Housing needs supply-siders

Unfortunately, American supply-side economists and pundits who would normally be advocates for the regulatory reforms we need in housing have been neutered by a series of spurious conclusions, reaching back decades.

There has been a spurious correlation between declining interest rates and rising home prices for 40 years. That led many supply-siders to falsely believe that a secular decline in mortgage rates was due to cyclical interest rate cuts by the Fed and to vastly overestimate the effect of lower mortgage rates on home prices. The variance in home prices across time and business cycles, across regions, and across income levels within regions, driven by similar variances on all those margins in rents, has vastly more explanatory power than interest rates do in home price outcomes.

Rising rents leading to rising home prices with wide variance across time and space is the bat signal for a supply-side problem. And, while supply-side economists and pundits don’t deny this problem, they, unfortunately, were so taken by their demand-side complaints about monetary policy and federal mortgage programs, that they became handmaidens for the most devastating supply-side crisis of my lifetime.

The mortgage crackdown in 2008 created the financial crisis and put a cliff in funding for new housing that was so severe, it led to a loss of something like 15 million homes.

And, all the supposed supply-siders had to say about it was that the recession was cleansing, the foreclosures were inevitable, interest rates should be higher, growth should be slower, and mortgage access should be tighter.

Think of “Fragile by Design,” by Charles Calomiris and Stephen Haber, blaming the Community Reinvestment Act, bank deregulation, and the federal mortgage agencies for high prices. Or, “This Time is Different” by Reinhart and Rogoff, about financial bubbles. Or, “Rethinking Housing Bubbles” by Vernon Smith and Steven Gjerstad. They are all obsessed with a housing “bubble” - a market where high prices aren’t justified by the fundamentals of homes’ rental value.

Don’t get me wrong. The Erdmann Housing Tracker attributes about 25% price inflation before 2008 to cyclical and lending trends. That’s not nothing!

But, the problem is that the topic of rental value doesn’t appear in the index of any of those books. Inflated rents and household expenditures to deal with them were already the most important factor driving housing markets before 2008. Most of the price inflation that the Erdmann Housing Tracker associates with “cyclical” trends before 2008 was from hundreds of thousands of families moving away from California and the Northeast to Florida and Arizona where the rents wouldn’t bankrupt them. It was a bubble in displacement, with the odd side effect that if you were a homeowner in California and the Northeast, your displacement came with a capital windfall.

So, many of the keepers of the supply-side point of view are rudderless - floating in a windless sea of their own misdirection.

Here I will highlight two recent examples of supply-sider fails.

High costs

One recent example was a Washington Post article by Julia Cartwright at the American Institute for Economic Research. Her mistake isn’t as bad as all that, but it is an example of counting the number of times the basketball was passed while the gorilla marches through the scene.

She notes that new home construction has been weighted much more toward high end homes in recent years. She blames high labor costs and low construction productivity growth - the Baumol effect. “Something structural is preventing the market from responding to the demand, and that something, more than zoning or permitting or interest rates, is the cost of labor.”

She suggests welcoming more immigrant labor, repealing prevailing wage laws, and deregulating operational licensing.

All of these observations are true. But, they miss the point. Housing is expensive because urban land is expensive. So, her accurate observations are sifted through a misinterpretation. “This dynamic reshapes what gets built. When fixed costs rise, developers shift toward projects that can justify the higher costs. Lower-cost housing becomes increasingly difficult to finance. The result is that new affordable units are almost never built at all; instead, existing units become affordable as properties age.”

High costs don’t change what housing gets built. In the distant past, housing was more expensive, in terms of square feet per average hourly income, for instance. And, when that was the case, the market reaction was to build smaller homes that cost less than the homes we build today. The reason the market has only been building more expensive high-tier homes is because the land is more expensive, and so those are the only homes that pencil after the land costs. And, she has filtering backwards, too. Homes used to filter down. Now they are filtering up in most places. They are filtering up because urban land use regulations and capacity constraints stemming from the post-2008 mortgage crackdown have left us with a huge shortage, so land prices are inflating. The future tenant of a given house will be paying more for it than the current tenant, and they will have a higher income, because of the musical chairs game we now have in housing.

