No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
On the Validity of the Regression Discontinuity Design for Estimating Electoral Effects: New Evidence from Over 40,000 Close Races
... We validate our quasi-experiment following the procedures established by Imbens and Lemieux (2008) and Eggers, Fowler, Hainmueller, Hall, and Snyder (2015). We adopt three procedures. ...
... We established four procedures to test the robustness of our main results. First, we ran our main results again with different bandwidths (Eggers et al., 2015;Imbens & Lemieux, 2008). Second, we ran our main results with the polynomial of different orders (Eggers et al., 2015;Imbens & Lemieux, 2008). ...
... First, we ran our main results again with different bandwidths (Eggers et al., 2015;Imbens & Lemieux, 2008). Second, we ran our main results with the polynomial of different orders (Eggers et al., 2015;Imbens & Lemieux, 2008). Third, we observed if there were any differences between right-wing and centrist leaders. ...
... In short, the RDD identify a local average treatment effect (LATE) around the cut-off. Close electoral races are widely employed in political science and political economy, as results are as-if-randomly distributed, which gives rise to a quasi-natural experimental setting(Eggers et al., 2015). Nevertheless, there are a few qualifications to make. ...
The present paper investigates whether highly skilled politicians, or experts, display different post-electoral behaviour than less skilled ones. Drawing from existing literature on the professionalisation of politics, the essay distinguishes party loyalists from experts, who have higher educational, professional outcomes and pre-electoral income. Subsequently, it employs open data by the European Parliament (EP) on the Members of the European Parliament (MEPs), combined with MEPs’ speeches in plenary sessions, to create a novel dataset of 777 total entries, composed of 563 individuals speaking 123340 times in the 2009-2024 period. The speeches are then used to estimate MEPs’ speech-based policy positions on the R package quanteda. Through linear regressions and a regression discontinuity design (RDD), the essay tests three hypotheses on MEPs’ post-electoral behaviour, including on their absence rate, leadership appointments, and speech-based policy positions. The paper finds that expertise is associated with lower absence rates in Parliament, but a heightened chance of resigning voluntarily from the party or the EU faction. Contrary to theoretical predictions, although politicians hold distinct policy positions, experts do not differe themselves from loyalists in their speeches. Ultimately, the paper sheds light on two aspects of experts’ post-electoral behaviour. First, there are potential trade-offs of employing experts, as they are significantly more likely to leave the party during their term in office. Second, that experts’ policy preferences do not differ per se from party loyalists, which yields heterogeneous but overlapping policy preferences across parties.
... Eggers et al. 2015. ...
Does government spending on public goods affect the vote choice of citizens? On one hand, prior research has characterized voters as fiscal conservatives who may turn toward conservative parties when government spending goes up. On the other hand, increased spending may signal that the economy is doing well, which makes progressive parties a more viable option. To adjudicate between both hypotheses, this article draws on a natural experiment, which created exogenous variation in government spending. A discontinuity in the 2011 German census meant that some municipalities saw an unforeseen increase in budgets. Using a regression discontinuity design, the authors show that the increase in budgets and subsequent spending on public goods benefited left-leaning parties but had no detectable effect on incumbent support. To parse out the causal channel, the authors rely on panel evidence and demonstrate that treated residents viewed their economic situation more favorably than did untreated residents, which led the former to espouse progressive parties.
... Para uma discussão específica e mais completa sobre a validade do pressuposto de aleatoriedade do resultado das eleições acirradas, verEggers et. at. (2015). ...
Neste estudo, investigamos se prefeitos que possuem incentivos de reeleição recorrem ao uso da máquina pública para aumentar suas chances de serem reeleitos. Utilizando dados da RAIS sobre empregos e dados eleitorais do TSE em nível municipal, os resultados apontam que municípios liderados por prefeitos com incentivo de reeleição empregam, em média, 51,6 pessoas a mais para cada 100 mil habitantes no poder executivo municipal em comparação com aqueles liderados por prefeitos sem incentivo de reeleição. Além disso, nos anos eleitorais os prefeitos também empregam mais pessoas no poder executivo, independentemente de possuírem incentivo de reeleição. Análise de robustez aponta que essa tendência não se estende ao poder legislativo municipal, indicando que o resultado principal é, de fato, oriundo do incentivo à reeleição do prefeito. Além disso, o uso de desenho de regressão de descontinuidade mostra que, para uma amostra de municípios com eleições equilibradas, não existe relação entre incentivo de reeleição e número de empregados no poder executivo, sugerindo uma associação entre competição eleitoral e menor incidência de práticas clientelistas.
... In first-past-the-post elections in which the winner takes all, there is a sharp discontinuity at the zero vote margin between the top two candidates. In this setting, the identity (and hence gender) of the winner can be considered quasi-random (Lee, 2008;Eggers et al., 2015;Imbens & Lemieux, 2008). Comparing constituencies in which a woman wins against a man by a narrow margin ('treated') with those in which a man wins against a woman by a narrow margin ('control') can thus isolate the causal influence of legislator gender. ...
There has been a phenomenal global increase in the proportion of women in politics in the last two decades, but there is no evidence of how this has influenced economic performance. We investigate this using data on competitive elections to India’s state assemblies, leveraging close elections to isolate causal effects. We find significantly higher growth in economic activity in constituencies that elect women and no evidence of negative spillovers to neighbouring male-led constituencies, consistent with net growth. Probing mechanisms, we find evidence consistent with women legislators being more efficacious, less corrupt and less vulnerable to political opportunism.
... If this assumption is violated, we will be less likely to detect an effect in elections where the Democratic candidate won, which makes for a conservative test of Prediction 3. I find that participating in an upstream election where the Democratic candidate won was associated with persistent downstream effects among compliers four years later and no turnout persistence in states where the Republican won (see Table 5). One concern is that the election outcome may not be truly exogenous, but Eggers et al. (2015) found compelling evidence that the outcomes of close elections are as-if random. 24 Three of the four states in this analysis were top-of-the-ticket tossups (IL 2014, MI 2014, NC 2016 and Iowa 2014 had a competitive down-ballot race. ...
I apply a new theoretical framework to voting to more cohesively bridge the economic cost-benefit model of voting with the psychology-motivated voting-as-a-habit literature. This new theoretical frame gives greater clarity as to how a vote in one election might beget a vote in another election, while yielding testable predictions as to which circumstances are more favorable for developing turnout persistence. To test these predictions, I make use of a novel dataset consisting of nine large-N, door-to-door voter mobilization field experiments in various election contexts (with ∼1.8 million voters in total). Consistent with prior empirical research, my analysis finds that being nudged to vote in one election leads to increased turnout four years later. But the main contribution of this paper is that the theoretical framework’s predictions and the corresponding empirical results make sense of turnout persistence heterogeneities that have been detected in certain prior empirical studies but not others.
