Wall Street posts three-week losing streak as Iran war batters sentiment
First, I aligned the real Case-Shiller home price index with a measure of aggregate US residential investment / personal income. They both demonstrate the infamous trend of moving in a flat line for decades, punctuated by a sudden spike far outside the long-term range. Their close parallel from 1945 to the 2008 crisis suggests that the aggregate real estate wealth measure also had been flat since at least the 19th century.
Then, I noted that the spike to total value well above the long-term trend didn’t coincide with a building boom. Quite the opposite. It coincided with a marked decline in residential investment. This shows up in the BEAs estimate of real consumption of housing, which lags the trend in total expenditures now by more than 20% relative to 1991. And, at the same time, rent inflation has accumulated to nearly 40% above the change in the general price level.
It also shows up in net residential investment as a portion of GDP (starting with the BEA’s total estimate of residential investment and subtracting brokers’ commissions and the BEA’s estimate of the annual depreciation of the existing stock of homes). This was actually negative for several years after the Great Recession.
Another check is real housing expenditures per capita. It follows a similar pattern. Before 1980, Americans improved the state of our housing at a rate that roughly matched rising real incomes. Then, from 1980 to 2007 growth in the quality and size of our housing slowed down. Then, in this measure also, it actually backtracked a bit after 2008 and remained flat for a decade.
Since 1980, the real value of the US housing stock has declined by 29% compared to aggregate income but the market value of those homes has increased by 37% more than aggregate income. They are worth more because inflationary rents on existing homes has increased where there has been no improvement in the homes themselves.
In Figure 6, I indexed real aggregate home values to the total value in 1980 and tracked real housing expenditures from there. I regressed the remaining value against rent inflation and labeled the residuals as cyclical changes.
There is a potential issue here in that both cumulative rent inflation and interest rates have followed somewhat linear non-stationary trends over most of those 43 years. The recent spike in mortgage rates has helped, and will continue to help, confirm that the rent inflation has been the key issue there, not mortgage rates. The big cyclical drop in Figure 5 after 2007 is due to reduced mortgage access. Since 2007, the lack of mortgage access has interacted with rents because the price drop led to a collapse in construction which led to an acceleration in rents.
From 1980 to 2023, a 1% increase in rent inflation was associated with a 1.68% increase in aggregate price/income. As I write about often here, wherever rents are higher - over time, across metropolitan areas, or within metropolitan areas, 1% rent increases are always associated with more than a 1% price increase. There are several reasonable explanations for that. I think the most reasonable one, and probably most important one, is that land trades at a higher price/rent multiple than structures do, and rent inflation is unrelated to changes in structures.
The trends in basic national aggregates tell the basic story that the analysis of prices between and within metro areas will address in more detail. Under current conditions, the less we invest in homes, the more families pay for housing. I make a couple points here:
If increased demand were responsible for any of the excess rent inflation, it would imply both higher demand elasticity and lower supply elasticity. In other words, the more one assigns rising prices to rising demand, the worse supply constraints must also have been to keep relative production of new housing declining in the face of it.
In other words, you can’t be a supply skeptic in 2025 America. Our baseline condition has been a deep decline in real investment and real value of residential real estate. If part of the elevated value of American residential real estate is due to demand stimulus (low interest rates, federal subsidies, tax benefits, etc.), then that position has to be paired with a position that supply conditions are even worse than they seem.
And:
Most asserted sources of increased demand have been buyer subsidies (mortgage tax deduction and capital gains exemptions for homeowners, federal mortgage conduits, and others). Homebuyer subsidies increase demand for homeownership. Homeowners are both suppliers and demanders of housing, so homeowner subsidies can increase both supply and demand. From 1940 to 1970, New Deal homeownership programs took hold. Homeownership rose by 20 percentage points—an unrepeatable scale. Real housing consumption increased and rent inflation was low during that period. We have clear historical evidence of the effect of federal programs meant to stimulate homeownership. They were not associated with inflated values when they were significantly boosting homeownership.
Mortgage rates were low for most of that period, too.