Effects of Income on Health Expenditure in Iran Using Logistic Smooth Transition Regression Model

This paper investigates non linear effects of income on health expenditure (as out of pocket payment). The explanatory variables in health expenditure equation include urbanization, age, insurance, employment of household's head, the number of people employed, earned income and the number of persons aged fewer than 12 and over 60 years in the household. Database used was chosen from the household income & expenditure surveys in 1386. Database contains 31277 rural and urban households. Linear and non-linear econometric models based on logistic smooth transition regression (STR) are estimated. One of the most important findings is that health spending shows asymmetric behavior relative to household income or total expenditure (as a transition variable) so that the impact of explanatory variables on health expenditure in the high and low income regime is different. Also the results show that health expenditure is a necessary good for households, because the coefficient of the logarithm of income in the demand function is positive and much smaller than unity. In addition, out of pocket payments for health care services in high income groups are more than low income ones. Based on the results, with increased urbanization and for high aged people (beyond sixty years of old), health expenditure increase but if family members are covered by insurance, out of pocket payment for health expenses will reduce. So, it is important to develop the social security system and increase its efficiency to cover more health expenditure, especially for poorer households which lead to justice in health areas and increase in welfare of society.