Synthetic controls with imperfect pretreatment fit

We analyze the properties of the Synthetic Control (SC) and related estimators when the pre‐treatment fit is imperfect. In this framework, we show that these estimators are generally biased if treatment assignment is correlated with unobserved confounders, even when the number of pre‐treatment periods goes to infinity. Still, we show that a demeaned version of the SC method can improve in terms of bias and variance relative to the difference‐in‐difference estimator. We also derive a specification test for the demeaned SC estimator in this setting with imperfect pre‐treatment fit. Given our theoretical results, we provide practical guidance for applied researchers on how to justify the use of such estimators in empirical applications.

Food insecurity patterns before and after initial receipt of Supplemental Security Income

ObjectiveTo assess patterns of food insecurity before and after initial receipt of Supplemental Security Income (SSI) benefits.DesignWe analysed data from a nationally representative sample. We estimated two difference-in-difference models comparing food insecurity patterns among eventual SSI recipients with patterns among eligible non-recipients during two time frames. The first model assessed changes in food insecurity immediately before SSI benefits were first received and the second model assessed changes in food insecurity after programme entry.Setting2008 panel of the Survey of Income and Program Participation.ParticipantsNon-institutionalized population of the USA.ResultsThe percentage of eventual SSI recipients experiencing food insecurity rose from 18 to 30 % in the year before programme entry, compared with a change from 17 to 18 % for eligible non-recipients. Adjusting for sociodemographic covariates, the difference-in-difference estimator for this comparison was statistically significant (P=0·01). Additionally, the percentage of recipients experiencing food insecurity fell from 28 % in the year before programme entry to 16 % in the year after entry, compared with a change from 16 to 17 % for eligible non-recipients. Adjusting for sociodemographic covariates, the difference-in-difference estimator for this comparison was marginally significant (P=0·07).ConclusionsFood insecurity rises prior to SSI entry but may be alleviated by programme benefits. Greater nutritional supports for SSI applicants awaiting decisions may reduce the burden of food insecurity in this population and improve health outcomes.

Conditional estimation and inference to address observed covariate imbalance in randomized clinical trials

Background Baseline covariate imbalance (between treatment groups) is a common problem in randomized clinical trials which often raises questions about the validity of trial results. Answering these questions requires careful consideration of the statistical implications of covariate imbalance. The possibil ity of having covariate imbalance contributes to the marginal variance of an unadjusted treatment difference estimator, which can be reduced by making appropriate adjustments. Actual observed imbalance introduces a conditional bias into an unadjusted estimator, which may increase the conditional size of an unadjusted test. Methods This article provides conditional estimation and inference procedures to address the conditional bias due to observed imbalance. Interestingly, it is possible to use the same adjusted treatment difference estimator to address the marginal variance issue and the conditional bias issue associated with covariate imbalance. Its marginal variance estimator tends to be conservative for conditional inference, and we propose a conditionally appropriate variance estimator. We also provide an estimator of the conditional bias in an unadjusted treatment difference estimator, together with a conditional variance estimator. Results The proposed methodology is illustrated with real data from a stroke trial and evaluated in simulation experiments based on the same trial. The simulation results show that covariate imbalance can result in a substantial conditional bias and that the proposed methods deal with the conditional bias quite effectively. Discussion We recommend that the proposed methodology be used routinely to address the observed covariate imbalance in randomized clinical trials.

Does Mediation-Arbitration Reduce Arbitration Rates? Evidence from a Natural Experiment

To study the effect of the introduction of mediation-arbitration as a dispute resolution procedure on interest arbitration, the authors exploit a natural experiment in the arbitration institutions for police and firefighter sectors in the Canadian province of Ontario. They obtain estimates using a difference-in-difference estimator. Results show that the introduction of mediation-arbitration is significantly associated with increased use of arbitration by firefighters relative to the police. The article also draws on interviews with stakeholders to help explain the mechanisms that contribute to the increase in arbitration rates.

Evaluation of IMF programmes on employment in the EU

The International Monetary Fund (IMF) has undergone notable changes by starting collaboration with the European Union (EU). Hence, this paper seeks to estimate the effects of IMF programmes on employment, with data from the EU-28 between 1993 and 2013. In order to control for selection on observable and unobservable variables, the study employs Propensity Score Matching in combination with the Differences-in-Difference estimator. Next, the robustness of the findings is checked by applying four different matching algorithms. Our paper concludes that employment decreases once a country resorts to the IMF, and this impact is still measurable after two years of programme initiation.