Americans giving up on homeownership

Tyler Cowen cited a new paper titled, “‘Giving Up’: The Impact of Decreasing Housing Affordability on Consumption, Work Effort, and Investment”. Add this to the “Housing Theory of Everything” discourse. Tyler calls it “one of the best, most interesting, and most important papers I have seen of late”.

I appreciate this paper. From my own work, coming at this from a reduced form, 30,000 foot view, I would estimate that the mortgage crackdown in 2008 permanently removed about 15 million households, give or take a few million, from the pool of potential homeowners. The authors, Seung Hyeong Lee and Younggeun Yoo, estimate that “the cohort born in the 1990s will reach retirement with a homeownership rate roughly 9.6 percentage points lower than that of their parents’ generation.”

With a bit more than 130 million households, that would eventually settle out at about 13 million fewer homeowners.

I agree with the need to focus on this problem. I think this paper is an interesting discussion of the problems that are downstream of unaffordable housing and declining homeownership. They note, “as households’ perceived probability of attaining homeownership falls, they systematically shift their behavior: they consume more relative to their wealth, reduce work effort, and take on riskier investments. We show empirically that renters with relatively low wealth exhibit the same patterns. These responses compound over the life cycle, producing substantially greater wealth dispersion between those who retain hope of homeownership and those who give up.”

This is important. Yet, this paper has the recurring problem. They are entirely focused on cost of ownership. They appear to be completely ignorant of the role of the 2008 mortgage crackdown on homeownership.

Yes, this is going to be that rant again. My critiques here aren’t critiques of these authors. They are critiques of the entire academy. This paper, which is pointedly about the sharp decline in homeownership since the 1990s, was discussed by sixteen individual peers listed under the abstract, and discussed at seminars.

At the end of the list of discussants and advisors, they note “All errors are our own.” No. The error is the academy’s. How in the world, in the year 2025, can an entire cadre of economists review and comment on a paper about a steep decline in homeownership without, even accidentally, even as a side thought, (“Huh, isn’t it weird that homeownership declined sharply at a time when new homes were at historically easy levels of affordability, given interest rates? Should you look into that?”) anyone suggesting to look at mortgage access rather than affordability.

The sharp disappearance of mortgage access to millions of households is obvious if you look at any mortgage access index, such as the Urban Institute’s Housing Credit Availability Index or the Mortgage Bankers Association Index, or at the New York Fed’s Household Debt and Credit report. It’s not subtle.

In Figure 1, see if you can spot it. Look closely!

Figure 1

How do we estimate the scale of this change? Here’s what I would suggest. Remember that crazy lending bubble from 2003 to 2007? The one that required the global economy to crash? The speculators and NINJA loan originators that defined the financial course of generations? Well, the Urban Institute’s index of default risk had ranged between about 12 and 14 from 1999 to 2003. It peaked at around 17 in 2006. Today it’s at 5. So, in 2006, it was 3 points out of the previous range. Today it is 7 points out of the previous range.

So, start with your estimate of how crazy the mortgage market was in 2006 and multiply that by -7/3.

There are a lot of topics to discuss. Not every paper has to be about everything. But I think this demands to be mentioned in a paper about the collapsing homeownership rate and its implications. Not least of which because reduced credit access lowers home prices while lowering homeownership. That is a confounding issue that needs to be addressed in a paper purporting to estimate the effect of housing affordability on homeownership. I estimate that the average home has lost about 24% of its value, relative to 2002 prices because of tightened mortgage access.

They write, “We begin by documenting empirical patterns that reflect the sharp decline in housing affordability. The rapid rise in housing prices, alongside stagnant wage growth, has significantly eroded households’ ability to purchase homes.”

Figure 2 compares the homeownership rate and mortgage affordability since 2000.

Figure 2

Is affordability today probably pulling down homeownership a bit on the margin? Maybe. Should that be the sole focus of a paper about collapsing homeownership? For the love of Pete, no! The academy has failed.

Using Zillow data that is available since 2015, Figure 3 compares trends in the homeownership rate over time to trends in the mortgage/rent expense for the median home over time. The decline in homeownership of about 1% since 2022 is plausibly associated, in part, with higher mortgage costs. (Note, in Figure 1, even this period has been associated with further tightening of credit access.) But, homeownership had dropped by more than 5% from 2006 to 2016 when the mortgage on the median home declined from 32% to 19% of the median income and when the typical home across the country could be purchased, fully leveraged, for less than it cost to rent.

Apparently, not a single peer or mentor thought to wonder whether affordability wasn’t the primary issue here.

Figure 3

So, actually, I probably shouldn’t lean on taking their 9% drop in homeownership as any confirmation of my work. If it is modeled as purely an affordability issue, then the cow is probably too spherical for me to use to claim vindication.

They note that “House prices continue to rise rapidly while wages stagnate”. Yet nowhere in the paper do they associated rising home prices with rising rents. I wonder if some version of their model could be used to estimate the effect of rent affordability on renter household formation. Among the youngest households, there has been a sharp decline in headship rates (the number of heads of households, as opposed to roommates, etc.). More of the drop has been from fewer renters than from fewer homeowners. And, headship rates are down among all age groups. Among 25-34 year-olds, the decline in homeowners has flowed more, on net, to fewer heads of households than to more renters.

In the coming building boom (should we legally allow it to proceed), the marginal growth in units will be rental housing because for the past decade, home buyers have been outbidding investors for units in our capacity constrained housing stock, and so the 15 million or so households that haven’t formed since 2008 because of the housing shortage are mostly renters.

