Random Thoughts on Mortgages, Household Finances, and Supply Conditions

Edit on my complaints about the first-time homebuyer paper

First, a post-script on the earlier post where I questioned the conclusions of a paper that claimed lower mortgage rates lower hurt first-time homebuyers. They don’t and the paper has many of the blind spots common in the economics academy.

In the post, I noted that while the share of first-time homebuyers declined during the period studied, the decline was caused by more activity from repeat buyers increasing the denominator, not a decline in first-time buyer activity. I cited a Federal Reserve source that showed increasing first-time buyer activity from 2011 to 2021.

But, looking more closely at the Fed source, the data clearly differs from the data the paper uses. The Fed source shows relatively flat first-time homebuyer share. And it also shows relatively flat all-cash share, which seems a bit questionable.

So, the very specific point about rising first-time homebuyer activity might not hold up, or at least probably comes from a source that differs. At worst, first-time homebuyer activity was flat, and likely still increased a bit over that time period. And, rising repeat-buyer activity surely is the primary change driving the first-time homebuyer share down in the paper. So, I don’t think it changes any of my complaints, but the difference was notable enough that I thought it was worth mentioning.

Real Rents and Rent Inflation

The blogger/tweeter known as “Jesse Livermore” has a really cool tool that allows you to compare different kinds of expenditures over time.

In Figure 1, I have used it to compare expenditures on housing (the total rental value of our homes) to expenditures on everything else. In the left panel, the gray line is “Core personal consumption expenditures excluding housing” and the blue line is expenditures on housing.

Basically, if we consume more, per capita, the line moves to the right. If we don’t consume more but we have to pay more for it, the line moves up.

In the right panel, the red line is the ratio. If we increase our consumption of housing relative to everything else, the line moves to the right. If we have to pay more for housing than we have to pay for everything else (if rent inflation is higher than other inflation), then it moves up.

Figure 1

There are 4 clear eras. From 1930 to the end of World War II, consumption of housing increased at about the same pace of growth in the consumption of other goods and services, but rents (at least as they were measured at the time) didn’t rise nearly as much as prices on other goods and services, so the line moved down vertically.

Then, from World War II to the 1970s, the New Deal mortgage programs were put in place. Homeownership skyrocketed from the mid-40s% to the mid-60s%. Interest rates were generally also low. We consumed more housing without causing any inflation.

Then, from the 1970s to 2007, zoning started creating regional rent inflation. Homeownership remained level, the growth of our consumption of housing slowed down, and we had to pay more for the housing we did consume. The line moved up and to the left.

Then, after 2007, we killed mortgage access for millions of families, homeownership dropped, and both the slowdown in real consumption of housing and the rise in rent inflation accelerated as a result.

I’d say the most common position among pundits and economists is to believe that mortgage access and low interest rates are inflationary. That couldn’t be more at odds with a century of experience. I went into some of these trends in my recent Mercatus paper about how elevated real estate values that are really a regressive transfer are being mismeasured as aggregate wealth, and in real income estimates, the regressive transfer is being vastly undermeasured.

Figure 2 is also from Livermore’s website. This compares the price level of housing and of everything else to wages over time. When the line is going up, as general consumption is (the gray line), that means that average wages buy you more over time. Outside of housing, wages buy twice as much as they did 45 years ago.

Housing (blue) barely has risen at all in recent decades. I’ve actually seen economists claim that this is just fine. But what this means is that you move out of your parents house into a tiny little apartment that you can afford on a 25 year-old’s salary, and after a lifetime of 5% annual wage increases, you can still only afford the tiny little apartment.

Figure 2

Keep in mind, the flat line for housing is the national average, but under a shortage, rent inflation is very regressive. So for high income earners, the line is going up and for low income earners, the line is going down. Millions of Americans are being priced out of the homes they lived in. They have to run to stay still or accept displacement. In cities that are shrinking, like Los Angeles, the line has to go down for many of the existing residents. A hundred thousand families have to feel poor enough if they stay to accept that they need to be the next family to leave. That’s how the market clears in LA - the blue line goes down until housing demand goes negative through voluntary displacement.

Household size

Those who believe that low mortgage rates and mortgage access have been to blame for high home prices also tend to be supply skeptics, and, amazingly, in the year of our Lord 2026, some continue to spout the old line that homes have been getting larger and families getting smaller, so there can’t be a housing crisis. People just expect too much. And they assure us that everything is fine because the average household has higher income than they did 10 or 20 or 30 years ago. (Of course average incomes don’t reflect the very regressive nature of rent inflation under a shortage.)

As Figure 2 implies, homes haven’t been getting larger for 25 years. (Data on the square footage of new homes also follows that pattern. New homes have been getting smaller.) The skeptics usually cite 1970 or 1960 as a comparison, because households did get smaller in the 1970s and homes did get larger in the 20th century.

Figure 3 shows household size and family household size. Families are a subset of households in which at least 2 related people live - a married couple with or without children or a single parent with children.

