Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
140 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Inference on LATEs with covariates (2402.12607v2)

Published 20 Feb 2024 in econ.EM

Abstract: In theory, two-stage least squares (TSLS) identifies a weighted average of covariate-specific local average treatment effects (LATEs) from a saturated specification, without making parametric assumptions on how available covariates enter the model. In practice, TSLS is severely biased as saturation leads to a large number of control dummies and an equally large number of, arguably weak, instruments. This paper derives asymptotically valid tests and confidence intervals for the weighted average of LATEs that is targeted, yet missed by saturated TSLS. The proposed inference procedure is robust to unobserved treatment effect heterogeneity, covariates with rich support, and weak identification. We find LATEs statistically significantly different from zero in applications in criminology, finance, health, and education.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (38)
  1. Abadie, A. (2003). Semiparametric instrumental variable estimation of treatment response models. Journal of Econometrics 113(2), 231–263.
  2. Ackerberg, D. A. and P. J. Devereux (2009). Improved JIVE estimators for overidentified linear models with and without heteroskedasticity. The Review of Economics and Statistics 91(2), 351–362.
  3. One instrument to rule them all: The bias and coverage of just-id iv. Journal of Econometrics 0(just-accepted), 1–18.
  4. Angrist, J. D. and G. W. Imbens (1995). Two-stage least squares estimation of average causal effects in models with variable treatment intensity. Journal of the American Statistical Association 90(430), 431–442.
  5. Jackknife instrumental variables estimation. Journal of Applied Econometrics 14(1), 57–67.
  6. Angrist, J. D. and A. B. Krueger (1991). Does compulsory school attendance affect schooling and earnings? The Quarterly Journal of Economics 106(4), 979–1014.
  7. Angrist, J. D. and J.-S. Pischke (2009). Mostly harmless econometrics: An empiricist’s companion. Princeton University Press.
  8. Bekker, P. A. (1994). Alternative approximations to the distributions of instrumental variable estimators. Econometrica 62(3), 657–681.
  9. Jackknife instrumental variable estimation with heteroskedasticity. Journal of Econometrics 185(2), 332–342.
  10. Finite-sample instrumental variables inference using an asymptotically pivotal statistic. Econometric Theory 19(5), 744–753.
  11. When is TSLS actually LATE? Working paper, National Bureau of Economic Research.
  12. Card, D. (1995). Using geographic variation in college proximity to estimate the return to schooling. In L. Christofides, E. Grant, and R. Swidinsky (Eds.), Aspects of Labor Market Behaviour: Essays in Honour of John Vanderkamp. University of Toronto Press.
  13. Inference in linear regression models with many covariates and heteroscedasticity. Journal of the American Statistical Association 113(523), 1350–1361.
  14. Chao, J. C. and N. R. Swanson (2005). Consistent estimation with a large number of weak instruments. Econometrica 73(5), 1673–1692.
  15. Asymptotic distribution of JIVE in a heteroskedastic IV regression with many instruments. Econometric Theory 28(1), 42–86.
  16. Jackknife estimation of a cluster-sample IV regression model with many weak instruments. Journal of Econometrics 235(2), 1747–1769.
  17. Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal 21(1), C1–C68.
  18. Inference in instrumental variable models with heteroskedasticity and many instruments. Econometric Theory 37(2), 281–310.
  19. Dealing with limited overlap in estimation of average treatment effects. Biometrika 96(1), 187–199.
  20. Queens. Journal of Political Economy 128(7), 2579–2652.
  21. Inference in instrumental variables analysis with heterogeneous treatment effects. Working paper, Princeton University.
  22. Diagnostics for exclusion restrictions in instrumental variables estimation. Working paper, Princeton University.
  23. Frölich, M. (2007). Nonparametric IV estimation of local average treatment effects with covariates. Journal of Econometrics 139(1), 35–75.
  24. Gelbach, J. B. (2002). Public schooling for young children and maternal labor supply. American Economic Review 92(1), 307–322.
  25. Estimation with many instrumental variables. Journal of Business & Economic Statistics 26(4), 398–422.
  26. Variance estimation with one unit per stratum. Journal of the American Statistical Association 64(327), 841–851.
  27. Instrumental variable estimation with heteroskedasticity and many instruments. Quantitative Economics 3(2), 211–255.
  28. Imbens, G. W. and J. D. Angrist (1994). Identification and estimation of local average treatment effects. Econometrica 62(2), 467–475.
  29. Kitagawa, T. (2015). A test for instrument validity. Econometrica 83(5), 2043–2063.
  30. Kleibergen, F. (2005). Testing parameters in GMM without assuming that they are identified. Econometrica 73(4), 1103–1123.
  31. Double robust inference for continuous updating GMM. Working paper, arXiv preprint arXiv:2105.08345.
  32. Kolesár, M. (2013). Estimation in an instrumental variables model with treatment effect heterogeneity. Working paper, Princeton University.
  33. Lee, S. (2018). A consistent variance estimator for 2SLS when instruments identify different LATEs. Journal of Business & Economic Statistics 36(3), 400–410.
  34. A conditional linear combination test with many weak instruments. Journal of Econometrics 283(2), 105602.
  35. A jackknife Lagrange multiplier test with many weak instruments. Econometric Theory 0(just-accepted), 1–24.
  36. Inference with many weak instruments. Review of Economic Studies 89(5), 2663–2686.
  37. Weak identification with many instruments. Working paper, arXiv preprint arXiv:2308.09535.
  38. Słoczyński, T. (2020). When should we (not) interpret linear IV estimands as LATE? Working paper, arXiv preprint arXiv:2011.06695.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now