State of the Art: Economic development through the lens of paintings (with Gorin and Heblich)
This paper analyzes paintings to glean information about their context. Relying on a large, open-access repository of art collections, we develop an algorithm to classify the emotions conveyed through paintings. We then project these emotions onto: the characteristics of artists; their influences and styles; and their context (a location within a country, in a given year). Our main object of interest is this context-specific residual of emotions, specifically its variation across locations and over time. In a first step, we use the context-specific vector of emotions to predict economic development and political change where these measures are readily available. In a second step, we extend the prediction to cover most of Europe from the 14th century onward. Our predicted measures of economic change exhibit large fluctuations around key, well-known economic transformations of Medieval Europe, the Renaissance, the later Reformation, and the Enlightenment. The prediction however uncovers interesting, and so far overlooked, geographic variation around more localized historical events inducing significant economic uncertainty.