Campaign finance levels as coordinating signals in three-way, experimental elections

If (often costly) election campaigns are simply advertising, they do not increase social welfare directly. Given this, should we limit campaign expenditures? We propose that costly campaigns can inform voters about the strength of candidates. This may increase welfare indirectly by helping voters avoid coordination failures. In laboratory elections, we study campaign finance levels as coordinating signals and compare them with our earlier work on polls. Both coordinate majority voters effectively, allowing them to stop Condorcet losers from winning. Finance levels were rational in that the total benefits of coordination exceeded the costs. Further, benefits of typical incremental contributions exceeded costs ex-post, while the next typical increment's benefits would not have.