Positional goods are the zero-sum goods — and “zero-sum” is a property of a good, not of an era. That identity is the page. The popular story (history was zero-sum until ~1800, then innovation moved us to a growing pie) is a temporal frame doing a categorical frame’s job: nothing about 1800 changed land, which was zero-sum then and is zero-sum now. What changed is that a vast non-positional sector grew up alongside it and swamped it by volume. So we never exited the zero-sum world — we grew a bigger one next to it, and the residual is exactly where housing, rank, and admission live, which is why the growth is real and nobody feels rescued. A positional good is one whose value depends on your share relative to others, so producing more cannot help everyone at once. The vault’s addition — Chris’s — is that positionality has two sources, and only one of them is real: goods that are naturally fixed (rank, status, prime location) and goods that are manufactured positional by policy that caps a supply which could otherwise grow (zoning, licensing, accreditation). The empirical payoff is sharp: compare 1963 to now — measured in hours of median labor, not a price basket (why) — and the goods that got dramatically cheaper are the non-positional ones, while the ones that got dramatically dearer — housing, healthcare, education, childcare — are almost exactly the restricted list. So the cost-of-living complaint is not evidence against growing-pie. It is evidence of where growth was blocked.
Links: Economics, Subsistence vs. Participation (the normative twin — this page is about goods, that one about needs), Inflation (Baumol’s cost disease — the rival explanation for the expensive list, §4), Measuring Inflation (the basket fight this page sits underneath), The Weighting Problem, Absolutes and Differentials, The Productivity–Pay Gap, Poverty in America Is a Sign of Exploitation (prep), The Inequality Optimum (the normative layer — this page says where inequality is unavoidable, that one asks what it is for), Theory of the Firm, Scope Confusion
Thesis page (portable). Origin: the Econ Nerds inequality review, where “inequality is only bad if the world is zero-sum” needed exactly one qualification and got none. Twice discussed in conversation before this and never persisted — the harvest this page recovers.
Growing-pie survives. The Industrial Revolution really did make more of nearly everything: a farmer fed five people in 1800 and feeds over 150 today, using less land, water, and labor. Anyone arguing that the pie doesn’t grow is refuted by the food supply alone, and this page is not that argument.
Everything else here follows from one equation, so state it before anything else:
“Zero-sum” is not a description of an era. It is a property of a good. And the goods that are permanently zero-sum are exactly the positional ones.
This is where the video’s framing goes wrong, and it is a temporal frame doing the work a categorical frame should do. Its story is that history was zero-sum (land, conquest, the DNA record) until ~1800, when innovation moved us into a growing-pie world. That reads as a claim about time: we used to be in the zero-sum era, now we’re in the other one.
But zero-sum-ness was never a fact about the calendar. It was a fact about land — the one resource that then dominated, and a resource that is still zero-sum today. Nothing about 1800 changed land. What changed is that land stopped being the dominant good, because a huge class of non-positional goods (food, cloth, fuel, tools, later computation) became mass-producible and swamped it in economic weight.
So the correct statement is not “we left the zero-sum world” but:
We did not exit the zero-sum world. We grew an enormous non-zero-sum sector alongside it, and the zero-sum sector — land, rank, prime location, fixed slots — is still there, unchanged, and now constitutes most of what people actually fight about.
That reframing does two things at once. It rescues growing-pie (the sector really did appear, and it is most of the economy by volume), and it explains why nobody feels rescued — because the residual zero-sum sector is precisely where housing, status, and admission live. §3 shows this is not a metaphor: the goods that stayed expensive are the goods that stayed zero-sum.
But “more for me is not less for you” is a claim about a good, not about the economy. It holds for pizza, cars, calories, clothing, computation. It fails for:
For these, relative wealth is the allocation mechanism no matter how large the pie gets. The pizza analogy silently assumes every good is non-positional; it is the load-bearing premise of every growing-pie argument and it is almost never stated.
The correct formulation is therefore not “inequality is fine because the pie grows,” but:
Inequality is not necessarily bad — a universal claim dies to one counterexample — and it is not automatically benign either, because for positional goods the distribution is the allocation.
