has no vote on the tariff, with the firm deciding randomly where to invest. Then γ(F ^m +g)+(1−γ)(−b). If (1−γ)(F ^m +g+b) is smaller than −(F−nC), the inefficiency is not reversed.

We see x ₁ as a more appealing equilibrium than x ₂, because it is plausible that as legislation becomes more inefficient it is less often approved. There are further reasons to focus on x ₁. Consider n=3 and suppose that x ₁ and x ₂ exist. Denote by A(x) and B(x) the expected payoffs from voting for and against the tariff, respectively, when the two other legislators vote for the tariff with probability x. We have that A′(x)<0, A″(x)>0, B′(x)<0, B″(x)<0, and A(0)>B(0)>A(1)>B(1). Hence, A(x) and B(x) intersect twice (at x ₁ and at x ₂), and expected payoffs are strictly higher at x ₁. Another reasoning could be based on a simultaneous version of Cournot’s tatonnement process, adapted to symmetric mixed-strategy equilibria. Consider an equilibrium \(\hat{x}\) to be (locally) stable if given a collective mistake in which everyone mixes with probability \(\hat{x}+m\), where m∈{−ϵ,ϵ} with ϵ>0, the following holds for all legislators: \(A(\hat{x}+\nobreak m)>B(\hat{x}+m)\) if m<0, and \(A(\hat{x}+m)<B(\hat{x}+m)\) if m>0. Similarly, consider an equilibrium \(\hat{x}\) to be (locally) unstable if for some legislator the opposite inequality holds. Given how A(x) and B(x) intersect, x ₁ is stable, whereas x ₂ is unstable.

If all districts with priority offer no incentives, we assume that all districts have the same win probability 1/h, where h is the number of districts with priority.

Contrary to the benchmark, when u(−b,l)=0 it is now an equilibrium for all legislators to vote for the tariff. A legislator who opposes the investment might vote for the tariff because his district does not compete actively at the lobbying stage and, thus, will not attract the firm’s investment. Anticipating this, he is indifferent between voting for and against the tariff.

Notice that even if at the last stage competition is very fierce and the firm can appropriate all the rent from matching to a district favoring the investment, it is still an equilibrium for a legislator who favors the investment to vote for the tariff. If the lobbying stage is modeled as a first-price auction, in the undominated pure strategy Nash equilibrium a district that favors the investment obtains the expected payoff u(g,l)=0; see Alcalde and Dahm (2011). That is, any one legislator gains nothing from voting against the tariff when a majority of other legislators vote for it, and might as well vote for it.