Université de Fribourg

The Finite Sample Performance of Estimators for Mediation Analysis Under Sequential Conditional Independence

Huber, Martin ; Lechner, Michael ; Mellace, Giovanni

In: Journal of Business & Economic Statistics, 2016, vol. 34, no. 1, p. 139-160

Using a comprehensive simulation study based on empirical data, this article investigates the finite sample properties of different classes of parametric and semiparametric stimators of (natural) direct and indirect causal effects used in mediation analysis under sequential conditional independence assumptions. The estimators are based on regression, inverse probability weighting, and...

Université de Fribourg

It's never too LATE : A new look at local average treatment effects with or without defiers

Dahl, Christian M. ; Huber, Martin ; Mellace, Giovanni

In: Discussion papers on business and economics, 2017, vol. 2, p. 1-69

In heterogeneous treatment effect models with endogeneity, identification of the LATE typically relies on the availability of an exogenous instrument monotonically related to treatment participation. We demonstrate that a strictly weaker local monotonicity condition identifies the LATEs on compliers and on defiers. We propose simple estimators that are potentially more efficient than 2SLS,...

Université de Fribourg

Testing exclusion restrictions and additive separability in sample selection models

Huber, Martin ; Mellace, Giovanni

In: Empirical economics, 2014, vol. 47, no. 1, p. 75-92

Standard sample selection models with non-randomly censored outcomes assume (i) an exclusion restriction (i.e., a variable affecting selection, but not the outcome) and (ii) additive separability of the errors in the selection process. This paper proposes tests for the joint satisfaction of these assumptions by applying the approach of Huber and Mellace (Testing instrument validity for LATE...

Université de Fribourg

Sharp IV bounds on average treatment effects on the treated and other populations under endogeneity and noncompliance

Huber, Martin ; Laffers, Lukas ; Mellace, Giovanni

In: Journal of Applied Econometrics, 2017, vol. 32, no. 1, p. 56-79

In the presence of an endogenous binary treatment and a valid binary instru- ment, 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...

Université de Fribourg

Testing instrument validity for LATE identification based on inequality moment constraints

Huber, Martin ; Mellace, Giovanni

In: Review of economics and statistics, 2015, vol. 97, no. 2, p. 398-411

We derive testable implications of instrument validity in just identified treat- ment effect models with endogeneity and consider several tests. The identifying assump- tions of the local average treatment effect allow us to both point identify and bound the mean potential outcomes (i) of the always takers under treatment and (ii) of the never takers under non-treatment. The point identified...

Université de Fribourg

Why do tougher caseworkers increase employment? : The role of programme assignment as a causal mechanism

Huber, Martin ; Lechner, Michael ; Mellace, Giovanni

In: Review of economics and statistics, 2017, vol. 99, no. 1, p. 180-183

Previous research found that less accommodating caseworkers are more success- ful in placing unemployed workers into employment. This paper explores the causal mechanisms behind this result using semiparametric mediation analysis. Analysing rich linked jobseeker-caseworker data for Switzerland, we find that the positive employment effects of less accommodating caseworkers are not driven by a ...

Université de Fribourg

Sharp bounds on causal effects under sample selection

Huber, Martin ; Mellace, Giovanni

In: Oxford bulletin of economics and statistics, 2015, vol. 77, no. 1, p. 129-151

In many empirical problems, the evaluation of treatment effects is complicated by sample selection so that the outcome is only observed for a non-random subpopulation. In the absence of instruments and/or tight parametric assumptions, treatment effects are not point identified, but can be bounded under mild restrictions. Previous work on partial identification has primarily focused on the...