This paper investigates the impact of labour force composition on productivity in EU arable farming. We test the heterogeneity of family and hired labour for a set of eight EU member states. To this end, we estimate augmented production functions using FADN data for the years 2001-2008. The results reject the notion that hired labour is generally less productive than family workers. In fact, farms with a higher share of hired workers are more productive than pure family farms in countries traditionally characterised by family labour, namely France, West Germany and Poland. Here, an increase in reliance on hired labour or the shift of family labour to more productive tasks could raise productivity. This finding calls into question a main pillar of the received family farm theory. In about half the countries, there are no statistically different effects of both types of labour. For the United Kingdom, we find the classical case with family farms being more productive than those relying on hired labour. In this situation supervision by family members could increase productivity. As a side result, we find little evidence of non-constant technical returns to scale.