Consistency of propensity score matching estimators hinges on the propensity score's ability to balance the distributions of covariates in the pools of treated and nontreated units. Conventional balance tests merely check for differences in covariates' means, but cannot account for differences in higher moments. For this reason, this paper proposes balance tests which test for differences in the entire distributions of continuous covariates based on quantile regression (to derive Kolmogorov-Smirnov and Cramer-von-Mises-Smirnov-type test statistics) and resampling methods (for inference). Simulations suggest that these methods are very powerful and capture imbalances related to higher moments when conventional balance tests fail to do so
The 2005 reform of the German welfare system introduced two competing organizational models for welfare administration. In most districts, a centralized organization was established where local welfare agencies are bound to central directives. At the same time, 69 districts were allowed to opt for a decentralized organization. We evaluate the relative success of both types in terms of integrating welfare recipients into employment. Com-pared to centralized organization, decentralized organization has a negative effect on employment chances of males. For women, no significant effect is found. These findings are robust to aspects of internal organization common to both types of agencies.
The nonparametric identification of the local average treatment effect (LATE) hinges on the satisfaction of three instrumental variable assumptions: (1) unconfounded assignment of the instrument, (2) no average direct effect of the instrument on the outcome within compliance types (exclusion restriction), and (3) weak monotonicity of the treatment in the instrument. While (1) often appears plausible in experiments when using randomization as instrument for actual participation, (2) and (3) may be controversial. For this reason, this paper proposes easily implementable sensitivity checks to assess the robustness of the LATE to deviations from either the exclusion restriction or monotonicity. An empirical illustration based on female labor supply data is also provided.
This paper demonstrates the identification of causal mechanisms of a binary treatment under selection on observables, (primarily) based on inverse probability weighting. I.e., we consider the average indirect effect of the treatment, which operates through an intermediate variable (or mediator) that is situated on the causal path between the treatment and the outcome, as well as the (unmediated) direct effect. Even under random treatment assignment, subsequent selection into the mediator is generally non-random such that causal mechanisms are only identified when controlling for confounders of the mediator and the outcome. To tackle this issue, units are weighted by the inverse of their conditional treatment propensity given the mediator and observed confounders. We show that the form and applicability of weighting depend on whether some confounders are themselves influenced by the treatment or not. A simulation study gives the intuition for these results and an empirical application to the direct and indirect health effects (through employment) of the U.S. Job Corps program is also provided.
We study whether the number of signatures collected to qualify a popular initiative affects the probability of reforming the status quo.
The initiative process is modeled as a sequential game under uncertainty: petitioners make an entry decision and collect signatures to qualify the initiative. Politicians decide about a political compromise - a counter proposal - after which petitioners have the option to withdraw the initiative before the vote.
In equilibrium, politicians infer the initiative's popularity from the number of signatures and collection time. The more the initiative is perceived as a threat to the status quo, the more likely politicians come up with a counter proposal. Under certain conditions, petitioners have the incentive to collect more signatures than required for qualification to demonstrate high success probability.
We test model predictions using the data set of all Swiss constitutional initiatives at the federal level between 1891 and 2010. Overall, we find supporting evidence for the model mechanisms. Fast signature collection is associated with a higher probability of reform. The effect is mediated through a higher probability of provoking a counter proposal. Ultimately, counter proposals are key to amending the status quo. Restricting the signature collection time reduces the informative mechanism of the signature collection process considerably.
In 2014 the Swiss Upper House introduced an electronic voting system, which would make it easier to monitor the voting behavior of its legislators. In this system, individual decisions on specific exogenously defined vote types are published automatically, while all other votes are not publicly disclosed. The present paper uses this institutional change to determine, in a quasi-experimental setting, whether the monitoring of parliamentary voting influences legislators? incentives to participate in floor votes. In addition, video recordings of all sessions are used to determine pre- and post-reform attendance rates during secret votes.
Attendance rates increase once legislators are subject to monitoring. This result cannot be explained by anticipation effects of the reform, the introduction of an electronic voting system, or election cycles. Attendance rises more among legislators who depend more heavily on their political career (full-time politicians, those with few interest groups, and incumbents running for re-election) than among their peers with better outside career options. Moreover, when voting is monitored, legislators abstain less and vote more often in line with their party majority.
This paper studies the electoral connection in the domain of cultural policy spending by analyzing direct democratic referendums on financing of the performing arts. Previous empirical investigations into voter preferences for the arts are based on the (strong) assumption of full voter turnout. I show theoretically and empirically that capturing the turnout decision leads to a different interpretation of the regression coefficients.
By using a 2009 referendum in the Swiss canton St.Gallen to move fiscal responsibility of the local theater from the city to the canton, I show that accounting for the actual turnout decision yields quantitatively different results than previous estimates. In particular, I find that coefficients of variables simultaneously enhancing (or simultaneously deteriorating) turnout and preferences for the arts, e.g., income or the share of the old, are over-estimated. However, my analysis supports the view that socio-economic variables and preferences for the arts are highly correlated.
Analyzing post-ballot surveys of federal votes which took place on the same day as the St.Gallen referendum, reveals that voters were indeed on average richer, better educated and older than the total population. While politicians are typically thought of as spending more for the arts than citizens desire, my analysis shows that voters likely have stronger preferences for the arts than the average citizen.
As any empirical method used for causal analysis, social experiments are prone to attrition which may flaw the validity of the results. This paper considers the problem of partially missing outcomes in experiments. Firstly, it systematically reveals under which forms of attrition - in terms of its relation to observable and/or unobservable factors - experiments do (not) yield causal parameters. Secondly, it shows how the various forms of attrition can be controlled for by different methods of inverse probability weighting (IPW) that are tailored to the specific missing data problem at hand. In particular, it discusses IPW methods that incorporate instrumental variables when attrition is related to unobservables, which has been widely ignored in the experimental literature before
The decomposition of gender or ethnic wage gaps into explained and unexplained components (often with the aim to assess labor market discrimination) has been a major research agenda in empirical labor economics. This paper demonstrates that conventional decompositions, no matter whether linear or non-parametric, are equivalent to assuming a (probably too) simplistic model of mediation (aimed at assessing causal mechanisms) and may therefore lack causal interpretability. The reason is that decompositions typically control for post-birth variables that lie on the causal pathway from gender/ethnicity (which are determined at or even before birth) to wage but neglect potential endogeneity that may arise from this approach. Based on the newer literature on mediation analysis, we therefore provide more attractive identifying assumptions and discuss non-parametric identification based on reweighting.
This study tests the two assumptions underlying popularly held notions that
maternal employment negatively affects children because it reduces time spent with parents: (1) that maternal employment reduces children's time with parents, and (2) that time with parents affects child outcomes. We analyze children's time-diary data from the Child Development Supplement of the Panel Study of Income Dynamics and use child fixed-effects and IV estimations to account for unobserved heterogeneity. We find that working mothers trade quantity of time for better "quality" of time. On average, maternal work has no effect on time in activities that positively influence children's development, but it reduces time in types of activities thatmay be detrimental to children's development. Stratification by mothers' education reveals that although all children, regardless of mother's education, benefit from spending educational and structured time with their mothers, mothers who are high school graduates have the greatest difficulty balancing work and childcare.We find some evidence that fathers compensate for maternal employment by increasing types of activities that can foster child development as well as types of activities thatmay be detrimental.Overall,we find that the effects ofmaternal employment are ambiguous because (1) employment does not necessarily reduce children's time with parents, and (2) not all types of parental time benefit child development.