If costs were the problem, builders wouldn’t be willing or able to pay elevated prices for the land. And, when the construction of homes at lower prices collapsed, the prices of existing homes were at generational lows. Builders went from selling more than 800,000 new homes at prices below $300,000 in 2004 to selling fewer than 300,000 every year since 2008. For most of that time, existing homes were much less expensive than they had been in 2004. Rising costs didn’t collapse that market. Existing homes were too cheap where families had lost access to mortgage funding.

The land under existing homes in low-tier neighborhoods went negative after the 2008 mortgage crackdown, so builders couldn’t compete with homes on new lots. That condition remained in effect for so long that by the time rental values on those homes had risen enough to make new homes on new lots competitive, the capacity of the industry, nationwide, had been cut for so long that builders couldn’t open new communities fast enough to meet the pent up demand for household formation. So, prices of marginal residential lots went from negative up, right past zero, to inflated levels in cities with high demand. And, that’s where we are now.

So, she has a number of accurate observations, and a number of reasonable policy suggestions, but they are irrelevant to the current condition of the American housing market.

Our problem is that a mortgage crackdown temporarily cratered our capacity to build single-family homes, and urban land use regulations prevented multi-unit construction from filling the gap. These are 2 related factors where governments at all levels are preventing families and producers from making economic decisions. We could use some non-confused supply-siders.

A more thorough example of supply-siders gone mad is Norbert J. Michel and Jerome Famularo, who have been writing about housing at the CATO Institute. CATO actually does some good work on housing, but I think that among their scholars, pundits, and followers, there are a good number of supply-siders that are confused in the way I described above, and Michel & Famularo scratch their itch for them.

In one ironic post, titled “How Measurement Choices Shape the Housing Debate-and the Charts in the President’s Economic Report” they mangle a number of empirical points.

Home Size

They claim that homes are more expensive because they are larger and nicer than they used to be. First, this claim just breaks down with the slightest observational curiosity. Go to the southside of Boston, the Bronx, East Los Angeles. In neighborhoods with below median incomes, in which the structures haven’t changed in decades, you’ll find home prices that have doubled or tripled after adjusting for inflation over the past 30 years. That was never the case, historically. This is a novel and economically important development.

To make the case, statistically, they create a metric for “price-to-income at constant size”. In other words, if the median new home today was the same size as the median new home in 1987, what would its price/income ratio be? And, their metric says that it would be about the same today as it was in 1987.

So, think about what they are saying. Real median household income is up about 30% over that time. But, those 30% higher incomes can’t buy any larger homes than the 30% lower incomes could 40 years ago. This reflects 2 very different periods of time. The average size of new homes was rising roughly in line with incomes until 2006, and it has been shrinking for 20 years. Taking all types of housing into account, new homes today are about the same size as they were in the 1990s. The would-be supply-siders are pressing into new frontiers on how long you can keep referencing an outdated claim.

We’re only about 5 years into low interest rates being an outdated claim as a cause for high home prices. I suppose we’ll still be hearing about that in 2050 when housing costs are still elevated and mortgage rates are 8%.

This metric has other problems, too. Since the shortage inflates the value of land under older, aging neighborhoods, most of the inflated cost of housing applies to older, existing homes and aging apartments. Builders and buyers of new homes make adjustments to keep new homes affordable, such as building them in less expensive locations on smaller lots, in less expensive regions. So, for instance, the highly inflated housing costs in the Bronx, East Los Angeles, etc., that they failed to notice observationally also fail to influence the metric they are using because there aren’t very many new homes in New York City and Los Angeles. In fact, the reason New York City and Los Angeles are expensive is because they don’t permit enough homes, so Americans have to migrate by the millions out of those cities to other places that will permit more homes, and thus don’t have such inflated land prices.

I have walked through a comparison of existing vs. new homes before, to make this point. But, you can avoid all the compositional issues they are creating by only looking at new single-family homes by just comparing the Case-Shiller price index to median household income. It tracks the value of existing homes. It’s up 50% more than median household income since 1987.