... Measuring closeness is important for two broad strands of the literature. One focuses on close elections for identifying causal effects of winning elections using regression discontinuity (RD) designs both at the party and the candidate level (Caughey and Sekhon 2011;Eggers et al. 2015;Fouirnaies and Hall 2014;Lee 2008). The other considers closeness as an important dimension of electoral competitiveness (Cox, Fiva, and Smith 2020;Grofman and Selb 2009). ...
We provide closed-form solutions for measuring electoral closeness of candidates in proportional representation (PR) systems. In contrast to plurality systems, closeness in PR systems cannot be directly inferred from votes. Our measure captures electoral closeness for both open- and closed-list systems and for both main families of seat allocation mechanisms. This unified measure quantifies the vote surplus (shortfall) for elected (nonelected) candidates. It can serve as an assignment variable in regression discontinuity designs or as a measure of electoral competitiveness. For illustration, we estimate the incumbency advantage for the parliaments in Switzerland, Honduras, and Norway.
... 12 https://cces.gov.harvard.edu/data. 13 While this assumption has been disputed in a small number of particular cases (Caughey and Sekhon 2011), it holds under the majority of cases studied (Eggers et al. 2015). ...
In the United States, elections are often administered by directly elected local officials who run as members of a political party. Do these officials use their office to give their party an edge in elections? Using a newly collected dataset of nearly 5,900 clerk elections and a close-election regression discontinuity design, we compare counties that narrowly elect a Democratic election administrator to those that narrowly elect a Republican. We find that Democrats and Republicans serving similar counties oversee similar election results, turnout, and policies. We also find that reelection is not the primary moderating force on clerks. Instead, clerks may be more likely to agree on election policies across parties than the general public and selecting different election policies may only modestly affect outcomes. While we cannot rule out small effects that nevertheless tip close elections, our results imply that clerks are not typically and noticeably advantaging their preferred party.
... Para que o desenho seja válido, é preciso atender ao pressuposto de continuidade ao redor do ponto de corte, ou seja, nesse caso, que candidatos não possam determinar as próprias votações em eleições acirradas. As evidências disponíveis apontam que esse pressuposto é atendido para o caso brasileiro (Eggers, Fowler, Hainmueller, Hall, & Snyder, 2015). Para o meu conjunto de dados, realizei o teste de densidade proposto por McCrary (2008), para confirmar o uso adequado do desenho. ...
Resumo Municípios partidariamente alinhados com o presidente ofertam mais políticas públicas? A entrega de serviços públicos é uma das atividades centrais dos governos. Uma vez que são os partidos políticos que controlam o Poder Executivo tanto no Governo Federal quanto no governo local, é razoável esperar que seus interesses partidários e eleitorais influenciem o curso da implementação de serviços. Neste artigo, analiso a cobertura da atenção básica à saúde como indicador de oferta. Dado o forte compartilhamento de responsabilidades entre os entes federados, argumento que o Governo Federal implementa políticas públicas de forma estratégica, aumentando a oferta de serviços em municípios partidariamente alinhados. Para testar empiricamente essa relação, estimo o efeito causal do alinhamento por meio de um desenho de regressão descontínua para eleições acirradas. Os resultados indicam que os municípios partidariamente alinhados com a Presidência da República têm cobertura da atenção básica à saúde, em média, 3% maior que outros governados por partidos oposicionistas. Em uma cidade com 10 mil habitantes, por exemplo, isso significaria 300 pessoas a mais sendo adequadamente atendidas por equipes de saúde.
... For the design to be valid, it must meet the assumption of continuity around the cutoff point, that is, in this case, that candidates cannot determine their own votes in close elections. The available evidence indicates that this assumption is met in the Brazilian case (Eggers, Fowler, Hainmueller, Hall, & Snyder, 2015). For my dataset, I performed the density test proposed by McCrary (2008) to confirm proper use of the design. ...
This research examines whether local governments politically aligned with the president’s party offer more public policies. The delivery of public services is one of the central activities of governments. Since the political parties control the executive branch in both federal and local government, it is reasonable to expect that their partisan and electoral interests will influence policy implementation. This study analyzes the coverage of primary healthcare as an indicator of policy delivery. Given the strong sharing of responsibilities among federated entities, I argue that the federal government implements public policies strategically, increasing the supply of services in partisan-aligned municipalities. To empirically test this relationship, I estimate the causal effect of alignment using a regression-discontinuity design for close elections. The results indicate that partisan-aligned municipalities have, on average, 3% higher coverage of primary healthcare than others governed by opposition parties. In a city with 10,000 inhabitants, for example, this would mean 300 more people receiving healthcare.
Many IT professionals seek to improve their job prospects by engaging as leaders of online communities, for example, by serving as a moderator or admin. We investigate whether such community leadership leads to (causal) improvements in individuals’ careers. We assemble a data set, including job histories of IT professionals who have sought election as moderators (mods) in Stack Exchange question-and-answer communities, between 2010 and 2020. We estimate the career benefits of moderatorship under two complementary identification strategies: difference-in-differences (DID) and regression discontinuity (RD). We observe qualitatively consistent results under each design, finding that election to a moderator role has a significant, causal, positive effect on job mobility. We estimate that moderatorship increases the probability of a job change by between 4.7 and 12.3 percentage points over the two years following the election. We also report a series of secondary analyses that speak to associated salary increases and show evidence consistent with the notion that social capital and signaling play a role. Our findings help us understand the benefits of online community leadership, and they extend our understanding of the motivations for online community engagement.
This paper was accepted by Chris Forman, information systems.
Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2019.03252 .
This study exploits the correlation between Yellow River flooding and human capital accumulation using county-level data from Anhui, Hebei, Henan, Jiangsu, and Shandong provinces in China. Employing a spatial regression discontinuity approach, we compare the differences in human capital accumulation within and beyond the Yellow River floodplain areas (YRFA). Empirical results show human capital accumulation in the YRFA is at least 12.1 percent lower than outside the YRFA. Furthermore, our results demonstrate intergenerational transmission and gender differences in the negative impact of the Yellow River flooding on human capital accumulation. The baseline specifications of this study are not affected by drought and overall natural disasters. This paper documents historical human capital accumulation, clan power, and social trust, through which Yellow River flooding has a long-term impact.
