Nathan J. Kelly - Department of Political Science




Ongoing Projects:

Income Inequality
My work contributes to a growing body of research that points to the importance of politics and public policy as important causes of inequality. By linking theoretical insights from American and comparative politics, I place distributional outcomes at the core of the American political system. I show that partisan and ideological politics shape income inequality, with inequality declining when Democratic policymakers are in office and liberal policies are enacted. My most important contribution in this work is an analysis of the mechanisms than link politics to income inequality. I focus on two mechanisms - traditional redistribution and the conditioning of market outcomes. I find that traditional redistribution is less responsive to political dynamics than market outcomes, and that politics has a larger impact on income inequality by conditioning markets than through explicit transfer programs. The results of this portion of the project are reported in two journal articles (2004 American Politics Research and 2005 American Journal of Political Science) and a book (2009 Cambridge University Press).

More recently, I have extended my analysis of the macro political theory of distributional outcomes to other contexts. One project (with Chris Witko) along this line examines economic inequality and other economic outcomes in the American states. A second project (with Tom Volsho) in this vein focuses on top income shares as the primary outcome of interest. I have also applied (with Jana Morgan) my model of economic inequality to distributional outcomes in Latin America and the Caribbean. While some recent studies have assessed the impact of politics on cross-national differences in inequality across Latin American countries, none have assessed the market conditioning mechanism that I focus on in my work.

A related but distinct component of the inequality project seeks to better understand the political consequences of inequality and public opinion toward redistributional policies. In my first paper on this topic (2010 AJPS with Peter Enns), I show that, contrary to traditional economic models of democracy and redistribution, public opinion becomes more conservative in response to income inequality. Moreover, both the rich and the poor respond similarly to changes in inequality. I have also authored a paper (2010 Poverty and Public Policy with Jana Morgan) in which I seek to explain individual attitudes toward redistribution in Latin American countries. I am currently working on a series of papers (with Tom Volscho) that explore how economic inequality feeds back on the political system. Finally, my latest book project examines in detail what I call "America's Inequality Trap," which describes a situation in which high levels of economic inequality generate a system of positive feedback that makes it increasingly difficult to reverse recent trends toward greater income inequality.

Assessing the Quality of Time Series Estimators

A portion of this project (with Luke Keele) uses Monte Carlo analysis to assess the nature and degree of bias in several common time series estimators in the presence of autocorrelation. More recently, I have worked (with Peter Enns and Takaaki Masaki) to examine the situations in which Error Correction Models can be utilized effectively for hypothesis testing in the social sciences. We show that many current applications of ECMs have not been applied appropriately and develop an easily implemented appraoch to ensure that future applications of these models can produce valid inferences.

Completed Projects:

Policy Production in the United States

I (with J. Tobin Grant) create measures of legislative, presidential, and judicial policy production that go back in time to the enactment of the current constitution in 1789. We also explore the factors that determine the path of these series over time. The central questions are: 1) How has policymaking in the three constitutional branches of government varied over time? 2) How does policymaking in each branch respond to policymaking in the other branches? 3) What factors explain variation in the volume of policy production in each branch? Our work contributes to understanding the role of divided government, the distribution of preferences within and across the branches, ideological polarization, mass preferences, and technological change in the creation of public policy.

Religion and Latino Politics

This project (with Jana Morgan Kelly) examines recent changes in the religious composition of the Latino population and the political implications of these changes. Specifically, the decline of Roman Catholicism among Latinos has lessened the Democratic advantage in this increasingly important group.
The Consequences of Survey Mode

Much attention has been given to the negative consequences of using telephone surveys to study political attitudes and behavior. This project (with Brian Fogarty and H. Whitt Kilburn) assessed the degree to which survey mode affects the substantive conclusions of research on political attitudes and attitude structure in the United States using the 2000 mixed mode National Election Studies dataset.