Periodic Reporting for period 2 - HMTG (Historical Migrations, Trade, and Growth)
Période du rapport: 2021-12-01 au 2023-05-31
As of the writing of this report, there has not yet been any journal publication emanating from my research project. The data on immigration shocks is publicly available through a dedicated website at https://www.immigrationshock.com/.
This project aims at bringing together insights from economic history, social psychology, and economics, to enhance our understanding of the impact of migrations on economic growth, on prejudice and altruism, and on urban growth. It is articulated around several distinct but related projects. During the first phase of this research project, progress has been made on three of those projects.
“Immigration, Innovation, and Growth” quantifies the impact of immigration on innovation and growth. We leverage 130 of historical migrations to the United States and show how to use historical settlements to identify contemporaneous immigration shocks. A structural model of endogenous migrations and growth suggests that the large inflow of foreign migrants to the US since 1965 may have contributed to an additional 8% growth in innovation.
“The Immigrant Next Door: Long-Term Contact, Generosity, and Prejudice” quantifies how decades-long exposure to individuals of foreign descent shapes natives’ attitudes and behavior toward that group. We show that exposure to the presence of population of foreign origins increases generosity towards foreign causes through charitable donations, reduces prejudice against foreign cultures, in particular towards Arab-Muslims, and increases contact with and knowledge about foreign cultures.
“Very Long Run Growth” aims at quantifying the process of technical progress at the core of economic growth over very long periods (millennia) and over large geographic areas (continents). To identify the timing, location, and subsequent diffusion of individual innovations, we collect systematic information on tens of millions of historical artefacts stored in dozens of art and history museums throughout the world. Leveraging tools from Machine Learning and Natural Language Processing, we are able to match the textual description of historical artefacts to a precise time, location, and to historical technology classes identified by historians. A combinatoric model of innovation allows us to identify innovations, and to quantify the size of innovative leaps. Our detailed information on space and time further allows us to track the geographic diffusion of innovations.
• What is the problem/issue being addressed? This project aims at quantifying the impact of historical and contemporaneous migrations on innovation, growth, social norms, and urbanization. Doing so, it addresses both empirical and theoretical challenges. On the empirical side, the literature in economics has long recognized that due to the endogeneity of migration choices, identifying the causal impact of migrations is elusive. We leverage information on historical migrations, and in particular historical shocks to migrations that are plausibly quasi-random, to make progress on this identification concern. Furthermore, the literature on economic growth and innovation has so far focused primarily on recent time periods, for which detailed information on patents and firms’ innovative strategy is widely available. We propose an original method to harvest the information collected by historians and museum curators to extend this analysis by several millennia. Finally, we develop simple and tractable theoretical models of the endogenous formation of preferences subject to social influences, and by carefully modelling both immigration and innovation choices.
• Why is it important for society? Many countries, in particular in Europe, face growing challenges surrounding immigrant integration and native backlash. Understanding the qualitative and quantitative impact of migrations, both in the short and long run, is therefore of primary importance. While recent work has shown that intergroup contact (e.g. between natives and migrants) has short run impacts on attitudes and behavior, little is known about the long run consequences of immigration. We propose to make progress on this important question, quantifying the impact of immigration over different horizons on economic and political outcomes.
This project aims at bringing together insights from economic history, social psychology, and economics, to enhance our understanding of the impact of migrations on economic growth, on prejudice and altruism, and on urban growth. It is articulated around several distinct but related projects. During the first phase of this research project, progress has been made on three of those projects.
“Immigration, Innovation, and Growth” quantifies the impact of immigration on innovation and growth. We leverage 130 of historical migrations to the United States and show how to use historical settlements to identify contemporaneous immigration shocks. A structural model of endogenous migrations and growth suggests that the large inflow of foreign migrants to the US since 1965 may have contributed to an additional 8% growth in innovation.
“The Immigrant Next Door: Long-Term Contact, Generosity, and Prejudice” quantifies how decades-long exposure to individuals of foreign descent shapes natives’ attitudes and behavior toward that group. We show that exposure to the presence of population of foreign origins increases generosity towards foreign causes through charitable donations, reduces prejudice against foreign cultures, in particular towards Arab-Muslims, and increases contact with and knowledge about foreign cultures.
