Unmeasured confounding is a key challenge for causal inference. In this paper, we establish a framework for unmeasured confounding adjustment with negative control variables. A negative control outcome is associated with the confounder but not causally affected by the exposure in view, and a negative control exposure is correlated with the primary exposure or the confounder but does not causally affect the outcome of interest. We introduce an outcome confounding bridge function that depicts the relationship between the confounding effects on the primary outcome and the negative control outcome, and we incorporate a negative control exposure to identify the bridge function and the average causal effect. We also consider the extension to the positive control setting by allowing for the nonzero causal effect of the primary exposure on the control outcome. We illustrate our approach with simulations and apply it to a study about the short-term effect of air pollution on mortality. Although a standard analysis shows a significant acute effect of PM2.5 on mortality, our analysis indicates that this effect may be confounded, and after double negative control adjustment, the effect is attenuated toward zero.

References

• Andrews, D. W. K. (1991). Heteroskedasticity and autocorrelation consistent covariance matrix estimation. Econometrica, 59(open in a new window)(3(open in a new window)), 817–858. https://doi.org/10.2307/2938229 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Andrews, D. W. K. (2017). Examples of L2-complete and boundedly-complete distributions. Journal of Econometrics, 199(open in a new window)(2(open in a new window)), 213–220. https://doi.org/10.1016/j.jeconom.2017.05.011 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Angrist, J. D., Imbens, G. W., & Rubin, D. B. (1996). Identification of causal effects using instrumental variables. Journal of the American Statistical Association, 91(open in a new window)(434(open in a new window)), 444–455. https://doi.org/10.1080/01621459.1996.10476902 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Baker, S. G., & Lindeman, K. S. (1994). The paired availability design: A proposal for evaluating epidural analgesia during labor. Statistics in Medicine, 13(open in a new window)(21(open in a new window)), 2269–2278. https://doi.org/10.1002/sim.v13:21 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Bang, H., & Robins, J. M. (2005). Doubly robust estimation in missing data and causal inference models. Biometrics, 61(open in a new window)(4(open in a new window)), 962–973. https://doi.org/10.1111/biom.2005.61.issue-4 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Bennett, A., & Kallus, N. (2021). Proximal reinforcement learning: Efficient off-policy evaluation in partially observed Markov decision processes. Operations Research, 72(open in a new window)(3(open in a new window)), 1071–1086. https://doi.org/10.1287/opre.2021.0781 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Cornfield, J., Haenszel, W., Hammond, E. C., Lilienfeld, A. M., Shimkin, M. B., & Wynder, E. L. (1959). Smoking and lung cancer: Recent evidence and a discussion of some questions. Journal of the National Cancer Institute, 22(1), 173–203. 

   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar

• Cui, Y., Pu, H., Shi, X., Miao, W., & Tchetgen Tchetgen, E. (2023). Semiparametric proximal causal inference. Journal of the American Statistical Association, 119(open in a new window)(546(open in a new window)), 1348–1359. https://doi.org/10.1080/01621459.2023.2191817 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Darolles, S., Fan, Y., Florens, J. P., & Renault, E. (2011). Nonparametric instrumental regression. Econometrica, 79(open in a new window)(5(open in a new window)), 1541–1565. https://doi.org/10.3982/ECTA6539 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Davey Smith, G. (2008). Assessing intrauterine influences on offspring health outcomes: Can epidemiological studies yield robust findings? Basic & Clinical Pharmacology & Toxicology, 102(open in a new window)(2(open in a new window)), 245–256. https://doi.org/10.1111/pto.2008.102.issue-2 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Davey Smith, G. (2012). Negative control exposures in epidemiologic studies. Epidemiology, 23(open in a new window)(2(open in a new window)), 350–351. https://doi.org/10.1097/EDE.0b013e318245912c 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• D'Haultfœuille, X. (2011). On the completeness condition in nonparametric instrumental problems. Econometric Theory, 27(open in a new window)(3(open in a new window)), 460–471. https://doi.org/10.1017/S0266466610000368 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Dukes, O., Shpitser, I., & Tchetgen Tchetgen, E. (2021). Proximal mediation analysis. arXiv preprint arXiv:2109.11904. 

   (Open in a new window)Google Scholar

• Flanders, W. D., Strickland, M. J., & Klein, M. (2017). A new method for partial correction of residual confounding in time-series and other observational studies. American Journal of Epidemiology, 185(open in a new window)(10(open in a new window)), 941–949. https://doi.org/10.1093/aje/kwx013 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Freeman, M. F., & Tukey, J. W. (1950). Transformations related to the angular and the square root. The Annals of Mathematical Statistics, 21(open in a new window)(4(open in a new window)), 607–611. https://doi.org/10.1214/aoms/1177729756 
  • View
   (Open in a new window)Google Scholar
• Gagnon-Bartsch, J., Jacob, L., & Speed, T. P. (2013). Removing unwanted variation from high dimensional data with negative controls (Technical Report 820). Dept. Statistics, Univ. California. 

