Taxation and unemployment: (systematic) evidence from panel data analysis

Nickell and Layard (1999) conclude that "[t]here appear to be no differential tax effects on unemployment but there is evidence that overall labor tax rates do influence labor costs in the long run and hence raise unemployment." This paper asks whether this conclusion still holds, by approaching the question from a rather systematic perspective. Besides, we also explore an alternative measure of income taxes, based on a representative-agent approach instead of macroeconomic aggregates. estimation of panel models for 20 OECD countries using different methods, including SUR, fixed-effects models, Pesaran's PMG and MG, as well mixed fixed- and random effects models; there are also Granger causality tests Our (preliminary) results support the judgement of Nickell and Layard. There is a positive effect running from the overall tax burden to unemployment. Establishing this result in a multicountry setting, taking econometric issues seriously, has proven, however, difficult.