Their metric is almost comically concocted in order to not measure the point of the article. How Measurement Choices Shape the Housing Debate. Truly. Truly.

Homeownership

Next, they tackle homeownership rates, and they claim that the current homeownership rate is near where it was in the 1990s - both for all households and for young households. I have written about this quite a bit, and I think this is an issue that is generally now understood by most housing pundits. The rise in homeownership since the bottom in 2015 is mostly because the housing shortage has blocked household formation of renters. The reported homeownership rate is inflated because the housing shortage has slowed down household formation and “household” is the denominator.

“How Measurement Choices Shape the Housing Debate.”

Housing Starts per capita

They also mention that the decline in fertility since 2008 may not be fully attributable to the decline in housing starts. That isn’t a ridiculous assertion.

Relatedly, with fewer children, families require more units per capita. Fewer marriages also affects this. I have written about this before. It is a legitimate factor, but it doesn’t have anything to do with the housing affordability problem. In the 20th century, families and individuals made adjustments on their real consumption of housing to keep nominal spending comfortable. There is nothing about cultural changes in household formation that would prevent us from continuing to do that. The reason housing has become nominally expensive is that new supply has been so throttled that the lowest-tiers of neighborhoods in expensive cities accumulate rent inflation faster than their residents can grow their incomes, and families are forced to make difficult decisions about what to give up or whether to accept being displaced from locations with personal endowment value. That is the only context where families systematically spend more on housing than is comfortable.

As noted above, the average new home currently is about 300 square feet smaller than the average new home in 2006 was. It is the homes that aren’t being built that are causing the crisis.

Also, as I have written about several times, the decline in children per household has been almost perfectly offset by the rise in adults per household. Children don’t form households. Adults do. Homes per capita is at all-time lows, but homes per adult had been rising steadily from World War II to 2006, and in the 2 decades since then, that trend has been totally reversed. There are about 14 million more adults living in family households than if we were still on the post-WW II trend.

The authors compare housing starts per capita to housing starts per additional resident. This is a reasonable attempt at normalizing the metric. If population growth is slower, then housing starts per capita will decline. However, I don’t think it successfully makes their point. The housing starts per additional resident in their chart for 17 of the 26 years from 1990 to 2015 was lower than any year from 1962 to 1989. And, given that the drop in population growth was related to post-2008 birthrates, which don’t have an immediate effect on household formation, as noted above, that is especially concerning. In more recent years, the high death rate of the pandemic and Trump’s immigration crackdowns have reduced population growth enough to push the housing starts per new resident metric back up to the historically normal range. In fact, I would agree that given current population trends, new construction is somewhere near a sustainable level. But, we need at least 15 million homes to meet pent up household formation, to correct elevated urban land prices, and to reverse the significant rent inflation that has accumulated during the shortage.

They note:

Zoning rules, permitting delays, and other restrictions make housing more costly. However, if rising housing costs partly reflect larger homes and higher incomes, then policies aimed at forcing down prices (such as mass deportations and bans on institutional investors) could have unintended consequences.

Sure enough, recent research from the San Francisco Fed suggests that faster income growth, not supply constraints, explains much of the differences in house price trajectories across metro areas.

The first paragraph here hits the right notes. They do support deregulation, and they correctly view the administration’s attack on investors in rental housing as toxic and unhelpful. But, their claim that high home costs are related to high incomes is just not supported by the facts. Rents in those neighborhoods they didn’t notice in East LA and the Bronx are rising faster than the residents’ incomes. And, to the extent that they are not, it’s because the main way those markets equilibrate is through mass displacement of the poorest families. And, that mass displacement is what raises the average local income. The methods used in that San Francisco Fed paper aren’t able to pick that up. The housing shortage is compositionally pushing up local incomes. They have the causality backwards. But they only look at metro area averages. They don’t even know about the systematic regressive rent inflation that the shortage creates within metro areas.