   (Open in a new window)Google Scholar

• Goldberger, A. S. (1972). Structural equation methods in the social sciences. Econometrica, 40(open in a new window)(6(open in a new window)), 979–1001. https://doi.org/10.2307/1913851 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Hall, A. R. (2005). Generalized method of moments. Oxford University Press. 

   (Open in a new window)Google Scholar

• Hamilton, J. D. (1994). Time series analysis. Princeton University Press. 
  • View
   (Open in a new window)Google Scholar
• Hansen, L. P. (1982). Large sample properties of generalized method of moments estimators. Econometrica, 50(open in a new window)(4(open in a new window)), 1029–1054. https://doi.org/10.2307/1912775 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Hill, A. B. (1965). The environment and disease: Association or causation?. Proceedings of the Royal Society of Medicine, 58(open in a new window)(5(open in a new window)), 295–300. https://doi.org/10.1177/003591576505800503 
  • View
   (Open in a new window)PubMed(Open in a new window)Google Scholar
• Hu, Y., & Shiu, J.-L. (2018). Nonparametric identification using instrumental variables: Sufficient conditions for completeness. Econometric Theory, 34(open in a new window)(3(open in a new window)), 659–693. https://doi.org/10.1017/S0266466617000251 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Kallus, N., Mao, X., & Uehara, M. (2021). Causal inference under unmeasured confounding with negative controls: A minimax learning approach. arXiv:2103.14029. 

   (Open in a new window)Google Scholar

• Kuroki, M., & Pearl, J. (2014). Measurement bias and effect restoration in causal inference. Biometrika, 101(open in a new window)(2(open in a new window)), 423–437. https://doi.org/10.1093/biomet/ast066 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Li, K. Q., Shi, X., Miao, W., & Tchetgen Tchetgen, E. (2022). Doubly robust proximal causal inference under confounded outcome-dependent sampling. arXiv:2208.01237. 

   (Open in a new window)Google Scholar

• Lipsitch, M., Tchetgen Tchetgen, E., & Cohen, T. (2010). Negative controls: A tool for detecting confounding and bias in observational studies. Epidemiology, 21(open in a new window)(3(open in a new window)), 383–388. https://doi.org/10.1097/EDE.0b013e3181d61eeb 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Mamun, A. A., Lawlor, D. A., Alati, R., O'callaghan, M. J., Williams, G. M., & Najman, J. M. (2006). Does maternal smoking during pregnancy have a direct effect on future offspring obesity? Evidence from a prospective birth cohort study. American Journal of Epidemiology, 164(open in a new window)(4(open in a new window)), 317–325. https://doi.org/10.1093/aje/kwj209 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Miao, W., Geng, Z., & Tchetgen Tchetgen, E. (2018). Identifying causal effects with proxy variables of an unmeasured confounder. Biometrika, 105(open in a new window)(4(open in a new window)), 987–993. https://doi.org/10.1093/biomet/asy038 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Miao, W., Shi, X., & Tchetgen Tchetgen, E. (2018). A confounding bridge approach for double negative control inference on causal effects. arXiv: 1808.04945. 

   (Open in a new window)Google Scholar

• Newey, W. K., & Powell, J. L. (2003). Instrumental variable estimation of nonparametric models. Econometrica, 71(open in a new window)(5(open in a new window)), 1565–1578. https://doi.org/10.1111/ecta.2003.71.issue-5 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Newey, W. K., & West, K. D. (1987). A simple, positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica, 55(open in a new window)(3(open in a new window)), 703–708. https://doi.org/10.2307/1913610 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Qi, Z., Miao, R., & Zhang, X. (2022). Proximal learning for individualized treatment regimes under unmeasured confounding. Journal of the American Statistical Association, 119(open in a new window)(546(open in a new window)), 915–928. https://doi.org/10.1080/01621459.2022.2147841 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Robins, J. M. (1994). Correcting for non-compliance in randomized trials using structural nested mean models. Communications in Statistics-Theory and Methods, 23(open in a new window)(8(open in a new window)), 2379–2412. https://doi.org/10.1080/03610929408831393 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Rosenbaum, P. R., & Rubin, D. B. (1983a). Assessing sensitivity to an unobserved binary covariate in an observational study with binary outcome. Journal of the Royal Statistical Society: Series B, 45(open in a new window)(2(open in a new window)), 212–218. https://doi.org/10.1111/j.2517-6161.1983.tb01242.x 
  • View
   (Open in a new window)Google Scholar
• Rosenbaum, P. R., & Rubin, D. B. (1983b). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(open in a new window)(1(open in a new window)), 41–55. https://doi.org/10.1093/biomet/70.1.41 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Rubin, D. B. (1973). The use of matched sampling and regression adjustment to remove bias in observational studies. Biometrics, 29(open in a new window)(1(open in a new window)), 185–203. https://doi.org/10.2307/2529685 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Shi, X., Miao, W., Hu, M., & Tchetgen Tchetgen, E. (2021). Theory for identification and inference with synthetic controls: A proximal causal inference framework. arXiv:2108.13935. 