We ask whether firms behave differently depending on the political party in charge, above and beyond responding to any actual differences in policy. We use the pollution abatement behaviour of U.S. Steam Electric Power Plants under the Clean Water Act as our case study. Exploiting the variation provided by the outcome of tightly contested gubernatorial elections, we provide causal evidence that large firms respond to the political ‘colour’ of the governor in the state they operate, even when neither the stringency nor the enforcement of the rules depend on it. Within a theoretical model of the interaction between the regulator and the regulated firms, we show that multiple equilibria arise, and the outcomes of the election provide an effective coordination device. This unexpected behaviour has real-world consequences and leads to significant differences in pollution levels.
Contemporary election reforms that are purported to increase or decrease turnout tend to have negligible effects on election outcomes. We offer an analytical framework to explain why. Contrary to heated political rhetoric, election policies have small effects on outcomes because they tend to target small shares of the electorate, have a small effect on turnout, and/or affect voters who are relatively balanced in their partisanship. After developing this framework, we address how the findings bear on minority voting rights. We then show that countermobilization from political parties cannot explain the small effects of election laws. We explain that even when a state passes multiple policies at the same time, the reforms will still only have a marginal effect on turnout and an ambiguous effect on who wins. Finally, we explain what policies should raise alarm about affecting outcomes.
The objective of this work is to estimate the effects of political turnover on municipal health indicators. For this purpose, we used a regression discontinuity design applied to a pool of data on close municipal elections that occurred in 2008 and 2012. The outcome variables tested were municipal health indicators observed 1 and 3 years after the elections. We assume that the change of parties in the leadership of the municipal executive power alters the managers and public health policies and, therefore, affects the indicators of this sector. Our results indicate that health indicators linked to administrative issues appeared to benefit from political turnover in the period immediately after the elections (1 year). These results were corroborated by our additional estimates (robustness). We highlight that the number of health workers and coverage of primary care remained better 3 years after the elections in municipalities that suffered political turnover, indicating that this improvement resulting from the change in management may not be transient. Finally, through averaging tests, we found that in municipalities where the incumbent lost the election and there was a turnover, health managers were more qualified, and there was greater creation of basic health units.
Housing policy is one of the most important areas of local politics. Yet little is known about how local legislatures and executives make housing policy decisions and how their elections shape policy in this important realm. We leverage housing policy data and a new data source of 15,621 city council elections and 3,261 mayoral elections in large cities in the United States and a regression discontinuity design to examine partisan divides in housing policy among the mass public as well as the impact of local leaders’ partisanship on housing policy. We provide robust evidence that electing mayors from different political parties shapes cities’ housing stock. Electing a Democrat as mayor leads to increased multifamily housing production. These effects are concentrated in cities where councils have less power over land use changes. Overall, our article shows that politics influences local housing policy, and it contributes to a larger literature on local political economy.
Political influence on bank credit allocation is often viewed as being necessary to address social problems like income inequality. We hypothesize that such influence elicits bank capital responses. Our hypothesis yields three testable predictions for which we find supporting evidence. First, when banks observe election outcomes that suggest greater impending political credit-allocation influence, they reduce capital to increase fragility and deter political influence. Second, banks subject to greater political influence nonetheless increase lending that politicians favor, and household consumption consequently increases. Third, these banks exhibit poorer post-lending performance. Our study has implications for the interaction between politics, household consumption, and bank risk through a specific channel—the interplay between credit-allocation regulation and bank capital structure.
This paper was accepted by Victoria Ivashina, finance.
Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.04056 .
Does leader age matter for the likelihood of interstate conflict? Many studies in biology, psychology, and physiology have found that aggression tends to decline with age throughout the adult lifespan, particularly in males. Moreover, a number of major international conflicts have been attributed to young leaders, including the conquests of Alexander the Great and the ambitious military campaigns of Napoleon. However, the exact nature of the relationship between leader age and international conflict has been difficult to study because of the endogeneity problem. Leaders do not come to power randomly. Rather, many domestic and international factors influence who becomes the leader of a country, and some of these factors could correlate with the chances of interstate conflict. For instance, wary democratic publics might favor older leaders when future international conflict seems likely, inducing a relationship between older leaders and interstate conflict. This article overcomes such confounding by using a regression discontinuity design. Specifically, it looks at close elections of national leaders who had large differences in age. It finds that when older candidates barely defeated younger ones, countries were much less likely to engage in military conflict. Its sample is also fairly representative of democracies more broadly, meaning that the findings likely hold true for cases outside the sample. The results demonstrate the important role that individuals play in shaping world politics. They also illustrate the value of design-based inference for learning about important questions in the study of international relations and peace science.
Candidates' placements in polls and past elections can be powerful coordination devices for parties and voters. Using a regression discontinuity design in French two-round elections, we show that candidates who place first in the first round are more likely to stay in the race and win than those who placed second. These effects are even larger for ranking second versus third, and also present for third versus fourth. They stem from allied parties agreeing on which candidate should drop out, voters coordinating their choice, and the bandwagon effect of wanting to vote for the winner. We find similar results across 19 other countries. (JEL D72, D83, D91, K16)
Allocation decisions are vulnerable to political influence, but it is unclear in which situations politicians use their discretionary power in a partisan manner. We analyze the allocation of presidential disaster declarations in the United States, exploiting the spatiotemporal randomness of all hurricane strikes from 1965-2018 along with changes in political alignment. We show that decisions are unbiased when disasters are either very strong or weak. Only after medium-intensity hurricanes do areas governed by presidents' co-partisans receive up to twice as many declarations. This hump-shaped political bias explains 8.3 percent of overall relief spending, totaling about USD 400 million per year.
Previous research finds a positive association between electoral fraud and post-election protests, violence, and civil conflict. This article contends that the effect of electoral fraud on peace can be heterogeneous. I investigate elections after civil wars that stalemated. My contribution is to present a theory and suggestive evidence that, in this context, electoral fraud by unpopular incumbents can be peace-promoting. An important policy implication is that international efforts to prevent electoral fraud (e.g. electoral observers), in this specific setting, may increase the odds that a civil war continues.
Existing theories offer conflicting perspectives regarding the relationship between career specialization and labor market outcomes. While some scholars argue it is better for workers to specialize and focus on one area, others argue it is advantageous for workers to diversify and compile experience across multiple work domains. We attempt to reconcile these competing perspectives by developing a theory highlighting the voluntary versus involuntary nature of worker–firm separations as a theoretical contingency that alters the relative advantages and disadvantages associated with specialized versus generalized careers. Our theory is rooted in the notion that the characteristics of involuntary worker–firm separations (i.e., job displacement) simultaneously amplify the disadvantages associated with specialized careers and the advantages associated with generalized careers, thereby giving displaced generalists a relative advantage over displaced specialists. We find support for our theory in the context of U.S. congressional staffing, using administrative employment records and a regression discontinuity identification strategy that exploits quasi-random staffer displacement resulting from narrowly decided congressional reelection bids. Our theoretical contingency is further supported in supplemental regressions where correlational evidence suggests that while specialists tend to be relatively penalized in the labor market after involuntary separations, specialists appear to be relatively privileged when separations are plausibly voluntary.
