Two-stage residual inclusion estimation: addressing endogeneity in health econometric modeling.

The paper focuses on two estimation methods that have been widely used to address endogeneity in empirical research in health economics and health services research-two-stage predictor substitution (2SPS) and two-stage residual inclusion (2SRI). 2SPS is the rote extension (to nonlinear models) of the popular linear two-stage least squares estimator. The 2SRI estimator is similar except that in the second-stage regression, the endogenous variables are not replaced by first-stage predictors. Instead, first-stage residuals are included as additional regressors. In a generic parametric framework, we show that 2SRI is consistent and 2SPS is not. Results from a simulation study and an illustrative example also recommend against 2SPS and favor 2SRI. Our findings are important given that there are many prominent examples of the application of inconsistent 2SPS in the recent literature. This study can be used as a guide by future researchers in health economics who are confronted with endogeneity in their empirical work.

[1]  Halbert White,et al.  Estimation, inference, and specification analysis , 1996 .

[2]  William H. Greene,et al.  Gender Economics Courses in Liberal Arts Colleges: Further Results , 1998 .

[3]  J. Kadane.  Structural Analysis of Discrete Data with Econometric Applications , 1984 .

[4]  K. McGeary,et al.  Illicit drug use and emergency room utilization. , 2000, Health services research.

[5]  G. Gowrisankaran,et al.  Estimating the quality of care in hospitals using instrumental variables. , 1999, Journal of health economics.

[6]  D. Cox,et al.  An Analysis of Transformations , 1964 .

[7]  D. Rivers,et al.  Limited Information Estimators and Exogeneity Tests for Simultaneous Probit Models , 1988 .

[8]  Sangho Moon,et al.  DO HMO PLANS REDUCE HEALTH CARE EXPENDITURE IN THE PRIVATE SECTOR , 2006 .

[9]  Kirsten Axelsen,et al.  Impact of statin copayments on adherence and medical care utilization and expenditures. , 2006, The American journal of managed care.

[10]  N. J. Burnett,et al.  Gender Economics Courses in Liberal Arts Colleges , 1997 .

[11]  J. Terza,et al.  Estimating the moral-hazard effect of supplemental medical insurance in the demand for prescription drugs by the elderly. , 1995, The American economic review.

[12]  S. Ennett,et al.  Controlling for the endogeneity of peer substance use on adolescent alcohol and tobacco use. , 1998, Health economics.

[13]  P. Deb,et al.  Provider choice and use of mental health care: implications for gatekeeper models. , 1998, Health services research.

[14]  Joseph V. Terza,et al.  Estimating count data models with endogenous switching: Sample selection and endogenous treatment effects , 1998 .

[15]  Estimating endogenous treatment effects in retrospective data analysis. , 1999, Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research.

[16]  Lung-fei Lee,et al.  Identification and Estimation in Binary Choice Models with Limited (Censored) Dependent Variables , 1979 .

[17]  E. Norton,et al.  Inter-vivos Transfers and Exchange , 2006 .

[18]  Whitney K. Newey,et al.  Efficient estimation of limited dependent variable models with endogenous explanatory variables , 1987 .

[19]  J. Hausman.  Specification tests in econometrics , 1978 .

[20]  Donald R. Williams,et al.  Labor Market Effects of Marijuana and Cocaine Use among Young Men , 1992 .

[21]  Jeffrey M. Woodbridge.  Econometric Analysis of Cross Section and Panel Data , 2002 .

[22]  Jeffrey S. DeSimone.  Illegal Drug Use and Employment , 2002, Journal of Labor Economics.

[23]  Controlling for systematic selection in retrospective analyses: an application to fluoxetine and sertraline prescribing in the United Kingdom. , 1999, Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research.

[24]  R. Dunn,et al.  Fluoxetine and dothiepin therapy in primary care and health resource utilization: Evidence from the United Kingdom. , 1999, International journal of psychiatry in clinical practice.

[25]  C. Propper,et al.  Maternity Rights and Mothers' Return to Work , 2008 .

