Spillover Effects of Financial Incentives for COVID-19 Vaccination: A Population-level Cluster Randomized Controlled Trial

Importance: Prior studies show that financial incentives have small positive direct effects in increasing COVID-19 vaccinations, however estimates of social and temporal spillovers are almost entirely absent from the existing literature. Objective: Building on a previously-successful COVID-19 financial incentive, we sought to assess whether incorporating estimates of social and temporal spillover in a research design meaningfully affects big-picture evaluations of policy effectiveness. Design: In a population-level, address-cluster RCT, participants were randomized to treatment and control, with one resident in each address-cluster randomly selected to be an address-cluster representative. The address-cluster representatives in treatment addresses received the treatment letter; all others received the control letter. Setting: Late 2021, Ravensburg, Germany. Participants: All adult residents of Ravensburg, Germany: 10,032 address-clusters or 41,548 residents in total. Intervention: The control letters informed recipients about seven upcoming free COVID-19 vaccination events. The treatment letters were identical to the control letters, except they also offered recipients 40 Euros should they get vaccinated at one of the vaccination events. Main Outcome(s) and Measure(s): Primary and booster vaccinations were recorded during the events. Results: We estimated the direct effect of the treatment letter on the recipient, the social spillover effect (i.e., impact on the cohabitants of treatment letter recipients), and the overall (address-level) effect. The direct, spillover, and overall effects on primary vaccinations were all indistinguishable from zero (0.1pp, 95% CI [-0.22, 0.33], p=.69; 0.0pp, [-0.22, 0.26], p=.86; 0.0pp, [-0.17, 0.23], p=.76, respectively). For boosters, the direct and overall effects were negative (-0.3pp, [-0.77, 0.14], p=.17; -0.3pp, [-0.51, -0.09], p<.01, respectively), and the spillover effect was negative and statistically significant (-0.3pp, [-0.53, -0.06], p=.01). Conclusions and Relevance: Financial incentives for vaccination can have negative impacts on those who don’t receive the incentives. Based on the timing of the effect, there is suggestive evidence that individuals who resided with a recipient of the financial incentive chose to postpone booster vaccination. Trial Registration: DOI:10.1186/ISRCTN59503725