The Effect of Income and Income Growth on Home Prices

In a recent post, I highlighted how regressive rent inflation has been. From 2015 to 2021, the growth of incomes after rent in the richest ZIP codes outpaced the growth of incomes in the poorest ZIP codes by about 50%. And almost all the difference was due to regressive rent inflation. That is a massive spike in economic inequality. It’s a direct, linear, systematic result of inadequate homebuilding.

Figure 1

The can’t be very many 6 year periods that could have been nearly as bad. This is hundred-year-flood level stuff. I assume that very few people even have any idea of how bad this data is. It won’t show up in national averages. About the only other way to see it would be to do some sort of longitudinal survey data where you follow individual families over a long period of time.

It would take decades of progressive income trends favoring lower incomes to reverse this.

The data that underlies the Erdmann Housing Tracker tells a similar story with prices rather than rents. The price appreciation related to supply constraints is highly regressive. In cities that are more expensive, it is the poorest neighborhoods that have appreciated the most. Again, in a quite systematic and linear way.

I happened upon something interesting in my data from the 30 largest metropolitan areas. I was looking at how home prices appreciated from 2011 to 2021, depending on the starting income and income growth over that period.

The basic analysis is in the left panel of Figure 2. The x-axis is the starting income in the ZIP code. The y-axis is the average home price appreciation from 2011 to 2021. And each line represents a different income growth over those 10 years. (edit: The dashed lines show the price appreciation that would match income growth.)

On these regressions, the r-squared is about 35% and the standard error is about 22%. All changes are expressed as continuously compounded percentages (or natural logs. In other words, think of a 50% gain as 50 individual 1% gains, compounded upon the others).

So, in a ZIP code that started with $25,000 average incomes (about as low as there is in the data), homes appreciated by just under 100% in ZIP codes where incomes only increased by 10% to a bit more than $27,500 by 2021 (the left end of the orange line). In a ZIP code that started with $25,000 incomes and increased 70% to a bit more than $50,000, home prices appreciated by over 120%.

In a ZIP code that started with $125,000 average incomes, homes appreciated by about 20% in slow-growing ZIP codes and about 50% in fast-growing ZIP codes. On average, in rich ZIP codes, prices basically appreciated in proportion to incomes.

Figure 2

But, then, I added an interaction variable between starting income and income growth. That’s in the right panel. That didn’t change the rich ZIP codes a lot. It made the growth rates match up a little more closely with home price appreciation. So, in rich ZIP codes where incomes only increased by 10%, home prices also only appreciated by about 10%. In rich ZIP codes where incomes increased by 70%, home prices appreciated by about 60%.

In that regression, in rich ZIP codes, prices appreciated in proportion with incomes, more evenly across all growth rates.

But, in poor ZIP codes, price appreciation was much less sensitive to income growth. Whether incomes increased by 10% or 70%, home prices appreciated by 100% to 120%.

Implications

As in all of these analyses I do, the story of the last 20 years is that if you’re poor, your cost of living is mostly at the mercy of the regressive macro patterns that have been created by the housing depression.

This is only basic analysis. Any conclusions here are speculative. There are a lot of moving parts. A ZIP code’s income could rise because of a good economy for its residents. It could rise because of intra-city migration. It could rise because of what locals call gentrification.

In a regression between starting income and income growth, gross income growth is somewhat stronger among the ZIP codes that began with higher incomes. There aren’t a lot of ZIP codes that started with very low incomes and then had income growth far above average.

I suspect that there are a lot of currents happening within a city that don’t show up well in linear analysis. I suspect that typically, there are sets of middle to lower-middle income ZIP codes that are taking on local newcomers with higher incomes. If those shifts aren’t accommodated with new homes, they trigger migration out of those ZIP codes that either flows out of the city altogether or consolidates into poorer neighborhoods.

In any case, I didn’t see any odd sub-trends in this data. But, the broader point, which I think is interesting, is that the top end of the market has basically behaved like a market with elastic supply and/or demand. Home prices track with incomes.

As one moves down through ZIP codes with lower incomes, both supply and demand become more inelastic in a supply constrained environment. Increasingly as you go down (moving to the left in Figure 2), home price trends don’t reflect the conditions of the neighborhood. They reflect the conditions of the rest of the city, and how much those conditions bear down on poorer neighborhoods. Increasingly, looking into the lower income ZIP codes, the price appreciation is a given, and income trends are likely a product of migration changes and compositional shifts in the remaining ZIP code population.

These trends suggest that gentrification is overly feared. It is a visible and bothersome development for someone who has seen housing costs double or triple. But it is that doubling or tripling which is the real cause of distress. There are debates about the effects of gentrification, but I haven’t seen data from any angle that attributes an effect on local housing costs anywhere close to the background cost those neighborhoods have been burdened with in any case.

Figure 3

Anyway, by the looks of Figure 3, it appears that rising incomes of their neighbors is a far cry from the most pressing problem the poorest ZIP codes are generally facing.