Historical sheep statistics from Mitchell are first adjusted to match the FAO statistics in 1961 (FAO presents statistics from 1961-present time). Then, the most historical data found in Mitchell is converted in a simple per-capita number, and this number is kept constant further back in time. Apparently for Europe this will result in an underestimate (evidence from Broadberry et al suggest that in Medieval times the per capita sheep number was higher), but for other regions this is probably not the case (Americas and Australia/New Zealand didn’t have livestock at all until the early 16th century resp. late 18th century).

Please consult the working paper for a description of the data and sources.