State of the Art: Economic Development Through the Lens of Paintings
Joint with Gorin and Heblich.
This paper analyzes 630,846 paintings produced since 1400 to study how past societies perceived major socioeconomic transformations. We develop computer vision algorithms to (i) predict the sentiment conveyed by each painting and (ii) extract visual representations of material living standards. We validate the distinct signals of emotional and material welfare shared across artworks produced in the same location and year by mapping them respectively to external indicators of happiness and economic output. Our empirical analysis first identifies the emotional response to economic development, uncertainty, and inequality within artists' oeuvres and conditional on painting sub-genres. Second, we exploit well-identified shocks to climate, trade, technology, knowledge production, and the geopolitical environment to trace how populations experienced such socioeconomic transitions.
.png&w=3840&q=75)