The time-series approach used in the minimum wage literature essentially aims to estimate a treatment effect of increasing the minimum wage. In this paper, we employ a novel approach based on aggregate time-series data that allows us to determine if minimum wage changes have significant effects on employment. This involves the use of tests for structural breaks as a device for identifying discontinuities in the data which potentially represent treatment effects. In an application based on Australian data, the tentative conclusion is that the introduction of minimum wage legislation in Australia in 1997 and subsequent minimum wage increases appear not to have had any significant negative employment effects for teenagers.