ABSTRACT
Much of the research on behavioural preferences as predictors of compliance with regulations aimed at reducing the transmission of COVID-19 has focused on developed countries, with very little consideration of African countries. We conduct an online survey (n = 1503) considering beliefs, and individual and social preferences as predictors of compliance with prophylactic measures and lockdown regulations in South Africa. We use incentivized experimental measures of individual (risk and time) preferences and social preferences (cooperativeness and altruism). We also consider survey measures of risk tolerance, patience and trust. We find that beliefs about others’ behaviour are highly predictive of reported behaviour. We also find that greater patience and cooperativeness are predictive of high compliance with prophylactic measures and lockdown regulations. Encouragingly, respondents report higher compliance at higher lockdown levels, suggesting responsiveness of behaviour to the level of risk of infection.
Acknowledgements
We gratefully acknowledge funding from the Science for Africa Foundation and its partners (grant number SARSCov2-2-20-004) and excellent research assistance from Fadzayi Chingwere.
Ethical clearance
Ethical clearance was obtained from the University of Pretoria Economic and Management Sciences Ethics Committee: EMS104/20.
Notes
1 Since internet access is limited in rural areas of South Africa, the online sample means that most respondents would be in urban parts of the country.
2 Details of South Africa’s lockdown level system can be accessed at https://www.gov.za/covid-19/about/about-alert-system.
3 1% of respondents chose not to indicate race, and 7% of respondents chose not to report household income. Census 2011 data reports gender splits of 51% Female and 49% Male; and race splits of 79% Black, 9% Coloured, 2% Indian/Asian and 10% White. Household income details from the South African Labour Force Survey (2017) adjusted for inflation suggest that 25% of South African households have income below ZAR 2,000 (approximately USD 150 at the time of data collection) per month; while 25% have income above ZAR 8,900 per month.
4 Almost 50% of the probability mass was at the upper limit for this question. We therefore report the Tobit measure in our results. As a robustness check, we also ran Ordered Logit regressions for this measure, with very similar results. These results are available from the corresponding author on request.
5 For the Social Preferences component, the factor loadings are: Risk tolerance, 0.55; Altruism, 0.53, Cooperativeness, 0.63.
6 For the Patience component, the factor loadings are: Present bias, 0.71; Patience, 0.70.
7 As a robustness check, we also consider the continuous measure of risk tolerance (where the amount invested in the risky option is captured on a scale from 1 to 11). The results do not change from those reported in .