Last week, a Swiss Think Tank published this interview about the development in trade in a city. I want to utilize this as a good example of the challenges and common misconception that we face when thinking about the future urban area.

First of all, it follows the concept of ceteris paribus (other things equal). This concept relies on the idea, that only one (or a few) variables are changed when simulating a situation instead of having all variables in movement. While this was highly beneficial for economists in the past due to limited tools (no computers) and limited computation power, it now longer upholds and more and more complex economic models arrise. In the interview, we see it first of all for the assumption of the past general trend of consolidation in retail that will go on and immediately forms the possible scenarios.

Secondly, the interviewees rely on the global mega trends that will, in their opinion, induce a “specialized market” in the urban area without any reflection on its validity in a model. The reason for that path of argument is very simple, but also problematic: Many models have difficulties when imposing changing energy functions, or effects of digitalization. The effect is often a jump function that could have very binary outcomes. For instance, energy or transport costs could immediately devalidate their idea of urban specialization and main retail in the agglomeration since costs to receive goods would be too high.

Third, most people looking at the city as a field of study avoid disruptive innovations or new technical concepts. While this was perfectly ok for the city planners of the 20th century since it took at least 10 to 20 years until a technology was making any disruptive effect, it is not true today anymore. The internet, the smartphone and possible the drone and the autonomous vehicle have within a, relatively, short timespan changed some of our core activities.

So what to do? I do not think there is a simple answer to that question. Probably we should stop giving meaningless interviews, also we should rather focus on the core mechanics of the individual and retry to make an aggregate out of that. Furthermore, we need to communicate that all of our simulation is very limited and inherently flawed, this way we can have a two way discussion: on the model, and on the results.