Abstract We apply a regression discontinuity approach to determine incumbency advantages in the Canadian Parliament, finding that incumbents enjoy a 9.4–11.2% increased probability of winning over non-incumbents. Owing to the presence of multiple parties, an incumbency advantage in terms of vote share does not always translate to an increased probability of winning, because incumbents do not necessarily obtain votes from their closest opponent. Also, under the assumption that strategic exit is not an issue, we are able to split the incumbency advantage into party incumbency and individual candidate incumbency components, finding that the advantage is almost entirely due to the individual.
On utilise une approche par la méthode de régression avec discontinuité pour déterminer les avantages qu’il y a àêtre l’élu en place dans les élections au Parlement du Canada. On découvre que les élus en place ont une probabilité de gagner de 9.4 à 11.2% de plus que leurs adversaires. A cause de la présence de multiples partis politiques, l’avantage de l’élu en place en termes de proportion du vote ne se traduit pas toujours par une probabilité accrue de gagner parce que les élus en place ne capturent pas nécessairement les votes de leur plus proche adversaire. De plus, en présumant qu’il ne s’agit pas de sortie stratégique, on peut départager l’avantage de l’élu en place en composantes correspondant à l’effet de parti et l’effet de l’individu. On découvre que l’avantage est presque complètement attribuable à l’individu.
Utilizing data on criminal charges lodged against candidates to the Fourteenth and Fifteenth Lok Sabha, India’s lower house of representatives, we study the conditions that resulted in approximately a quarter of members of parliament elected in 2004 and in 2009 facing or having previously faced criminal charges. Our results document that Indian political parties are more likely to select alleged criminal candidates when confronting greater electoral uncertainty and in parliamentary constituencies whose populations exhibit lower levels of literacy. We interpret the decisions of political parties to enlist known criminals as candidates as a function of the capacity of these candidates to intimidate voters. To substantiate this, we show that criminal candidates depress electoral turnout. In addition, our results suggest that India’s well-known incumbency disadvantage stems from the superior electoral performance of allegedly criminal candidates, who drive parliamentary incumbents from office. Our study raises questions for democratic theory, which claims that electoral competition improves accountability, and for the future of the Indian polity, which is experiencing a growing criminalization of the national political arena.
This study presents a method of testing casual hypotheses, called regression-discontinuity analysis, in situations where the investigator is unable to randomly assign Ss to experimental and control groups. The Ss were selected from near winners—5126 students who received certificates of merit and 2848 students who merely received letters of commendation. Comparison of the results obtained from the new mode of analysis with those obtained when the ex post facto design was applied to the same data. The new analysis suggested that public recognition for achievement tends to increase the likelihood that the recipient will receive a scholarship but did not support the inference that recognition affects the student's attitudes and career plans. From Psyc Abstracts 36:01:1AF09T. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
This paper applies the forecasting models of Klarner and Buchanan (2006a) for the U.S. Senate and Klarner and Buchanan (2006b) for the U.S. House of Representatives to the upcoming 2008 elections. Forecasts are also conducted for the 2008 presidential race at the state level. The forecasts presented in this article, made July 28, 2008 (99 days before the election), predicted an 11-seat gain for the Democrats in the House of Representatives, a three-seat gain for the Democrats in the Senate, and that Barack Obama would obtain 53.0% of the popular vote and 346 electoral votes. Furthermore, Obama was forecast to have an 83.6% chance of winning the White House and an 85.9% chance of winning the popular vote.
This article reviews research on electoral fraud—clandestine and illegal efforts to shape election results. Only a handful of works classify reports on electoral fraud to identify its nature, magnitude, and causes. This review therefore looks at the larger number of historical works (as well as some ethnographies and surveys) that discuss ballot rigging. Its conclusions are threefold. First, fraud takes on a panoply of forms; it ranges from procedural violations of electoral law (that may or may not intend to distort results) to the outright use of violence against voters. Second, even when ballot rigging is an integral part of electoral competition, it is infrequently decisive. Fraud, nevertheless, undermines political stability because, in close races, it can be crucial. Third, political competition shapes the rhythm and nature of electoral fraud. Efforts to steal elections increase with inequality, but competitiveness—which institutions help to shape—determines the ballot-rigging strategies parties adopt.
Many recent studies show that firms profit from connections to influential politicians, but less is known about how much politicians financially benefit from wielding political influence. We estimate the returns to serving in Parliament using original data on the estates of recently deceased British politicians. Applying both matching and a regression discontinuity design to compare MPs with parliamentary candidates who narrowly lost, we find that serving in office almost doubled the wealth of Conservative MPs but had no discernible financial benefits for Labour MPs. Conservative MPs profited from office largely through lucrative outside employment they acquired as a result of their political positions; we show that gaining a seat in Parliament more than tripled the probability that a Conservative politician would later serve as a director of a publicly-traded firm – enough to account for a sizable portion of the wealth differential. We suggest that Labour MPs did not profit from office largely because trade unions collectively exerted sufficient control over the party and its MPs to prevent members from selling their services to other clients.
We reconsider the effect of very low birth weight classification on infant mortality. We demonstrate that the estimates are highly sensitive to the exclusion of observations in the immediate vicinity of the 1,500-g threshold, weakening the confidence in the results originally reported in Almond, Doyle, Kowalski, and Williams (2010).
Despite extensive public infrastructure spending, surprisingly little is known about its economic return. In this paper, we
estimate the value of school facility investments using housing markets: standard models of local public goods imply that
school districts should spend up to the point where marginal increases would have zero effect on local housing prices. Our
research design isolates exogenous variation in investments by comparing school districts where referenda on bond issues targeted
to fund capital expenditures passed and failed by narrow margins. We extend this traditional regression discontinuity approach
to identify the dynamic treatment effects of bond authorization on local housing prices, student achievement, and district
composition. Our results indicate that California school districts underinvest in school facilities: passing a referendum
causes immediate, sizable increases in home prices, implying a willingness to pay on the part of marginal homebuyers of $1.50
or more for each $1 of capital spending. These effects do not appear to be driven by changes in the income or racial composition
of homeowners, and the impact on test scores appears to explain only a small portion of the total housing price effect.