The reason rising prices have been associated with rising homeownership is that household formation is being limited now by the limited completion of units. We are capacity constrained on the long, hard path from a decimated construction industry in 2012 to an industry that can sustainably meet demand for new homes. Since 2016, the supply of new homes has barely been able to meet the demand for owner-occupied homes, so the growth of renter households has been very slow. Effectively, housing markets have been clearing against a vertical short-term supply curve. Rents have had to rise (at unprecedented rates relative to general inflation) until the necessary number of renter households don’t form. New completed homes have just recently risen to a level that might be close to sustainably meeting the need for marginal demand for new household formation. And there are 15 million households waiting to form because of the low construction of previous years - mostly renters.

Again, while I appreciate this work, the homeowner affordability framing doesn’t sit comfortably with recent trends.

Figure 4

5 years from now, completions are going to be above 2 million units annually (if new rental housing is still legal), the homeownership rate will be lower than it is today (because the additional building will facilitate renter household formation), and rents and prices will be moderating.

Imagine the downstream social issues that come from losing hope, not just about being a homeowner, but about being a head of household. Expanding mortgage access can, fortuitously, solve both those problems. But, trying to understand that through a lens of owner affordability is going to be deeply confusing. For instance, rising housing supply funded by investors will lower rents and home prices. But, it will statistically appear to mechanically reduce homeownership. It will, in practice, reduce the homeownership rate because it will allow renter households to form. That will improve homeowner affordability.

I don’t think the approach of this paper will be helpful in understanding what is happening. The part of the paper that discusses the issues downstream of declining homeownership will still be important. But, the academy will be at a loss to explain how to fix it.

Should we ban investors? Subsidize homebuyers? Or allow households and bankers to choose mutually beneficial arrangements that turn renters into owners?

The building boom will be associated with declining homeownership and moderating prices. The downstream issues these authors identify for families locked out of housing are real, but since they are being mentored in an academy that is apparently incapable of noticing the most important issue, they and their peers will be confused by that development. They might even be opposed to it.

The authors conceive of a subsidy not aimed at the poorest families, but aimed at families that are on the margin of possible homeownership who are in danger of losing hope. They estimate that this subsidy would raise the homeownership rate 3.4%.

This reminds me of the Conor Dougherty piece in the New York Times about Kalamazoo where he recommends giving rent subsidies to families in Michigan with incomes of more than $100,000 because rents are too high.

We can now say, after this 17 year experiment in credit austerity, that the main effect of generous mortgage access was to increase the supply of homes, which lowered the rental yield on rental homes. The most harmed families from the mortgage crackdown were renters in previously affordable homes where rents have frequently risen by 40% or more as a result of the mortgage crackdown and its effect on supply.

I suppose subsidizing marginal homeowners, as a side effect, could improve supply again and lower rents on the families that wouldn’t receive the subsidy. But, this seems like an oddly regressive approach. The subsidy would probably have some marginal benefits, but if we simply returned to 20th century lending standards at the federal agencies, many more households would be willing and able to buy homes without any subsidy.

Green Shoots

The paper has over 100 citations. They did not skimp on the literature review. Did none of those 100 citations point them toward considering mortgage access?

I did find one cited paper that treated the tightening of lending standards as an economic factor leading to measurable changes in housing markets, including lower homeownership rates. But, rather than suggest loosening access, it also suggests targeted subsidies.

Finally, a citation in that paper states the issue clearly: “This paper identifies the impact of borrowing constraints on homeownership in the U.S. in the aftermath of the 2008 financial crisis. The existence of credit rationing in the U.S. mortgage market means that some households for whom it would be optimal to choose to be homeowners may not be able to do so. Borrowers with certain wealth, income and credit characteristics are unable to obtain a loan even if they are willing to pay a higher cost of credit… Credit supply eased and then became considerably more restricted in the wake of the Great Recession… The homeownership rate in 2010-2013 is predicted to be 5.2 percentage points lower than it would be if the constraints were at the 2004-2007 level and 2.3 percentage points lower than if the constraints were set at the 2001 level.”

And, they cite a paper from the Urban Institute, which, finally, several layers into the cited literature, lays out the evidence of tighter lending and expresses a normative statement about it. “This research suggests that policymakers should continue and strengthen efforts to improve access to credit; the problems are particularly acute for borrowers with less-than-pristine credit scores. Resolving the uncertainly surrounding agency repurchases, the high costs of servicing delinquent loans, and other factors contributing to extraordinarily elevated lending standards will be critical for cultivating and sustaining a more robust recovery in the housing and mortgage markets, and for improving equity in access to mortgage credit for traditionally underserved communities.”

The Urban Institute does great work on this. I probably don’t do a great job of detailing the nuts and bolts of the problem. Here is one concise explanation from their paper of one aspect of the problem. “One such factor is lender overlays due to repurchase risk. About 80 percent of the loans made in the past few years have been bundled into securities guaranteed either explicitly or implicitly by the US government. For these loans, the government retains the right to put the credit risk back on a lender if the agency finds a mistake in the underwriting of the loan. Because a great deal of uncertainty has existed over how government agencies enforce this right, lenders have reduced their risk with their own credit overlays, lending only to borrowers with far better credit than is required by the agencies.”

Anyway, there are dusty corners of the literature where the right questions have been asked about the most important housing policy shift of my lifetime. But, they somehow fail to maintain the attention of the academy. Paper after paper, delving into these important topics where mortgage access is at the center of the issue, ignore it completely.

I should note that I am very pleased to have been a source for a recent article in the Economist, which takes the problem seriously. It was titled, “America’s huge mortgage market is slowly dying”. At least some non-profits and serious news media are shining a light on the issue. The academy has failed to lead on this. It’s long past time to at least follow.

PS. Here’s a second Economist article highlighting problems with mortgage regulation.