ABSTRACT
Questions such as whether the instrumental variable (IV) estimates of return to schooling are sensitive to the violation of the exclusion restriction often draw empirical concerns. We employ a Bayesian method to compare the IV estimate using either mother’s or father’s education as an instrumental variable under different degree of violation of the exclusion restriction. We find that a moderate violation of the exclusion restriction does not have a substantial influence on the estimation results. Particularly, the IV estimate using mother’s education as an IV is more robust than that based on father’s education.
Acknowledgments
The authors would like to thank the Editor Mark Talyor, and two anonymous referees for their insightful comments. We also thank Cynthia Bansak, McKinley Blackburn, Belton Fleisher, Anu Gill, Patrick McCarthy, Xiaojun Wang and Allan Zebedee for their helpful comments in the development of this paper.
Disclosure statement
No potential conflict of interest was reported by the authors.
Supplementary materials
Supplemental data for this article can be accessed here.
Notes
1 We are grateful to an anonymous referee for pointing out this issue.
2 We also conduct the analysis using data from year 2010 and 2014, and the results remain consistent (see the online appendix).
3 The widely known Armed Forces Qualifications Test (AFQT) score is not used here, since it is calculated from selected subtests and biased by schooling and age differences (Blackburn Citation2004).