AI study reveals England’s productivity divide is far more complex than North-South
Researchers at The University of Ƶ have used artificial intelligence to uncover a complex picture behind England’s long-running productivity puzzle, challenging the idea that the country’s economic performance can be explained by a simple North-South divide.
In a major study published in the journal, and applied ‘GeoAI’ techniques - combining geography and artificial intelligence - to analyse how productivity varies across local authorities in England between 2010 and 2022.
Productivity, measured as Gross Value Added (GVA) per hour worked, is a key driver of wages and living standards. Since the 2008 financial crisis, UK productivity growth has lagged behind other major economies, fuelling debate among economists and policymakers.
The research shows that the national picture hides a complex local story. While London and the South-East still contain many of the highest-productivity areas, performance within the region varies. Some traditionally strong local authorities have experienced stagnation or decline over the past decade - and several lower-productivity areas in the Midlands and northern England have recorded faster growth, albeit from a lower starting point.
The study found that nearly half of England’s local authorities performed below the national average on both productivity level and growth rate between 2010 and 2022. Fewer than one in five achieved both high productivity and strong growth.
Using GIS and machine learning models, the team identified factors most strongly linked to high productivity - a high concentration of information and communication sector jobs, higher wages, and proximity to other high-productivity areas known as “spillover effects.” The findings show being near a productive neighbour can boost performance, but only once certain thresholds are reached. Agglomeration effects are real, but not automatic or evenly shared.
The study also found that some widely cited drivers, including regional R&D investment and infrastructure, were less influential in explaining productivity differences than expected.
The researchers classified England’s 296 local authorities into 12 productivity types, ranging from vulnerable labour markets with weak industrial bases to specialised information and finance centres with very strong output per hour worked. The results show no single policy solution will work everywhere. Some places need to strengthen their industrial mix, others would benefit from stronger links to economic hubs, and in some areas improving health and workforce resilience could make a difference.
The findings come as debates around devolution, regional growth, and the future of the UK economy intensify. The researchers argue that national productivity strategies must take local spatial dynamics into account, as policies designed at broad regional scales may overlook variations within them.
“The usual headline story of a ‘North-South divide’ is too simplistic - when we look closely, we see a patchwork of places moving at different speeds,” said Professor Wong. “The productivity puzzle can be interpreted as a new ‘hare and tortoise story’ - many high performers are losing ground in the race, when some poor performers are trying hard to catch up.”
If we want to raise living standards across the country, we need to understand how productivity works in different places. AI gives us a new lens to see spatial dynamics more clearly.
DOI: