The Housing Shortage by Metro Area

Here is an estimate of the number of homes each major metropolitan area is short.

For now, this is a very broad first pass at the answer.

To get a little more serious about this, I need to add things like controls for density and property taxes. But, I realized I could pretty easily take a broad stab at the number with the Erdmann Housing Tracker data.

I tend to overthink this question. For instance, I think the elevated rents and home prices associated with housing shortages are mostly a product of the frictions associated with displacement. Families are willing to pay more to avoid being displaced. But, in the very long term, those frictions are temporary. Eventually, after generations, families will make decisions about migration and location that reduce the very specific signal of inadequate supply that the tracker picks up on.

So, these numbers (after applying important control variables) probably understate the quantity of homes that should have already been built in the Closed Access cities, because millions of families have already moved away, lost the endowment value of having ties in those cities, and started building endowment value in their new cities. Their displacement helped permanently lower demand for housing in the cities they were displaced from.

The significant price appreciation negatively correlated with incomes that has characterized American housing for 30 years is mostly due to the problem being so bad that rents are rising faster than incomes in some neighborhoods. Even if supply remains restrained, if it loosens up enough to be merely bad instead of catastrophic, families will slowly make adjustments on other margins that will lower costs.

In other words, the supply component of the tracker isn’t a supply and demand equilibrium. It is a very slowly reverting disequilibrium.

Anyway, in much more simple terms, just think of these numbers as the number of homes that would need to be constructed within a decade or two in order to reverse the excess price appreciation associated with inadequate housing.

I simply have assumed here that price appreciation associated with inadequate supply has a linear correlation with the supply component of the tracker (which I think is a reasonable assumption). I have chosen a moderate target slope for the relative price/income level across incomes in each metro area that would be associated with reasonable supply.

I assumed, for the sake of this exercise, that the number of units short for the US, based on other analysis I have posted about, is 10% of the housing stock, or 14.7 million units, and I calibrated the number of units short for each metro area to that, based on the supply component of the track for each metro area.

In Tampa, the price/income in a ZIP code with $75,000 income tends to be about 1.4 points higher than in a ZIP code with $150,000 income. I associate that with a 10% unit shortage. In Dallas, it’s 1.5 points higher, so I associate that with a 7.6% unit shortage.

As I said, this is a very broad first brush. I hope to someday try to estimate a little more carefully. But, I thought this might be of interest to readers. (Here is a follow-up post with more numbers and slight revisions.)

The first column is the number of extra homes that will be constructed in each city as long as it is legal to fund and build them. To prevent these quantities of homes from being constructed, they will have to be politically obstructed in some way. Of course, as of today, they are, either because of land use regulations or mortgage regulations.

The second column is that number as a percentage of the metro area’s current stock of homes. For the worst contenders, this seems about right. If they had allowed a bit less than 1% additional annual population growth over the past 30 years, they might well be affordable.

The third column is the average number of new housing permits each metro area has issued in the past 5 years, as a percentage of their stock of homes.

The fourth column is the number of years, at the current building rate, it would take to make up the shortfall. In other words, if construction doubled tomorrow, how many years would it take to achieve broad affordability.

Of course, Austin comes out looking the best. Could there have been any doubt?

And, the cities I refer to as the Closed Access cities all show up dependably at the bottom of the list, with Miami on the margin. The number of years of production needed, based on current rates, in closed access cities is basically at least twice as high as any other city.

Nationally, about 1.6 million new homes are being constructed annually. Three million homes a year for a decade might normalize rents. Again, this is not building that needs subsidized or stimulated. This is building that will naturally happen where it is legal. The homebuilders will only be demand constrained where demand is legally blocked.

Current construction activity is basically a floor, and the rate that the 15 million extra homes get built above that floor depends on mortgage access, YIMBY wins, legal acceptance of build-to-rent, etc.

If we did build those 15 million units, it would take roughly $8 trillion of investment. Let’s assume that all those obstructions are eliminated, and even the closed access cities have building booms.

Today, residential real estate has a total value of about $55 trillion. About half of that is excess land value because of inadequate supply. So, $8 trillion of investment would be associated with a loss of about $24 trillion in land value, and total residential real estate would decline to only about $39 trillion. That would put the real estate/income ratio back at the 20th century norm.

And we would be wealthier for it.

Believe it or not, those big checks that you don’t mail to yourself each month in order to rent your house from yourself haven’t actually made you richer, even if they are twice what they were a decade ago. But the dude living in the tent down at the park is poorer for it.

Housing wealth is wealth. Extra housing wealth is poverty.

This would not be disruptive. At realistic rates of new construction, the average prices of existing homes will generally flatten out in nominal terms and slowly decline in real terms. Similar to home price trends in late 2022 and early 2023. That’s at more than 3 million new units annually. Would that we could have such problems.

That might seem like a high price sensitivity (6% price effect from 1% more supply). If you think it is too high, then you need to adjust your estimate of the shortage so that it is higher than 15 million homes.

*Detroit, Philadelphia, Pittsburgh, and St. Louis aren’t included because regressive property tax rates cause some nonlinearities in the tracker data when I don’t control for them. Also, they have relatively low demand and generally have smaller estimated shortages, or none at all.