Xan, a second year economics PhD student at the University of Chicago, opened my eyes to the crucial roles played by constraints.
First, he illustrates how they shape worldviews:
A lot of people's politics and worldview can be explained by what constraints they allow to operate when they are proposing a solution to a given problem. But constraints are part of the definition of a problem, so different constraints actually correspond to different problems, with different solutions. A lot of disagreements come directly from the fact that the parties are actually solving different problems but giving it the same name. If people realized this, the debate would go to the important question of which constraints we should be allowing to characterize the problem.
The answer to that depends on what reality is really like, and what question we are really trying to answer. I'm always happy to entertain different assumptions for theoretical purposes, but when it's time to solve the problem of What Should I Actually Do, then my goal is to realize which features of reality I can control, and respect the ones I can't. In particular, I am constrained by what I can get other people to do.
Frequently, people believe that different constraints are operating in reality, without explicitly saying so. People who realize and focus on this are much better at cutting through what others are saying and getting to the source of their position. For example, here is Tyler Cowen explaining the *fundamental* reason Paul Krugman is a liberal:
Many of Krugman's current false (modal) predictions stem from his claims that if left-wing politicians would "get tough" and take their case directly to the public, good progressive results will follow. I view that claim as a move into a non-scientific mode of thought. While it is sometimes true, usually it is not, and there is plenty of political science literature on how hard it is to form a winning political strategy through rhetoric.
Without such a view, however, Krugman would have to entertain the possibility that moderate outcomes, or sometimes observed outcomes, are more likely second-, third-, or fourth-best efficient than he would like to admit. If you took away this one rather weak prop of his worldview, he could quite readily turn into a conservative, of course in the literal rather than the right-wing partisan sense.
And then he illustrates their crucial in economics and life more generally:
I think the perspective falls right out of economics once you dig deep enough, if you are the kind of person who likes to step back (which not every economist is). I think one of the major differences between undergrad and grad econ for me has been the overwhelming emphasis on constrained optimization. In some sense that's really what economics is about. It's sort of a generalization of "optimal resource allocation in the face of scarcity," recognizing that not all constraints necessarily come from some sort of scarcity. Maybe you just don't want to do what I want you to do. That constrains me too.
Undergrad econ has stuff like "max utility subject to having only $1,000". Grad econ has budget constraints, but also feasibility constraints, information constraints, "incentive compatibility constraints"....really anything that keeps you from just having things the way you'd really like them. Constraints are ubiquitous in real life, where we rarely get to have things just the way we'd like them. Constraints are also what make problems (and hence real life) interesting.
