This paper evaluates two approaches to work-sharing by examining both within the same macro model. The standard approach involves imposing a quantity constraint on labour market participants (a maximum number of standard hours for each worker). This approach is compared to a revenue-neutral employment subsidy financed by a tax on overtime hours ? an initiative intended to harness market incentives. The paper shows that the second approach brings much preferred results ? it involves lower unemployment, higher investment, and no reduction in the wage earnings of those already employed. The analysis suggests that policymakers should not reject work-sharing just because they are (justifiably) skeptical of mandated reductions in hours. The model involves the following features: (i) it is optimization-based (so there is a well-defined reason for labour market failure); (ii) it facilitates the investigation of trade-offs (so it can be determined whether improvements in unemployment must be accompanied by reductions in productivity, investment, average hours or wage rates); (iii) it involves a small open economy (so concerns about the limits to independent policy in this setting are respected); and (iv) it can be readily calibrated (so empirically relevant quantitative results are derived).