Assessing the effectiveness of COVID-19 vaccine lotteries: A cross-state synthetic control methods approach

Abstract

Vaccines are the most effective means at combating sickness and death caused by COVID-19. Yet, there are significant populations within the United States who are vaccine-hesitant, some due to ideological or pseudo-scientific motivations, others due to significant perceived and real costs from vaccination. Given this vaccine hesitancy, twenty state governors from May 12th to July 21st 2021 implemented some form of vaccination lottery aiming to increase low vaccination rates. In the aftermath of these programs, however, the critical question of whether these lotteries had a direct effect on vaccination remains. Previous literature on financial incentives for public health behaviors is consistent: Financial incentives significantly increase incentivized behaviors. Yet, work done specifically on state vaccine lotteries is both limited in scope and mixed in its conclusions. To help fill this gap in the literature, we use synthetic control methods to analyze all 20 states and causally identify, for eighteen states, the effects of their lotteries on both first-dose and complete vaccination rates. Within those eighteen states, we find strong evidence that all but three states' lotteries had positive effects on first-dose vaccination. We find for complete vaccinations, however, over half the states analyzed had negative or null effects. We explore possibilities related to these mixed results including the states' overall partisanship, vaccine hesitancy, and the size of their lotteries finding null effects for each of these explanations. Therefore, we conclude that the design of these programs is likely to blame: Every state lottery only incentivized first-doses with no additional or contingent incentive based on a second dose. Our findings suggest that the design of financial incentives is critical to their success, or failure, but generally, these programs can induce an uptake in vaccination across diverse demographic, ideological, and geographic contexts in the United States.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.