How big is the housing shortage?

Two teams of researchers recently released estimates of the U.S. housing shortage – and they differ by a factor of five. Is the national shortage 20 million homes or just 4 million? With a range that big, both published by pro-housing groups, you’d be forgiven for thinking this is an exercise in futility.

But look under the hood and each estimate is asking and answering a different question. Together they offer a useful parallax on the current cost crisis. To oversimplify, here’s how I’m thinking about the studies:

• America needs to find space for about 4 million more households.
• City housing deficits add up to about 20 million dwellings, but the total is less than the sum of the parts.
• If we deregulated everywhere, high-priced places would build much more housing than Up For Growth predicts. And moderate-priced places would build much less housing than the JEC predicts.

JEC: 20 million

The easier report to understand is that produced at the Joint Economic Committee by the classically-named duo Kevin Corinth and Hugo Dante, the latter a GMU/Mercatus alumnus. They use a straightforward supply and demand framework with some simple, defensible assumptions about the housing production function and the demand curve within each county.

If I were going to calculate a housing shortage, I would do it along these lines. In this framework, the shortage tells us the gap between the number of houses that exist and the number that would be demanded at the cost of construction plus a modest land cost.

Pros: The JEC model gets the geography right. The biggest housing shortfalls are in high-cost, high-regulation (and/or land-constrained) places: Hawaii, DC, California, and Massachusetts. The magnitude of the shortages there – 30 to 35 percent of the housing stock – are believable given the magnitude of the problem.

Cons: The model treats counties as closed markets. This runs into two problems. A minor one is that they overstate the impact of the proposed HOUSES Act – which would imitate the excellent Southern Nevada Public Land Management Act for other regions – because some flat federal land in urban counties is too far from employment centers to be worth developing. That’s easy to fix. The bigger problem is that housing demand in one county is measured holding all other counties fixed. If Los Angeles built a million new homes, housing demand in Las Vegas, Phoenix, and San Bernardino would drop. That means – thankfully – that the national shortage is less than the sum of all the local shortages.

Up For Growth: 4 million

Up For Growth, the most effective pro-housing organization in Washington, branded its novel approach to the homes gap: Housing Underproduction™. The name points to the policy solution. But the model’s guts are on the demand side. As its helpful graphic below shows, Housing Underproduction™ is an estimate of how many households there ought to be. Like JEC, Up For Growth’s researchers, including Michael Wilkerson of ECONorthwest, used reasonable assumptions to estimate a gap between how households currently match up to the housing stock and how a healthier distribution of household sizes and vacancy rates would look.

Pros: Because it counts people, not prices, Up For Growth’s local estimate add up to a national number without double-counting. And its methodology identifies housing need on a more sociological basis. Notably, low-income places with large Hispanic populations, such as San Antonio, have very high rates of Housing Underproduction™, because crowding is so high. (That’s an area for further research).

Cons: Up For Growth’s local estimates miss the places where housing needs are blatantly enormous. They reflect the places where people who need more housing are, not the places they would like to be. Up For Growth finds bigger gaps in San Bernardino and McAllen than San Jose and Austin. That’s not because people love low wages and scorching summers; it’s because crowding and moving to cheaper cities are coping strategies that people pursue in tandem.

A general equilibrium thought experiment

Both methodologies ultimately fall short on two sides of the same coin: people and prices are linked across the country. Up For Growth identifies shortages of housing where people are (and calls it Housing Underproduction™); JEC identifies the underproduction of housing where prices are high (and calls it a Housing Shortage).

What would actually happen if we unleashed a massive, deregulated building boom?

• Places with the highest prices, like Los Angeles, would build the most & densest housing.
• Vacancies would rise and rents would fall.
• At lower rents, crowded Angelenos would move into bigger homes, happily sliding down their demand curve.
• Some residents of other cities, like San Bernardino, would also move to Los Angeles.
• In San Bernardino, building would have also commenced. But before long, the combined effect of declining rents in LA and at home would staunch demand for new housing.
• LA would build much more housing than Up For Growth estimates; San Bernardino would build much less housing than JEC estimates.
• Some of the people who Up For Growth identified as needing houses in San Bernardino will satisfy that need in LA (or play musical chairs in a way that accomplishes the same thing).
• Some of the housing demand that JEC observes in San Bernardino would evaporate in a world where rents are lower in Los Angeles.

This mental model also hints at an aspect of Full YIMBY that we don’t like to think about: stagnant cities would lose population. If construction were instantaneous, you might see 7 million new homes built, mostly in high-demand cities, and 4 million new households formed. The excess would be made up by 3 million new vacancies in declining cities.

(Declining cities’ defense mechanisms are that construction takes time and housing supply is extremely inelastic in decline. Thus, prices fall much more in Rochester than they rise in Raleigh when people move south.)

An academic paper that comes close to reconciling the two sides is Duranton and Puga’s. It features sci-fi levels of depopulation in small cities. Unfortunately, given what we’re learning about the modest gains from moving to a high-wage city, Duranton and Puga’s powerful agglomeration effect is probably too strong. A version of their model sans agglomeration might come closer to the truth.