Or how pathetic planners are:
bsky.app/profile/mich...

If doing a mass upzoning, then IZ needs to be calibrated to the equilibrium land lift for all parcels (here, 0.8).

Eg. in a two-zone model, if upzoning one parcel at a time, the land lift shrinks with each upzoning, so the IZ requirement needs to adjust each time.

michaelwiebe.com/blog/2025/07...

Yes, funded IZ works by capturing new land value. As the new land value shrinks, so too must the IZ requirement.

Why increasing supply reduces prices:

When the supply curve shifts right, and demand is unchanged, there is excess quantity supplied at the original price: at P1, suppliers produce Q2, but demand is Q1. To clear the market, the price must fall to P2, where supply = demand = Q3.

But 2D AMM is overpowered for this result. We can use a single-location model; the simplest case is expanding the UGB, ie. 'upzoning' from agricultural to residential, which increases the stock of residential land and decreases its price. Stay tuned...

7/7

Greenaway-McGrevy extends AMM to a 2D disk. This allows us to increase the share of apartment-zoned land at a given distance (partial upzoning!), and calculate the cross-parcel effect: dr_H(x)/dω_H(x).

bsky.app/profile/mich...

6/

Diamond et al. (2025) run a counterfactual where the city is uniformly upzoned for triplexes. This doesn't capture the cross-parcel effect, where upzoning one additional parcel for triplexes reduces the price of existing triplex-zoned land.

z-y-huang.github.io/assets/pdf/U...

5/

Kulka et al. (2026) model a uniform zoning constraint in neighborhood-zone nj, where each nj is a separate market. So upzoning affects all parcels in the market, and there's no cross-parcel effect.

drive.google.com/file/d/1CSqS...

4/

Measuring the cost of land inputs to housing construction We propose a new market indicator to track land input costs in housing: Land Cost to Floor Area (LCFA), defined as the ratio of land price to the permitted floor area ratio (FAR) set by local land ...

The problem: existing models of upzoning only do total upzoning.

Standard AMM assumes a single density constraint at distance x; so upzoning is relaxing the constraint for all parcels at that distance. But we want to upzone only some parcels.

www.tandfonline.com/doi/full/10....

3/

This 'partial upzoning' captures both effects.

In contrast, 'total upzoning' models only the own-parcel effect: if we go from all house-zoning to all apartment-zoning, there are no initially apartment-zoned parcels, so P1A is undefined and we can't calculate P2A - P1A.

2/

When we upzone one parcel from house- to apartment-zoning, there are two effects:
1) own-parcel: the upzoned parcel increases in value from P1H to P2A.
2) cross-parcel: increasing the stock of apartment-zoned land reduces its price from P1A to P2A.

1/

Coincides with rise of inclusionary zoning?

Can you measure the increase in land value due to upzoning? This tells us how stringent the zoning constraint was.

wordpress.clarku.edu/juzhang/wp-c...

This is per sqft?

Yup

bsky.app/profile/mich...

@pushtheneedle.bsky.social

But when zoning is a constraint, upzoning has an option value effect, so land values would increase with allowed density. In this case, the right image would have land prices increasing in density, though land cost per home still falls.

4/4

When zoning is not a constraint, the option value of further upzoning is zero: upzoning does not increase land values. Here, all parcels are already zoned for 5-story apartments, and apartment-developers and house-builders pay the same amount for land.

3/

When demand for land is perfectly elastic, increasing the supply of land doesn't change the price. But this is not true in general, and so land values increase less-than-proportionally compared to density. Hence, land cost per sqft falls.

2/

Meme from Patrick Condon claiming that increasing density by 4x merely increases land prices by 4x, with no saving in land cost per sqft of housing.

Image from pushtheneedle, showing that when land prices are constant as density increases, the land cost per home falls with density, which reduces the price per home.

Two incorrect models of land and housing
Left: land values and density increase 1-1, so land cost per sqft of housing is constant. This assumes perfectly elastic demand for land.
Right: land prices are constant as density increases. This assumes zoning is not a constraint.

1/

The parcel-level view also misses the cross-parcel effect: upzoning one parcel increases the stock of apartment-zoned land, which reduces the scarcity premium of existing apartment-zoned land.

9/

If option value and density both increase 1-1, then V/N is constant. With the market-level view, we can see that this is assuming perfectly elastic demand for land. This is what Condon is doing:

michaelwiebe.com/blog/2025/09...

8/

For land value V and number of homes N, land cost per home is V/N. On the parcel-level view, we can say that V increases and N increases; but we don't know that V/N decreases. That requires a supply and demand model of land, ie the market-level view.

7/

Yup. One wrinkle here is that supply may be more inelastic right now, due to the Iran conflict.

Two views of upzoning:
1) parcel-level: upzoning lets you build more homes on the same land, which ⬇️ land costs per home.
2) market-level: upzoning ⬆️ the stock of apartment-zoned land, making it cheaper, which ⬆️ the supply of apartments.

1/

Tax incidence for Canadian consumers depends on relative elasticities for Canadian consumers and producers.

Car crashes

No, that's the point: if they really believed in their zoned capacity numbers, they wouldn't need any obstructions at all.

The logically consistent NIMBY: