The market always gives society exactly what they want within the constraints of what is possible — but humans may not be smart enough to live in a free market.
Links: Economics, Value and Profit, US Spending Per Student, K-12 Instruction vs. Administration, The Limits of Utopia, Scope Confusion, Civilizational Cycles, Accounting Identities as Domain-Matching — the meta-pattern (GDP-as-welfare-metric across tariffs and immigration), The GDP-Identity Tariff Fallacy — the symmetric form on tariffs
The market always gives society exactly what they want within the constraints of what is possible.
This is not a defense of outcomes. It’s a structural claim about the mechanism. The market is a distributed computation engine that solves supply and demand simultaneously through the price signal. It processes information as fast as reality allows. What economists call “market failures” are mostly just information that hasn’t propagated yet — and any intervention assumes the intervener has better information than the distributed market, which is almost always wrong (Hayek’s knowledge problem).
The market responds to effective demand — want backed by purchasing power. Dollar-weighted demand is not a flaw; it’s the signal. The weighting is the information. Those who generate more value for others accumulate more purchasing power, which gives their preferences more weight in future allocation. This is a feature of the mechanism, not a bug.
The teacher example: We say we want better teachers. We pay for nicer clothes. The market believes the credit card, not the survey. Revealed preference is the only honest measure of what people actually value. The uncomfortable truth is that the market is right to believe the credit card — that’s the revealed truth about what people value at the margin.
But: a person who puts their own money into private school is sending the correct market signal — public education isn’t delivering value proportional to cost. And then they rightly get angry at the ~$20k per student the government spends on their behalf in a system they’ve already rejected with their wallet. They can’t opt out. They can’t redirect. The market signal is being drowned out by compulsory revenue collection. Full enrollment in public schools isn’t demand — it’s a captive audience.
A persistent failure in economic analysis is evaluating only supply or demand, never both simultaneously. The market never makes this error — price is the intersection of both sides. But humans analyzing the market constantly isolate one variable and tell a story about it.
The immigration example: Most economists say immigration is good because GDP goes up. This only looks at the supply side. Immigrant workers also put demand on the economy — they buy things, use services, occupy housing. More demand and more supply is not automatically better.
The fix is simple: GDP per capita — the economist’s own definition of standard of living. If an immigrant contributes more than this measure, they are increasing the standard of living (GDP/cap goes up). If they don’t, they are decreasing it. The market, absent coercive distortions, would naturally select for the first type — employers would recruit productive workers because it’s profitable. Immigration policy that substitutes political criteria for economic ones is itself an anti-market force.
The formal version that makes the supply-demand symmetry explicit:
Immigrant's net welfare contribution = (marginal production) − (marginal consumption)
≈ (their productive output) − (their per-capita share of resource demand)
The three cases:
| Marginal production vs. GDP/capita | Effect on aggregate GDP | Effect on per-capita GDP | Effect on existing residents |
|---|---|---|---|
| Above the line (high-skilled, integrated, working-age, immediate productive contribution) | Up | Up | Better off on average |
| At the line (matching median resident profile) | Up | Flat | Unchanged on average |
| Below the line (low-skilled without integration, non-working dependents, language/credential barriers, service-heavy household composition) | Up | Down | Worse off on average |
Aggregate GDP rises in all three cases. Per-capita GDP — the actual standard-of-living measure — rises only in the first case, holds in the second, and falls in the third. The mainstream economist case (“immigration is good because GDP rises”) collapses these three by using only the aggregate metric. It treats the second and third cases as equivalent to the first, which they aren’t.
This is the symmetric form of the GDP-identity tariff fallacy: same identity-misreading, applied to the supply-side of immigration instead of to the −M term of the trade identity. Both arguments use aggregate GDP as the welfare metric when per-capita or distributional measures would tell a different story. See accounting-identities-as-domain-matching.md §”GDP-as-welfare-metric” for the meta-pattern; gdp-identity-tariff-fallacy.md for the tariff-side version.
The GDP/capita test handles the aggregate welfare question. The distributional layer is separate and also matters: immigration that increases per-capita aggregate may still hurt specific deciles (bottom-decile workers facing wage competition; lower-income residents facing housing competition; service-strained communities) while benefiting other deciles (high-end employers, asset-holders benefiting from housing-price appreciation, consumers of immigrant-provided services). The aggregate-vs-distribution distinction the vault has been building (see Lyn Alden trade-deficit analysis) applies here cleanly: a policy can raise aggregate per-capita welfare and concentrate the gains in specific deciles while concentrating the costs on others. Both observations can be true; treating either alone as the complete answer is the framing mistake.