This paper studies gender interactions within hierarchical organizations using a large data set on the duration of Italian
municipal governments elected between 1993 and 2003. A municipal government can be viewed as a hierarchy, whose stability
over time depends on the degree of cooperation between and within ranks. We find that in municipalities headed by female mayors,
the probability of early termination of the legislature is higher. This result persists and becomes stronger when we control
for municipality fixed effects as well as for non-random sorting of women into municipalities using regression discontinuity
in gender-mixed electoral races decided by a narrow margin. The likelihood that a female mayor survives until the end of her
term is lowest when the council is entirely male and in regions with less favourable attitudes towards working women. This
evidence is suggestive that group dynamics are an important factor in driving the gender difference. Other interpretations
receive less support in the data. Our results may provide an alternative explanation for the underrepresentation of women
in leadership positions.
Economic impacts of unionization on employers are difficult to estimate in the absence of large, representative data on establishments
with union status information. Estimates are also confounded by selection bias, because unions could organize at highly profitable
enterprises that are more likely to grow and pay higher wages. Using multiple establishment-level data sets that represent
establishments that faced organizing drives in the United States during 1984–1999, this paper uses a regression discontinuity
design to estimate the impact of unionization on business survival, employment, output, productivity, and wages. Essentially,
outcomes for employers where unions barely won the election (e.g., by one vote) are compared with those where the unions barely lost. The analysis finds small impacts on all outcomes that we examine; estimates for wages are close to zero. The evidence
suggests that—at least in recent decades—the legal mandate that requires the employer to bargain with a certified union has
had little economic impact on employers, because unions have been somewhat unsuccessful at securing significant wage gains.
Are cities as politically polarized as states and countries? “No” is the answer from our regression discontinuity design analysis,
which shows that whether the mayor is a Democrat or a Republican does not affect the size of city government, the allocation
of local public spending, or crime rates. However, there is a substantial incumbent effect for mayors. We investigate three
mechanisms that could account for the striking lack of partisan impact at the local level, and find the most support for Tiebout
competition among localities within metropolitan areas.
This dissertation presents a comparative analysis of incumbency advantage across developed democracies. In particular, it examines two variants of incumbency advantage: 1) the extra electoral benefits that political parties gain from fielding incumbent candidates (incumbency advantage for political parties) and 2) the electoral advantage that individual incumbents enjoy over non-incumbent candidates of the same party (incumbency advantage for individual candidates).
For each type of incumbency advantage, this dissertation offers three distinctive contributions. First, it provides comparable estimates of both types of incumbency advantage across different electoral systems. The existing literature lacks appropriate estimates of either type of incumbency advantage that are comparable across different electoral systems. This dissertation furnishes fully comparable estimates ??? the first of its kind ??? and makes it possible to conduct a systematic comparative analysis of incumbency advantage.
Second, it develops a theory of electoral systems' impact on the magnitude of incumbency advantage. This theory is partly based on the theory of personal-vote incentives, since the personal vote is one of the critical sources of incumbency advantage. However, the theory developed in this dissertation highlights an important departure from the personal-vote theory, because the personal-vote incentives do not always translate into actual electoral gains from personal-vote building activities. This new theory of comparative incumbency advantage advances our knowledge of the consequences of electoral systems and illuminates the important distinction between the personal-vote incentives and the actual electoral gains.
Third, it provides an elaborate multiple-country empirical analysis, based on the newly compiled dataset of district- and candidate-level election results in ten developed democracies (Austria, Belgium, Canada, Finland, Ireland, Italy, Japan, New Zealand, Norway, and the United Kingdom). This empirical analysis constitutes by far the most extensive cross-national analysis of incumbency advantage based on the detailed aggregate election data.
The dissertation also presents a few significant implications for the relationship between electoral systems and accountability. In particular, the clear distinction made between incumbency advantage for political parties and incumbency advantage for individual candidates makes it possible to derive specific implications for the collective accountability of parties and the individual accountability of legislators, respectively.
Drug trade-related violence has escalated dramatically in Mexico since 2007, and recent years have also witnessed large-scale efforts to combat trafficking, spearheaded by Mexico's conservative PAN party. This study examines the direct and spillover effects of Mexican policy toward the drug trade. Regression discontinuity estimates show that drug-related violence increases substantially after close elections of PAN mayors. Empirical evidence suggests that the violence reflects rival traffickers' attempts to usurp territories after crackdowns have weakened incumbent criminals. Moreover, the study uses a network model of trafficking routes to show that PAN victories divert drug traffic, increasing violence along alternative drug routes.
Why do parties and governments cheat in elections they cannot lose? This book documents the widespread use of blatant and excessive manipulation of elections and explains what drives this practice. Alberto Simpser shows that, in many instances, elections are about more than winning. Electoral manipulation is not only a tool used to gain votes, but also a means of transmitting or distorting information. This manipulation conveys an image of strength, shaping the behavior of citizens, bureaucrats, politicians, parties, unions, and businesspeople to the benefit of the manipulators, increasing the scope for the manipulators to pursue their goals while in government and mitigating future challenges to their hold on power. Why Governments and Parties Manipulate Elections provides a general theory about what drives electoral manipulation and empirically documents global patterns of manipulation.
Monte Carlo methods are used to study the efficacy of multivariate matched sampling and regression adjustment for controlling bias due to specific matching variables when dependent variables are moderately nonlinear in . The general conclusion is that nearest available Mahalanobis metric matching in combination with regression adjustment on matched pair differences is a highly effective plan for controlling bias due to .
Following David Lee's pioneering work, numerous scholars have applied the regression discontinuity (RD) design to popular elections. Contrary to the assumptions of RD, however, we show that bare winners and bare losers in U. S. House elections (1942-2008) differ markedly on pretreatment covariates. Bare winners possess large ex ante financial, experience, and incumbency advantages over their opponents and are usually the candidates predicted to win by Congressional Quarterly's pre-election ratings. Covariate imbalance actually worsens in the closest House elections. National partisan tides help explain these patterns.Previous works have missed this imbalance because they rely excessively on model-based extrapolation.We present evidence suggesting that sorting in close House elections is due mainly to activities on or before Election Day rather than postelection recounts or other manipulation. The sorting is so strong that it is impossible to achieve covariate balance between matched treated and control observations, making covariate adjustment a dubious enterprise. Although RD is problematic for postwar House elections, this example does highlight the design's advantages over alternatives. RD's assumptions are clear and weaker than model-based alternatives, and their implications are empirically testable.
