Winning Proposal to the Pierce Scholar's Fund: 2008 Competition

Project Title: Empirical Estimation of the Electoral Value of the Party Label in Developed Democracies

Faculty Sponsor and PI: Robert J. Franzese, Jr.

Ph.D. Student: Kenichi Ariga

Project Description

This project aims to develop the systematic measurement of "the electoral value of the party label" for 18 developed democracies over the post-WWII period from multivariate statistical analyses on district- and candidate-level voting-data in these countries. The "electoral value of the party label" refers to the extent to which collective party reputation ("party label") is important for the reelection of individual legislators. A systematic and cross-time and cross-country (i.e., cross-political-system) comparable measure of the electorally-motivated incentives of politicians to cohere as a party and "compete as a team" is valuable in itself, and it also has great potential as an important explanatory variable for the variation in a wide range of policy outcomes across developed democracies.

The comparative politics and political economy literatures today are replete with theoretical claims and empirical studies that emphasize the impact of political institutions on policy outcomes. Theories illuminate how various macro-political institutions—presidential vs. parliamentary regimes, proportional vs. majoritarian electoral rules, federal vs. unitary systems, etc.—shape the preferences and incentives of politicians and, thereby, the variation in policies and outcomes across countries. Empirical studies stress differences in institutions as key explanatory variables in tests of hypotheses from these theories. The literature is right to give an important role to political institutions in shaping the incentives of politicians and so policies and outcomes. However, using as key explanatory variables fixed institutional features that either do not vary or are not measured sufficiently precisely to record any variation over time, existing studies are too often too limited in capturing important variation in policymakers' incentives within political systems and even more limited in studying how the effects of these institutions on incentives may vary across contexts depending on other important elements of the political environment.

For example, one popular account for the cross-country variation in the balance between public-goods and particularistic pork-barrel spending stresses the different electoral incentives of parties competing in majoritarian vs. in proportional representation (PR) contests. The account predicts that majoritarian rules induce greater (lesser) spending on particularistic programs (public goods) while the PR rule results in the opposite, because governing parties under the majoritarian rule win reelection by winning a majority in a majority of districts and so can win with the support of much smaller proportion of the electorate than parties competing under the PR rule who, essentially, must win a proportionate majority of the population. The exclusive focus on the majoritarian-PR difference in electoral rules, however, fails to capture the effect of elections on the balance between these programs fully. Some other institutional rules or informal conventions concerning electoral and partisan competition that also vary across countries—such as those conspiring to enhance partisan electoral unity—importantly shape politicians' incentives differently across political systems even under the same electoral rules. The changing nature of the competitive environment from one election to another would also imply changing policymaker incentives over time within the same country. Moreover, these electoral incentives could differ across political parties even within system at the same time, reflecting differences in those parties' natures, rules, and conventions. The exclusive focus on the difference of these broad electoral rules misses these important variations. Furthermore, we strongly suspect that the effects of those broad electoral rules vary depending on the degree of such electoral unity.

Accordingly, this project proposes to measure systematically and comparably the electoral value of party label felt by incumbent legislators, with the aim later to use it also as a prominent explanatory variable in a theoretical model more fully capturing the variation in the electoral incentives of policymakers. The current draft of that theoretical model concludes that, as the collective party reputation becomes more important for the reelection of individual legislators, they would prefer to provide public goods, which would enhance their parties' reputations, and, as its importance declines, they would prefer particularistic programs, which would enhance their personal reputation. We suspect next iterations of this model to show how this partisan electoral unity interacts with electoral-rule proportionality to shape more fully these distributive vs. public-goods spending incentives. (This variable also has potential to explain variation in other important policy outcomes, of course.)

Following Cox and Rosenbluth, we will measure the electoral value of party label by the estimated effect of the national party-swing on the reelection probability of individual legislators. However, we generalize and extend this strategy to PR-system multimember districts and the full range of empirically existing electoral systems among the developed democracies. If the national party-swing greatly affects the reelection probabilities of legislators, their electoral fate is more closely tied together with the collective fate of their party, and therefore, the collective party reputation is more important for them. The statistical model estimates the winning probability of individual legislators using actual, district- and candidate-level election data. Among the most difficult methodological challenges for this estimation is that multi-party voting data cannot be considered as i.i.d. data, especially under electoral rules that permit vote-pooling among candidates from the same party, and standard econometric probability-models are not applicable. The statistical model must accommodate the particular natures of dependency among observations arising at district-level (and across districts within nations). In addition, deriving systematic, comparable estimates of the same substantive quantity—the value of the party label to winning/retaining one's seat—across systems using such widely varying electoral rules across countries requires extreme care and attention to empirical and theoretical detail.

Kenichi is now developing an original dataset of district- and candidate-level election-results across developed democracies over the post-war era that is the most detailed and broadest-coverage such dataset to date. Upon its completion, we expect that Kenichi needs to spend the entire summer to develop and refine the appropriate statistical model(s) to estimate the electoral value of party label in 18 advanced industrial democracies. We apply to the Roy Pierce Scholar's Fund to support this work conducted by Kenichi during the summer.

The project is one of the major parts of the dissertation research conducted by Kenichi, titled "Electoral Rules and the Electoral Value of the Party Label: Intra-Party Bargaining and Public Policymaking in Parliamentary Democracies," but also part of collaborative work with me on the degree and nature of electoral and partisan "budgeteering" (manipulation of budget). In this project, cross-system and over-time variations in the nature and sharpness of partisan competition for seats and government, which includes "representational party unity" centrally, are important explanatory variables for the relative emphases on distributive and redistributive policy. Kenichi's "electoral value of the party label" is or relates very closely to "representational party unity".

Kenichi's project is genuinely comparative and truly at the cutting edge. It will greatly enhance our knowledge of the effects of electoral competition and electoral rules on policymakers' incentives and, thereby, on policies and outcomes. Therefore, I believe it perfectly matches the funding objective of the Roy Pierce Scholar's Fund.