Following Phillips's original work on the UK, applied research on unemployment and wages has been dominated by the analysis of highly aggregated time-series data sets. However, it has proved difficult with such methods to uncover statistically reliable models. This paper adopts a different approach. It uses microeconomic data on 175 000 British workers from 1973 to 1990 to provide evidence for the existence of a negatively sloped relationship linking the level of pay to the local rate of unemployment. The paper casts doubt on standard ideas in macroeconomics, regional economics and labour economics