Microeconomics at Booth: what it means to think at the margin
The real lesson from Richard Hornbeck's microeconomics course was not supply and demand. It was a discipline for deciding what actually matters at the edge of any decision.
Work in progress
This essay is being actively written. Check back soon.
Richard Hornbeck teaches microeconomics the way a good surgeon operates: precise instruments, no wasted motion, no tolerance for fuzzy reasoning. By the end of his course, I felt like something had been permanently adjusted in how I think through decisions.
The official content is standard MBA micro. Consumer surplus. Labor supply and demand. Competitive equilibrium. Deadweight loss from taxation and subsidies. Long-run pricing and market entry. If you have taken undergraduate economics, you recognize the topics.
But Hornbeck is not interested in whether you can recite the topics. He is interested in whether you have internalized a particular style of thought.
Thinking at the margin means asking: what changes by one unit?
That sounds simple. It is not, in practice.
One of the earliest problems we worked through involved a company producing in two plants — one in Detroit, one in Mexico City. The question was how to allocate production across the two locations to minimize cost. The answer is not to put everything in the cheaper location. The answer is to equalize marginal costs across both plants. As long as one more unit is cheaper to produce somewhere, produce it there. Stop when the costs at the margin are equal.
The same logic runs through consumer choice, hiring decisions, pricing, market entry, and capital allocation. The question is never "what is the average?" The question is always "what does the next unit cost or return?"
Once you start asking that question cleanly, a lot of conventional business reasoning falls apart. Sunk costs stop mattering. The relevant comparison is always forward-looking. Historical investment is irrelevant to current decisions except where it creates real constraints on future options.
The problem sets were hard. Eight of them across the quarter, each requiring you to apply the frameworks to novel situations without being told which framework to use. The midterm was closed-book and 90 minutes. The pressure was calibrated to separate students who understood the logic from students who had memorized steps.
What I found most useful was the section on externalities and subsidies. The classic case of deadweight loss from a tax is well-known. What sharpened my thinking was the reverse: when a market has a negative externality, a corrective tax can actually reduce deadweight loss, not create it. Efficiency is not the same as laissez-faire. The framework is not "markets good, intervention bad." It is about where the equilibrium falls relative to where it would be if all costs were internalized.
That nuance matters enormously when thinking about AI markets, data markets, and platform dynamics. The externalities in those contexts are not small.
The course did something harder to describe too. It gave me a common language for talking about trade-offs with precision. When I now work through a product decision or evaluate a business model, I have a more reliable vocabulary for describing where value gets created, where it gets captured, and who pays costs that do not show up in the price.
That vocabulary does not answer questions on its own. But it makes the questions cleaner. At Booth, that turns out to be most of the work.