WikiSummary, the Social Science Summary Database

Ansolabehere, de Figueiredo, and Snyder: Why is there so little money in US politics

From WikiSummary, the Free Social Science Summary Database


Ansolabehere, de Figueiredo, and Snyder. 2003. Why is there so little money in US politics? Journal of Economic Perspectives 17:105-130.

In Brief

People donate to political campaigns not as an investment, but merely for the consumption utility they get from doing some they love. They aren't trying to buy anything with their money--they just like participating in politics.

Main Argument

In most theories of campaign contributions, models assume that donations are investments--i.e. that donors expect candidates to act in the donors' interests should the candidate win. The authors argue that such is not the case, however; as evidence, they present a two-stage least squares regression showing that legislators do not change their voting habits in response to donations. Instead, they respond to their party ID and district partisanship.

If donations aren't investments, then what are they? The authors argue that donations are consumption, not investment. Just as we donate to charities because we feel good doing so (that is, as an act of consumption and not an investment), we donate to politicians for analogous reasons. Donating is simply another means of democratic participation. And since donating is a consumption good, the best predictor of donating is personal wealth.

The authors could have clarified their point via and analogy to Riker and Ordeshook's calculus of voting. Essentially, Ansolabehere et al. are saying that the p and b terms (from Riker and Ordeshook) implied by investment theories of campaign donations are erroneous; the emphasis ought to be on the d term.

While it is true that people donate because of ideological motivations, they also donate simply out of interest in elections and for the utility they derive from donating to like-minded candidates. Donating is simply another means of democratic participation. And since donating is a consumption good, the best predictor of donating is personal wealth. And, sure enough, income is the strongest predictor of donating.


  • H1: Since donating is a consumption good, personal income affects level of donation.
  • H2: You channel your donation for maximum utility by putting it toward close elections.

Empirical Evidence

Data: Proving that Contributions are Not Investments

  • Source: Campaign contributions' influence on legislators' roll call votes in US House, 1978-1994
  • Y: Legislative votes that favor business interests
  • X: Corporate Contributions, Labor Contributions, Member Party ID, District Partisanship, Southern dummy

Across models, corporate and labor contributions have relatively small effects on members' roll call votes compared to members' party affiliation and their districts' partisanship.

Data: Campaign Spending Hasn't Risen as Fast as You Think

Sure, campaign spending has risen dramatically (with inflation controlled), but it's steady (as percent of GDP). We're only spending more because we're wealthier.

Data: Proving that Income Predicts 'Consumption' of Donating

See Table 3 in the text.

  • Y: Log of Campaign Spending in Gubernatorial Elections, per capita
  • X: Log Income per capita (H1), Log State Government Spending per capita, Log of population, closeness of election (H2), closeness of primaries, incumbency dummies,

Results: Personal income per capita is positive and significant and coefficient is approximately 1--that is, contributions are almost perfectly elastic with respect to income. Closeness of the election has the strongest effect (an effect almost three times stronger than the effect of personal income), while closeness in the first and second primary are almost perfectly elastic (all significant).

Comments and Criticisms

Are the investment and consumption theories mutually exclusive? In particular, doesn't the fact that we channel our donations toward close races imply that investments matter too? Some theories of voter turnout have sought to dismiss Riker and Ordeshook's p, b, and c terms as irrelevant, focusing only on the d term--an approach analogous to the argument here. However, without the other terms from the calculus of voting, we cannot explain why citizens our strategic with their participation. While consumption utility may explain why some people vote and others do not--or, more to the point, why some contribute and others do not--it is not sufficient to explain why we are strategic about how and where to contribute.