Studying Policy Design Quality in Comparative Perspective

Expected effects on policy portfolio diversity in the environmental sector.


This article is a first attempt to systematically examine policy design and its impact on policy effectiveness in a comparative perspective. We begin by providing a novel concept and measure of policy design. Our Average Instrument Diversity (AID) index captures whether governments tend to reuse the same policy instruments and instrument combinations or produce policy solutions that are carefully tailored to the policy problem at hand. Second, we demonstrate that our AID index is a valid and reliable measure of policy design quality with a strong explanatory power for the outcome variables tested. Analyzing the composition of environmental policy portfolios in 21 OECD countries, we show that higher levels of AID are positively associated with a country’s policy effectiveness in environmental matters. Based on this finding, we analyze, in a third step, the factors that lead countries to adopt more or less diverse policy portfolios. We find that the policy design quality is significantly improved when policy-makers are not bound by high institutional constraints and, more importantly, are backed by well-equipped bureaucracies.

American Political Science Review

Online appendix

The online appendix contains an extension of the procedures and results presented in the paper; the JAGS code for the statistical model, and the ggmcmc output for convergence diagnostics of the model parameters.

Code and replication material

The whole set of code and data for the replication is available at the Harvard dataverse (doi: 10.7910/DVN/M5SDCH).

Local copies are available for:

  • Report (4.1 Mb, 193 pgs), containing the main report in PDF with all the code.
  • Report (39 Mb), containing the main report in HTML with all the code.
  • Material (27 Mb), compressed file containing all the files needed to perform the analysis and generate the final report.

Dataset and accompanying functions

The paper is completed with a set of functions in the PolicyPortfolios R package, and its documentation.