John Mullahy

Estimating Log Models: To Transform or Not to Transform?

Health economists often use log models to deal with skewed outcomes, such as health utilization or health expenditures. The literature provides a number of alternative estimation approaches for log models, including ordinary least-squares on ln(y) and generalized linear models. This study examines how well the alternative estimators behave econometrically in terms of bias and precision when the data are skewed or have other common data problems (heteroscedasticity, heavy tails, etc.). No single alternative is best under all conditions examined. The paper provides a straightforward algorithm for choosing among the alternative estimators. Even if the estimators considered are consistent, there can be major losses in precision from selecting a less appropriate estimator.

Specification and testing of some modified count data models

Abstract This paper explores the specification and testing of some modified count data models. These alternatives permit more flexible specification of the data-generating process (dgp) than do familiar count data models (e.g., the Poisson), and provide a natural means for modeling data that are over- or underdispersed by the standards of the basic models. In the cases considered, the familiar forms of the distributions result as parameter-restricted versions of the proposed modified distributions. Accordingly, score tests of the restrictions that use only the easily-computed ML estimates of the standard models are proposed. The tests proposed by Hausman (1978) and White (1982) are also considered. The tests are then applied to count data models estimated using survey microdata on beverage consumption.

Much ado about two: reconsidering retransformation and the two-part model in health econometrics.

In health economics applications involving outcomes (y) and covariates (x), it is often the case that the central inferential problems of interest involve E[y/x] and its associated partial effects or elasticities. Many such outcomes have two fundamental statistical properties: y > or = 0; and the outcome y = 0 is observed with sufficient frequency that the zeros cannot be ignored econometrically. This paper (1) describes circumstances where the standard two-part model with homoskedastic retransformation will fail to provide consistent inferences about important policy parameters; and (2) demonstrates some alternative approaches that are likely to prove helpful in applications.

Instrumental-Variable Estimation of Count Data Models: Applications to Models of Cigarette Smoking Behavior

As with most analyses involving microdata, applications of count data models must somehow account for unobserved heterogeneity. The count model literature has generally assumed that unobservables and observed covariates are statistically independent. Yet for many applications this independence assumption is clearly tenuous. When the unobservables are omitted variables correlated with included regressors, standard estimation methods will generally be inconsistent. Though alternative consistent estimators may exist in special circumstances, it is suggested here that a nonlinear instrumental-variable strategy offers a reasonably general solution to such estimation problems. This approach is applied in two examples that focus on cigarette smoking behavior.

Employment, Unemployment, and Problem Drinking

The misuse of alcohol is estimated to result in enormous economic costs, composed largely of reduced labor market productivity. However, there has been debate on this issue. The purpose of this paper is to help to resolve this debate by presenting sound structural estimates of the relationship between measures of problem drinking and of employment and unemployment. The analysis is based on the 1988 Alcohol Supplement of the National Health Interview Survey. We find that for both men and women, problem drinking results in reduced employment and increased unemployment.

Alcoholism, Work, and Income

This article reports on an empirical analysis of the relationships between alcoholism and income and working. We show that the relationships between alcoholism and labor market success have important age or life-cycle dimensions. We present evidence that alcoholism may affect income more by restricting labor market participation than by affecting the wages of workers. Finally, we demonstrate that the effects of alcoholism on earnings depend on the extent to which one controls for other covariates associated with alcoholism; as such, we suggest that there may be important indirect as well as direct effects of alcoholism on labor market success.

Heterogeneity, Excess Zeros, and the Structure of Count Data Models

This paper demonstrates that the unobserved heterogeneity commonly assumed to be the source of overdispersion in count data models has predictable implications for the probability structure of such mixture models. In particular, the common observation of excess zeros is a strict implication of unobserved heterogeneity. This result has important implications for using count model estimates for predicting certain interesting parameters. Test statistics do detect such heterogeneity-related departures from the null model are proposed and applied in a health-care utilization example, suggesting that a null Poisson model should be rejected in favour of a mixed alternative.

Variation in Chinese population health related quality of life: results from a EuroQol study in Beijing, China.

The purpose of this study is to measure Chinese population health related quality of life (HRQoL) using European quality of life (EQ-5D) instrument, to examine the validity of EQ-5D in measuring Chinese population HRQoL, to explore the relationships between EQ-5D and other health determinants, and to display the similarities and differences of HRQoL between the Chinese population and the populations of other countries. The data used in this study includes 2994 respondents whose age are 12 years and older, which is from the 2000 Beijing Household Health Survey. Univariate and bivariate analyses have been used to examine the level of HRQoL and the relationships between HRQoL and other variables. Multi-variate analyses have been used to explore the relationships between the EQ-5D Visual Analogue Scale (VAS) and the EQ-5D five dimension indicators. There are four principal findings from this study. First, the EQ-5D instrument is a valid measure for Chinese HRQoL, but with a significant ceiling effect. Second, Pain/Discomfort and Anxiety/Depression are the major Chinese HRQoL problems and the extents of these problems differ in subgroup populations. Third, typically mean scores are lower for older age group; this is observed at lower ages in the Chinese population than in populations from developed countries. Fourth, Chinese HRQoL has strong association relationship with population socio-economic status (SES), which might imply that issues brought on by the rapid economic transition have both positive and negative impacts on Chinese HRQoL.

Factors influencing decisions regarding influenza vaccination and treatment: a survey of healthcare workers.

Surveys conducted in our healthcare facility evaluated factors associated with acceptance of influenza vaccination and opinions regarding influenza prevention and treatment and willingness to pay. Avoiding lost work and low risk were primary reasons for vaccine recipients and non-recipients, respectively. One-third of vaccine recipients would refuse vaccination if asked to pay at least

Econometric Modeling of Health Care Costs and Expenditures: A Survey of Analytical Issues and Related Policy Considerations

10.