Stewie and Peter Griffin on Economic Epistemology
The most important concept in social science
If you are chronically online you have surely seen the viral AI-generated videos of Stewie and Peter Griffen from Family Guy discussing erudite topics with Minecraft parkour in the background. The videos are educational content that typically assumes a FAQ format to explain subjects like computer science or math. A recent Instagram reel from the account @economics.ib features one of these conversations. Instead of scrolling past, it caught my attention because it succinctly describes one of the most foundational problems in social science. The problem is that “science” does not really work in social science and the belief that does, leads to broken conclusions.
In the reel, Stewie suggests that with future advancements in computational power and data collection, we will have plenty of information on transactions and prices, implying that we can predict market movements. Though Stewie himself doesn’t make the argument in the reel, a common second implication from those who hold this view is that with an adequate model, we can solve resource allocation problems.
Stewie and Peter’s exchange teaches a profound truism: society is not engineered like a machine. As polymath F.A. Hayek put it, “The physical sciences [study] fundamentally much more simple phenomena.” In a bout of wisdom, Peter refutes Stewie’s idea, explaining that the economy is not a Newtonian physics problem. Instead, economies are evolved orders, dynamic systems where people respond to things, changing their behavior in ways that make many determinations impossible.
In physical science there is a presumption which informs the background of scientific thought. The presumption is that the rules of the physical world are plausibly the same everywhere. This principle can be stated in different ways, and scientists have added many details to it. The main point is that if you control the conditions, you can do scientific experiments again and get the same results.
However, society and the economy, in the context of this video, are what are called complex adaptive systems that mutate and change over time. Peter correctly points out that Stewie is mistaken in thinking the main issue is just having enough data. People in society are not governed by fixed “laws” of the universe. Instead, they react to changing goals and limits that are filtered through their culture and psychology. “Society” is a non-deterministic system.
Some behaviors in people only happen when they see others behaving in a certain way. In other situations, groups of people can make chaotic decisions. Examples of the latter include the garbage can model, Hobbesian traps, and the security dilemma from political science, just to name a few. In game theory, there is the concept of “dynamic inconsistency,” which describes when a player’s preferred moves set them up for an unpreferred outcome down the road. This concept shows how our “current selves” can betray our “future selves.”
Obviously, there are logistical problems, too. Having price and transaction data doesn’t mean you can measure people’s thoughts and decision formulation. We generally can’t perform controlled experiments in social science like we can in the physical sciences. Of course, we have “natural experiments,” big data observations and case studies that we can have more confidence in due to statistical inference. For Harvard economist Raj Chetty, this is enough to lump economics in with the physical sciences. This seems to be a stretch and is a view not shared by most in the field from my experience.
Sure, to create a demand curve for a product, we need statistical measurements. However, this doesn’t mean that the quantity demanded or price of the product is fixed or follows rule based patterns. Additionally, the advent of behavioral economics, the scientific study of psychological factors that inform decisions can’t be appreciated enough. While advancements in data science will surely improve our ability to analyze trends, they will not overcome the underlying problem of it being impossible to model the economy as we do with physical systems because we are dealing with the subjective feelings and emotions of billions of people who are often times motivated to be contrarian. If people knew that there were models out there predicting certain outcomes, they might alter their behavior to capitalize on the predictions, an observation related to what is called the Lucas Critique in macroeconomics.
A case in point is the stock market. There are a hundred years of pricing data and incredible amounts of data on global transactions down to the milliseconds. There are countless participants, information on companies is easily accessible, there is data from all over the world, there is free entry and exit, and securities are homogeneous in that one share of Google is as good as another share of Google. It’s not unreasonable to think that the stock market should be the easiest thing to model in economics by a large margin. Yet, nobody can accurately predict transactions, especially in the long term, despite investors’ motivation to figure it out due to the obvious advantages (maybe if Max Cohen finished Euclid?). Renaissance Technologies’s Medallion Fund run by the late James Simons and a group of mathematicians affiliated with UC Berkeley is the closest success story in my estimation.
This fundamental distinction between simple and complex phenomena usually rears its head in popular politics. When people view economics or public policy as a physics problem, they are likely to attempt to “scientifically manage society” and produce results from the top down. Government policies that rely on centralized information face a major problem, namely, the information needed is not tangible. Politicians, bureaucrats, and policymakers try to make decisions for their constituents, assuming they can read the data or survey the public and make informed choices. However, doing this replaces the evolutionary pressures of the system with human decision-making. This blunts the system’s ability to adapt and evolve, similar to how natural selection works in Darwin’s theory or creative destruction in economist Joseph Schumpeter’s theory.
If Peter Griffin can teach us anything, it is that society, or the economy in this instance, is not a physics problem. Understanding this distinction is essential for anyone interested in the philosophy of science, economics, public policy, or the role of artificial intelligence in decision-making.