“Very Long Run Growth” aims at quantifying the process of technical progress at the core of economic growth over very long periods (millennia) and over large geographic areas (continents). To identify the timing, location, and subsequent diffusion of individual innovations, we collect systematic information on tens of millions of historical artefacts stored in dozens of art and history museums throughout the world. Leveraging tools from Machine Learning and Natural Language Processing, we are able to match the textual description of historical artefacts to a precise time, location, and to historical technology classes identified by historians. A combinatoric model of innovation allows us to identify innovations, and to quantify the size of innovative leaps. Our detailed information on space and time further allows us to track the geographic diffusion of innovations.
• What is the problem/issue being addressed? This project aims at quantifying the impact of historical and contemporaneous migrations on innovation, growth, social norms, and urbanization. Doing so, it addresses both empirical and theoretical challenges. On the empirical side, the literature in economics has long recognized that due to the endogeneity of migration choices, identifying the causal impact of migrations is elusive. We leverage information on historical migrations, and in particular historical shocks to migrations that are plausibly quasi-random, to make progress on this identification concern. Furthermore, the literature on economic growth and innovation has so far focused primarily on recent time periods, for which detailed information on patents and firms’ innovative strategy is widely available. We propose an original method to harvest the information collected by historians and museum curators to extend this analysis by several millennia. Finally, we develop simple and tractable theoretical models of the endogenous formation of preferences subject to social influences, and by carefully modelling both immigration and innovation choices.
• Why is it important for society? Many countries, in particular in Europe, face growing challenges surrounding immigrant integration and native backlash. Understanding the qualitative and quantitative impact of migrations, both in the short and long run, is therefore of primary importance. While recent work has shown that intergroup contact (e.g. between natives and migrants) has short run impacts on attitudes and behavior, little is known about the long run consequences of immigration. We propose to make progress on this important question, quantifying the impact of immigration over different horizons on economic and political outcomes.
Two projects (“Immigration, Innovation, and Growth” and “The Immigrant Next Door”), have been submitted to a journal, and both are at the Revise and Resubmit stage (at the American Economic Review). The data and code accompanying “Immigration, Innovation, and Growth”, has been made publicly available at https://www.immigrationshock.com/.
Substantial progress has been made on a third project (“Very Long Run Growth”). The raw data has been collected, and algorithms have been deployed to extract systematic information on the dating, location, and technology class of historical artefacts. We aim to submit preliminary results on the machine learning techniques used to classify historical artefacts into technology classes to a computer science conference in the coming months.
The other projects are currently at a preliminary stage.
Substantial progress has been made on a third project (“Very Long Run Growth”). The raw data has been collected, and algorithms have been deployed to extract systematic information on the dating, location, and technology class of historical artefacts. We aim to submit preliminary results on the machine learning techniques used to classify historical artefacts into technology classes to a computer science conference in the coming months.
The other projects are currently at a preliminary stage.
The results expected until the end of the project are primarily a series of publications in general audience economic journals. Two projects are already reaching completion (“Immigration, Innovation, and Growth”, and “The Immigrant Next Door”). In addition, several medium sized conferences will be organized to disseminate the results from this project, and to bring together a diverse set of scholars working on topics related to this project. Finally, this project is strengthening a community of young researchers in Paris working on topics related to economic history, trade and growth (junior faculty, pre-doctoral students, graduate students, and post-doctoral researchers).
The project is going beyond the state of the art in several dimensions. It offers novel tools to collect, organize, and use novel sources of historical data, using state-of-the-art methods in machine learning and natural language processing. Finally, it shows how to combine experimental, observational, and structural estimation techniques together to shed light on unconventional questions.
The project is going beyond the state of the art in several dimensions. It offers novel tools to collect, organize, and use novel sources of historical data, using state-of-the-art methods in machine learning and natural language processing. Finally, it shows how to combine experimental, observational, and structural estimation techniques together to shed light on unconventional questions.