College of Business and Behavioral Science


Thomas Mroz


Office: 200C Sirrine Hall
Phone: 656-3142
Personal Website:

 Educational Background

1984 Economics, minor Statistics
Stanford Univiersity Ph.D

1976 Economics
Vassar College AB

 Courses Taught

Economics 816 Labor Economics
Economics 981
Economics 982
Economics 406
Economics 606
Economics 807


Thomas Mroz’s current research explores a variety of topics in labor, health, development, and econometrics. One project with Sergey Mityakov examines how our inability to measure all of the chosen inputs to any household production function should affect the interpretation of estimated effects. Under some plausible conditions, estimable effects can provide bounds on true marginal effects, but the bounds can sometimes be non-informative. That project also explores the precise forms of control variables that one needs to use in an empirical production function specification for the estimated effects to have a valid economic interpretation. A surprising result is that no existing study of household production functions, in the ubiquitous case where there are unmeasured chosen inputs, has used an empirical specification compatible with an economic model of optimizing individuals. Another set of projects explores how best to use imperfect information about outcomes and control variables when estimating economic relationships. In a study of the impacts of untreated diabetes on diabetes progression, amputations and death, he and co-authors Gabriel Picone, Frank Sloan and Arseniy Yashkin find that the longer an individual has undiagnosed diabetes the more likely would be diabetic complications including amputations and death. More frequent screening for diabetes among the elderly significantly reduce the duration of time a diabetic individual would be undiagnosed and would result in economically significant and cost-effective reductions in medical costs and increases in the value of life. Another project, with Gou (Grace) Li, examines how best to use imperfect information about explanatory variables to uncover better estimators of the impacts of those variables. For example, oftentimes researchers only know limited information about an individual’s geographic location (e.g., region, state, or county), but are interested in how more specifically geographically-targeted factors (e.g. at the town or school district level) affect outcomes of interest. Commonly used approaches that consider only the effects using geographically averaged factors perform quite poorly. Maximum likelihood approaches that explicitly incorporate the implicit aggregation in the available data set often perform quite well, even in comparison to hypothetical situations where perfect geographically targeted information is available. Other work in progress, with Paul Wilson, examines how best to select the functional form of an econometric model when one is concerned about the estimation of effects rather than choosing models based upon more abstractly defined measure of goodness of fit. Standard model selection procedures such as BIC and F-tests, for example, often select models that fail to capture important features of economic relationships. Bayesian approaches appear to provide better inferences about meaningful effects. Recently, he and Sergey Mityakov and Kevin Tsui have begun to examine how one can use detailed geographically aggregated information to construct instrumental variables to explore the impacts of becoming unemployed on health and family well-being. Preliminary results suggest that the existing literature examining the effects of unemployment on alcohol abuse outcomes has significantly underestimated the adverse effects of becoming unemployed.

 Research Publications

“The declining prevalence of overweight among Russian children: Income, diet, and physical activity behavior changes,” with Lisa Jahns, Linda Adair, and Barry Popkin, Economics and Human Biology Vol. 10, pp. 139-146, 2012.

“Caregiving and the Elderly’s Health in Mexico,” with Antonio J. Trujillo, Claudia Piras, Nhan T. Tran, ,and Gustavo Angeles, International Journal of Health Services, Vol. 42, No. 4, pp. 667-694, 2012.

“A Simple, Flexible Estimator for Count and Other Ordered Discrete Data,” Journal of Applied Econometrics, Vol. 27, pp. 646-665, 2012.

“Determinants of Gender Differences in Health among the Elderly In Latin America,” with Antonio J. Trujillo, Claudia Piras, John A,. Vernon, and Gustavo Angeles, World Health and Population, Vol. 11, No. 3, pp. 24-43, 2010.

“Mother’s Work, Migration, and Child Achievement,” with Haiyong Liu and Wilbert Van der Klaauw, Journal of Econometrics, Vol. 156, No. 1, pp. 212-228, May 2010.

“Heterogeneous Impact of the Social Programme Oportunidades on Contraceptive Methods Use by Young Adult Women Living in Rural Areas,” with Héctor Lamadrid-Figueroa, Gustavo Ángeles, José Urquieta-Salomón, Bernardo Hernández-Prado, Aurelio Cruz-Valdez, and Martha M. Téllez-Rojo, Journal of Development Effectiveness, Vol. 2, Issue 1, pp. 74-86, March 2010.

“Parental Compensatory Behaviors and Early Child Health Outcomes in Cebu, Philippines,” with Haiyong Liu and Linda Adair, Journal of Development Economics, Vol. 90, No. 2, pp. 209-230, November 2009.

“Impact of Oportunidades on Skilled Attendance at Delivery in Rural Areas”, with Jose Urquieta, Gustavo Angeles, Hector Lamadrid-Figueroa, and Bernardo Hernández, Economic Development and Cultural Change, Vol. 57, No. 3, pp. 539-558, April 2009

“Arbitrarily Normalized Coefficients, Information Sets, and False Reports of ‘Biases’ in Binary Outcome Models,” with Yaraslau Zayats, Review of Economics and Statistics, Volume 90, Number 3, pp. 406-413, August 2008.