Role of risk attitude in the decisions to enroll in micro health insurance plans in a low-income country

Tania Wahed, The University of Texas School of Public Health


Background: Lack of financial resources is a major barrier for access to health care for the poor. A financing system favoring the poor would include prepayment for health care through taxes or insurance with contributions tied to a person’s ability to pay rather than to use of services. Micro health insurance (MHI) was specifically designed for the poor, priced within low-income households’ ability to pay, and to improve access to health care. MHI can play a significant role in promoting the recent Universal Health Care (UHC) initiatives in Bangladesh by providing the foundation on which Governments can promote UHC initiatives. The main weaknesses of micro health insurance schemes are low enrollment and the low level of revenue, which threatens viability of insurance provision. Studies have reported risk attitude as a significant determinant of demand for micro health insurance (30-32, 34-36). A key policy interest is to identify the determinants of demand, emphasizing risk attitude, for micro health insurance programs. Objective: The first aim was to measure risk attitude of the rural people in Bangladesh. The second aim was to estimate the role of risk attitude as a determinant of demand for micro health insurance in Bangladesh. Method: A pilot experiment and a cross sectional survey were undertaken in Chakaria, a rural sub district in Bangladesh, from September, 2015 to June 2016. A micro insurance program, Amader Shasthya, launched in four unions of Chakaria in 2012, offered two options: a basic package which covered outpatient care, and an inpatient package which covered hospitalization costs at a local private clinic. This offered the opportunity to measure risk attitude for both large losses with a small probability (inpatient costs) and smaller losses with higher probabilities (outpatient costs). The sample for the study was drawn using a stratified random sampling design from households surveyed in the Chakaria Health and Demographic Surveillance System (CHDSS). The households were stratified into three groups by enrollment status; i.e. enrolled in the two types of insurance schemes offered and the third not enrolled in either of the schemes. In the pilot experiment, a small sample of 90 randomly selected heads of households, 30 from each enrollment status, were invited to come to a central location at a particular time and date. Risk attitudes of 84 rural respondents were obtained using Dohmen and colleagues’ general and health risk scales and Tanaka, Camerer and Nguyen’s (TCN) paid lottery experiment. The cross sectional study sample included 625 randomly selected heads of households, which included 200 individuals from each of the two insurance groups and 225 households from the non-enrolled group. Risk attitudes of 579 respondents were determined using the simple general and health risk scales developed by Dohmen and colleague. Results: The mean estimated values for risk aversion parameter (σ) and probability weighting function (α) were .57 and .60 respectively. The estimated average value of λ was 4.29 following Tanaka and colleagues’ method. The estimated average value of λ was 6.5 using Liu’s method. From Dohmen’s scale 53% of the respondents were risk averse. Respondents, on average, identified themselves as 6.96 on a scale of 1-11. Both Dohmen’s scale and Tanaka and colleague’s method were valid measures of risk attitude however, Dohmen’s scale was easier to understand and administer and was the preferred measure of risk attitude for the larger cross sectional field survey. From the cross sectional survey risk attitude measured with Dohmen’s general risk scale was not significant at the 5% alpha level (p=.06). The relationship between enrollment and risk aversion was negative for the inpatient package, but positive for the basic package. Trust in the insurer and the insurance product (p<.001) and distance to the nearest facility (p=.048) were significant predictors of micro health insurance demand. A reversal in the direction of effect for the inpatient and basic packages was observed. This can be explained by contextual factors concerning the ‘riskiness’ of the insurance itself. Conclusion: The majority of the rural sample from Chakaria was risk averse with a high level of loss aversion. The risk behavior of the rural population followed Prospect Theory of choice under uncertainty. Both risk attitude measures were valid measures, but Dohmen’s scale was easier to comprehend and use. Trust in the insurer and the insurance schemes and distance to the nearest facility included in the insurance scheme were the only factors significant in the cross sectional analysis of demand for micro insurance. (Abstract shortened by ProQuest.)^

Subject Area

Public health|Health care management

Recommended Citation

Wahed, Tania, "Role of risk attitude in the decisions to enroll in micro health insurance plans in a low-income country" (2016). Texas Medical Center Dissertations (via ProQuest). AAI10249750.