Supply and Demand: Simultaneous Equations, not Simultaneous Causation

Typical causal graphs, i.e, Directed Acyclic Graphs (DAGs), don’t allow for cycles as this would allow a variable to be the cause of itself. Whether such a thing is possible is unclear and falls within the purview of the metaphysical inquiry known as “simultaneous causation.” However, simultaneous equations where variables are both cause and effect do exist, with the canonical example being supply and demand, and this is often interpreted to imply that DAGs are an inadequate description of reality. In this article, I show that the classical model of supply and demand, while certainly an example of simultaneous equations, is hardly an example of simultaneous causation, and furthermore that the statistical properties of the system can be explained by selection bias (not endogeneity bias). At the time scales that supply and demand are actually changing, a DAG could presumably represent key features of the system including price endogeneity. <h1>On a variable as the cause of itself</h1> Time and again, confusion arises as to whether two variables can cause each other at the same time. This is not two variables at time t causing each other at time t+1. This is two variables simultaneously causing each other at the same time. An example of this came up on X (aka Twitter), where the conversation focused on an author who had published separate papers concluding that creativity causes dishonesty and dishonesty causes creativity. The following exchange occurred (links to posts provided to respect users’ right to delete): <a href="https://medium.com/@baogorek/supply-and-demand-simultaneous-equations-not-simultaneous-causation-5252cf44fb29">Read More</a>