This study investigated income elasticity of household health expenditures and differences by income level from 1998 through 2003. Data from Korean Labor and Income Panel Study was used for empirical analyses. To estimate the income effects on health expenditure, the two-part model was employed: a logistic regression for any health expenditure-first part-and a Ordinary Least Square regression for health expenditure conditional on any spending-second part. To estimate income elasticity, both health expenditure and income were log transformed in the second part. In addition, the random effects(RE) model was used for a longitudinal panel which was continuously followed from 1998 through 2003 to estimate income effects on health expenditures controlling for within and between unobservable household characteristics. Furthermore, difference in income effects on health expenditure across income level was investigated. Although income slightly increased odds of any health expenditure, there was not no table differences across income level. Income significantly increased health expenditures during study period(overall income elasticity: about 0.2) and the highest 20% income group presented higher income elasticity than the lowest 20% income group.