Experimenters often use post-stratification to adjust estimates. Post-stratification is akin to blocking, except that the number of treated units in each stratum is a random variable because stratification occurs after treatment assignment. We analyse both post-stratification and blocking under the Neyman—Rubin model and compare the efficiency of these designs. We derive the variances for a post-stratified estimator and a simple difference-in-means estimator under different randomization schemes. Post-stratification is nearly as efficient as blocking: the difference in their variances is of the order of 1/n 2 , with a constant depending on treatment proportion. Post-stratification is therefore a reasonable alternative to blocking when blocking is not feasible. However, in finite samples, post-stratification can increase variance if the number of strata is large and the strata are poorly chosen. To examine why the estimators' variances are different, we extend our results by conditioning on the observed number of treated units in each stratum. Conditioning also provides more accurate variance estimates because it takes into account how close (or far) a realized random sample is from a comparable blocked experiment. We then show that the practical substance of our results remains under an infinite population sampling model. Finally, we provide an analysis of an actual experiment to illustrate our analytical results.
Research in political economy emphasizes the tendency of elites to persist and reproduce their power over time, potentially undermining the effectiveness of institutional reforms. One particular form of elite persistence is illustrated by the existence of political dynasties. A natural question is whether certain political reforms can break dynastic patterns and open up the political system. In this paper I study the extent to which the introduction of term limits by the 1987 Philippine Constitution effectively broke the hold of incumbent families on power. The ability of term limits to dismantle political dynasties is not obvious, as term-limited incumbents may be replaced by relatives or may run for a different elected office. Whether these strategies undermine the direct effects of term-limits in reducing the time an individual can hold office is an empirical question. I find no evidence of a statistically significant impact of term limits on curbing families' persistence in power. Moreover, term limits deter high-quality challengers from running prior to the expiration of an incumbent's term. Challengers prefer to wait for the incumbent to be termed-out and run in an open-seat race. As a consequence, incumbents are safer in their early terms prior to the limit. These results suggest that political reforms that do not modify the underlying sources of dynastic power may be ineffective in changing the political equilibrium.
This article studies gubernatorial midterm slumps in U.S. state legislative elections. We employ a regression discontinuity design, which allows us to rule out the hypothesis that the midterm slump simply reflects a type of reversion to the mean generated by simple partisan swings or the withdrawal of gubernatorial coattails or anticipatory balancing. Our results show that the party of the governor experiences an average seat-share loss of about 3.5 percentage points. We also find evidence suggesting that a large share of the variation in gubernatorial midterm slumps can be accounted for by (1) crude partisan balancing and (2) referendums on state economic performance, with approximately equal weight given to each.
The ability of matched sampling and linear regression adjustment to reduce the bias of an estimate of the treatment effect in two sample observational studies is investigated for a simple matching method and five simple estimates. Monte Carlo results are given for moderately linear exponential response surfaces and analytic results are presented for quadratic response surfaces. The conclusions are (1) in general both matched sampling and regression adjustment can be expected to reduce bias, (2) in some cases when the variance of the matching variable differs in the two populations both matching and regression adjustment can increase bias, (3) when the variance of the matching variable is the same in the two populations and the distributions of the matching variable are symmetric the usual covariance adjusted estimate based on random samples is almost unbiased, and (4) the combination of regression adjustment in matched samples generally produces the least biased estimate.
A fundamental tenent of representative democracy is that votes should be counted in a fair and unbiased manner. Contrary to this premise I …nd that in U.S. House elections this is not the case. I employ a novel approach to detect electoral manipulation by looking at extremely close elections where the outcome between any two candidates should be random. I …nd that in extremely close elections the incumbent wins markedly more often than would be expected. This suggests that the vote counting process is biased in favor of incumbents in a manner that is independent of the underlying incumbency advantage.
Using a non-parametric regression discontinuity design that compares candidates who barely win an election to those who barely lose, this paper estimates the effect of incumbency on a candidate's electoral prospects in India. Starting in 1991, I estimate that, rather than being at an advantage, incumbents are actually fourteen percent less likely to win an election than similar non-incumbents. This is a large disadvantage. To over-come it, an incumbent would have had to have won by an additional five and a half percent of the popular vote, a change equivalent to an incumbent moving from the first to the thirty-fifth percentile of elected officials ranked by margin of victory. This disadvantage contrasts with the advantage that incumbents enjoy in other countries as well as the advantage that Indian incumbents enjoyed prior to 1991. The disadvantage is also general with almost all incumbents faring equally poorly regardless of experience and party affiliation. While the available data prevent a formal test, the dominance of a single political party (the Indian National Congress) before 1991 may have provided a framework in which experience was valuable because incumbents who gained experience under the Congress system would interact with the same system when reelected. Starting in 1991, however, no party could be counted on to control parliament, making experience under the previous regime potentially less valuable.
This paper examines both theoretically and empirically whether the common practice of using OLS multivariate regression models to estimate average treatment effects (ATEs) under experimental designs is justified by the Neyman model for causal inference. Using data from eight large U.S. social policy experiments, we find that estimated standard errors and significance levels for ATE estimators are similar under the OLS and Neyman models when baseline covariates are included in the models, even though theory suggests that this may not have been the case. This occurs primarily because treatment effects do not appear to vary substantially across study subjects.
Elections between black and white candidates tend to involve close mar-gins and high turnout. Using a novel dataset of municipal vote returns during the rise of black mayors in U.S. cities, this paper establishes new facts about turnout and competition in close interracial elections. In the South, but not the North, close black victories were more likely than close black losses, involved higher turnout than close black losses, and were more likely than close black losses to be followed by subsequent black victories. These results are consistent with a model in which the historical exclusion of Southern blacks from politics made them disproportionately sensitive to mobilization efforts by political elites, leading to a black can-didate advantage in close elections. The results contribute to a growing body of evidence that the outcomes of reasonably close elections are not always random, which suggests that detailed knowledge of the electoral context is a precondition to regression discontinuity analyses based on vote shares.
This paper explores the interdependency of political institutions from the voter’s perspective. Specifically, we are interested in: Does the partisan identity of the mayor influence the voter’s decision in the subsequent town council election?; Does this partisan identity influence the vote in ensuing higher level elections?; and Do voters condition their vote for the mayor on the result of the last council election? We rely on a regression discontinuity design focusing on close election outcomes based on municipal level data for Germany. We find that the party of the mayor can receive a bonus of 4-6 percentage points in vote share in the next town council election (depending on the timing of the local elections). The mayor partisan identity does not affect federal or European election outcomes within the same municipality. And, we show that voters punish mayor candidates of parties that performed strongly in earlier council elections. Throughout the paper, we explore how the findings can be related to an incumbency externality effect and to the theory of voter preferences for divided government.
