Is geography destiny? What is the role of first-nature geography in determining prosperity? This paper estimates the effect of randomly removing and introducing favorable first-nature geography to a specific region using a difference in difference design. In 1825 a storm created a new natural navigable waterway, bringing trade and prosperity to the otherwise relatively isolated northwestern Denmark. 700 years prior, the same event happened in reverse, when a previous channel closed up between 1086 and 1208. The elasticity of geography-induced market access is estimated to be 1.6, corresponding to 26.7 percent population growth within a generation of the event. Demonstrated mechanisms include trade, fertility, fishing, and the rise of manufacturing. The central finding is replicated in reverse in a register of dated archaeological sites. The 1086-1208 closing caused fewer buildings and sites containing coins. The general insight is the same: First-nature geography determines the levels and location of prosperity. 

The paper won the New Researcher prize, at the Economic History Society's conference in 2023. 

https://arxiv.org/abs/2408.00885 

I am in huge intellectual debt to historian Bo Poulsen. Read one of his latest articles on the event here: https://vbn.aau.dk/da/publications/between-adaptation-and-mitigation-the-nineteenth-century-north-se 

See replication files here

https://github.com/christianvedels/A_perfect_storm_replication 

YouTube video: https://youtu.be/f7Xo7BO_Xqk?si=xMAS9viHL34BGbWt 

With Christian Møller-Dahl and Torben Johansen

This paper introduces a new tool, OccCANINE, to automatically transform occupational descriptions into the HISCO classification system. The manual work involved in processing and classifying occupational descriptions is error-prone, tedious, and time-consuming. We finetune a preexisting language model (CANINE) to do this automatically thereby performing in seconds and minutes what previously took days and weeks. The model is trained on 14 million pairs of occupational descriptions and HISCO codes in 13 different languages contributed by 22 different sources. Our approach is shown to have accuracy, recall and precision above 90 percent. Our tool breaks the metaphorical HISCO barrier and makes this data readily available for analysis of occupational structures with broad applicability in economics, economic history and various related disciplines.

Read more in the working paper on arxiv: https://arxiv.org/abs/2402.13604

For instructions on how to use OccCANINE, slides, and much more see:

https://github.com/christianvedels/OccCANINE