YIMBY research wanted: upzoning and land prices

Does upzoning reduce the price of existing high-density land?

Upzoning reduces housing prices by increasing the supply of high-density land, which reduces high-density land prices; this reduces land input costs and hence shifts the supply curve for housing. So to evaluate whether upzoning works, the first step is to check whether existing high-density land becomes cheaper. This gives us a simple empirical test: take the set of high-density land that existed before upzoning, and track its price over time. However, no one has done this (yet).

For example, Auckland had 389 parcels zoned for terraced housing and apartments (THA) whose zoning status was unchanged by the 2016 upzoning.¹ The THA zoning category increased in size to 13,341 parcels, as lower-density zones were upzoned to THA. Intuitively, the land price of the initial 389 THA parcels will fall, as competition from the thousands of new THA parcels gives developers bargaining power over landowners. So we can see whether upzoning worked by testing if increasing the supply of high-density land made it cheaper.²

São Paulo also upzoned in 2016, changing the allowed density on a block-by-block basis. On average, the built area ratio (BAR)³ increased by 35%, with about half of blocks being upzoned. We can use this to test whether upzoning reduced the price of existing higher-density land at a fine-grained level. For example, if upzoning increases the size of the BAR=4.0 category, then we expect the price of the initial BAR=4.0 land to decrease. For categories that decreased in size, we might expect land prices to increase, as the zoning category is more scarce.⁴ Here’s a hypothetical graph:

If you have access to data on land prices or assessed land values in Auckland or São Paulo, please reach out! Existing research has not looked at this question, and it is straightforward to answer. Showing that upzoning reduces high-density land prices would give us empirical evidence for upzoning that doesn’t rely on complicated statistics. Even laypeople understand that land is an input into housing production, so lower land prices mean lower housing prices.

1. There were initially 1880 THA parcels; 389 remained THA and 1491 parcels were downzoned. See Table 2 in Cheung, Monkkonen, and Yiu (2024).

2. Similarly, there were 3461 Mixed Housing Suburban (two storeys, three dwellings) parcels and 4880 Mixed Housing Urban (three storeys, three dwellings) parcels with unchanged zoning status. Since upzoning increased the size of the MHS and MHU zoning categories, we predict that the land price of the initial parcels decreased.

3. The ratio of a building’s total floor area to the land area. At BAR=1, a 1-story building could cover the entire lot, or a 2-story building could cover half of the lot.

4. The effect could be positive or negative. If high-density housing is a substitute for low-density housing, then increased supply of the former would decrease the price of the latter, which would translate to lower low-density land prices.