A Geo-Spatial Analysis of Agricultural Production and Productivity for Southeast Europe, 1840-1940

M. Erdem Kabadayi, Koc University
Grigor Boykov, Sofia University / Koc University, Istanbul
Piet Gerrits, Koc University
Uygar Karaca, Koc University
Nick Rakovski, Koc University

This paper introduces a novel geo-spatial model to overcome a major difficulty in historical economic geography. Generally, published census data, be it contemporary or historical, is territory-based and available in aggregated levels for administrative units with varying spatial resolution. Pre-census micro level individual economic data on the other hand, is normally point-based. In geo-spatial terms, we can attribute census data to polygons, but pre-census individual data to points. We have developed a method to curate commensurable panel data for agricultural production for regions, covering 100 years, starting from pre-census Ottoman era in the 1840s and ending in the 1940s, with available series of national censuses. During the gradual collapse of the Ottoman Empire new nation states emerged and regions broke away from the Ottoman polity in Southeast Europe. We will focus on three regions experienced this break off and constituted successor states: Ruse and Plovdiv becoming regions of Bulgaria in 1878 and 1885, and Bitola of Serbia in 1912. New states in Southeast Europe conducted national population, agricultural, and industrial censuses. Starting from the 1920s all these regions were encompassed by national censuses. We developed a geo-spatial predictive modelling based upon soil quality and depth data, ruggedness, and connectivity for the 19th century and then we geo-sampled 5% of the villages (200 in total) covering 5% total population (around 20,000) in three regions. We have been extracting data on agricultural mix, production levels, and cultivated area from household-based Ottoman tax registers from the 1840s for the sampled villages. In this paper we will aggregate data on agricultural production, harvested through our sampling model to the level of census territories and bridge the dataset gap between pre-census and census era to conduct an unprecedented longitudinal analysis of changes in population geography and agricultural structure for Southeast Europe.

No extended abstract or paper available

 Presented in Session 80. Maps and geospacial data