Direct influence over communication media is a potent resource during electoral campaigns, and politicians have an incentive to gain control of the airwaves to advance their careers. In this article, we use data on community radio license applications in Brazil to identify both the causal effect of incumbency on politicians’ ability to control the media and the causal effect of media control on their future electoral prospects. Using a regression discontinuity design, we compare city council candidates who barely won or barely lost an election, showing that incumbency more than doubles the probability of an application’s approval by the Ministry of Communications. Next, using genetic matching, we compare candidates who acquired community radio licenses before an election to similar politicians who did not, showing that a radio station substantially increases one’s vote share and probability of victory. These findings demonstrate that media control helps entrench local political power in Brazil.
Incumbents are highly likely to win reelection at all levels of government, but scholars continue to debate the extent to which serving in office has a causal effect on winning. For city council elections it is unclear whether or not we should predict a causal effect at all. City councilors may not regularly seek reelection, and any apparent advantage could be entirely attributable to preexisting qualities rather than incumbency. This article uses a regression discontinuity design to provide evidence that city council incumbents are more likely to run and win their next elections because they served a term in office.
U.S. cities are limited in their ability to set policy. Can these constraints mute the impact of mayors’ partisanship on policy outcomes? We hypothesize that mayoral partisanship will more strongly affect outcomes in policy areas where there is less shared authority between local, state, and federal governments. To test this hypothesis, we create a novel dataset combining U.S. mayoral election returns from 1990 to 2006 with city fiscal data. Using regression discontinuity design, we find that cities that elect a Democratic mayor spend a smaller share of their budget on public safety, a policy area where local discretion is high, than otherwise similar cities that elect a Republican or an Independent. We find no differences on tax policy, social policy, and other areas that are characterized by significant overlapping authority. These results suggest that models of national policymaking are only partially applicable to U.S. cities. They also have implications for political accountability: mayors may not be able to influence the full range of policies that are nominally local responsibilities.
This paper estimates the incumbency effects in elections to the House of Representatives of 45 states in the United States using a quasi-experimental research method, regression discontinuity design (RDD). This design isolates the causal effect of incumbency from other contemporaneous factors, such as candidate quality, by comparing incumbents and non-incumbents in close contests. I find that incumbents in state legislative elections have a significant advantage, and this advantage serves as a strong barrier to re-entry of challengers who had previously been defeated. However, the incumbency advantage estimated using the RDD is much smaller than are the estimates using existing methods, implying a significant selection bias in the latter.
The view that the returns to public educational investments are highest for early childhood interventions stems primarily from several influential randomized trials - Abecedarian, Perry, and the Early Training Project - that point to super-normal returns to preschool interventions. This paper implements a unified statistical framework to present a de novo analysis of these ex- periments, focusing on two core issues that have received little attention in previous analyses: treatment effect heterogeneity by gender and over-rejection of the null hypothesis due to mul- tiple inference. The primary finding of this reanalysis is that girls garnered substantial short- and long-term benefits from the interventions. However, there were no significant long-term benefits for boys. These conclusions would not be apparent when using "naive" estimators that do not adjust for multiple inference.
This paper uses a regression discontinuity design to compare the incumbency advantage enjoyed by freshmen and non-freshmen incumbents. The results show that compared to freshmen incumbents that barely won the last election, non-freshmen incumbents that barely won get 2.3 percentage points more in the next election. Further results suggest that the ability to deter high quality challengers is an important source of that advantage.
In regression discontinuity (RD) designs for evaluating causal effects of interventions, assignment to a treatment is determined at least partly by the value of an observed covariate lying on either side of a fixed threshold. These designs were first introduced in the evaluation literature by Thistlewaite and Campbell [1960. Regression-discontinuity analysis: an alternative to the ex-post Facto experiment. Journal of Educational Psychology 51, 309–317] With the exception of a few unpublished theoretical papers, these methods did not attract much attention in the economics literature until recently. Starting in the late 1990s, there has been a large number of studies in economics applying and extending RD methods. In this paper we review some of the practical and theoretical issues in implementation of RD methods.
Using panel data from US states over the period 1941–2002, I measure the impact of gubernatorial partisanship on a wide range of different policy settings and economic outcomes. Across 32 measures, there are surprisingly few differences in policy settings, social outcomes and economic outcomes under Democrat and Republican Governors. In terms of policies, Democratic Governors tend to prefer slightly higher minimum wages. Under Republican Governors, incarceration rates are higher, while welfare caseloads are higher under Democratic Governors. In terms of social and economic outcomes, Democratic Governors tend to preside over higher median post-tax income, lower post-tax inequality, and lower unemployment rates. However, for 26 of the 32 dependent variables, gubernatorial partisanship does not have a statistically significant impact on policy outcomes and social welfare. I find no evidence of gubernatorial partisan differences in tax rates, welfare generosity, the number of government employees or their salaries, state revenue, incarceration rates, execution rates, pre-tax incomes and inequality, crime rates, suicide rates, and test scores. These results are robust to the use of regression discontinuity estimation, to take account of the possibility of reverse causality. Overall, it seems that Governors behave in a fairly non-ideological manner.
In this paper we demonstrate empirically that incumbency is a source of spillover effects in Germany's mixed electoral system. Using a quasi-experimental research design that allows for causal inferences under a weaker set of assumptions than the regression models commonly used in the electoral systems literature, we find that incumbency causes a gain of 1.4–1.7 percentage points in PR vote shares. We also present simulations of Bundestag seat distributions to show that spillover effects caused by incumbency are sufficiently large to trigger significant shifts in parliamentary majorities.
This paper provides an introduction and "user guide" to Regression Discontinuity (RD) designs for empirical researchers. It presents the basic theory behind the research design, details when RD is likely to be valid or invalid given economic incentives, explains why it is considered a "quasi-experimental" design, and summarizes different ways (with their advantages and disadvantages) of estimating RD designs and the limitations of interpreting these estimates. Concepts are discussed using examples drawn from the growing body of empirical research using RD. ( JEL C21, C31)
The ability of matched sampling and linear regression adjustment to reduce the bias of an estimate of the treatment effect in two sample observational studies is investigated for a simple matching method and five simple estimates. Monte Carlo results are given for moderately linear exponential response surfaces and analytic results are presented for quadratic response surfaces. The conclusions are (1) in general both matched sampling and regression adjustment can be expected to reduce bias, (2) in some cases when the variance of the matching variable differs in the two populations both matching and regression adjustment can increase bias, (3) when the variance of the matching variable is the same in the two populations and the distributions of the matching variable are symmetric the usual covariance adjusted estimate based on random samples is almost unbiased, and (4) the combination of regression adjustment in matched samples generally produces the least biased estimate.
