Alternative Approaches to Evaluation in Empirical Microeconomics

Abstract

Structural mean models (SMMs) were originally formulated to estimate causal effects among those selecting treatment in randomised controlled trials a¤ected by non-ignorable non-compliance. It has already been established that SMM estimators identify these causal e¤ects in randomised placebo-controlled trials where no-one assigned to the control group can receive the treatment. However, SMMs are starting to be used for randomised controlled trials without placebo-controls, and for instrumental variable analysis of observational studies; for example, Mendelian randomisation studies, and studies where physicians select patients'treatments. In such scenarios, identi…cation depends on the assumption of no e¤ect modi…cation, namely, the causal e¤ect is equal for the subgroups de…ned by the instrument. We consider the nature of this assumption by showing how it depends crucially on the underlying causal model generating the data, which in applications is almost always unknown. If its no e¤ect modi…cation assumption does not hold then an SMM estimator does not estimate its associated causal e¤ect. However, if treatment selection is monotonic we highlight that additive and multiplicative SMMs do identify local (or complier) causal e¤ects, but that the double-logistic SMM estimator does not without further assumptions. We clarify the proper interpretation of inferences from SMM estimators using a data example and simulation study.

Do unemployed workers benefit from enterprise zones? The French experience

“…3 Figures from the 1999 Census of the Population indicate that the 9 enterprise zones in the Paris region hosted about 220,000 inhabitants, i.e. 2% of the population of the region.…”

“…This second stage uses conditional matching techniques to address possible issues of treatment selectivity (see Blundell and Costa-Dias, 2009, for a recent survey). Given our …ne control of composition and right censoring biases in the …rst stage, and given the way selection into treatment was implemented, we argue that, conditional on the variables that a¤ect treatment probability, trends in unemployment durations in treated and control municipalities would have been on average similar in the absence of treatment.…”

Entry Regulation and Entrepreneurship: Empirical Evidence from a German Natural Experiment

“…With such a group, I can contrast the changes in the average outcomes of occupational choices before and after the reform in each of the groups affected by the policy with the changes in the average outcomes of occupational choices before and after the reform in the control group. The difference in these changes, known as the difference-in-differences (DID) estimator, represents the average treatment effect on the treated (ATT), which equals the average causal effect of the reform (e.g., Blundell and Dias, 2009). …”

Implications of global budget payment system on nursing home costs

“…As pointed out by Blundell and Dias (2009), the choice of the most appropriate policy evaluation method relies on the nature of the policy change, as well as the research question and data availability. In our study, the policy change concerns all NHs in the sample at the same time.…”