Across the country, since 2008, rent inflation and subsequent price inflation have been highly negatively correlated with incomes. Cross-sectionally, rising housing costs are definitely not being driven by high or rising incomes. It is the homes in the worst locations where residents with the lowest incomes live that have become the most inflated. The entire Erdmann Housing Tracker model is built on measuring that regressive price inflation. If housing costs weren’t deeply, systematically, negatively correlated with incomes, my model would have nothing to measure.

To the extent that these scholars are concerned about regulations, they are really stepping on their own feet here. They are citing a paper that asserts that local supply constraints aren’t important. And, of course, the way in which they are mostly in their own way, and in favor of over-regulation that hurts families and has been responsible for the post-2008 shortage of 15 million homes, is the topic that isn’t going to show up in their writing at all - over-regulated mortgage access.

Amenities

In a follow-up post, titled, “Don’t Be Fooled by Sticker Shock—There’s No National Housing Affordability Crisis”, they write, “Proximity to high-wage labor markets, desirable schools, and local amenities is itself a form of housing quality. Rising prices in these locations therefore do not necessarily imply a proportional decline in affordability, although local land-use laws certainly can worsen it.”

In this case, they have simply been failed by the economics academy, and it is understandable that they make this error. Interestingly, they mix up geographies here. Proximity to high-wage labor markets operates generally on a commuting zone scale, while schools and local amenities are generally more at the neighborhood level. Systematically, rising housing costs have burdened neighborhoods with the worst amenities, the worst schools, and the poorest neighbors, in cities with the lowest rates of home construction. Those markets only appear to be high-wage labor markets because poor residents move away. But, if you don’t know about the cantilevers driving American housing costs, it’s hard to notice that issue. And this is one way that scholars tell on themselves. They assume that rising home prices are positively correlated with rising incomes and amenities, and so they think it doesn’t matter if it is at the neighborhood level or the metro area level. I described how Joe Stiglitz told on himself in this way in my recent Mercatus paper. They think that a lot of rich newcomers are driving up the prices of high-tier homes, but what’s really happening is rising costs of low-tier homes are driving away poor existing residents. If you know what evidence to check on, the story is clear. But, they don’t know that they need to check on it, so they announce their mistaken assumption.

Mortgage costs

In this post, the authors focus on homeownership, and they only discuss affordability, because the most important regulatory burden imposed in generations is not even in the peripheral vision of our confused supply-siders. This is related to the homeownership claims above. To get back to normal, we would need at least 15 million new homes. Roughly counting, about 2 million would be owned, 8 million would be rented, and 5 million would be vacant. We lack renter households. The shortage has nothing to do with the cost of homeownership. In fact, homeowners are advantaged because the mortgage crackdown lowered the cost of ownership relative to renting. From 2009 to 2020, homeownership was historically affordable. And, for the past decade, since homeowners who can get mortgage funding are so advantaged in this market, every year, they have been buying all the new homes and some of the existing homes that used to be rentals. Renter household formation flatlined.

So, first, we created millions of renters by locking families out of mortgage access. (They couldn’t even get approved for mortgages before 2020 when the cost of ownership for many homes was a small fraction of their rental cost.) Then, the subsequent shortage of new construction caused rent inflation to skyrocket on those very families.

The cost of homeownership has jack-all to do with any of that. But, the problem of mortgage access, because of federal regulations that have cratered housing supply, is so far off the radar of our confused supply-siders that the political battleground for appropriate housing and housing finance regulation is handicapped. We can’t even have a conversation about the most important thing. Those who should be the loudest defenders of deregulation of the most important thing will only be engaged in motivated reasoning to avoid being useful. It’s not even motivated reasoning intended to misstate the problem. It’s motivated reasoning to be ignorant of the problem itself. Motivated reasoning to be ignorant of the empirical evidence that would point to the trends that would point to some questions that might lead to a glimpse of the most important thing.

The shortage of housing is about renters. We could solve it by allowing more rental units or allowing more homeowners.

It is especially galling to see CATO Institute scholars fumbling around with their hands over their eyes going, “Land rents? What land rents?”

We need to find new supply-siders.