[larsiusprime]

It's not just zoning, though that exacerbates it. City centers are where economic activity tends to be concentrated so even without zoning you would see these exponential land value effects as you approach the centers of economic activity (and indeed, we did see these same relative patterns prior to the passing of modern zoning laws).

[AnthonyMouse]

Weren't transportation and communication a lot slower if you go that far back? Zoning laws started getting imposed not long after the invention of the telephone and the internal combustion engine and before either of them were in the possession of the majority of people.

The value of land in an urban area is obviously going to be higher than it is in a rural area just as a matter of scarcity, but it's not at all obvious that the reason it's currently so much more expensive in Manhattan than in the places directly adjacent to it isn't primarily a result of zoning just because that might not have been the reason in 1890.

[larsiusprime]

If you're talking about the Manhattan effect specifically, I'm sure that zoning has a large effect on the extremeness of it. If you're talking about "land gets exponentially more valuable in the city center", then you see that pattern everywhere and indeed we saw it in the historical pre-zoning period as well. Agglomeration effects are sticky regardless of what time period you're in, what changes is the coefficients on the exponential equation.

[AnthonyMouse]

You keep saying "exponentially" but let's look at the median rent for e.g. a 1 bedroom in Manhattan vs. the Bronx. It's around twice as much in Manhattan, because now you're measuring prices per-unit of indoor space rather than per-acre of land.

In a formal sense you can call this exponential. It's twice as much in Manhattan as the Bronx and twice as much in the Bronx as in some location even further out. But if you're trying to explain the >100x difference in the price of land between the Bronx and Manhattan, the ~2x difference in how much people value living there because of agglomeration effects is not the dominant effect.

And you would expect the densest places like Manhattan to have the strongest agglomeration effects. You can ratchet up permissively zoned land costs through zoning just by increasing the level of restrictions elsewhere, e.g. ban >2 story buildings instead of banning >5 story buildings, but the willingness of people to pay more in order to live near things is proportional to the number of things they would be paying more to live near.

On top of that, the zoning restrictions exacerbate the agglomeration effects. If you could build taller buildings in the places you currently can't then the premium commanded by housing in a dense area would go down by increasing the supply of it.

Suppose you have three neighborhoods. A has 10 story buildings and a median rent of $2000, B and C both have 2 story buildings and a median rent of $1000. You then rezone B with the result that new construction happens and it now has 5 story buildings. Local housing supply just increased by more than 20%. Rents are now $1700 in A, $1300 in B and $900 in C, because supply increased and some people moved from both A and C to B, which became more desirable. People then say "look, you've increased rents in B by $300" and blame agglomeration effects and try to argue against doing it. But the average rent in the city went from ~$1700 to ~$1500 and >20% more people now have housing.