We use ONS micro data for large UK establishments in the production industries in the period 1997-2008 to study the relationship between their productivity and the presence of substantial RandD activities, either at the production unit itself, or at other UK reporting units owned by the same enterprise group. We estimate that total factor (revenue) productivity is on average about 14% higher at the establishments which have substantial RandD themselves, compared to those with no RandD. Among the establishments with no RandD themselves, we estimate that productivity is on average about 9% higher at those which belong to enterprise groups which do have substantial RandD elsewhere in the UK in the same sub-sector. For the establishments with substantial RandD themselves, we also estimate a significant positive relationship between current productivity and past RandD expenditure using dynamic specifications which allow for both establishment-specific `fixed effects' and a serially correlated error component.
R&D
,productivity