Iran conflict latest: Trump pauses Iran energy plant strikes by 10 days
The St. Louis Fed published a blog post last month that is a depressing reminder of how out to lunch economists and federal economics staffers are on housing.
It’s titled When Houses Outrun Paychecks: The Lost Decades of Housing Affordability and it is a discussion about how unaffordable housing is creating a host of social and economic problems. “In most places, underlying home values have risen far faster than local incomes, reinforcing the affordability squeeze documented above.”
Under the subheading “Factors Driving Up House Prices” they note:
Demand shifted outward against a relatively inelastic supply curve. Zoning rules, land-use regulations, and capacity constraints in construction limited new supply, especially in high-productivity metro areas, so shocks to demand translated primarily into higher prices rather than more units.
There were several factors that shifted housing demand outward. From 2000 to 2021, a sustained decline in mortgage rates lowered the user cost of housing and relaxed borrowing constraints, effectively raising households’ housing “budget” even when wages grew slowly. Credit expansions in the 2000s further amplified this effect, and in the post-2008 environment, institutional investors and global capital increasingly treated U.S. housing as a quasi-safe asset, adding an investment-demand component on top of local fundamentals.
At the same time, amenity and productivity premiums were capitalized into land values: Counties with strong job growth, high-quality schools and attractive amenities saw disproportionate appreciation. Finally, the widening gap in household incomes meant that much of the aggregate income growth accrued to higher-income households, who could bid up prices for a fixed or slowly growing housing stock, thereby pushing equilibrium prices away from what median earners can afford.
Everything there is true. In fact, they are batting above average to note the importance of capacity constraints in construction.
But, of course, the elephant in the room is the mortgage crackdown in 2008. A big, honking negative shock to buyer demand, and, thus, new home sales. And, so they are left to try to make all the pieces make sense without knowing the critical motivating factor at their center.
Prices in 2023 (and today) are high because of low mortgage rates that ended 4 years ago? A credit expansion that ended 20 years ago? Institutional investors that started buying homes at 50% discounts?
So, they end up in that last paragraph making the classic error that I described in my recent Mercatus paper. It is true that metropolitan areas with higher demand have seen disproportionately high price appreciation. But within those markets - within every market - it is the cheapest homes in the poorest neighborhoods with the worst amenities that have appreciated the most. County level data may not be detailed enough to see that.
In a way, you could argue that price appreciation is being pushed by families with higher incomes. Price appreciation, in the aggregate, has been associated with an increase in owner-occupier families that has been outpacing the capacity to build. But, it isn’t helpful to pin that on a widening gap in incomes or on zoning.
Yet, some of the effects of ignoring the mortgage crackdown are subtle. You could say zoning explains everything, because when family purchase demand cratered in 2008, developers were unable to fill the gap with new multi-unit housing. But, zoning didn’t change in 2008. Mortgage access did.
They conclude, “This is not just a story about a few “superstar” coastal cities. Tight land-use rules, inelastic supply, falling mortgage rates, credit cycles, amenity premiums and widening income differences have all interacted to push prices up relative to earnings across a broad swath of the country.”
But, by “credit cycles” they mean every change in credit access in the past 30 years except for the elephant in the room - the large overcorrection to very tight standards.
The authors cite their own interesting paper where they analyze the effect of investors on home prices. They find that small investors, not large, out-of-state investors, have driven up home prices since 2009. But, since they don’t think they need to adjust for the mortgage crackdown, their findings are all funhouse mirror findings.
Somebody has to own every home, so their variable of interest - the rise of small and medium investors - is, itself, a proxy for the mortgage crackdown.
They find that where investor activity increased, prices increased from 2009 to 2017, and they especially increased in low-tier neighborhoods.
Mostly, prices increased in proportion to the scale at which they had decreased before the investors put a bottom in the market. Investors are a proxy for the mortgage crackdown.
Among the controls they used in their regressions, they mention, “whether the location is sensitive to large house price movements measured by the average real housing price growth during the 2000–2006 boom and the 2006–2007 bust.” If you don’t know that the mortgage crackdown was the most important factor in housing markets after 2007, that probably seems like a reasonable control.
Figure 1 shows the price/income path of 3 low-tier ZIP codes, and the time periods where the authors applied control variables and where they estimated the effect of investor activity.
 |
They also claim that more investor activity leads to more multi-family construction. They write, “The tightening of credit standards after the financial crisis coincided with a period of adjustment of the households’ balance sheets (i.e., see Garriga and Hedlund, 2020, for a detailed quantitative analysis using a model of household purchases and endogenous house prices). The declining demand of owner-occupied housing changed the type of newly constructed units.”
You would think that these means that they see it. But, I think, what is going on is that the tightening of credit access in 2008 is observed, a priori, as a return to normal conditions. So, even when they see it, they can’t see it.
There was more of an increase in multi-family construction where there were more investors because both of those trends were caused by the permanent denial of mortgage access to millions of families who would have purchased single-family homes.
They also attempt to control for credit conditions. “The first column controls for the mortgage denial rate over 2009–2017 in each MSA.” Again, they are looking everywhere except where the elephant is. Denial rates were low by 2009-2017 because low-tier borrowers had learned not to apply by then.
Across the country, home prices were too low, especially in neighborhoods with lower incomes. Builders couldn’t profitably build. So, supply was inelastic in those markets until prices increased back to the equilibrium. I would bet dollars to donuts if they replaced their investor variables with the denial rate on mortgages in 2006-2008, they would find stronger results across the board.
The authors write, “We conduct omitted variable bias tests based on the work of Altonji et al. (2005) and Oster (2019), which we outline in Supporting Information Appendix C… The results strongly reject that the effect of the share of investors on housing prices is driven by omitted variables. Thus, these tests alleviate concerns of omitted variable bias.”
I don’t know enough statistics to reverse engineer these robustness tests, but I blush a bit every time I see these notes. Are any robustness tests in economics useful at all?
