Working Paper Series

Launched in summer 2022, the Faculty Affiliate Working Group meets monthly to share and discuss their research on democracy.  Browse the features below to read their work in progress.

Polarization When Voters Demand Competence

Zuheir Desai, Anderson Frey and Scott A. Tyson

While the documentation and measurement of polarization has become exceptionally good in recent years, the substantive factors that drive it are relatively less well-understood. We develop a theory that identifies the salience of competence as a driving force for platform polarization. An important ingredient of our theory is uncertainty about a district’s median voter, which depends on its ideological leaning, and is more pronounced in centrist districts. We show that platform polarization increases in the centrism of the electoral district and that increased voter demand for competence exacerbates polarization but only in ideologically extreme districts. We examine these implications using data from Brazilian municipality elections. Pre-COVID-19, we find that platform polarization is higher in more centrist municipalities. Using a difference-in-differences methodology we then show that COVID-19 leads to increased political polarization in Brazil, concentrated in municipalities that strongly lean to the left or right.


Competition and Free-Riding in Electoral Campaigns with Outside Spending

Brenton Kenkel and Mellissa Meisels

Why do outside groups spend money on electoral campaigns? How do incentives to free-ride or, conversely, to compete for policy influence affect independent expenditures? We develop a game theoretical model of elections as contests with competition both between and within coalitions of spending committees. Each member of the coalition (including the candidate herself) benefits from the candidate winning, but may also value access to policy favors that are distributed in proportion to campaign expenditures. We structurally estimate the model parameters using data on election outcomes, independent expenditures, and campaign spending in U.S. House elections from 2004 to 2020. Our estimates allow us to interpret electoral and spending data in terms of an underlying utility model for voters and spending committees. Observed spending patterns are consistent with intra-coalition competition among party organizations, 501(c) “dark money” groups, and traditional political action committees (PACs)—but not by Super PACs.

Read The Full Paper Here

Health and Populism: A longitudinal study of individual health and support for populism over three decades

Nolan M. Kavanaugh and Anil Menon

We examine whether declining health drives people toward right-wing populist parties. Using the British Household Panel Survey and the Understanding Society survey, collected between 1992 and 2020, we follow the health and political behavior of thousands of individuals in the United Kingdom over nearly three decades, yielding 175,663 respondent-years. Our findings show that declining health was associated with both increased populist attitudes and support for populist parties. Specifically, a decline in self-reported general health from “fair” to “poor” increased populist support by 0.5 percentage points, or 10% compared to the base rate of 5% populist support. In a cross-sectional analysis, we also find that individuals with worse health were considerably more likely to vote for Brexit. We then utilize 20 years of European Social Survey data to explore the mechanisms undergirding this relationship. Our findings suggest that policies designed to improve public health might help protect against growing threats to democracy.


The Direct Cost to Voters of Polling Site Closure and Consolidation

Scott Abramson and Sharece Thrower

Restrictive voting laws are an increasingly salient and contested feature of American politics. However, estimating their unbiased direct impact on voter turnout is challenging given the strategic and often simultaneous actions of political actors who can both impose and mitigate the costs of these laws. Using individual-level data from Davidson County, Tennessee, we leverage variation induced by an early-morning tornado on Super Tuesday 2020 to estimate the direct causal effect of polling-site closures and consolidations on voter turnout. We find that moving to a new polling station decreases in-person turnout by 6.4% on average. Specifically, travel costs, rather than search costs, drive almost all of this decrease. Additionally, we find that voting at a consolidated polling-site only decreases turnout when the number of voters assigned to a given station increases by more than 140%. This finding suggests prolonged wait-times and overcrowding may also dampen in-person turnout.

Selected Works: Race, Gender and Academic Publishing

Kevin A. Clarke and Lawrence Rothenberg

Despite growing concerns over diversity, the extent to which the race and gender imbalance in academic publishing is a function of a direct form of discrimination or discrimination related to area of study or methodology is not well-documented or understood.  We investigate whether the disparities in political science journals exhibit patterns of taste-based discrimination or statistical discrimination.  Our empirical strategy is to apply a correlated topic modeling to a unique dual data set.  The first data set comprises assistant professors who have published in at least one of the three "top" journals in political science between 2015 and 2017.  The second data set comprises all those who reasonably could have published during that period;  that is, those who completed their PhDs between 2008 and 2018 at top departments (defined widely).  We find that the topics addressed by published authors and dissertations written by women and other underrepresented minorities are different.  These results suggest that some form of statistical discrimination is at work.


The Public's Response to Democratic Backsliding

Katherine Clayton

Self-enforcing democracies require citizens to identify threats to democracy when they occur and punish those responsible by voting them out of office. However, democratic backsliding can proceed incrementally and stealthily, with elected officials gradually subverting democratic institutions under the guise of democratic values. Can citizens identify the actions that lay the groundwork for future power grabs as threats to democracy? I answer this question via large, representative survey experiments conducted in the United States. Consistent with expectations, citizens view “groundwork” actions that make antidemocratic power grabs possible as substantially less threatening than the power grabs themselves. When co-partisan elites justify these groundwork actions as democracy-enhancing, citizens’ threat perceptions are even lower, often translating into majority support for actions that enable in-party power grabs. These findings suggest that citizens’ weak ability to anticipate authoritarian power grabs may explain how elites get away with undermining democracy.

Read the paper here

When Strong Institutions Undermine Strong Democracies

Gretchen Helmke and Jack Paine

This brief essay describes a new type of regime change in which strong institutions undermine strong democracies. Although scholars and pundits have begun to recognize the variety of ways in which constitutions can be weaponized to hollow out democracy, the process has not been adequately conceptualized, nor well theorized. In our view, constitutions may not only permit backsliding, they can incite it. Strong institutions can unravel seemingly strong democracies. When institutions are designed to solve one problem but persist even after circumstances change, new actors who benefit from specific provisions can weaponize the old rules. Democratic erosion occurs not in spite of institutional strength, but because of it.

Read the paper here

The Effect of Local News on Political Knowledge

Scott Abramson and Sergio Montero

The quality of democratic representation is often thought to be predicated upon a robust fourth estate. In many models, this operates through political knowledge: the news media provides voters information about the identity and performance of elected officials, which voters use to hold them accountable. Under weak monotonicity assumptions, we nonparametrically identify sharp bounds for the effect of local news availability on voters' general political knowledge, accounting for knowledge spillovers due to social interactions. We show that increasing the number of local news sources unambiguously improves voters' factual knowledge about incumbent politicians. Furthermore, we show this is not specific to a media type, finding qualitatively and quantitatively similar results concerning the competitiveness of local television and local print news markets.

Paper coming soon!