This session explored the geography of productivity, examining what place-based industrial strategy can realistically achieve and what conditions are needed for local economies to thrive.
The session positioned regional productivity as the outcome of several interacting systems: the health and skills of workers, the geography of labour markets, public investment in digital innovation, the movement of people and knowledge across places, and the spatial concentration of inventive activity. A consistent theme was that productivity cannot be understood solely within fixed administrative boundaries. Commuting flows, labour mobility, inter-regional connectivity and innovation clusters all shape how benefits and constraints travel across places. The discussion also highlighted the importance of distinguishing between local effects and spillovers, and of recognising that regional productivity gaps may reflect structural differences between places as much as changes within individual regions over time.
The first paper examined the relationship between local health and local labour productivity, focusing on whether health capital contributes to productivity not only where people live, but also where they work. Using 2019 local authority district data, the ONS Health Index and commuting flows from the 2011 Census, the analysis linked residential health conditions to workplace-based productivity measures such as GVA per job and GVA per hour.
The main finding was that the “healthy people” domain was positively associated with GVA per job through indirect effects: the health of people commuting into an area appeared more strongly related to workplace productivity than the health of residents within the same area. The discussion raised questions about scale, data power and whether effects may be different at smaller geographies. Participants also asked whether functional labour markets should be differentiated by characteristics such as rurality, urban form, commuting intensity and access to health services.
The second paper considered whether public funding for digital R&D is followed by increased regional demand for advanced digital skills and greater digital patenting. The research used Gateway to Research project data, Lightcast job postings and patent data to identify digital projects, classify skill requirements and examine regional relationships between public investment, skills demand and innovation outputs.
The presentation reported evidence of a positive association between public digital R&D funding and demand for advanced digital skills, particularly higher-level specialist, programming, AI and machine-learning skills. There was also evidence of a relationship between funding and digital patenting, although the links between skills demand and digital innovation activity appeared weaker. Discussion focused on regional variation, the concentration of public R&D spending, and whether place-based connectivity helps explain how public support translates into productivity gains.
The third paper analysed how labour market dynamics relate to structural change, knowledge diffusion and productivity. It developed the idea that regions may become trapped in particular industrial structures, either by failing to diversify into more productive industries or by losing productive specialisms. Drawing on business structure data and employee mobility data, the work examined how regional specialisation and structural change have evolved over time.
The paper highlighted a sustained fall in structural change after the financial crisis, with levels remaining below pre-crisis patterns even before the pandemic. Inward labour mobility was associated with higher structural change before the post-crisis period, but this relationship weakened afterwards. The discussion explored whether firm mobility, sectoral mobility, within-industry productivity change and the time lag between losing and gaining specialisms should be incorporated into the analysis.
The fourth paper examined knowledge decay within OECD regions by mapping patenting activity at a much more granular spatial scale than conventional regional analysis. The work used geocoded patent applications to identify one-kilometre cells of inventive activity and measure travel times between them by public transport, driving or combined modes. This approach allowed the analysis to move from broad regional comparisons to within-region patterns of clustering and accessibility.
The presentation showed that innovation activity can be highly concentrated within administratively defined regions, with different knowledge types displaying different spatial decay patterns. Policy discussion focused on how granular evidence might inform choices between strengthening existing clusters and improving connections to places currently less engaged in innovation. Questions also explored commercialisation, citation networks, non-patented knowledge, rural constraints and the role of case studies in understanding how innovation works in specific places.