How can we reduce the impacts of serious illnesses on family consumption?

Responsible researcher: Viviane Pires Ribeiro

Paper Title: Insuring Consumption Against Illness

Authors: Paul Gertler and Jonathan Gruber

Intervention Location: Indonesia

Sample Size: 3,933 households

Big topic: Health

Variable of Main Interest: Health insurance

Type of Intervention : Government measures to mitigate the impacts of serious illnesses on family consumption

Methodology: Fixed effects models

across two islands: east of Bali; and East Kalimantan (KalTim), located on the east coast of the island of Borneo. Together, they represent around six million inhabitants. KalTim has the third highest per capita among all 27 provinces, while NB is in twenty-second place.

Methodology Details

The sample used by Gertler and Gruber (2002) was collected for the period 1991 and 1993, allowing us to examine changes in health, income and consumption over a two-year period. Data are from a stratified random database of households, consisting of approximately 20 households per village (216 villages). They were collected for each household at the same point of the year, so that the effects of seasonality were conditioned in the fixed effects model. The response rate in the first round of the survey was relatively higher at 91%, and the attrition between the first and second rounds was low at about 7%.

The sample is a panel of 3,933 households, consisting of all households that participated in the survey in both rounds, whose head from the first round was in the second round sample, and which have non- proxy and complete data on health measures. To this end, the IRMS questionnaire was used, which was developed based on a detailed list: labor supply, consumption and existing health research modules, and was increased when found to be incomplete. The research team used focus groups and extensive pilot testing to ensure the questions fit the cultural context. To minimize measurement error, all adults in the household were interviewed directly, rather than interviewing one individual and using proxy responses for the remainder.

The authors specified and implemented a reduced-form model of households' abilities to insure consumption against disease. The model is a fixed effects specification and as such controls for unobserved heterogeneity. In particular, the first difference eliminated the correlation of omitted unobserved individual characteristics (such as preferences and health resources) that confound the identification of the effect of illness on labor market outcomes. We also controlled for an important source of spurious correlation, shocks to the local community economy such as weather that affect both changes in permanent income and changes in health, including a set of community fixed effects.

Results

Using reliable and valid measures of health problems that distinguish between varying degrees of severity, Gertler and Gruber (2002) identified that Indonesian families are not able to insure the economic costs of serious illnesses. It is estimated that 35% of these costs are not covered by other sources available to these families. It was also observed that the more serious the disease, the less families are able to purchase a health plan. However, families can fully insure the economic costs of illnesses that do not affect physical functioning. Furthermore, they can cover 71% of the costs resulting from illnesses that moderately limit an individual's physical capacity, but only 38% of the costs of illnesses that severely limit physical capacity.

The analysis indicates that the disease is associated with a drop in consumption of 0.84 percent from baseline. This is a non-trivial effect given the low frequency of serious illnesses that cause serious health limitations. Furthermore, this underestimates the total cost of lost well-being that is associated with illness, for at least two reasons. First, there are additional welfare costs arising from the variability of uninsured consumption in addition to the reduced level of consumption. Second, there are costs to those resources used to smooth consumption when family members become ill.

Therefore, it was found that the ability of families to take out insurance drops drastically with the severity of the disease shock. That is, families are able to cover less than 40% of the loss of income due to illnesses associated with a very serious loss of physical capacity.

Public Policy Lessons

Although the results obtained from the study carried out by Gertler and Gruber (2002) indicate that families are able to cover the costs of frequent and less serious illnesses, they also suggest that families are unable to cover the costs of rare and serious illnesses. Which indicates that there may be a significant welfare cost to increasing user rates in public hospitals in order to shift subsidies to primary and preventive care (which is affordable).

So governments considering increasing hospital user rates should consider how to secure the costs of healthcare for serious illnesses, for example, caps on hospital admission rates or planning pre-payment schemes in conjunction with reducing subsidies.

The authors cite previous studies that suggest another form of social insurance in developing countries, which is to finance public health care through payroll taxes, and; allow beneficiaries to purchase health care from private providers. In other words, the argument is that low-income countries have restricted abilities to tax, so the resources available for social insurance are severely limited. This results in a trade-off between coverage that only covers serious illnesses, with high coverage, but unlimited; and coverage that covers all expenses starting from the first dollar, but a low ceiling on total expenses covered. Other studies highlight that many low-income countries prefer the latter strategy, providing minimum benefits for all diseases rather than full insurance for rare and high-cost diseases.

According to Gertler and Gruber (2002), this choice occurs out of concern that low-income groups will not be able to “pay” the deductible and, therefore, would not benefit from the plan. However, if families can insure small health shocks, consequently benefits limited to a first dollar have little contribution to increasing insurance.

Regarding formal disability insurance, the authors suggest that there are gains from the introduction of this plan in countries such as Indonesia. For in Indonesia and (possibly) many other developing countries, most of the cost of illness is associated with lost income rather than spending on health care. However, in developed countries, publicly funded disability insurance programs can be quite costly in terms of administrative costs, moral hazard and marginal cost.

References

GERTLER, Paul; GRUBER, Jonathan. Insuring consumption against illness. American economic review , vol. 92, no. 1, p. 51-70, 2002.