[26]  S. Ettner,et al.  Differences between generalists and mental health specialists in the psychiatric treatment of Medicare beneficiaries. , 1999, Health services research.

[27]  P. Zarembka.  Frontiers in econometrics , 1973 .

[28]  P. Alexandre,et al.  Illicit Drug Use, Employment, and Labor Force Participation , 2001 .

[29]  Marshall K. Gramm.  The Case for Regulatory Rent-Seeking: CRA Based Protests of Bank Mergers and Acquisitions , 2003 .

[30]  Jeffrey M. Wooldridge,et al.  Quasi‐Likelihood Methods for Count Data , 2008 .

[31]  Kenneth Train,et al.  Control Function Corrections for Unobserved Factors in Differentiated Product Models , 2006 .

[32]  W. Newey,et al.  Large sample estimation and hypothesis testing , 1986 .

[33]  J. Terza.  Estimation of policy effects using parametric nonlinear models: a contextual critique of the generalized method of moments , 2006, Health Services and Outcomes Research Methodology.

[34]  L. Hansen.  Large Sample Properties of Generalized Method of Moments Estimators , 1982 .

[35]  J. Cawley.  An instrumental variables approach to measuring the effect of body weight on employment disability. , 2000, Health services research.

[36]  J. Mullahy.  Instrumental-Variable Estimation of Count Data Models: Applications to Models of Cigarette Smoking Behavior , 1997, Review of Economics and Statistics.

[37]  J. Terza,et al.  Estimating the effects of prescription drug coverage for Medicare beneficiaries. , 2007, Health services research.

[38]  J. Terza,et al.  Assessing the impact of drug use on hospital costs. , 2009, Health services research.

[39]  Jeffrey M. Wooldridge,et al.  Some Alternatives to the Box-Cox Regression Model , 1992 .

[40]  Richard Blundell,et al.  An Exogeneity Test for a Simultaneous Equation Tobit Model with an Application to Labor Supply , 1986 .

[41]  J. Terza.  Alcohol abuse and employment: a second look , 2002 .

[42]  D. McFadden.  Conditional logit analysis of qualitative choice behavior , 1972 .

[43]  T. Mcguire,et al.  The Productivity of Outpatient Treatment for Substance Abuse , 2002 .

[44]  M. Hashem Pesaran,et al.  Handbook of Applied Econometrics Volume II: Microeconomics , 1999 .

[45]  M. Fox.  Medical student indebtedness and the propensity to enter academic medicine. , 2003, Health economics.

[46]  A Test of the Structure of PAC Contracts: An Analysis of House Gun Control Votes in the 1980s , 2000 .

[47]  D. Howard,et al.  The impact of waiting time on liver transplant outcomes. , 2000, Health services research.

[48]  Donald J. Wright,et al.  Moral hazard and adverse selection in Australian private hospitals: 1989-1990. , 2003, Journal of health economics.

[49]  Joseph V. Terza,et al.  Parametric Nonlinear Regression with Endogenous Switching , 2009 .

[50]  A. Goldberger.  A course in econometrics , 1991 .

[51]  Richard Blundell,et al.  Simultaneous Microeconometric Models with Censored or Qualitative Dependent Variables , 1993 .

[52]  Richard Blundell,et al.  Estimation in a class of simultaneous equation limited dependent variable models , 1989 .

[53]  Joseph C. Gardiner,et al.  Testing and Correcting for Non-Random Selection Bias due to Censoring: An Application to Medical Costs , 2003, Health Services and Outcomes Research Methodology.

[54]  W. Greene,et al.  计量经济分析 = Econometric analysis , 2009 .

[55]  Harry H. Kelejian,et al.  Two-Stage Least Squares and Econometric Systems Linear in Parameters but Nonlinear in the Endogenous Variables , 1971 .

[56]  B. Weisbrod,et al.  Do Religious Nonprofit and For-Profit Organizations Respond Differently to Financial Incentives? The Hospice Industry , 2007, Journal of health economics.

[57]  Donald Kenkel,et al.  The effect of physician advice on alcohol consumption: count regression with an endogenous treatment effect , 2001 .