The Econometric SocietyEdited by: Guido W. Imbens • Print ISSN: 0012-9682 • Online ISSN: 1468-0262
This supplement contains an additional example that illustrates Proposition 4 in the paper. We also describe the programs and data used to derive the empirical results.
Zip file containing data and program files.
In this appendix to ?Robust Mechanism Design? we provide a set of sufficient conditions which encompass the quasilinear environment. With these conditions, we can generalize the equivalence results presented in Proposition 4 for common prior full support payoff type spaces. The first condition replaces the compactness condition, the second and third conditions generalize the important features of the quasilinear utility model. The appendix ends with an example which is meant to illustrate that the conditions, in particular, the condition on bounded allocation differences, is not easily dispensed with.
This is a an appendix specifying the theoretical model outlined in the text.
1. Extra-figureA shows the SSP impacts (treatment minus control group means) for IA for all individuals and for 2 subgroups: those with less than a high school education, and those with a high school education or higher. This is created by treat_byed.sas reading from wide54r data set 2. Extra-figureB shows the simulated SSP impacts from the normal heterogeneity model (column 4 of Table 6) for two "types": people with a random effect equal to the 20th percentile of the heterogeneity distribution (which is N(0, 1.18)); and people with a random effect equal to the 80thpercentile of the heterogeneity distribution. See sim8020.sas 3. Extra-figureC shows the actual hazard rate into SSP eligibility, and the predicted rate from the the normal heterogeneity model (column 4 of Table 6). See extra-figc.xls
This archive contains subdirectories with various materials: 1) additional 2) dataprep Includes programs to read the original data, and create table 2 of the paper. 3) wages Includes programs to construct table 3 and figures 4 and 5. 4) Files on data preparation. Programs and output for figures 4, 5, 7, 8 and 11, and tables 3-7.
