Patents, machine learning, and war: invention during the British Industrial Revolution

  • Stephen Billington

Student thesis: Doctoral ThesisDoctor of Philosophy


This thesis investigates the incentives to invent and innovate, with a particular focus on the role of Britain's patent system in the first half of the nineteenth century, during the Industrial Revolution. The research focuses on three important questions for economics and economic history: (1) does it matter how we classify historical patent data? (2) how did Britain's patent fees affect patenting behaviour? and (3) was Britain's involvement in the French Revolutionary and Napoleonic wars harmful to industrialisation? Using a combination of machine learning techniques, econometric analysis, and new archival evidence, this thesis finds that: (1) the results of statistical analysis of historical patent data depends on the taxonomy used to classify that data; (2) Britain's high patent fees were unlikely to have harmed invention, as the most economically valuable inventions were patented in multiple British patent jurisdictions, and were produced by individuals of a higher social status; and (3) the Napoleonic War coincided with a permanent increase in economically valuable invention, which likely aided industrialisation. Together, these findings lead to the following conclusions: The historical innovation literature likely yields inconsistent results and interpretations because of the choice of their patent taxonomies; inventors were motivated by economic incentives, suggesting patents can encourage invention; and war can have positive effects on stimulating demand for new technologies. Overall, the evidence supports the recent positive reassessment of the British patent system in relation to the British Industrial Revolution, and considers the patent system as beneficial to innovation.
Date of AwardDec 2019
Original languageEnglish
Awarding Institution
  • Queen's University Belfast
SupervisorChris Colvin (Supervisor) & Christopher Coyle (Supervisor)

Cite this