This paper uses new micro-level US data to re-examine productivity leadership in cotton spinning c. 1900. We find that output aggregation problems make the Census unreliable in this industry, and that Lancashire, not New England was the productivity leader for almost every type of yarn. This is...

This paper uses new product-specific, micro-level US data to show that New England had lower levels of productivity in cotton spinning than Lancashire, c. 1900, contradicting results derived by Broadberry from the Censuses of Production. The discrepancy stems from the Censuses’ poor methods of...