Innovation Upends Extrapolation: Urbanization

One of my favorite example of a mindless extrapolation was a headline I saw a few years back that said “By 2100 We Will All Live In Cities” (sadly I can’t find it anymore and failed to bookmark it). It may have been slightly exaggerated for dramatic purpose, but it is easy to find extrapolations that say by 2100 more than 80% of the global population will live in cities. Here is an example from the World Economic Forum:


You will readily get there if you simply extrapolate charts like this one from Our World in Data.


But extrapolation is dangerous (much more so than interpolation). In extrapolation you are assuming that the trends driving the observed changes will continue in their present form. That’s largely safe when you are dealing with simple physical systems like the trajectory of a single tennis ball. But when you are dealing with a system of massive complexity such as human societies, it is generally a terrible idea, especially over more than a few years.

There are many confounding factors, but the one I am most interested in is innovation. Suppose you had gone back to the year 1800 and extrapolated urban population based on the last 300 years, you might have come up with an estimate for the year 1900 of say 10% urban population and been off a fair bit (actual number 16%). Then in 1900 again even if you had allowed for significant further acceleration you would have (Read more...)