“Enough Police, but Not Enough Police Work: An Institutional Explanation of Under-Policing in the U.S.” with Robert Mickey and Daniel Ziblatt. While there is a growing consensus that aggressive criminal sentencing does not tend to reduce crime in the U.S., a popular view is that the country is ``under-policed.'' Although police officers per capita are above the median among OECD countries, the U.S. is an outlier in having few police per homicide. We argue that the U.S.'s low police-per-homicide rate is explained not by police staffing levels but by high levels of bureaucratic shirking, characterized by inefficient allocation, low effort, and the avoidance of bodily harm. We show that U.S. officers are paid significantly more (in relative and absolute terms) while not being at greater risk in the line of duty than their international counterparts. We argue that comparatively high agency loss in policing is explained by institutional features of American federalism, which make possible both the cartel power of American police and the policy pathologies that follow from its deployment.
“Intersectional or Inseparable? Connecting Theory to Estimands in Studies of Identity” with Aaron R. Kaufman and Chris Celaya. In studies of inequality and discrimination, researchers often attempt to isolate the relationship between social identities and social outcomes. However, theories of race, class, intersectionality, and racial capitalism suggest that some identities are causally downstream from other identities, or that multiple identities are conceptually intertwined. Focusing on race and class, we investigate what these distinct theories imply for quantitative studies. Specifically, we distinguish five estimands that correspond to different theories about whether race (or class) is developmentally upstream of class (or race), or that race and class are so developmentally intertwined as to be analytically inseparable: 1) a race-first estimand, 2) a class-first estimand, 3) a factorial estimand, 4) an intersectional estimand, and 5) an inseparable estimand. Using three correspondence audit and audit-like experiments in different contexts, we illustrate the importance of distinguishing these estimands when respondents impute information about an individual's class identity from their racial identity.
“The Insulation of Local Governance from Black Electoral Power: Northern Cities and the Great Migration” with Rob Mickey and Dan Ziblatt. Why has America’s democratization remained incomplete? Democratic institutions in the U.S. are highly decentralized, and lower level governments have the potential to counteract national level democratization. In this paper, we argue that, in response to the threat of growing Black electoral power resulting from the Great Migration, Northern cities moved to insulate governmental institutions from their diversifying electorates. Using a shift-share instrument, we find that greater migration of Black Americans from the South between 1940 and 1970 led cities to switch from mayor-council to city manager systems, shifting the administration of local governmental functions such as budgeting and the hiring of bureaucratic personnel in the office of an appointed manager. We then illustrate how the Great Migration shaped the decision of city elites to switch to city-manager government through a case study of Santa Monica, CA. Our findings show how, at a critical juncture in the course of the country’s national democratization, local governments acted to stymie it.
“The Electoral Effects of Candidate Ideology in the Trump Era” with Adam Bonica. US politics and institutions operate differently in the Trump era. Has the relationship between candidate ideology and election outcomes changed, as well? Leveraging new data, advances in ideological scaling, and multiple research designs, we test whether more ideologically extreme congressional candidates receive lower general election vote shares or affect voter turnout in the 2016-2024 period. We find no evidence of an extremism penalty in general election vote shares. However, in novel analyses that leverage within-voter variation and disaggregate effects by party, we find some evidence of partisan mobilization: more leftwing Democratic candidates increase turnout of registered Democrats, but more rightwing Republican candidates do not increase Republican turnout. National partisan polarization, declining social trust, new information technology, inflation, pandemics, and authoritarianism may have reduced the relevance of unidimensional ideological positioning over the past decade.
“Are Large Corporations Politically Moderate? Using Money in Politics to Infer the Preferences of Business” with Paul Pierson (coverage in The Guardian): The political preferences of business interests has been a topic of debate for decades. Recent measures of the political preferences of large corporations have focused on campaign contributions to legislative candidates from corporate political action committees (PACs). We investigate an alternative source of evidence: contributions to politically-engaged intermediary organizations. We argue that these expenditures—often substantially larger than traditional PAC expenditures—are important sources of information about corporate political preferences. Compared with traditional analyses, they suggest a corporate community that is both more conservative and more closely aligned with the Republican Party.
“When Governments Learn from Copartisans: Partisan Policy Diffusion” (winner of Best Paper in Public Policy at APSA 2018): Louis Brandeis' theory of states as laboratories of democracy suggests that governments engage in learning, emulating successful policies from other states and rejecting unsuccessful ones. However, Brandeis' theory did not address the role of parties. Politicians have incentives to avoid implementing successful outpartisan policies, as this may improve the outparty brand. Furthermore, organizations, such as party-aligned interest and expert groups that provide policymakers with information, may bias institutional learning against outpartisan policies. In turn, state governments may not converge on politically or economically successful policies. This article tests theories of partisan policy learning using a large dataset of policies in the U.S. states. Emulation of successful policies is more likely to occur between governments controlled by the same party. Consistent with the nationalization of party coalitions, success interacts more weakly with geography. The findings have implications for our understanding of the incentives of federalism in the context of partisan polarization.
“Testing City Limits: The Rise of Healthy San Francisco.” Dominant theories argue that cities have a hard time passing redistributive policies, so what explains the development of Healthy San Francisco, a local public health option and employer mandate to provide health care to uninsured workers?