Here is where most public argument goes wrong in both directions. The expensive-and-scarce list is treated as a natural fact by one side and as pure greed by the other. It is mostly neither. Ask one question:
Could we make more of it if we were allowed to?
| Good | Naturally fixed? | What is actually capped |
|---|---|---|
| Land | Yes | geography |
| Housing | No | zoning, height limits, permitting, lot minimums — you can build up |
| Being a doctor | No | licensure, residency slot caps, scope-of-practice rules |
| A degree | No | accreditation, campus capacity, prestige-by-exclusion |
| Elite status | Yes | arithmetic |
| Childcare | No | staffing ratios, licensing, facility codes |
Land is fixed; housing is not. That single distinction dissolves a large fraction of the housing debate. The scarcity that people experience is overwhelmingly in the manufactured column — the supply could grow and is prevented from growing. Which means:
This is Olson’s institutional sclerosis and regulatory scope creep showing up as a price. Incumbents and regulators converge on higher entry barriers for entirely different reasons — rents on one side, legibility and blame-avoidance on the other — which makes it a clean specimen of convergent interests: nobody needs to have planned a housing crisis for one to be produced.
The claim is checkable, and the pattern is the whole argument. Run the comparison in time-prices — how many hours of median labor buy the thing — not in a price basket. The basket apparatus fails here for reasons Measuring Inflation already documents (there is no correct index; the basket re-populates rather than re-weights; new goods break it — the BLS telecom index was off 0.8–1.9pp/yr over 1988–97 purely from omitting cell phones). Time-prices need no deflator, no basket and no base period, which is why they can settle this:
| Got dramatically cheaper | Got dramatically more expensive |
|---|---|
| Food, clothing, appliances | Housing |
| Communications, computation | Healthcare |
| Entertainment, travel | Higher education |
| Durable goods generally | Childcare |
The left column is the non-positional basket. The right column is the restricted basket. That is the finding, and it explains the thing that otherwise looks like mass irrationality: people consume more than any generation in history and feel squeezed. Both halves are true, and they are true of different goods.
It also reframes the CPI-basket fight (Measuring Inflation). Whether “inflation” reads as mild or catastrophic depends on how heavily the basket weights the restricted sectors — so the basket dispute is downstream of this page, not independent of it. Weight the abundant goods and everything is cheap; weight housing and healthcare and nothing is affordable. Both baskets are honest. Neither is neutral. That is the Weighting Problem in its most consequential application.
Baumol’s cost disease (Inflation §) is the innocent account of the same right-hand column: in sectors where productivity cannot rise — because the service is a person’s time — costs rise relative to sectors where it can. That is real, it needs no villain, and it predicts exactly this list.
So the two explanations overlap and must be separated:
| Baumol predicts | Manufactured positionality predicts | |
|---|---|---|
| What drives cost | labor-intensity — cost tracks how irreducibly human the service is | entry barriers — cost tracks how restricted supply is |
| Discriminating case | a labor-heavy sector with free entry should still inflate | a labor-light sector with heavy restriction should inflate anyway |
Housing is the discriminating case, and it goes against Baumol. Construction is not a stagnant-productivity handicraft; prefabrication and modern methods have improved substantially. Yet housing costs exploded — because the binding constraint is the land and the permit, not the labor. Baumol cannot explain a cost rise concentrated in the entitlement rather than the building.
Conversely, live performance and elder care fit Baumol cleanly and are not heavily restricted. Both mechanisms are real; the mix differs by sector, and the test is whether cost tracks labor-intensity or entry barriers.
Stated as a claim rather than a question, because it follows directly from §1 and Chris has now confirmed it twice:
Positional goods are defined by relative share. A good defined by relative share cannot be made abundant. Therefore no level of production ends scarcity — scarcity is not a quantity of stuff but a **category of good, and that category does not empty.**
Chris: “since there are positional goods, there will always be scarcity as these are defined as such — not everyone gets the house on the lake!”
The lake house is the right image because it is not a technology problem. Build more houses, build better houses, build houses more cheaply — there is still one shoreline, and the allocation among claimants remains. Contrast calories, computation, or fabric, where “more” is a straightforwardly available answer.
Consequences worth stating:
Where it could fail: if the salience of positional goods is itself variable — if a society can care less about rank without becoming poorer — then positional scarcity persists while ceasing to bind. That is a claim about culture, not production, and it is the one route around the argument. Nothing here rules it out.