Could NYC really build more without zoning?

I am arguing on Twitter about whether New York City (where I live) could really build a significant amount of new housing if zoning was less restrictive. One possible argument runs something like this: “New York is so built out that even if zoning was liberalized and more housing could be built, we could never build at 1920s levels again.”

Is there any empirical way to test this sort of argument? It occurred to me one possible technique might be to look at Sunbelt regions with a reputation for permissive zoning policies. Even in those places, central cities are presumably far more built-out than suburbs. So if the existence of a spare land was a major constraint, central cities would build at anemic New York City levels, and suburbs would build far more. So I looked at data for Austin, a boomtown with lots of new housing.

There were about 115,000 new units built in the city of Austin between 2010 and 2019. Given that Austin had 795,000 people in 2010, that means 145,000 units per million residents over a decade, or around 14,500 per year. There were about 170,000 units built in the rest of metro Austin (which had about 933,000 people) during this period, or 182,000 units per million residents per decade or around 18,200 per year. (See Census Tables S0101 for 2010 data, B25034 for housing data). This does not seem like a huge difference to me- so if Austin has much less spare land than its suburbs, that does not matter very much. Perhaps the higher demand for close-in land balances out the higher supply of suburban land.

How does New York and its suburbs compare? New York had 8.184 million residents and just over 241,000 housing units in the 2010s, so that’s about 29,000 housing units per million residents during the decade, or 2900 per year. In other words, if the city of New York had built at the same level as the city of Austin, it would have had over 1.17 million new housing units in the 2010s(145,000 times 8.1). What about the suburbs? The rest of the metro area had about 10.73 million residents in 2010, and built about 263,000 housing units in the 2010s, so that’s about about 24,000 units per million residents. In other words, the suburbs actually allowed less housing than the city, despite having more spare land. And if New York’s suburbs had allowed as much housing as Austin’s suburbs, there would have been over 1.9 million housing units built (182,000 x 10.7).

In other words: if spare land was what mattered, NYC suburbs would have allowed more housing than New York City. Instead, they allowed less housing.

And if spare land was what mattered, the differences between city and suburb would have been greater than regional differences (assuming all suburbs have more spare land than all cities).*

But instead, regional differences mattered more. Austin allowed five times as much housing as New York City but only 20 percent less than its suburbs.

*One major qualification: I have no means of proving or disproving this assertion. But the amount of suburban growth in Austin’s suburbs suggest that the city of Austin is fairly built out, and New York suburbs are generally so non-dense that I find it easy to believe that there is spare land in, for example, Suffolk County).