there's a phrase you hear constantly in political and business discourse — "let the market decide" — that sounds like a concrete statement about process but is actually empty without knowing what market, with what rules, enforced by whom. every market is a set of rules. "let the market decide" means "let these specific rules determine the outcome." changing the rules changes the outcome. this is not controversial, but people treat it as if it were.
the kidney exchange market is the canonical illustration of this. for decades, kidney transplants faced a matching problem that killed people. two couples: one has a willing donor who is biologically incompatible with their family member's recipient, the other couple has the opposite problem. each couple has half of what the other needs. you'd think they could trade — one donor gives a kidney to the other recipient, the other donor does likewise. this is straightforwardly beneficial for everyone. in practice, it mostly didn't happen, because there was no mechanism for it. no infrastructure, no clearing house, no legal framework, no way to ensure both transplants happened (the coordination problem if one pair defected). alvin roth (nobel prize 2012) built the mechanism. he designed the matching algorithm, worked out the legal structure, built the chain-formation logic that makes large multi-hospital swaps possible. kidney exchanges now save thousands of lives a year. the transplants were always mutually beneficial. the market required a designer before it could work.
school choice is less dramatic but equally instructive. cities have to assign kids to schools. parents have preferences. schools have capacity constraints. the question isn't whether to have a "market" in school seats — you have to allocate them somehow — the question is by what mechanism. the boston mechanism, used for decades, worked like this: families ranked their preferred schools, and the algorithm gave each student their first choice if space was available, otherwise their second, and so on. sounds reasonable. the problem, which economists eventually showed formally, is that it creates incentives for families to not honestly report their preferences. if your first choice is competitive and your second choice is nearly full, ranking your true first choice risks getting shut out of both. the rational strategy is to rank the school you're likely to get over the school you actually want. families with better information about the game — typically wealthier, more educated families — do better. the mechanism was systematically disadvantaging the people it was nominally designed to help. boston eventually switched to a deferred acceptance algorithm that is strategy-proof — you can't do better than honest reporting — and the outcomes improved.
the lesson is not that boston was evil. it's that they had a market — an allocation mechanism — that had been running for years with nobody asking whether the mechanism was actually doing what they thought. the mechanism had its own logic, its own incentive structure, and that structure had consequences that were invisible until someone bothered to look.
spectrum auctions are the same story in a high-dollar context. for most of the 20th century, the FCC allocated radio spectrum through administrative hearings — companies would apply, regulators would evaluate them, spectrum would be assigned. this was obviously inefficient (it took years, invited lobbying, and failed to put spectrum in the hands of whoever valued it most) but it took until 1993 for congress to authorize auctions. the early auctions immediately generated billions of dollars, but the more interesting story is what happened when the FCC tried different auction formats and got different results. simultaneous multi-round auctions — where all the spectrum goes up at once, bidders can watch what's happening and adjust — produce different outcomes than sequential single-round auctions. different rules. different winners. different prices. different efficiency. "let the market decide" means nothing unless you specify which auction.
all of this should be familiar to anyone building two-sided platforms, which is to say most startups. uber's matching algorithm is market design. so is its surge pricing formula. when uber suppresses surge pricing during certain events (airports, crises), it's choosing to subsidize rider experience at the cost of driver supply — a policy choice, not a neutral decision. airbnb's review system is market design. the choice to make reviews public vs. private, to allow responses, to weight recent reviews more, to remove low-volume reviews — each of these affects what behavior the mechanism incentivizes and who wins and loses in the platform. google's ad auction is one of the most sophisticated pieces of market design ever deployed at scale: second-price auctions, quality scores, reserve prices, advertiser controls — economists worked on this. it's not "an algorithm." it's an institution, and the institution reflects choices about what outcomes to prioritize.
the phrase "let the market decide" is usually invoked by people who benefit from the current rules and don't want to discuss changing them. that's fine! that's politics. but it's worth being clear about what they're actually saying, which is: "let these specific rules, which currently favor me, continue to determine outcomes." markets don't emerge from nature. they're built by people, reflecting choices about who participates, what they're allowed to do, and what happens when the rules are violated. the good ones are designed well. the bad ones cause the same chaos as any other badly designed institution, and then everyone blames "the market" for failing them.