The study of elections to the U.S. House of Representatives has long struggled
with the basic tension between national partisan and local district forces. Each of the 435
campaigns for a voting seat in the House can be both a proxy for the larger political
questions of the day and a contest between individual candidates with their own strengths
and weaknesses. This distinction between national and local is important because it holds
implications for the House as a policymaking institution. If each race substitutes for
larger conflicts between the national parties, individual members of Congress have every
incentive to obey the will of their party and to help it project a positive image. After all,
their electoral fates are dependent on the success of the party as a whole (Cox and
McCubbins 1993). By contrast, if individual members are responsible for their own
electoral fates, they have little need for the party and can largely chart their own
legislative course. They will assemble their own coalitions of interests at the local level
and will probably respond largely to those coalitions.
Of these two descriptions of congressional elections, the literature has focused
mainly on local district variation as the most important phenomenon to explore.
Individual candidates are believed to determine much of their own fates: they choose
when to run, they raise their own money, and they win by selecting useful campaign
issues and communicating them effectively to voters. National political forces are
considered secondary in importance and surmountable through money and skill.
In this paper, I use a multi-level modeling approach relatively new to political
science, as well as data on House campaigns over a 26-year time period, to examine some
of these assumptions more carefully. I take as a starting point that the literature has been
right to emphasize the importance of spending and candidate experience to House
elections. Rather than ignore the importance of these factors, I explicitly test the strength
and consistency of national political forces in the context of local competitive dynamics.
This approach allows me to examine whether competitive campaigns act as mediators
between national forces and local results, and whether national forces have consistent
effects or are felt only sometimes, in some districts.
Once all these factors are explicitly modeled together, it becomes clear that
national political forces play a decisive role in congressional elections. Not only do they
have a significant effect on individual campaigns, it appears this effect touches most
Convention