Vaccinations and Welfare
Paper Session
Sunday, Jan. 8, 2023 10:15 AM - 12:15 PM (CST)
- Chair: Brandyn F. Churchill, University of Massachusetts Amherst
Predicting COVID-19 Vaccination Intention: Using A Machine Learning Approach
Abstract
This study examines the factors that influence people’s decision to be vaccinated and their intentions to vaccinate their children. We conducted a cross-sectional online survey that was distributed through the Lumbee Tribe of North Carolina from the end of July 2021 to the beginning of August 2021. The survey questions include sociodemographic questions, questions about people’s testing behaviors, vaccination status and vaccination intentions, COVID-19 knowledge, attitude, and beliefs. We employed four different Machine Learning models and find that the Random Forest algorithm best predicts the self-vaccination intention with a 96% of accuracy and children vaccination intention with a 86.5% of accuracy, respectively. The results also show that testing frequency, social responsibility and paid leave are strong predictors of people’s vaccination decisions for themselves and their children. Policymakers may consider shifting resources to benefit communities and the populations facing frequent testing needs as well as emphasizing the social responsibility of vaccination to amplify interventions in promoting vaccination intake.To Vaccinate or to Wait
Abstract
This paper models individual decisions under irreversibility of the vaccine using real options. Vaccine irreversibility increases the value in waiting to vaccinate and postpone vaccination, even if vaccination has positive net gain. The waiting value magnifies the vaccination cost ex ante in a rational framework. In this framework, we analyze the difference between the reward at vaccination, or equivalent tax on non-vaccinated for increasing the likelihood to vaccinate. For individuals, any subsidy at vaccination to reduce the vaccine cost is more effective on the likelihood to vaccinate relative to any equivalent taxes imposed to increase the cost of no vaccination. The factors that add to the value in waiting and postpone vaccination are increase in the uncertainty about the disease, and likelihood for the expected cost of the infection to go down.Vax Populi: The Social Costs of Online Vaccine Skepticism
Abstract
We quantify the effects of online vaccine skepticism on vaccine takeup and health complications for individuals untargeted by immunization. We collect the universe of Italian vaccine-related tweets for 2013-2018, label anti-vax stances through NLP, and match them with vaccine coverage and Vaccine-preventable hospitalizations at the most granular level (municipality and year). We propose a model of opinion dynamics on social networks that matches the observed data and shows that a vaccine mandate has two relevant effects on takeup: it increases the average vaccination rate, but it also increases the controversialness around the topic, endogenously fueling polarization of opinions among Twitter users. We then leverage the intransitivity in network connections with “friends of friends” to isolate exogenous source of variation for users’ vaccinerelated stances and implement an IV strategy. We find that a 10 pp increase in the municipality’s anti-vax stance causes a 0.43pp decrease in coverage of the Measles-Mumps-Rubella vaccine, 2.1 additional hospitalizations every 100k residents among individuals untargeted by the immunization (newborns, the immunosuppressed, pregnant women) and an excess expenditure of 7,311 euro, representing an 11% increase in health expenses.Additional Lives Saved during COVID-19? How Vaccination Affects Willingness to Go to the Doctor
Abstract
At the beginning of the current pandemic, many individuals who had not yet been vaccinated against COVID-19 skipped or postponed doctor visits for fear of exposure to the disease. This disruption in health care may have had a significant negative impact on their health. To test the hypothesis that vaccination reduced this hesitancy, we employ Census Household Pulse Survey data for January 2021-July 2021, control for selection, time trends, and demographics, and find that medical care avoidance increased with non-vaccinated status. We take this as evidence of additional adverse medical outcomes of the pandemic in addition to the virus itself.Short-run Risk Compensation after COVID-19 Vaccination
Abstract
Vaccination is essential to address public health risks during the COVID-19 pandemic. One concern is risk compensation or the Peltzman effect that people may weaken adherence to social distancing measures after receiving vaccines. If this response is sufficiently elastic, the preventive effect of vaccines in a scale-up setting would be significantly dampened compared to the effect proven in clinical trials. However, there has been little scientific research on risk compensation after COVID-19 vaccination.We study the causal impacts of COVID-19 vaccination on social distancing behaviors. The national vaccination schedule in South Korea, where vaccination eligibility dates differed by age group, provides us a unique opportunity to provide credible evidence on risk compensation. Specifically, we utilize a regression discontinuity design based on the difference in vaccination timing between ages 59 and 60 using exact birth date information.
We use large, high-frequency, administrative data as well as survey data to measure vaccine take-up and social distancing behaviors. We also obtain data on daily vaccine take-up rates of the entire population of ages 59 and 60 from the Korea Disease Control and Prevention Agency.
We find that the vaccine take-up rate increased sharply among those born just before the birth date cutoff as soon as they became eligible for vaccines. The difference in the take-up rate relative to those born just after the cutoff date remained large, more than 60 percentage points, until the control group also became eligible 45 days later.
Despite the large gap in the vaccination rate, however, we find no robust evidence of risk compensation in any of our measures of risk compensation behavior, such as offline credit card usage, domestic air travel, and self-reported practice of social distancing. Lastly, we explore some potential explanations for the lack of risk compensation in our study setting.
JEL Classifications
- I1 - Health