Essays on Selection in Health Survey Data

TitleEssays on Selection in Health Survey Data
Publication TypeThesis
Year of Publication2010
AuthorsObrizan, M
Academic DepartmentEconomics
UniversityUniversity of Iowa
CityIowa City, IA
Thesis TypeDissertation
Accession Number1142838
KeywordsEmployment and Labor Force, Health Conditions and Status, Healthcare, Income, Methodology, Other

In this paper I consider the effects of sample selection in a study of stroke occurrence among Medicare-eligible Americans using the survey on Assets and HEAlth Dynamics among the Oldest Old (AHEAD). First of all, I apply the propensity score method based on the assumption of data missing at random and find no substantial effects of the sample selectivity, similar to Wolinsky et al. (2009). Then I employ a weaker assumption of incidental truncation and extend the Heckman's (1979) sample selection model to the case of a discrete outcome variable (whether a stroke occurred after the baseline). This constitutes the main methodological contribution of the current paper, with potential applications in health, labor and applied economics, where the outcome variable is discrete and selection takes place along multiple dimensions. The sample selection model with binary outcome(s) extends the multivariate probit setup in Chib and Greenberg (1998) to allow some of the responses to be missing. I also use Cholesky factorization of the variance matrix to simplify the Gibbs algorithm in Chib and Greenberg (1998). I do not find strong selection effects driven by the data in my application to the stroke equation. The main substantive contribution of the paper is that there is no evidence of selection in the AHEAD data based on either propensity score or sample selection model. Government shares in total output are characterized by significant variation across countries. In this paper I use a neoclassical growth model to show that in equilibrium government shares decline non-linearly in the price of government services relative to private consumption. I calibrate the model to the US economy and show that it generates government shares which closely match those from the Penn World Table. In addition, I provide empirical evidence showing that relative price of government consumption increases in income and that richer countries tend to have smaller governments if measured at international prices. The systematic relationships between factor productivities, output composition and economic development have not been explored in the literature.

Endnote Keywords


Endnote ID


Citation Key6038