If my model laid out in my paper is correct, it has dramatic implications for non-partisan elections and party primaries.

1. Non-Partisan Elections Make Things Worse, Not Better

One major piece of the Progressive movement's response to big-city political machines was to institute non-partisan local elections. In a non-partisan election, no party preferences are listed on the ballot; candidate names appear by themselves. This is how local elections are conducted in most American localities.

If my story is right, these elections are even worse than partisan local elections. The nub of the story I laid out yesterday was that party heuristics -- the fact of party endorsement -- gave voters in local elections too little information to make an even-mildly informed decision about for whom they should vote. Given the low-information nature of individual city council races, party heuristics make up the lion's share, if not the entirety, of the information that voters receive in local elections. If the party heuristics don't give a great deal of information about what a candidate is likely to do, then voting won't produce a government that represents the preferences of locals on local issues.

Non-partisan elections take the little information contained in the party heuristic at the local level and throw it out the window. If voters in partisan elections are effectively poorly informed about the policy stances of candidates, voters in non-partisan elections are completely lost.

When compared with partisan elections, non-partisan elections generate lower voter turnout, strengthen the incumbency effect and make variables like candidate ethnic, gender and other candidate status variables (such as whether a candidate's name is preceded by an honorific) matter much more. Voters in non-partisan elections are not provided with any easy way to determine the stances of candidates on local policies and, as a result, use variables that don't track at all to policy stances or rely on the name they've heard. As I summarize the research on the field in the paper, "voters in nonpartisan system do not replace the heuristic provided by party label with other, better indicators of policy stances taken by candidates - they do not learn their policy position or rely on newspaper or civic group endorsements. Instead, for the most part, voters simply remain entirely uninformed."

Non-partisan elections make things worse, not better.

2. Majority party primaries do not replace the need for general election competition

As some of the comments have noted, most of the action in big city elections takes place in party primaries. This is true, but it takes little of the normative sting out of the problems caused by a lack of general election competition.

The problem with majority party primaries is the same as the problem with non-partisan elections -- voters are not given any information on the ballot. There are no parties internal to the Democratic or Republican Parties. Even where there are formal internal-to-a-party coalitions, there is no way to provide information on the ballot and hence voters cannot use them to engage in meaningful retrospective evaluation or to collect views of groups over time. Voters in party primaries are, hence as adrift about ideological stances of candidates as voters are in non-partisan elections. They have little information about the candidates and the ballot does not provide them any information to mitigate this ignorance.

As such, majority party primaries do not generate much ideological competition. Not surprisingly, city council primaries usually turn on incumbency, ethnicity and other candidate status variables, party organization support and money, with issue stances playing a very small role. Voters don't have the ability to vote on the issues because they have too little information, and as a result turn to less useful heuristics.

All of that said, it is better to have primary elections than not to have them (they provide some check on truly unpopular politicians). But they do not replicate, or even go much of the way towards replicating, the beneficial effects of general election competition. The conclusion I drew in the last post, that big city elections are unlikely to produce representative outcomes, is not changed by the fact of primary elections.