VII. METHODOLOGICAL CHALLENGES
Estimating the elasticity of demand for health insurance and health care is not a simple task.
Given the variety of factors involved in consumers’ decision-making process and the complexity
of the health care market, isolating the change in demand that can be attributed only to a change
in price or income is exceedingly difficult. This chapter highlights four common methodological
challenges in estimating the price or income elasticity of demand for health insurance or health
care services: unobserved price, endogeneity, omitted variables, and provider-induced demand.
We describe the impact of each problem on estimates of price elasticity and summarize the
approaches that researchers have taken to address them.
A. UNOBSERVED PRICE
Studies that examine the own-price elasticity of demand must consider not only the price of
insurance or care that is purchased, but also the price when it is not purchased. However,
surveys typically do not ask respondents (employers or employees) about the price they were
offered, if they ultimately did not buy coverage or care. For the purpose of estimating price
elasticity, the problem with such data is that observed prices are not random. For example, firms
that offer coverage to their employees may systematically have been offered lower premiums. If
so, data from firms that offer coverage cannot be used to accurately predict premiums for firms
that do not.
A simple approach to address this problem, often used in early studies, is to include in the
model both types of firms (with observed and unobserved prices) and estimate a Tobit model that
adjusts for the left-censored observations. However, the key assumption of a Tobit model is that
the same factors affect both the decision to offer coverage and the price of that coverage. Some
researchers have challenged this assumption and advocate use of a sequential decision model