Publication Type: Paper/Poster Proposal
Abstract: In the U.S. Consumer Expenditure (CE) Interview Survey, consumer units (roughly equivalent to households) are asked to provide item-level monthly reports of expenditures. For this survey, reporting rates (defined as the proportion of expenditure categories in which a consumer unit reports a purchase in a given month) are potentially useful indicators of data quality. Cho, Eltinge and Steinberg (2004) reported preliminary results on relationships between reporting rates and some characteristics of interviewers and the timing of the interview. The current paper extends this previous work in four ways. First, we present detailed results from logistic regression modeling of reporting rates with predictor variables associated with household characteristics (e.g., family size and composition, as well as prior-interview expenditure reports) as well as variables associated with the interview process (e.g., the number of respondents and the use of records by the respondent during the interview). Second, we supplement standard development and significance-testing results with diagnostics for the predictive power of the resulting models. These diagnostics are especially important for evaluation of the potential practical benefits of modification of field procedures suggested by the modeling results. Third, we extend these results to some specific groups of expenditure categories, and explore differences among the results for these groups. Finally, we compare and contrast these empirical results with previous literature on response and reporting rates in single-wave and panel surveys.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: The paper assesses the impact of fiscal restriction policies specifically Tax and Expenditure Limitations on states' expenditures by function during periods of fiscal stress.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Our main conjecture is that principal-agent problems will prevent most tax and spending limits from having their intended effect of reducing the size of state government. First, we take a two-group design where we compare revenue and expenditures in states that have experimented with TELs to a composite of other states that never enacted a TEL. Second, we employ an event study, or an interrupted time series design, where we examine each state one at a time, and test expected spending after a TEL to actual spending. Third, we explore why TELs don’t work, looking at changes in debt and shifts to charges and fees that occur following the enactment of a TEL. Our results do not prove sanguine, and support our conjecture.

Publication Type: Abstract
Abstract: Despite the fact that Brazil has a health system with universal access and care free of charge at point of delivery, several previous studies have shown that Brazil has a high number of households with large health expenditures. The primary objective of this research is to assess the effectiveness of the public health system and private health insurance in providing financial protection in health care to the Brazilian population.
This research utilizes data from a large national survey of households. The determinants of catastrophic health expenditures (CHE) were estimated by a probit regression with a Heckman selection adjustment, controlling for the need for health care.
According to our findings, the Brazilian health system provides a significant reduction (41%) in the probability that a household has CHE. On the other hand, private health insurance did not reduce out of pocket expenditures and, given that the premiums are quite expensive, made CHE more likely – by 37%.
Recommendations include further improvements in the quantity, accessibility, quality and reliability of public health providers and more extensive and appropriate provision of medicines by the public system. A tighter regulation of private health plans is also necessary, both on the high cost of the insurance premiums and on the restricted nature of the benefits.

Publication Type: Abstract
Abstract: Rationale: Health care expenditure regressions are used in a wide variety of economic analyses including risk adjustment and program and treatment evaluations. Two recent articles have demonstrated that generalized gamma models with heteroskedasticity (GGM-het) and extended general linear models (EGLM) provide flexible approaches to deal with a variety of data problems commonly encountered in expenditure estimation. To date, however, there have been few empirical applications of these models to expenditures.

Objective: We use nationally representative data from the first six panels of the U.S. Medical Expenditure Panel Survey (MEPS) to compare the bias and predictive accuracy of GGM-het and EGLM models with other regression models in a cross-validation study design.

Methodology: We estimate models of prescription drug, ambulatory and total health care expenditures conditional on having any expenditure. Models are estimated separately for the elderly and other privately insured adults. Since expenditure distributions vary by type of service and population, the appropriate functional form is also likely to vary. In estimating expenditures, we focus on two recently developed modeling approaches that flexibly accommodate skewness, kurtosis, heteroskedasticity and other data problems. The GGM-het model, proposed by Manning, Basu, and Mullahy (2005), uses a log-link like many standard GLM models. However, the GGM-het model is more flexible than standard models because the generalized gamma distribution has a scale parameter and two shape parameters and variance is explicitly modeled as a function of explanatory variables. In the EGLM model, proposed by Basu and Rathouz (2005), the link function is not specified prior to estimation. Instead, both the link and variance functions are simultaneously estimated along with the coefficients.

Our models use socioeconomic characteristics and condition information from the first year of each MEPS panel to predict expenditures in the second year. We use a split-sample cross validation design to compare results from GGM-het, EGLM, log OLS with heteroskedasticity (log-het), linear OLS, Poisson and Gamma models. We use the validation sample to test for over-fitting and to examine predictive ratios and mean prediction errors in the entire sample, in the tails of the distribution and for persons with chronic conditions.

Results: In our preliminary analysis we focused on total expenditures and estimated all types of models except EGLM. We found that the expenditure distribution for the elderly was more kurtotic than the distribution for other adults and the distributions varied in the extent of heteroskedasticity beyond simple functions of the mean. Overall, the GGM-het and log-het models fit the data for privately insured adults very well. However, none of our models was clearly superior for the elderly.

Conclusions: Our preliminary analysis confirms that GGM-het models are robust to a wide variety of common data problems. For some distributions, however, an even more flexible estimator, such as the EGLM model, may be required.