I missed this piece by Alon Levy, in which Alon takes issue with the notion that NIMBYism is really just a collective action problem. The theory is that no single neighborhood can, on its own, lower prices. But if residents of high-priced neighborhoods would cooperate on a broad, city-wide upzoning, home prices would fall. Everyone would be better off: the city would be a less expensive, more interesting, more vibrant place. Property owners would be better off, too. Land values would rise even as housing prices fell, because owners could build more stuff on their land.

Alon argues that if the problem was with metropolitan-wide regulations, then individual neighborhoods would have an incentive to seek upzonings. Neighborhood-level upzonings would not affect overall prices much and therefore would increase neighborhood land values, and so neighborhoods would have an incentive to “cheat”. In reality, land use regulations tend to be more permissive when promulgated at the metropolitan or provincial level. He argues that individual neighborhoods are actually choosing regulations that enable them to extract high rents from renters (particularly job seekers), and so have no incentive to upzone their own neighborhoods under any condition.

I basically agree with Alon but I think the point can be made simply by talking about elasticity of demand.

People loosely talk about city-wide or metropolitan-wide housing markets as if they were a unified market, but in reality a city’s housing market consists of a bunch of individual neighborhood housing markets, each with its own demand schedule. The reason is that amenities — the attractiveness of the housing stock, parks, school quality, proximity to job centers or retails centers, average resident income, etc. — vary wildly from neighborhood to neighborhood. We take it for granted that neighborhoods are effectively independent housing markets when we shop for housing, and are not surprised when home prices rise or fall a hundred thousand dollars just by crossing a street or railroad track.

Neighborhood housing markets are of course linked. The demand schedule for housing in one neighborhood depends not only on the amenities it offers, but also the availability of substitutes and metropolitan demand factors such as job growth. Some neighborhoods have good substitutes; some don’t.

If we recognize that a neighborhood housing market has its own demand curve, it makes sense to talk about the elasticity of that demand curve. The elasticity of that curve dictates whether homeowners have an economic incentive to support an upzoning.

If we assume homeowners want to maximize their land value, they will maximize (as Alon notes):

(P – C) * ρ

where P is the price of housing, C is the cost of construction, and ρ is density of units per area. This is the value of land or, equivalently, homeowner profit.

If neighborhood residents collectively act like a monopolist (i.e., they are a cartel), they will have an incentive to choose quantity so that marginal revenue is equal to marginal cost. That is,

P*(1 + 1/η) = C

where η is elasticity of demand. Demand is “elastic” — that is, large changes in quantity cause small changes in price — if ⌈η⌋ > 1. Demand is “inelastic” — that is, small changes in quantity cause large changes in price — if 0 < ⌈η⌋ < 1. (η is always negative).

The “elasticity of demand” need not be constant for a particular demand curve. Demand ordinarily becomes more elastic as you slide up the curve (higher price/lower quantity) and more inelastic if you slide down the curve (lower price/higher quantity).

If demand is inelastic, then marginal revenue is negative. That is, any increase in production will decrease total revenue. In fact, if demand is inelastic, a monopolist can increase total revenue (and profit) simply by reducing production. This is why monopolists are always assumed to be working on the elastic part of the demand schedule.

This is all very basic Econ 101 stuff. But the implication is simple: if a neighborhood is setting quantities of housing in order to maximize land value, it will oppose an increase in quantity (an upzoning) if it’s operating at the inelastic part of the demand schedule.

You can’t simply assume that a neighborhood, like any other monopolist, is facing elastic demand. A neighborhood cannot reduce quantity to raise prices. After all, neighborhoods can’t order homeowners to tear down their houses. So if a neighborhood finds itself confronting inelastic demand, it’s stuck there. All it can do is sullenly refuse to increase quantity, and complain that it’s already got too many people.

People do assume, though, that demand for neighborhood housing in expensive cities is elastic, probably because the neighborhoods are so terribly expensive and the neighborhoods seem to have good substitutes (namely, every other terribly expensive neighborhood). But if you assume that neighborhood residents are acting in their economic self interest, then perhaps you should start by assuming inelastic demand and insist on evidence it’s not true.

A corollary of inelastic demand is that you really can make an expensive neighborhood cheaper just by building housing in that neighborhood. (And the way to make a city more affordable is by making a bunch of neighborhoods more affordable.)

Let me raise one other nuance. Even if we assume that the demand for housing in a neighborhood is elastic at a particular price point and that MR > MC, that does not mean that MR > MC for any quantity of housing. It might be that increasing the quantity of housing by 10% would increase land value, but that increasing it by 100% would decrease land value.

This is where the cartel structure comes into play. If the neighborhood land was owned by a single owner, then the owner would just increase housing by 10%. But a neighborhood consists of, say, 1,000 different owners each on his own lot. A neighborhood can’t permit each owner to build one-tenth of a housing unit. In other word, a neighborhood can’t increase supply by a small amount and do so uniformly. Housing comes in discrete, indivisible bundles. A 10% increase means that a small percentage of owners will get to add a housing unit and make a lot of money, but others will just see a drop in the price of housing. There’s no way, consistent with the cartel structure, to implement a small increase in quantity.

So neighborhoods tend to oppose small upzonings.

The only sort of upzoning that would muster neighborhood support in light of internal cartel politics would be a large upzoning, but large upzonings are likely to reduce the value of neighborhood land.

So neighborhoods tend to oppose large upzonings.

(People systematically underestimate how much land SF-zoned neighborhoods contain. One might assume that the homeowners would make a fortune if their neighborhoods were upzoned to Barcelona densities, but the fact is that even a relatively modest and uniform increase in entitlements will double or triple the population of a single-family neighborhood. There’s a limit to the number of people who will pay a premium to live in a neighborhood, no matter how nice.)

Let me emphasize that this is not a welfare analysis. For starters, I haven’t considered the welfare of the very large number of people who are shut out of these neighborhoods. Econ 101, again, tells us that monopoly pricing reduces total welfare. My point is that homeowners might have powerful economic incentives to oppose upzonings no matter how modest and no matter the eloquence and passion of the appeal.


2 thoughts on “Housing supply and land values

  1. “A neighborhood can’t permit each owner to build one-tenth of a housing unit. In other word, a neighborhood can’t increase supply by a small amount and do so uniformly. Housing comes in discrete, indivisible bundles. A 10% increase means that a small percentage of owners will get to add a housing unit and make a lot of money, but others will just see a drop in the price of housing.”

    What about teardown + rebuilds, ADUs, additions, etc.? With small increases in allowed buildable ft2 and/or building footprint, wouldn’t these incremental supply increases be relatively widespread?

    From my knowledge of a couple Korean cities, when buildable ft2 or FAR increases by 2x, even in an entire neighborhood, land values absolutely have NOT declined. (Best example is Busan Haeundae, where FARs were jacked up from 4x to 6x in 2009 plus provisional FARs up to 9x are granted to “good” projects…land prices there are way up since, and the big loosening was a key catalyst; this is a huge neighborhood as well.)

    I am well aware that, in Haeundae, even plots with already-10-story buildings have appreciated significantly in value post the loosening.

    Another example from Korea is for Newstay apartments; Newstay is a new marketing term used by the government for corporate-owned rental highrises (similar to REIT-owned complexes except non-delinquent tenants must have the option to renew rental for up to 8 years with very small rental increases).

    Originally the Newstay FAR was 3.0x (vs. typical 2.5x Korean FAR) but it was increased to 5.0x recently. Right as this happened, land prices skyrocketed in the 8 large neighborhoods nationwide where Newstays will initially be allowed. In one neighborhood, supposedly land prices went up by 6x.


Comments are now closed.