Recent studies have confirmed that the COVID-19 lockdown has caused massive job losses. However, the impact of this loss on food security is not well-understood. Moreover, a paucity of evidence exists regarding social protection grants’ countervailing effects against such shocks. This study examined the effects of job loss (labour income loss) on child and household hungers (our two measures food insecurity) during COVID-19 pandemic in South Africa. It also ascertained whether these effect were offset by alternative social grant programs to document the protective role of the latter. We used South Africa’s National Income Dynamics Study (NIDS) and the Coronavirus Rapid Mobile Survey (CRAM) data. These data cover a nationally representative sample of 7073 individuals. We employed a probit model to estimate the effect of job loss and receipts of various social grants on child and households’ hungers. We also estimated the double-selection logit model to account for the model’s uncertainty surrounding the variable selection and treatment-effects estimation using lasso (Telasso) for causal inference of our analysis. Our analyses showed that households exposed to a labour market shock during the pandemic experienced a significant increase in our measures of food insecurity (child and household hungers). Specifically, we found that compared with households containing employed respondents, households with respondents who lost their jobs due to COVID-19 lockdown were 5.4% more likely to report child hunger and 2.6% more likely to report household hunger in the past seven days A receipt of child support grant reduces the likelihood of reporting child hunger and household hunger by 21.7%and 16.9% respectively among these households. A receipt of old age pension grant reduces the likelihood of reporting household hunger by 24% with no significant effect on child hunger. Conclusion The COVID-19 lockdown resulted in unprecedent job losses with significant implications for food insecurity. Job loss due to COVID-19 lockdown significantly increased food insecurity in South Africa. Receipts of social grants effectively offset this adverse effect. The protective effect of the social grant is heterogenous across its alternative programs (child support grant and old age pension grant) and food insecurity, suggesting the differences in the size of transfers and motivations for sending these transfers.