Bounding program benefits when participation is misreported: Estimation and inference with Stata

• Autores: Andy Lin, Denni Tommasi, Lina Zhong
• Localización: The Stata journal, ISSN 1536-867X, Vol. 24, Nº. 2, 2024, págs. 185-212
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • Instrumental-variables estimation is an approach commonly used to evaluate the effect of a program in case of noncompliance. However, when the binary treatment status is misreported, standard techniques are not sufficient to point identify and consistently estimate the effect of interest. We present a new command, ivbounds, that implements three partial identification strategies de- veloped by Tommasi and Zhang (2024, Journal of Econometrics 238: 105556) to bound the heterogeneous treatment effect when both noncompliance and misre- porting of treatment status are present. We illustrate the use of the command by reassessing the benefits of participating in the 401(k) pension plan on savings in the United States.