Entropy Balancing for Continuous Treatments

  • Interest in evaluating the effects of continuous treatments has been on the rise recently. To facilitate the estimation of causal effects in this setting, the present paper introduces entropy balancing for continuous treatments (EBCT) by extending the original entropy balancing methodology of Hainmüller (2012). In order to estimate balancing weights, the proposed approach solves a globally convex constrained optimization problem, allowing for much more computationally efficient implementation compared to other available methods. EBCT weights reliably eradicate Pearson correlations between covariates and the continuous treatment variable. This is the case even when other methods based on the generalized propensity score tend to yield insufficient balance due to strong selection into different treatment intensities. Moreover, the optimization procedure is more successful in avoiding extreme weights attached to a single unit. Extensive Monte-Carlo simulations show that treatment effect estimates using EBCT display similar or lower biasInterest in evaluating the effects of continuous treatments has been on the rise recently. To facilitate the estimation of causal effects in this setting, the present paper introduces entropy balancing for continuous treatments (EBCT) by extending the original entropy balancing methodology of Hainmüller (2012). In order to estimate balancing weights, the proposed approach solves a globally convex constrained optimization problem, allowing for much more computationally efficient implementation compared to other available methods. EBCT weights reliably eradicate Pearson correlations between covariates and the continuous treatment variable. This is the case even when other methods based on the generalized propensity score tend to yield insufficient balance due to strong selection into different treatment intensities. Moreover, the optimization procedure is more successful in avoiding extreme weights attached to a single unit. Extensive Monte-Carlo simulations show that treatment effect estimates using EBCT display similar or lower bias and uniformly lower root mean squared error. These properties make EBCT an attractive method for the evaluation of continuous treatments. Software implementation is available for Stata and R.show moreshow less

Download full text files

  • cepa21.pdfeng
    (800KB)

    SHA-1:33242c2ab84193506d624d95ccfbf8660fe9b933d1f5cacfd9ea72693b969cfd70fdc6a36fe30cb82c77c19addc30cf3fe0a5637daccc0eabf3b4f6723b95eba

Export metadata

Metadaten
Author:Stefan TübbickeGND
URN:urn:nbn:de:kobv:517-opus4-478950
DOI:https://doi.org/10.25932/publishup-47895
ISSN:2628-653X
Parent Title (English):CEPA Discussion Papers
Document Type:Working Paper
Language:English
Date of first Publication:2020/10/12
Year of Completion:2020
Publishing Institution:Universität Potsdam
Release Date:2020/10/12
Tag:Balancing weights; Continuous Treatment; Monte-Carlo simulation; Observational studies
Issue:21
Pagenumber:32
RVK - Regensburg Classification:QH 239, QH 235
Organizational units:Zentrale und wissenschaftliche Einrichtungen / Center for Economic Policy Analysis (CEPA)
Wirtschafts- und Sozialwissenschaftliche Fakultät / Wirtschaftswissenschaften / Fachgruppe Volkswirtschaftslehre
Dewey Decimal Classification:3 Sozialwissenschaften / 33 Wirtschaft / 330 Wirtschaft
JEL Classification:C Mathematical and Quantitative Methods / C1 Econometric and Statistical Methods: General / C14 Semiparametric and Nonparametric Methods
C Mathematical and Quantitative Methods / C2 Single Equation Models; Single Variables / C21 Cross-Sectional Models; Spatial Models; Treatment Effect Models; Quantile Regressions (Updated!)
C Mathematical and Quantitative Methods / C8 Data Collection and Data Estimation Methodology; Computer Programs / C87 Econometric Software
Peer Review:Nicht referiert
Licence (German):License LogoKeine Nutzungslizenz vergeben - es gilt das deutsche Urheberrecht