Foreign ownership and productivity in the United Kingdom - Some issues when using the ARD establishment level data

In a recent paper, Griffith (1999) reviews the scope and usefulness of the recently released micro data underlying the UK Annual Census of Production (the ARD). In particular, she produces estimates of the Cobb-Douglas production function that show that foreign owned establishments in the motor vehicle industry do not have higher levels of productivity, after taking account of different inputs. This is an important result, since much recent work has suggested that multinational firms have a significant impact on productivity and technical progress in host countries. However, there are a number of issues that can be raised regarding the Griffith study and which are important in any work involving the ARD. Firstly, there are problems with the roll, data used since a significant number uses unweighted data and of establishments have been omitted. Secondly, Griffith uses unweighted data and therefore her parameter estimates are not representative of the underlying population of establishments operating in the UK motor vehicle industry. There is also an issue of whether plant level data should be used as opposed to establishment data. Finally, and most importantly, I do find that foreign-owned plants are significantly more productive than UK-owned plants, when using the same econometric approach as that adopted by Griffith. For comparative purposes, the analysis is extended beyond motor vehicles to include the next jour largest and most important foreign-owned sectors in UK manufacturing.