This Knowledge Exchange fellowship has two strands i. The Engagement focus aims to increase government use of academic work, by increased engagement between BIS and NIESR through a) a series of seminars for policy people, b) review of processes for government research and data management, and c) improved communication of the policy perspective. ii. The Research focus aims to undertake analysis of firm level data and explore the potential to extend the findings in the literature in order a) to gain greater impact on policy development from academic knowledge, and b) identify and develop potential for further policy relevant research from these areas of NIESR expertise. The Engagement focus of the project contains different elements: a. Seminars. NIESR has agreed to arrange a series of 4 seminars on BIS policy topics over the 12 months of the fellowship. These will involve experts from the academic community, including ESRC centres such as SKOPE, and will be open to government officials from BIS and other departments. The topics are likely to include areas such as macro-economic projections, migration and the labour market, and sources of productivity gain. b. Research and data processes. The opportunity of the Fellowship will be used to improve academic links with government processes for research and data where possible. The aim is to develop and provide greater feedback to academic bidders on the features of weaker bids, and how to strengthen them. c. Policy context. As well as policy relevant seminars, other steps will be taken to ensure that BIS priorities are conveyed to the academic community. The approach will be to arrange a workshop for academics, hosted by NIESR, and invite policy leads from government to provide feedback on what has worked for them. The Research focus will use firm-level micro-data to explore the relation between the dynamic economy (firms closing and opening, growing and contracting, improving and declining) and productivity, the labour market, and one or more of skills, innovation, high growth firms, sectors and clusters. The approach is to explore the potential of these data for greater insight and impact, rather than direct data analysis - in recognition of the challenges presented by the patchy documentation, and inconsistent data records. The starting point is the finding that the great majority of productivity gain arises from 'external restructuring' of firms or plants (80-90% of Total Factor Productivity according to Disney et al (2003), and greater according to Harris and Moffat (2012). At first sight this seems surprising - it suggests that little productivity gain occurs within existing firms or plants. The implication is sometimes drawn that economic churn amongst firms is the route for growth rather than improvement within firms. The picture could in fact be different. The empirical literature appears to focus mainly on churn among plants rather than firms and there is little detail on how much of the gain occurs between plants but within firms, and how much between firms. This is the first question that will be explored. The second stage will consider how this decomposition varies over time, under different macro-economic circumstances. The literature suggests there is some but not much variation. One would expect the components of change to differ at times of growth from times of contraction or recession. The third stage will assess the feasibility of separating the decomposition by firm size, as churn is likely to operate rather differently for small firms than for middle or larger ones. Policy on investment in areas such as skills or innovation will be informed and influenced by the findings. The fourth stage will explore the potential for future work in one or more policy areas - skills and training (building on work by Haskel et al 2003, Dearden et al 2005, and Galindo-Rueda et al 2005), in innovation, or employment.