The combined test: does the policy improve per-capita aggregate welfare, AND is the distribution of gains and costs defensible? Either alone isn’t enough. The mainstream “immigration is good for GDP” claim is silent on both questions.
The pattern repeats everywhere:
The market is a simultaneous equation solver. Humans are narrative thinkers. We do one side at a time and often stop before we get to the other.
Two categories, both real:
Government price controls, subsidies, tariffs, monopoly grants, regulatory capture. These distort the price signal the market needs to function. The market can’t give society what it wants if the information mechanism is being overridden. This is straightforward and widely acknowledged (if not widely acted upon).
It is possible humans are not smart enough to live in a free market.
The market is a perfect machine that requires an imperfect operator. The theory works — but it requires participants who can:
Humans often can’t. This is the 99-problems / Chesterton’s fence issue. Modern society buries so much complexity that consumers can’t make good choices. They end up relying on outside forces — media, brands, political tribes, experts — to do their thinking. And those outside forces have their own incentives, which are often anti-market.
The failure cascade:
Even the professional class trained to understand markets — economists — fails this test. They analyze one side of the ledger and call it analysis. If the experts can’t process both supply and demand simultaneously, the average consumer has no chance.
The market doesn’t fail. The participants fail the market. And the gap between what the market requires of us and what we’re capable of — that gap is where every problem lives.
Whenever government guarantees revenue to an industry — through subsidies, loan backstops, mandatory participation, or tax-funded allocation — the same distortion plays out. The mechanism is general. Education, healthcare, housing, and retirement are all instances of the same law.
That’s the whole law. Every industry that receives guaranteed government revenue follows this path. The specifics differ; the mechanism doesn’t.
Education is the cleanest demonstration because we can see the same mechanism at two levels — K-12 and higher education — with decades of data.
School vouchers fix one problem and leave a bigger one untouched:
The three steps at K-12:
Vouchers redirect the money but don’t change the amount or the compulsion. You still end up with costs rising faster than outcomes — just with more competition around the margins.
Federal student loans are the college-level version, and we have 30+ years of data:
In the 1980s, public schools still guided students toward vocational training as early as high school. This was closer to a market outcome — it recognized that people have different comparative advantages and matched them accordingly.
That system was dismantled by a narrative: “everyone deserves college.” It sounded moral but ignored supply and demand. The trades now face severe labor shortages because we told an entire generation that plumbing is beneath them. Meanwhile a master plumber out-earns most liberal arts graduates and carries no debt.
The pattern: a well-intentioned narrative overrode the market signal, backed by government money that made the bad decision feel free at the point of purchase. By the time the bill came due, the decision was irreversible.
TODO: Add data and references — tuition inflation vs. CPI, vocational enrollment trends 1970s–present, trade salary comparisons, student loan default rates, comparative international models (German apprenticeship system).
Social Security is the purest instance of the three steps because there’s not even a pretense of consumer choice.
Social Security proves the law generalizes beyond industries that receive government money to any system where government intermediates between the earner and the outcome. The mechanism is identical: guaranteed revenue, capture by the operator, and the death of the price signal.
TODO: Add data — Social Security ROI vs. index fund returns by cohort, trust fund depletion projections, international comparisons (Singapore CPF, Chile AFP).
Removing these distortions through democratic means requires an electorate that understands markets well enough to vote for market solutions. But the current education system — the one shaped by the distortions — doesn’t produce that electorate.
This may be the deepest version of the human-limits problem: you can’t get to free markets from here through the democratic process, because the voters have been shaped by the very distortions you’re trying to remove.
The market will likely find solutions anyway — it always does, working around the distortions. But the path and the timeline are open questions.
The tempting answer is intervention — regulation, mandates, expert panels. But this runs headfirst into Hayek’s knowledge problem: the interveners know less than the distributed market, and they have their own incentive distortions.
The better answer, though it falls flat in practice: be a better consumer. Learn how to negotiate. The market rewards informed participants. The difficulty is that this advice requires exactly the information processing capacity that most people lack.
This is an open problem. The mechanism is sound. The inputs are corrupted. And the corruption is partially structural (complexity) and partially self-inflicted (outsourced thinking).