In my paper, I argue that the lack of competition in city council elections can only be explained by understanding the laws governing local elections and how they interact with voter behavior. I develop a model to explain the story.

(For those of you interested in the provenance of the thinking, the model is derived from two-stage entry models used to study tying in antitrust, and the general approach follows the "Politics as Markets" approach to studying election law rooted in the work of Rick Pildes and Sam Issacharoff).

Here's the basic concept: Assume that there are both national and local elections in a city and that both elections use first-past the post/single member district systems (this is true almost all American cities at this point). Next, assume my last post was right. People have at least different preferences about local issues than they do about national issues. That is, members of national parties do not form strongly coherent blocks at the local level -- they could not agree on a common local platform. Party can explain a little about local preferences, but not much more than that.

Given these basic assumptions, the ordinary assumptions of a Downsian model would suggest that two separate party systems would develop -- Republicans and Democrats would contest national elections and either two different parties would contest local ones or the Democrats and Republicans would change their positions on local issues to become competitive at the local level. The question is why this doesn't happen.

To explain, I need to make some assumptions about voters and to incorporate a series of common state, federal and constitutional laws. The two assumptions are that voters care more about national issues than local issues when making party identification decisions. and that individual city council races are low-salience -- voters have little to no independent knowledge about city council candidates. There is substantial evidence backing up both of these assumptions (look in the paper if you're interested ).

The laws are what I call the unitary party rules:

1. National Parties automatically receive ballot places in local elections, usually on the basis of how well they do in gubernatorial races.

2. States and political parties make it difficult to change parties between elections -- there are laws limiting the ability of voters to switch parties and still vote in primaries (for a period of time), and strong limits on the ability of candidates and activists to switch parties between elections.

3. As a matter of Constitutional law, national political parties have the right to participate and use their organizational and financial muscle in local elections (even if the elections are non-partisan).

4. Primary elections are used to select candidates.

Under these laws and the assumptions above, you can see why there is no partisan competition in local elections.

In the first instance, the vote in local elections will directly track the vote in national elections. Voters with little information will use the information that the law provides to them -- the party name on the ballot. If "Republican" and "Democrat" provide a non-zero amount of information about a candidate, a voter with no other information (by assumption) about the candidate will rationally use the national party heuristic to vote.

The question is why the minority local party doesn't modify its issue stances to become popular at the local level. By assumption, the only way it could do this is if it did so on a city-wide level -- individual candidates can't get enough attention. But using primary elections makes this impossible. Local party members don't have consistent preferences and the result is candidates that are all over the map on local issues. Further, they will have trouble attracting candidates and activists from the majority, as those candidates and activists (as well as the representative local voters who they'd have to attract to have primaries among a local issue preference-consistent group) would not choose to join the local minority party in local elections because they care more about national elections and don't want to be penalized.

This is an inefficient outcome -- local elections end up not having competition and hence don't end up with representative results. The question then is why is there not entry?

The assumptions and laws show that there are substantial barriers to entry. A local only third party cannot attract adherents, because people care more national issues than local ones and won't abandon their national party to contest local elections. Further, they face the ordinary limits on third parties -- Duverger's Law and the organizational muscle of the national parties. Under the model, the only way a local-only third party could get adherents is by competing in national elections as well as local elections, but national elections aren't uncompetitive. So there is no entry and no competition

The model is a bit dry, but perhaps it can be explained through a story. This is from my paper…

The dramatic effect of the lack of information on local city council elections can be seen if one considers the case of New York City's Councilmanic District Five on the Upper East Side of Manhattan. In the 2001 local election, Gifford Miller, a powerful and well-known Democratic incumbent who directly after the election would become Speaker of the City Council, faced a relatively unknown candidate named Robert Strougo. Not surprisingly, Miller won 68 percent of the vote to Strougo's 31 percent, neatly tracking the 2-1 dominance of Democrats in the district.

In 2005, a perfect storm of factors lined up to reverse this result. First, Miller could not run for reelection because of term limits. His aide, Jessica Lappin, who had never run for public office before, was the Democratic candidate. Second, Republican Mayor Michael Bloomberg reached new heights of popularity, particularly on the Upper East Side (he would end up winning 59 percent of the citywide vote and more than 80 percent of the vote on the Upper East Side). In District Five, the Republicans nominated Joel Zinberg, a former Democrat, cancer surgeon and Yale-educated lawyer, who built his candidacy around Bloomberg's popularity, declaring his goal as furthering the Mayor's agenda. The New York Times and the New York Post endorsed Zinberg, as did Bloomberg. In the face of this, Lappin's campaign simply sounded a single theme. When asked by a local paper what differentiated the candidates, she responded, "I'm a Democrat. I mean, that's sort of the most obvious difference between us... He's a Republican, and I'm proud to be a Democrat, and I think that certainly distinguishes us."

The result of the election was a near carbon copy of the 2001 race: Lappin received 65 percent of the vote to Zinberg's 35 percent. Thus, in a district which a Republican mayor won 80 percent of the vote, the Republican city council candidate devoted to exactly the same platform as the Mayor only won 35 percent, despite being endorsed by the mayor and the major newspaper and facing a political neophyte. The only factor that mattered was the 2-1 advantage Democrats had in registration.

The story of Joel Zinberg is the story of all city council candidates: what they say and who they are matter very little to those who will vote for them. It is their party status and the popularity of that party at the national level that defines them.