Disaggregated Bilateral Trade Data, 1870-2024

Description: The digitization of historical data has become vital for the analysis of long-term economic and social trends. The session aims to spotlight the pioneering efforts in digitizing and harmonizing disaggregated bilateral trade data from 1870 to today (data by the RICardo prpject even go back to the beginning of the 19th century). The session will explore the challenges and methodologies involved in making historical trade data comparable across time and space, especially considering territorial changes, the emergence/disappearance of polities as well as issues in classification, notably in the usage of the Standard International Trade Classification (SITC).

Relevance: The session's relevance is underscored by two significant trends: (1) The advent of AI and machine learning in digitization and analysis of historical data and (2) the heightened academic and policy interest in globalization's impacts.

Aim: Over the last ten years, various research teams have developed historical disaggregated trade databases. Researchers have all encountered challenges in the digitization and harmonization of data, which become increasingly complex the further back we go in time. There is great diversity in data recording methods, commodity description and classification methods, commodity diversification, heterogeneity of trading entities (changes in border, name, political status). This session aims to pool problems encountered, share best practices and innovative approaches. The session's primary aim is to share best practices and innovative approaches for digitizing and harmonizing product-level trade data.

Contribution: The session will highlight methods that make use of the latest technological advances, AI and machine learning, in digitization, harmonization, and analysis of historical data. It will foster dialogue, offer novel insights regarding the technical and methodological challenges, and provide an opportunity to compare alternative solutions. By bringing together scholars that have compiled and/or worked with such data, this session promises to foster dialogue and offer novel insights regarding the technical and methodological challenges of creating a comprehensive dataset of product-level trade from the first globalization to the present. Eventually, this session aims to contribute to a richer understanding of global economic integration and its implications.

Session Format:

1. Introduction (5 min, Hungerland): Overview of advances and challenges in trade data standardization

2. Short Talks (50 min, Daudin, Betran, Federico, Hungerland): max. 10-minute presentations recent advances regarding datasets, classifications, and/or research questions

3. Discussion (50 min, with Federico and Wolf as panelists): Best practices, AI tools, and future standardization 

Confirmed Presentations: 

• Guillaume Daudin (Université Paris Dauphine-PSL): "Leveraging geopolitical history to deal with taxonomic diversity of trade partners in RICardo"

• Markus Lampe (WU Wien) "TRADEDEPRESSION 10 years later"

• Concha Betrán (University of Valencia): "Disaggregated Bilateral Trade in Interwar Spain"

• Giovanni Federico (NYU Abu Dhabi), David Chilosi (King's College London) and Andrea Incerpi (Siena) "Intensive/extensive margin for Italy 1880-2020"

• Fabian Hungerland "Germany's messy data: Changing borders, hyperinflation, wars and division"