In this paper, we provide causal evidence of the immediate and near-term impact of stringent COVID-19 lockdown policies on employment outcomes, using Ghana as a case study. We take advantage of a specific policy setting, in which strict stay-at-home orders were issued and enforced in two spatially delimited areas, bringing Ghana's major metropolitan centres to a standstill, while in the rest of the country less stringent regulations were in place. Using a difference-in-differences design, we find that the three-week lockdown had a large and significant immediate negative impact on employment in the treated districts, particularly among workers in informal self-employment. While the gap in employment between the treated and control districts had narrowed four months after the lockdown was lifted, we detect a persistent nationwide decline in both earnings and employment, jeopardizing particularly the livelihoods of small business owners mainly operating in the informal economy.