Families stopped getting smaller 35 years ago. That’s why they have to cite 60 year old data.

Figure 3

And, since 2007, the stable size of the average family is a combination of fewer children and more adults - both of which are consequences of the housing crisis.

Figure 4 shows adults per household. The rise in adults per family in the last 20 years has erased 40 years worth of the previous downward trend.

Figure 4

This isn’t some mysterious, secret data. The Census bureau publishes these numbers every year in a simple table. Have the supply skeptics that have been citing declining household sizes and increasing home sizes been involved in a discussion for 20 years without ever taking 5 minutes to look at a simple Census Bureau table?

There are enough mouthpieces for these assertions at some think tanks that it is a common reaction by laypeople to say that households are getting smaller and homes are getting larger, so there can’t be a shortage. They don’t know that it’s a lie. I don’t think those pundits are consciously lying. I think they have a Herculean loyalty to ignorance about the supposed topic of their expertise. But at this point it has the same practical effect.

Is Supply Bringing Down Rents?

I’ve been seeing a lot of triumphalism about the effect of supply on recent rent trends. I have pushed back a bit on that because rents on class C apartments are down double digits in some markets, and current rates of construction just aren’t high enough to lower rents at that rate.

I think a lot of those recent declines in rents are associated with the reversal of some Covid-era migration that happens to be correlated with the growth rate of different regions. People moved to fast-growing areas, and now the pendulum is swinging a little bit past neutral on that growth.

But, the point still stands. The entire country is in a state of shortage. When you are in that state, anything that causes demand to be lower than supply - whether it is new construction or a temporary decline in population growth - is going to create highly progressive rent deflation or disinflation. Under shortage conditions, high-tier consumers trade away amenities in order to lower their costs and low-tier consumers are stuck in the worst housing they can stand and they pay rising rents to hold on to it until the situation becomes catastrophic for them and they have to lower their living standards or accept displacement.

So, when supply outpaces demand, there is pent up demand to reattain the amenities in high-tier homes and households moving up into those new units reduces the pressure on low-tier rents. That causes Class C apartment rents to decline much more than Class A rents do. This is why “we don’t need luxury homes, we need affordable homes” is such a toxic position to hold. Absolutely, the most effective way to lower Class C rents is to build a lot of Class A apartments. Everyone wins. The developers can keep building their luxury apartments without much of a hit to their rental value (because their rents were never as elevated as Class C apartments were). And the families hurt the most by the housing shortage get the most relief.

Call it filtering. Call it sales chains. Call it whatever you want, but don’t whisper it or apologize for it. Yell it from the mountaintops, or into a megaphone pointed at the face of those who can’t understand it.

Anyway, as a way to try to get at the cyclically neutral effect of new supply, I looked at the 100 largest metro areas, and I regressed the real rent inflation from 2020 to the present against their permitting rate per capita of multi-family homes from 2002 to 2006 and from 2019 to 2023.

Here’s my thinking. Cities were generally equally crappy about permitting multi-family homes before 2008. Then, in 2008, federal mortgage regulators created a uniform shock to housing supply that cratered single-family construction. Regions that have managed to increase the pace of multi-family construction after 2008 should have filled the gap left by the loss of single-family homes better.

The result is pretty noisy, but the effect size is surprisingly in line with the supply story. Each 1% increase in a region’s housing stock, through multi-family construction permitted from 2019 to 2023 is associated with a 2% decline in rents. That is right at typical estimates of housing demand elasticity and it is the number I use in my Metro Area Analysis packages.

Very roughly speaking, that means little decline in Class A rents but more like a 4% decline in Class C rents. So, a city experiencing 10% Class C rent deflation might be approving new homes at a rate of 2% of the housing stock annually while population trends may have temporarily turned slightly negative.

The reason I think we should be careful about these assertions is because population growth in cities like that is likely to return to more than 2%. Rent inflation will probably return, even with construction rates at current levels. And it will be chum for the skeptics who downplay the role of supply. But, I think we can say with confidence that after accounting for these temporary cyclical issues: Yes. New luxury, high amenity housing is probably the most progressive economic force in our economy right now. And, ignoramuses are going to fight it from 7 different angles, because it’s not intuitive to a layperson.

Mortgage Rates and Mortgage Access

Finally, data scientist Ryan Canavan shared Figure 5 with me recently. At first, it seemed like good evidence that the rise in mortgage rates in 2022 had caused low tier home purchase demand to decline sharply.

But, the more I sit on it, the more I think this reflects tightening in lending standards over this same period of time. I’m not sure it is possible to sharply identify the causality here - whether standards were tightened coincidentally when interest rates happened to rise, making it look like rising rates are the cause of less borrowing by lower-earning borrowers. If mortgage rates can manage to decline a bit, future trends in borrowing and in home prices might help clear it up.

Figure 5

I may get a chance to dig in to HMDA data shortly, and maybe I will think of a way to get more clarity here.