The objective of this project is to study the willingness to pay (WTP) for health insurance (HI) of individuals working in the informal sector in Sierra Leone, using a purposely-designed survey of a representative sample of this sector. We elicit the WTP using the Double-Bounded Dichotomous Choice with Follow Up method. We also examine the factors that are positively and negatively associated with the likelihood of the respondents to answer affirmatively to joining a HI scheme and to paying three different possible premiums, to join the HI scheme. We additionally analyze the individual and household characteristics associated with the maximum amount the household is willing to pay to join the HI scheme. The results indicate that the average WTP for the HI is 20,237.16 SLL (3.6 USD) per adult but it ranges from about 14,000 SLL (2.5 USD) to about 35,000 SLL (6.2 USD) depending on region, occupation, household and respondent characteristics. The analysis of the maximum WTP indicates that living outside the Western region and working in farming instead of petty trade are associated with a decrease in the maximum premium respondents are WTP for the HI scheme. Instead, the maximum WTP is positively associated to being a driver or a biker; having secondary or tertiary education (as opposed to not having any); the number of pregnant women in the household; having a TV; and, having paid for the last medical requirement. In summary, the various analyses show that a premium for the HI package could be set at approximately 20,000 SLL (3.54 USD) but also that establishing a single premium for all individuals in the informal sector could be risky. The efficient functioning of a HI scheme relies on covering as much of the population as possible, in order to spread risks and make the scheme viable. The impact of the various population characteristics raises the issue of how to rate premiums. In other words, setting a premium that may be too high for a big proportion of the population could mean losing many potential enrollees and might have viability consequences for the operation of the scheme.
The data were obtained by running a Discrete Choice Experiment (DCE) conducted in the Northern and Western regions in 2013. The DCE was used to learn households' preferences regarding a HIS by allowing them to choose to participate or not in a set of HISs described by their distinct attributes and their levels, including cost. Statistics Sierra Leone (SSL) designed the sample and selected locations for this study based on recent pre-census data containing settlement names, population, and household sizes. A two-stage stratified random sampling method was used to identify the households. The first stage involved stratifying the population by region/district, and the second by rural and urban location in each district. The purpose was to have a representative sample of informal sector households in both villages (rural areas) and major towns (urban areas). Since the decision to purchase health care is more often a household decision rather than an individual one, the household was chosen as the unit of reference. SSL calculated the sample size of 1670, considering an informal sector population of 1,380,110 individuals, with the objective to achieve a 95% confidence interval around participation in the HIS and a margin of error of 2.4%. Due to failure to pass the dominance test (understanding of the questions) and incomplete data during the questionnaire administration, 1,458 households' data were used for the final analysis, which provided 39,366 observations for the analysis of nine choice sets with three alternatives.
