In economic theory, one can distinguish between variety as a source of regional knowledge spillovers, called Jacobs externalities, and variety as a portfolio protecting a region from external shocks. It is argued that Jacobs externalities are best measured by related variety (within sectors), while the portfolio argument is better captured by unrelated variety (between sectors). A methodology based on entropy measures is introduced to compute related variety and unrelated variety. Using data at the NUTS 3 level in the Netherlands for 1996-2002, it was found that Jacobs externalities enhance employment growth, while unrelated variety dampens unemployment growth. Productivity growth can be explained by traditional determinants including investments and research and development expenditures. Implications for regional policy follow.