Although the presidential coattail effect has been an object of frequent study, the question of whether popular congressional
candidates boost vote shares in return for their parties’ presidential candidates remains unexplored. This article investigates
whether so-called “reverse coattails” exist using a regression discontinuity design with congressional district-level data
from presidential elections between 1952 and 2004. Taking incumbency to be near-randomly distributed in cases where congressional
candidates have just won or lost their previous elections, I find that the numerous substantial advantages of congressional
incumbency have no effect on presidential returns for these incumbents’ parties. This null finding underscores my claim that
the existing coattail literature deserves greater scrutiny. My results also prompt a rethinking of the nature of the advantages
that incumbents bring to their campaigns and may help deepen our understanding of partisanship in the United States.
There are two fundamentally different views of the role of elections in policy formation. In one view, voters can affect candidates' policy choices: competition for votes induces politicians to move toward the center. In this view, elections
have the effect of bringing about some degree of policy compromise. In the alternative view, voters merely elect policies: politicians cannot make credible promises to moderate their policies, and elections are merely a means to decide
which one of two opposing policy views will be implemented. We assess which of these contrasting perspectives is more empirically
relevant for the U. S. House. Focusing on elections decided by a narrow margin allows us to generate quasi-experimental estimates
of the impact of a “randomized” change in electoral strength on subsequent representatives' roll-call voting records. We find
that voters merely elect policies: the degree of electoral strength has no effect on a legislator's voting behavior. For example, a large exogenous increase in electoral strength for the Democratic party in a district does not result in shifting both parties' nominees
to the left. Politicians' inability to credibly commit to a compromise appears to dominate any competition-induced convergence
in policy.
A long-standing issue in political economics is to what extent party control makes a difference in determining fiscal and economics policies. This question is very difficult to answer empirically because parties are not randomly selected to govern political entities. This article uses a regression-discontinuity design, namely, party control changes discontinuously at 50% of the vote share, which can produce "near" experimental causal estimates of the effect of party control on economic outcomes. The method is applied to a large panel data set from Swedish local governments with a number of attractive features. The results show that there is an economically significant party effect: Left-wing governments spend and tax 2-3% more than right-wing governments. Left-wing governments also have 7% lower unemployment rates, which is partly due to that left-wing governments employ 4% more workers than right-wing governments. (JEL: C21, D72, D78, H71, H72) (c) 2008 by the European Economic Association.
Political dynasties have long been present in democracies, raising concerns that inequality in the distribution of political
power may reflect imperfections in democratic representation. However, the persistence of political elites may simply reflect
differences in ability or political vocation across families and not their entrenchment in power. We show that dynastic prevalence
in the Congress of the U.S. is high compared to that in other occupations and that political dynasties do not merely reflect
permanent differences in family characteristics. On the contrary, using two instrumental variable techniques we find that
political power is self-perpetuating: legislators who hold power for longer become more likely to have relatives entering
Congress in the future. Thus, in politics, power begets power.
This paper estimates the effect of a candidate’s incumbency status on his or her chances of winning using a large dataset on state legislative elections in India during 1975-2003. I use an innovative research design, called Regression Discontinuity Design (RDD), that provides unbiased estimate of the effect due to incumbency by comparing the candidates in closely fought elections, and find that incumbency has a significant negative effect on the fortunes of incumbent candidates in India and the incumbency effect has decreased further in the last decade. Also, the variation in the incumbency effects across Indian states depends on the differences in levels of public good provision such as the health facilities, rates of employment and poverty, and state per capita income.
Patients who receive more hospital treatment tend to have worse underlying health, confounding estimates of the returns to such care. This paper compares the costs and benefits of extending the length of hospital stay following delivery using a discontinuity in stay length for infants born close to midnight. Third-party reimbursement rules in California entitle newborns to a minimum number of hospital "days," counted as the number of midnights in care. A newborn delivered at 12:05 a.m. will have an extra night of reimbursable care compared to an infant born minutes earlier. We use a dataset of all California births from 1991-2002, including nearly 100,000 births within 20 minutes of midnight, and find that children born just prior to midnight have significantly shorter lengths of stay than those born just after midnight, despite similar observable characteristics. Furthermore, a law change in 1997 entitled newborns to a minimum of 2 days in care. The midnight discontinuity can therefore be used to consider two distinct treatments: increasing stay length from one to two nights (prior to the law change) and from two to three nights (following the law change). On both margins, we find no effect of stay length on readmissions or mortality for either the infant or the mother, and the estimates are precise. The results suggest that for uncomplicated births, longer hospitals stays incur substantial costs without apparent health benefits.
Ž. THE REGRESSION DISCONTINUITY RD data design is a quasi-experimental design with the defining characteristic that the probability of receiving treatment changes discontinuously as a function of one or more underlying variables. This data design arises frequently in economic and other applications but is only infrequently exploited as a source of identifying information in evaluating effects of a treatment. In the first application and discussion of the RD method, Thistlethwaite and Campbell Ž. 1960 study the effect of student scholarships on career aspirations, using the fact that awards are only made if a test score exceeds a threshold. More recently, Van der Klaauw Ž. 1997 estimates the effect of financial aid offers on students’ decisions to attend a particular college, taking into account administrative rules that set the aid amount partly on the basis of a discontinuous function of the students’ grade point average and SAT Ž. score. Angrist and Lavy 1999 estimate the effect of class size on student test scores, taking advantage of a rule stipulating that another classroom be added when the average Ž. class size exceeds a threshold level. Finally, Black 1999 uses an RD approach to estimate parents’ willingness to pay for higher quality schools by comparing housing prices near geographic school attendance boundaries. Regression discontinuity methods have potentially broad applicability in economic research, because geographic boundaries or rules governing programs often create discontinuities in the treatment assignment mechanism that can be exploited under the method. Although there have been several discussions and applications of RD methods in the literature, important questions still remain concerning sources of identification and ways of estimating treatment effects under minimal parametric restrictions. Here, we show that identifying conditions invoked in previous applications of RD methods are often overly strong and that treatment effects can be nonparametrically identified under an RD design by a weak functional form restriction. The restriction is unusual in that it requires imposing continuity assumptions in order to take advantage of the known discontinuity in the treatment assignment mechanism. We also propose a way of nonparametrically estimating treatment effects and offer an interpretation of the Wald estimator as an RD estimator.