Under some circumstances, heterogeneity in harm from a disease, learned after a disease is contracted, can lead revenue from a treatment to exceed revenue from a preventative. Calibrations suggest that skewness in the U.S. distribution of HIV risk would lead firms to earn only half the revenue from a vaccine as from a drug. Empirical tests are consistent with the predictions of the model that vaccines are less likely to be developed for diseases with substantial disease-risk heterogeneity.
Many public-health and industry experts believe that firms earn more revenue from disease treatments, such as drugs, than preventives, such as vaccines (see, e.g., Rosenberg 1999), and that stronger government support is needed for the development of preventive health technologies. These views are especially common in the case of HIV (human immunodeficiency virus) (see, e.g., Thomas 2002), and indeed governments have established special programs to support HIV vaccine research such as the International AIDS Vaccine Initiative (IAVI).
In this paper, we argue that time-varying consumer heterogeneity can drive a wedge between relative private and social incentives to invest in preventives and treatments. We showthat heterogeneity among consumers in disease risk will limit firms’ ability to extract consumer surplus from preventives, biasing firms’ R&D incentives away from preventives toward treatments compared to a social planner’s. Heterogeneity in harm from infection, on the other hand, can lead to the opposite bias, toward preventives. The model also suggests when these biases are likely to be quantitatively important. Firms’ bias against investing in preventives will be strongest for diseases with skewed distributions of disease risk (i.e., with high risk concentrated in a small segment of the population). Common diseases cannot exhibit much skewness as a mathematical principle, so the bias toward preventives has the most scope to affect R&D incentives for relatively rare diseases with risk heterogeneity. Calibrations for HIV suggest that the rates of partner change and other risk factors are sufficiently skewed that the bias against HIV preventives may be substantial. Empirically, we find that the relative likelihood of developing treatments compared to preventives is greater for diseases with heterogeneity in disease risk.
To see why heterogeneity in disease risk can lead to a bias against preventives, consider the following numerical example. Suppose a monopoly pharmaceutical manufacturer sells directly to 100 rational, riskneutral consumers, who will suffer harm quantified as $100 from contracting the disease. The firm can develop a treatment or a preventive; both are costless to manufacture, are perfectly effective, and have no side effects. Suppose first that the consumers are homogeneous, having the same 19% risk of contracting a disease. If the firm develops a treatment, it can sell to all people who contract the disease at a price (equal to the avoided harm) of $100. Expected treatment revenue is $1,900 because an expected 19 consumers contract the disease and buy the treatment. If the firmdevelops a preventive, it could sell to all 100 consumers at a price, equal to consumers’ expected harm of $19, for total revenue of $1,900. With homogeneous consumers, the firm’s expected revenue is $1,900, which represents full extraction of consumer surplus in the market, whether it sells a treatment or preventive.
Consider the same example except suppose now that consumers are heterogeneous in disease risk, with 90 having a 10% chance of contracting a disease while 10 have a 100% chance. Because the number of people expected to contract the disease is the same 19 as in the homogenous-consumer case, expected treatment revenue remains the same at $1,900. On the other hand, the firm’s revenue from a preventive falls.
The firm can either sell to the 10 high-risk consumers at their expected harm of $100, or sell to all consumers.