Evidence on the effectiveness of workfare as an anti-poverty program in developing countries is weak compared with the relatively well-established role of public works during economic crisis as a social safety net. This paper contributes to evidence building by examining the impact of a large-scale workfare program in Bangladesh, the Employment Generation Program for the Poorest. Taking advantage of the program's distinguishable feature of direct wage transfer to a person's bank account, the paper uses accessibility to local banks as an instrumental variable to identify the program's impacts on rural social assistance beneficiaries. Based on locality-by-time fixed effects models over two rounds of locality panel data, the analysis finds that the Employment Generation Program for the Poorest has contributed to increasing overall household consumption and reducing outstanding loans. In particular, expenditures on quality food and health care have significantly increased, which likely helps individuals continue to engage in income-generating activities in the labor market. However, the implementation costs and poor quality of public assets built through work projects could potentially undermine the program's efficiency. Moreover, further evidence is required on the impacts of work experience through workfare on subsequent labor market outcomes and the value of public assets, to assess the program's effectiveness compared with administratively simpler alternative instruments such as unconditional cash transfers.