Effectiveness of Vaccination Recommendations versus Mandates: Evidence from the hepatitis A vaccine

In the United States, immunization rates are persistently low for numerous vaccines, and recently there have been multiple outbreaks of vaccine-preventable diseases resulting from undervaccination. In response, a number of policies have been implemented in an attempt to achieve and maintain high population vaccination rates and reduce disease incidence. In this paper I provide novel evidence on the effectiveness of two vaccination policies – simple non-binding recommendations to vaccinate versus mandates requiring vaccination prior to childcare or kindergarten attendance– in the context of the only disease whose institutional features permit a credible examination of both: hepatitis A.  Using a difference-in-differences strategy that allows me to take advantage of plausibly exogenous variation across states in the timing of the policy introductions, I find that recommendations significantly increased hepatitis A vaccination rates among young children by at least 20 percentage points, while mandates increased rates by another 8 percentage points.  Together these policies explain approximately half of the change in vaccination rates over the sample period. These policies also significantly reduced population hepatitis A incidence, and similarly explain approximately half the change in disease incidence over the period of interest.  These results are robust to the inclusion of both a number of controls for state vaccine- and healthcare-related policies, and state-specific linear time trends. I also demonstrate that these results are not present for other vaccines and diseases, which is strong evidence that my findings are not being driven by unobserved state policies that generally increased vaccination or decreased disease incidence.

This paper further provides evidence that the effects of the recommendations and the mandates differ along several important dimensions. First, my results indicate that the mandates are effective at inducing individuals to complete the vaccine series, while the recommendations only significantly increase the probability that individuals initiate the series. Second, when I allow policy effects to vary over time, I find that individuals who are induced to vaccinate by the mandate respond  rapidly to the policy, whereas the recommendations have a more gradual effect that continues to increase for several years following implementation. Finally, for the mandates I find no significant heterogeneity in the effects of by ethnicity, mother’s education, or family income, although there is some evidence of a heterogeneous policy response for the recommendations. Overall, my results are informative about responses to vaccination recommendations versus mandates for young children, and suggest a range of policy options for addressing suboptimally low population vaccination rates.