Opponents of work obligations in return for transfer payments argue that workfare can crowd out private sector work, that workfare harms the welfare of the poor and thereby reduces a society’s welfare in general. This paper analyzes these objections against workfare in a discrete optimal income tax model. Workfare productivity emerges to be the crucial determinant. We identify a productivity threshold. If workfare productivity exceeds this threshold, then crowding out indeed happens, but it is second-best. Crowding out occurs entirely in the sense that workfare and private sector employment are mutually exclusive. Welfare increases due to the fact that out of the three effects workfare entails, beyond the detected threshold, the output effect and the incentive effect dominate the negative utility effect. Numerical simulations calibrated on the distribution of gross hourly wages in Germany reveal that on an individual level, the poor may even be better off under a tax-transfer scheme with workfare rather than without workfare. This is due to the fact that a higher degree of redistribution can be achieved.