LeeCoppock.com

I just came across a cool new paper in which the authors analyze the color palettes of paintings to learn more about the economic growth when the paintings were created. They can actually see strong correlations between economic growth and paint colors. But even more, they use color data to uncover more about actual economic growth in periods where data is is sketchy or incomplete. Wow, this is a cool paper. The title is Colors of Growth, and it is written by four economists: Lars Boener, Tim Reinicke, Samad Sarferaz, and Battista Severgnini.

They look at European paintings from 1600 to 1820 and they see significant variation in colors used. One reason is that certain colors are easy to attain even when an economy is poor (apparently red is widely available in nature), but other colors are hard to get (apparently a popular shade of blue required lapis lazuli from Afghanistan). In addition, the availability of indoor light comes along with economic growth. Thus, paintings that feature lighter colors indoors help to shown when a population is getting wealthier.

For example, Figure 4 from the paper (shown below) shows a 20-year moving average of different colors in from British oil paintings. Notice the significant increase in the lighter shades just after the middle of the 17th century. That period corresponds with a period of significant British productivity growth. But then, several political and economic crises led to a growth slowdown at the end of the 17th century and this is also reflected in the paint colors.

But the authors go on to use their model to actually uncover new evidence of growth and stagnation “that deviate from existing GDP estimates” for Germany and France. Their new color-based evidence is consistent with historical narratives of the period. This raises the possibility that the color analysis can help us find problems with historical economic data. Amazing.