Most money spent in US congressional campaigns comes from donors outside a race’s electoral district. Scholars argue that representatives accepting out-of-district donations become “surrogate representatives” for outside donors. Yet researchers neglect a critical question: how do geographic constituents react when their representatives accept money from out-of-district donors? We argue that geographic constituents feel forced to “share” their representatives with outside donors at the expense of their own representation. In an experiment during Georgia’s 2021 Special Senate Election, Georgians who learned about out-of-district donations to particular candidates expected representatives to spend less time and effort working for Georgians. A second experiment during the 2022 Senate elections identified the causal mechanism: local identity. Respondents whose local identity was primed and learned about out-of-district donations expected representatives to spend less time and effort working for them. Our findings highlight the rivalrous nature of representation and the tradeoffs accompanying out-of-district donations and surrogate representation. (PDF; Dataverse)

Local governments affect the lives of their citizens on a daily basis, and it matters who serves. Much work has shown that, at all levels, Black citizens tend to be descriptively underrepresented in government. We take up the question of Black descriptive representation at the level of the county legislature, gathering data on the composition of North Carolina's 100 county commissions. We propose an alternative measure of descriptive representation, termed "seats above expectation," and apply a counterfactual simulation approach to gauge the effects of at-large and ward-based elections. We find that Black citizens are underrepresented statewide: there are four fewer Black county commissioners than we would expect, conditional upon county board sizes, demographics, and institutional arrangements. Furthermore, we find that universal implementation of ward-based elections could increase the statewide total of Black county commissioners by as much as 20, a 17% increase over the baseline. (PDF; Dataverse)

A growing body of evidence attests that legislators are sometimes responsive to the policy preferences of citizens in single-party regimes, yet debate surrounds the mechanisms driving this relationship. We experimentally test two potential responsiveness mechanisms—elections versus mandates from party leaders—by provisioning delegates to the Vietnamese National Assembly with information on the policy preferences of their constituents and reminding them of either (1) the competitiveness of the upcoming 2021 elections or (2) a central decree that legislative activities should reflect constituents’ preferences. Consistent with existing work, delegates informed of citizens’ preferences are more likely to speak on the parliamentary floor and in closed-session caucuses. Importantly, we find that such responsiveness is entirely driven by election reminders; upward incentive reminders have virtually no effect on behavior. (PDF; Dataverse)

This research note builds upon recent experimental work in the Vietnamese National Assembly (VNA) to explore a critical qualification regarding potential responsiveness in authoritarian parliaments: delegates grow increasingly responsive as the number of peers possessing the same information rises. This reinforcement, or "safety-in-numbers," effect arises because speaking in authoritarian assemblies is an intrinsically dangerous task, and delegates are reluctant to do so without confidence in the information they would present. This logic contrasts sharply with the notion of performative responsiveness occurring in more democratic parliaments. Here we describe the saturation design for the original experiment, theorize safety-in-numbers behavior among authoritarian legislators, and test additional observable implications of the logic. (PDF; Dataverse)

We investigate whether communicating constituents' preferences to legislators increases the responsiveness of delegates to the VNA. Utilizing a randomized control trial, we assign legislators to three groups: (1) those briefed on the opinions of their provincial citizenry; (2) those presented with the preferences of local firms; and (3) those receiving only information on the communist party's objectives. Because voting data is not public, we collect data on a range of other potentially responsive behaviors during the 2018 session. These include answers to a VNA Library survey about debate readiness; whether delegates spoke in group caucuses, query sessions, and floor debates; and the content of those speeches. We find consistent evidence that citizen-treated delegates were more responsive, via debate preparation and the decision to speak, than control delegates; evidence from speech content is mixed. (PDF; Dataverse; J-PAL policy evaluation)

The emergence of automated text analysis has opened up possibilities for the systematic analysis of large quantities of legal documents on a previously unimaginable scale. This is a development that has the potential to allow social scientists to "take law seriously" in the sense of moving beyond the relatively simple measures that have dominated quantitative work in judicial politics to richer and more nuanced measures of the substantive content of legal texts. This paper provides an introduction to automated text analysis methods for scholars in judicial politics. In addition to laying out the intuition behind the most important methods, we provide an application of these methods to a prominent recent legal dispute (the U.S. Supreme Court's decision on the Affordable Care Act), and consider the promise and potential pitfalls of automated text analysis in the study of courts and legal systems. (PDF)

In recent decades, the American political system has become increasingly polarized. Has this trend affected the U.S. Supreme Court? In this chapter, we approach the question empirically through seven decades’ worth of data on the nomination, confirmation, law clerk hires, and voting behaviors of the justices. We find strong evidence of increased polarization in the perceived ideology of nominees and in the Senate’s confirmation process. However, polarization’s impact is less clear-cut where the behavior of the justices themselves is concerned. While new patterns such as ideological homophily in the clerk hiring process and partisan sorting in voting behaviors point toward greater polarization, network analysis of the voting coalitions reveals that moderate levels of polarization are not new to the Court. (PDF)

Congressional standing committees incentivize legislative expertise, aid in controlling the agenda, and serve as the drawing rooms for legislative proposals. Recognizing this, scholars have long sought to measure the relative prestige of congressional committees via analysis of intercommittee transfers. Yet by ignoring the accrued committee seniority of members who leave a committee, these methods have inadvertently treated all transfers as equal. I introduce a measure from network science which exploits the additional information that seniority contributes to produce more accurate estimates of committee prestige. I then demonstrate that the prestige of a legislator's committee portfolio predicts the political action committee (PAC) contributions she receives for the next campaign cycle. (Draft available upon request.)

Although political polarization is universally decried as a rising force in American politics, the concept is often conflated with -- and measured as -- partisan polarization. I offer a more behavioral conception and a network-based measure that (1) permits the distinction between partisan and nonpartisan polarization and (2) can be applied both to collegial courts and legislatures. I apply this measure to opinion-joining networks at the U.S. Supreme Court (1791-2017) and to cosponsorship networks in both chambers of the U.S. Congress (1973-2017), finding that while overall polarization has not risen appreciably in recent decades the partisan component of polarization has spiked dramatically. (Draft available upon request.)

We develop two new metrics -- one of individual-level outgroup exposure, the other of area-level segregation -- based on individuals' geocoded locations. Validating the exposure metric with an original network census from India, we find that the k-nearest-neighbors (knn) score accurately reflects social contact, which conditions intergroup relations. The knn score is also a better proxy for social contact than two existing measures. We also demonstrate that averaging over the knn score yields an accurate measure of residential segregation, and simulations confirm that it outperforms two commonly used metrics in discriminating between segregated and integrated areas. (Draft available upon request.)

Abstract in progress.

Abstract in progress.

Building on important recent work, I propose a new measure of judicial preferences that is rooted in a uniquely judicial behavior, acknowledges the centrality of precedent, and can be estimated for all federal appellate judges. The present paper motivates the measure, describes the necessary data and resultant challenges, and presents preliminary results---a doctrinal common space for federal judges. (Draft available upon request.)

With the ebb of democratization's "third wave" and ensuing "democratic recession", scholarly interest in authoritarian regimes has surged. Largely coeval with a boom in new institutionalist approaches to political science, the discipline's renewed attention to authoritarian regimes has frequently trained its sights upon the political institutions of dictatorship. This survey attempts to organize and synthesize the sub-literature on authoritarian legislatures. In the process, it raises several concerns, including the paucity of micro-level evidence, roughly operationalized concepts, issues of causal identification, epiphenomenality, and a narrow focus on stability and growth. Future avenues for fruitful future research are then proposed. (Draft available upon request.)