This study attempts to find out the determinants of household healthcare expenditure. The analysis uses a multi-equation recursive estimation procedure to study the determinants of healthcare expenditure. First, it uses a binary logit model to estimate the probability of being ill, which is then used as an independent variable in the second stage logit model for provider choice. Ordinary Least Squares (OLS) estimates are obtained instead of Tobit estimates for the parameters of healthcare expenditure model in the third stage. This study brings out several interesting findings. First, the level of income has a significant effect on peoples' choice of healthcare provider and on the amount of healthcare expenditure. Second, people with smoking habits and less access to safe drinking water and sanitary toilets are more vulnerable to diseases. Some of the factors ordinarily believed to affect the incidence of diseases like, education level of an individual, education level of the household head, and being male member of the family have a less pronounced effect on reporting illness and healthcare expenditure.

Analysis of expenditure for healthcare is extremely important for formulating policies and strategies for the health sector. In any developed or developing society, health-sector policies are set out for achieving five objectives: (1) To ensure provision of basic healthcare to the entire population; (2) To increase the allocation efficiency of resources of the sector; (3) To improve the quality of services; (4) To enhance equity of access to public facilities; and (5) To contain the costs of providing health services. Appropriate strategies required for achieving these objectives cannot be devised without adequate knowledge about the extent, determinants and elasticity of healthcare expenditure at the household level.

Analysis of expenditure for healthcare is especially important in a poverty-stricken country like Bangladesh, which is constantly striving to concomitantly accomplish financial sustainability and universal coverage of health services as rapidly as possible. To perform such tasks, one requires quick expansion of coverage, improvement of the quality of services and mobilization of sufficient funds from the public sector and households through adoption of various financing strategies, such as introduction of users' fees, implementation of insurances schemes, and community financing. Analysis of expenditure is also crucial in designing financing strategies.

Some scholars have long been admonishing that demand for healthcare is predominantly supplier-induced, because in the health sector, agents or health service providers themselves are also the suppliers of services and therefore, demand function may be distorted or truncated. Some argue that a demand function does not exist at all in the health sector. Despite these arguments, health sector demand functions exist independent of supply functions, that can be estimated if appropriate efforts are made to closely observe household behavior and carefully analyze the views of individuals. The estimates of demand function can then be effectively used in formulating national healthcare policies.

Applied Economics Journal, Ordinary Least Squares, OLS, Financial Sustainability, Financing Strategies, Poverty Reduction Strategy Papers, PRSP, Health and Population Sector Program, HPSP, Research Methodology, Healthcare Expenditure, Macro-Economy.