   (Open in a new window)Google Scholar

• Shi, X., Miao, W., Nelson, J. C., & Tchetgen Tchetgen, E. (2020). Multiply robust causal inference with double negative control adjustment for categorical unmeasured confounding. Journal of the Royal Statistical Society: Series B, 82(open in a new window)(2(open in a new window)), 521–540. https://doi.org/10.1111/rssb.12361 
  • View
   (Open in a new window)Google Scholar
• Sofer, T., Richardson, D. B., Colicino, E., Schwartz, J., & E. J. Tchetgen Tchetgen (2016). On negative outcome control of unobserved confounding as a generalization of difference-in-differences. Statistical Science, 31(open in a new window)(3(open in a new window)), 348–361. https://doi.org/10.1214/16-STS558 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Stuart, E. A. (2010). Matching methods for causal inference: A review and a look forward. Statistical Science, 25(open in a new window)(1(open in a new window)), 1–21. https://doi.org/10.1214/09-STS313 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Sverdrup, E., & Cui, Y. (2023). Proximal causal learning of heterogeneous treatment effects. arXiv:2301.10913. 

   (Open in a new window)Google Scholar

• Tchetgen Tchetgen, E. (2014). The control outcome calibration approach for causal inference with unobserved confounding. American Journal of Epidemiology, 179(open in a new window)(5(open in a new window)), 633–640. https://doi.org/10.1093/aje/kwt303 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Tchetgen Tchetgen, E. J., Ying, A., Cui, Y., Shi, X., & Miao, W. (2020). An introduction to proximal causal learning. arXiv:2009.10982. 

   (Open in a new window)Google Scholar

• Trichopoulos, D., Zavitsanos, X., Katsouyanni, K., Tzonou, A., & Dalla-Vorgia, P. (1983). Psychological stress and fatal heart attack: The athens (1981) earthquake natural experiment. The Lancet, 321(open in a new window)(8322(open in a new window)), 441–444. https://doi.org/10.1016/S0140-6736(83)91439-3 
  • View
   (Open in a new window)Google Scholar
• Wang, J., Zhao, Q., Hastie, T., & Owen, A. B. (2017). Confounder adjustment in multiple hypothesis testing. The Annals of Statistics, 45(open in a new window)(5(open in a new window)), 1863–1894. https://doi.org/10.1214/16-AOS1511 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Wright, P. G. (1928). Tariff on animal and vegetable oils. Macmillan. 

   (Open in a new window)Google Scholar

• Ying, A., Cui, Y., & Tchetgen Tchetgen, E. (2022). Proximal causal inference for marginal counterfactual survival curves. arXiv:2204.13144. 

   (Open in a new window)Google Scholar

• Ying, A., Miao, W., Shi, X., & Tchetgen Tchetgen, E. J. (2023). Proximal causal inference for complex longitudinal studies. Journal of the Royal Statistical Society Series B: Statistical Methodology, 85(open in a new window)(3(open in a new window)), 684–704. https://doi.org/10.1093/jrsssb/qkad020 
  • View
   (Open in a new window)Web of Science ®(Open in a new window)Google Scholar
• Zhang, J., Li, W., Miao, W., & Tchetgen Tchetgen, E. (2023). Proximal causal inference without uniqueness assumptions. Statistics & Probability Letters, 198(open in a new window), Article 109836. https://doi.org/10.1016/j.spl.2023.109836 
  • View
   (Open in a new window)PubMed (Open in a new window)Web of Science ®(Open in a new window)Google Scholar

To cite this article: Ailing Gu, Xuanzhen Zhang, Shumin Chen & Ling Zhang (30 Aug 2024): Robust optimal reinsurance-investment strategy with extrapolative bias premiums and ambiguity aversion, Statistical Theory and Related Fields, DOI: 10.1080/24754269.2024.2393062

To link to this article: https://doi.org/10.1080/24754269.2024.2393062