Search

In certain situation, only one unit is treated against a large sample of control units. In such a case, the synthetic control approach has proven to be useful as it combines control units in order to build a synthetic counterfactual. This chapter presents the general case of synthetic control alongside an extensive example of URBAN II in London.

This chapter theoretically discusses various approaches to ex-ante evaluation of regional development policies. A particular focus on the concept of “transportability” is presented and applied to meta-analysis, quasi-experimental methods and structural models.

This chapter introduces the identification and estimation of policy effects when outcome variables can be ordered according to a given variable and when the treatment occurs at a given point. The structural change occurring in the outcome variable is in such a case assumed to be the effect of the policy. The assumptions behind Regression Discontinuity Design are hence discussed alongside with extensions for heterogeneous effect.

When units are observed across time and space and treatment occurs at a certain point in time for only some units, difference-in-differences can be used to estimate the impact of a policy. This chapter discusses standard identifying assumption of the policy impact through this parametric technique.

This chapter presents the semi-parametric technique of propensity score matching. The standard approach is briefly discussed with an example and extensions to the case of multi-valued treatment and Oaxaca-Blinder regressions are also presented.