In heterogeneous treatment effect models with endogeneity, the identification of the local average treatment effect (LATE) typically relies on an instrument that satisfies two conditions: (i) joint independence of the potential post-instrument variables and the instrument and (ii) monotonicity of the treatment in the instrument. We show that identification is still feasible when replacing monotonicity by a strictly weaker local monotonicity condition (given marginal potential outcomes). We demonstrate that the latter allows identifying the LATEs on the (i) compliers (whose treatment reacts to the instrument in the intended way), (ii) defiers (who react counter-intuitively), and (iii) both populations jointly. Furthermore, (i) and (iii) coincides with standard LATE if monotonicity holds. We also present an application to the quarter of birth instrument.
We investigate the finite sample properties of a large number of estimators for the average treatment effect on the treated that are suitable when adjustment for observed covariates is required, like inverse probability weighting, kernel and other variants of matching, as well as different parametric models. The simulation design used is based on real data usually employed for the evaluation of labour market programmes in Germany. We vary several dimensions of the design that are of practical importance, like sample size, the type of the outcome variable, and aspects of the selection process. We find that trimming individual observations with too much weight as well as the choice of tuning parameters are important for all estimators. A conclusion from our simulations is that a particular radius matching estimator combined with regression performs best overall, in particular when robustness to misspecifications of the propensity score and different types of outcome variables is considered an important property.
In this paper, we assess the impact of firms introducing part-time work schemes for gradual labour market exit of elderly workers on their employees' labour market outcomes. The analysis is based on unique linked employer-employee data that combine high-quality survey and administrative data. Our results suggest that partial or gradual retirement options offered by firms are an important tool to alleviate the negative effects of low labour market attachment of elderly workers in ageing societies.
Using exceptionally rich linked administrative and survey information on German welfare recipients we investigate the health effects of transitions from welfare to employment and of assignments to welfare-to-work programmes. Applying semi-parametric propensity score matching estimators we find that employment substantially increases (mental) health. The positive effects are mainly driven by males and individuals with bad initial health conditions and are largest for males with poor health. In contrast, the effects of welfare-to-work programmes, including subsidized jobs, are ambiguous and statistically insignificant for most outcomes. Robustness checks that include a semi-parametric instrumental variable approach do not provide reasons for concern.
During the last decade, many Western economies reformed their welfare system with the aim of activating welfare recipients by increasing welfare-to-work programmes (WTWP) and job-search enforcement. We evaluate the short-term effects of three important German WTWP implemented after a major reform in January 2005
(‘Hartz IV'), namely short training, further training with a planned duration of up to three months and public workfare programmes (‘One-Euro-Jobs'). Our analysis is based on a combination of large-scale survey and administrative data that is rich with respect to
individual, household, agency level and regional information. We use this richness of the data to base the econometric evaluation on a selection-on-observables approach. We find that short-term training programmes on average increase their participants' employment
perspectives. There is also considerable effect heterogeneity across different subgroups of participants that could be exploited to improve the allocation of welfare recipients to the specific programmes and thus increase overall programme effectiveness.
Using a simulation design that is based on empirical data, a recent study by Huber, Lechner and Wunsch (2012) finds that distance-weighted radius matching with bias adjustment as proposed in Lechner, Miquel and Wunsch (2011) is competitive among a broad range of propensity score-based estimators used to correct for mean differences due to observable covariates. In this paper, we further investigate the finite sample behaviour of radius matching with respect to various tuning parameters. The results are intended to help the practitioner to choose suitable values of these parame¬ters when using this method, which has been implemented as "radiusmatch" command in the software packages GAUSS, STATA and the R package "radiusmatching".
In the presence of an endogenous binary treatment and a valid binary instrument, causal effects are point identified only for the subpopulation of compliers, given that the treatment is monotone in the instrument. With the exception of the entire population, causal inference for further subpopulations has been widely ignored in econometrics. We invoke treatment monotonicity and/or dominance assumptions to derive sharp bounds on the average treatment effects on the treated, as well as on other groups. Furthermore, we use our methods to assess the educational impact of a school voucher program in Colombia and discuss testable implications of our assumptions.
This papers proposes a simple method for testing whether non-compliance in experiments is ignorable, i.e., not jointly related to the treatment and the outcome. The approach consists of (i) regressing the outcome variable on a constant, the treatment, the assignment indicator, and the treatment/assignment interaction and (ii) testing whether the coefficients on the latter two variables are jointly equal to zero. A brief simulation study illustrates the finite sample properties of the test.
Sample selection and attrition are inherent in a range of treatment evaluation problems such as the estimation of the returns to schooling or training. Conventional estimators tackling selection bias typically rely on restrictive functional form assumptions that are unlikely to hold in reality. This paper shows identification of average and quantile treatment effects in the presence of the double selection problem (i) into a selective subpopulation (e.g., working - selection on unobservables) and (ii) into a binary treatment (e.g., training - selection on observables) based on weighting observations by the inverse of a nested propensity score that characterizes either selection probability. Root-n-consistent weighting estimators based on parametric propensity score models are applied to female labor market data to estimate the returns to education.
We test the validity of the sibling sex ratio instrument in Angrist and Evans (1998) using the methods proposed by Kitagawa (2008) and Huber and Mellace (2011). The sex ratio of the first two siblings is arguably randomly assigned and influences the probability of having a third child, which makes it a candidate instrument for fertility when estimating the effect of fertility on female labor supply. However, identification hinges on the random assignment of the instrument, an instrumental exclusion restriction, and the monotonicity of fertility in the instrument, see Imbens and Angrist (1994). We find that the instrumental variable tests of Kitagawa (2008) and Huber and Mellace (2011) do not point to a violation of these assumptions in the Angrist and Evans (1998) data (which can, however, not be ruled out even asymptotically as the tests cannot detect all possible violations).
