I promise to keep this short if you let me. I will not try to build all the arguments for city‑building. This is the kind of post you can read to understand a problem better and to think about the next move.
In this post I will keep the discussion simple, so that you can focus on the core of the problem: the cost of buildings outpaces income.
In the chart Phoenix was represented by ZIP codes with log density from 6–10 and log income from –1 to +1 on a log scale. The lowest price‑to‑income ratio in NYC ranges from 10% to 60% higher than in Phoenix, and the average income is about 10% higher in NYC. There is substantial evidence of agglomeration value in NYC, or at least of agglomeration that creates income for landowners.
We Made Cities Illegal
The first thing you might notice is the variance in price‑to‑income ratios across ZIP codes: in dense poor ZIP codes, the ratio is 2.5–3.5 points higher than in low‑density poor ZIP codes.
Cities across the country made that level of density illegal during the 20th century. Even New York had dense housing, but the densest parts have declined since the 1930s. To the extent the city continues to grow, less dense areas try to make up for the decline in the densest parts.
Queens has grown. Brooklyn and the Bronx have had strong construction, but the boroughs still lag when it comes to the supply of good new construction.
As I have discussed, the supply shortage is an issue for all ZIP codes, not just the supply of good construction where it matters for rent and cost.
The red solid lines in the figure highlight the density premium. In all three categories, the supply shortage is not small: it is a real loss of incomes.
The signals are there—people are willing to consider trading down to smaller units, longer commutes, etc.—but the overall affordability question is about scarcity.
Thoughts on Filtering
As I think about this, my focus has been on the affordability, not city‑building or quality of life. The real issue in the affordability question is the supply shortage.
I frequently criticize attempts to attribute too much importance to interest rates. Still, interest rates affect land values more than structures: the land value is sensitive to them. I think the best path forward is adequate construction so that land values remain small.
I may be hoping that Phoenix is becoming a city, some noise in the data. Could we hope to see medium‑high density ZIP codes where poor families pay 5x or 6x incomes for homes? That is the opportunity.
These are two different issues: the supply shortage in NYC remains severe. In PC 2022, low‑density poor neighborhoods sell for 12x income, mostly from scarcity.
Affordability Vs. City-Building
I write a lot about affordability, but the real question remains: the supply shortage fuels the difference in costs.
The issue with rent: people are trading down, but that doesn’t help the overall affordability question, which is defined by how much housing costs outpace incomes.
The zoning crisis forces the city to keep people out of the city. In Phoenix, the poor ZIP codes pay about double the price‑to‑income ratio that rich ZIP codes do, whereas in NYC, poor ZIP codes pay triple.
These data points point to the potential of Phoenix as an emerging city if it increases its supply of good construction.
It is an opportunity before us.
The Four Sources of Urban Home Values
So, with all that being said, the core takeaway is the affordability question: scarcity dominates as opposed to density, cyclical inflation, or agglomeration.
-
Agglomeration value
-
Cyclical inflation
-
Density
-
Supply shortage
1. Agglomeration Value
The agglomeration value in Phoenix is smaller because Phoenix didn’t lack housing before 2008. In NYC, the agglomeration value has little effect because it all goes to consumer surplus.
The lowest price‑to‑income ratio in rich ZIP codes is 10%–60% higher in NYC than in Phoenix.
2. Cyclical Inflation
From 1999 to 2005, price‑to‑income ratios increased 60–70% in Phoenix and the richest NY ZIP codes, a big cycle. Land values are sensitive to interest rates, and the cycle is linked to land value, not structure value. Real long‑term rates fell from the late 1990s to the mid‑2000s.
3. Density
Density value applies to cities such as NYC and Chicago. In high‑density poor ZIP codes, the price‑to‑income ratio is about 2.5–3.5 points higher than in low‑density poor ZIP codes.
In 2022, low‑density poor neighborhoods in NYC sell for 12x income, almost all from scarcity.
4. Supply Shortage
Supply shortage was sizeable in NYC by 2005 and remains significant in Phoenix. In NYC, poor ZIP codes pay roughly triple the price‑to‑income ratio that rich ZIP codes do, even without a density premium. In Phoenix, poor ZIP codes pay about double those of rich ZIP codes.
In 2022, Phoenix has some density value. Phoenix has a light‑rail line and growing young households in the denser parts of Tempe and Phoenix living without cars.
Conclusion
In the next post(s), I will look at a simple market such as Phoenix, because a simple market can help us understand why homes are selling for what they are and why we are (or aren’t) building them.
***
Technically, the 5 density categories represent ZIP codes with residences per square mile of 6, 7, 8, 9, and 10 on a natural log scale. Phoenix has no ZIP codes in this data set with log density of 6, 9, or 10. The income categories are the ZIP code with average log income equal to the metro area average, and ZIP codes with incomes 1 point above and below that, on a natural log scale. They are modeled as if the average property tax is 1% of market value in all ZIP codes.