• Was Alfred Marshall a good observer or a visionary? In the late XIX century, the British economist theorised about the existence of three sources of agglomeration economies: labour pooling, input sharing, and knowledge spillovers. Over the last two decades, an extensive literature has tested the existence of the three Marshallian forces in modern economies. However - besides the detailed descriptive case-studies provided by the author himself - there is limited quantitative evidence regarding the existence of such forces at the times of Marshall. To shed light on these issues, we exploit novel geo-localised census-level data on entrepreneurs and business proprietors retrieved from six consecutive UK Censues (1851, 1861, 1881, 1891, 1901, 1911), coupled with census-level workers’ data, information on patents recorded between 1851 and 1911 and historical IO tables. We use these data to estimate co-agglomeration models that estimate the importance of labour pooling, input sharing and knowledge spillovers in explaining clustering of economic activities at the times of British rapid industrialisation. Our preliminary results highlight a significant role for knowledge spillovers and labour pooling, but only a limited effect for the input sharing. We also investigate the location patterns and co-agglomeration drivers that characterised three traditional industries discussed by Marshall, namely: cutlery, textile and machinery. Interestingly, the relative strength of agglomeration is generally consistent with what has been found for modern economies - and to a large extent in line with Marshall’s predictions. Our evidence on the historical importance of the ’Marshallian trinity’ coupled with the richness of our data lays the foundations for an ambitious research agenda on the spatial evolution of economic activities in late XIX century England, and on the rise and fall of sectors and regions.

Slides (CEPR Economic History Annual Symposium)