Annual EBHS Conference, 39th Annual Economic and Business History Society Conference

Font Size: 
Land prices, Banks and Entrepreneurship. Explaining intraregional divergences in industrial activity. The case of Flanders (19th – 20th century).
Nicolas De Vijlder, Koen Schoors

Last modified: 2014-05-19


From the middle of the 18th to early 20th century, most countries in North-western Europe experienced a declining importance of the agricultural sector, to the benefit of the industrial and service sectors. Following the United Kingdom’s industrial development from 1750 onwards, Belgium was at the helm of this transition in continental Europe. Other countries, such as the Netherlands and Germany experienced similar processes, albeit only from the 1870s onwards during the second wave of the industrial revolution. Divergence however, not only occurred between states. Within the same national entity some regions were at the forefront of industrial development, while in others agriculture remained the prevailing economic activity. In this respect, Belgium provides an interesting example. Initial industrial activity was centered in the resource-rich Walloon area, whilst Flanders retained its rural/proto-industrial character throughout the 19th century. However, between 1900 and 1937, the secondary and tertiary sectors became increasingly important in certain areas of Flanders. Consequently, by 1937 industrial employment at the regional level, expressed as a percentage of the total workforce, was comparable throughout Belgium. However, large intraregional differences, especially in Flanders, still prevailed.

In our paper we investigate the presence of a ‘De Soto effect’ in Flanders between the 19th and the first half of the 20th century. The ‘De Soto effect‘ entails that landownership provides borrowers with an attractive form of collateral that, subject to the presence of institutionalized lending facilities (banks), can ease access to external finance.  Using a spatially lagged econometric model, we hence analyze to what extent the intraregional variation in industrial activity observed in 1937 can be explained by the availability of credit. In our analysis, we consider 1,190 municipalities in Flanders. Our dependent variable is the level of industrialization per locality, expressed as the ratio between the number of employees in the secondary and tertiary sectors and the locality’s population. Our explanatory variables are the average price for arable land in 1846 at the municipal level and the number of employees in lending institutions in 1910 and their interaction. These were lagged vis-à-vis the dependent variable to account for the interplay between lending institutions, property prices and levels of industrial and trade activity and to make sure we measure a causal relation. We allow spatial lags of industrial activity, and our base model was augmented with several control variables to account for demand-effects and the size of the municipality. All data were compiled using official industrial and population censuses.

The results do not allow us to reject the De Soto hypothesis. Neither high land values in the early 1800s, nor high banking activity in 1910 seem to be robustly and independently related to more industrial activity in 1937. But the interaction of these two variables turns out to be strongly and robustly related to industrial activity. Indeed, we find that 1937 industrial activity was especially high in those regions where both exogenous land prices were high and banking activity was strongly developed by 1910, which is in line with the De Soto hypothesis.