NAHB’s Housing Opportunity Index: Another Metric That Punctures the Charlottesville Affordability Crisis Myth

Proponents of upzoning have justified the urgency of radical changes on the basis of an alleged crisis of housing affordability that besets Charlottesville in a particularly acute fashion. They have showed much more vigor in pushing this as narrative than in demonstrating it as fact. We have looked hard at available data to assess whether Charlottesville indeed has a particularly unaffordable housing market. Regardless of the approach and data sources we have used, we find Charlottesville to be either about average or better than average in terms of the absolute level of affordability and better than average in terms of the decadal trend. You can see our earlier analysis using Census data in our first-ever presentation from over a year ago here; you can see our work on the affordability trend here; you can see fresher wage-based data here.

The National Association of Homebuilders publishes a data series called the Housing Opportunity Index. The HOI measures the percentages of homes sold in a given geography that would be affordable to the “median household” in that geography. We wondered how Charlottesville would look on that score. Unfortunately, NAHB does not regularly publish a result for the Charlottesville MSA because it is such a small one. NAHB, however, makes its methodology available, and it is not one that we found difficult to replicate. The basic idea is this:

  • Get all home sale prices for the geography and period
  • Calculate the prevailing mortage rate at the time of purchase to calculate the monthly payment for a 90% LTV loan on the property. Add in the monthly cost of property tax and insurance.
  • Assume that this cost should not exceed 28% of monthly income in order to be considered “affordable.”
  • Compare to the Area Median Income for the sale-year according to HUD’s AMI calculation for the MSA.

We were able to get all sales data for the city of Charlottesville for our time period of interest (2017 – 1H2022). We applied the similar filters used for our Repeat Sales Index, namely filtering out deed-of-gift transfers and looking only at units coded as Single Family, Single Family Detached, and the various categories of Single Family conversion. We used both the unadjusted AMI that HUD reports for Charlottesville’s MSA, and then also created an adjusted series where we scaled the HUD AMI by the difference between the city and MSA Family Median Income as reported in the American Community Survey (HUD’s AMI metric is based on Family Median Income, more details here). We used the same mortgage rates that NAHB uses, namely Freddie Mac’s 30-year mortgage rate index. We added in 0.95% of sale value as property tax and an additional 0.5% for insurance.

Our results are graphed below. We graph both the Charlottesville series that uses unadjusted HUD AMI and a series that uses the adjusted AMI. We compare that to the series for the entire USA and for a couple of comparison cities. We selected Houston because many upzoning proponents cite Houston, which has effectively no zoning code, as an example of an affordable market. We include Portland because the city’s CP consultant, RHI, mentions it as a model. And, finally, we include Boise as an example of a city that everyone would acknowledge has recently undergone a severe housing affordability crisis.

Cville affordability trend tracks USA closely, compares favorably to Houston, and far outperforms Portland and Boise.

The song remains the same: Charlottesville looks relatively affordable on yet another metric. Up until the second quarter of 2022, over 50% of housing sales were affordable at MSA AMI. And before the impact of pandemic-related distortions that hit in 2021, affordability sat at an even higher level. At City AMI, it was still close to 50%. We believe that the City AMI measure is not truly comparable to the MSA-level series. MSAs are geographically large and often contain “hinterlands” with relatively cheap real estate and fewer renters/more homeowners. NAHB does not publish the HOI at the city level, but we believe it is very likely the HOI for Portland city would be quite a lot lower than for the Portland MSA; likewise, the HOI for Houston MSA is flattered by high affordability in the Woodlands, Sugarland and other outlying, sprawling suburbs.

We also took a look at what proportion of houses would be available at lower income levels. We considered 60% and 80% of AMI. The FLUM narrative would have you believe that home ownership is entirely impossible for families at these income levels. Once again, we find that emotional claims on Twitter do not have much evidentiary value.

Significant fraction of homes sold at prices affordable to households well below 100% AMI

As it turns out, a not inconsiderable portion of homes sold in recent years have sold at prices that would count as affordable to households at 80% of AMI. Even for those at 60% of AMI, 10%-20% of homes have sold at realistically affordable prices. How do we square this result with the evident frustration that many evince when it comes to house-hunting in the city? Well, for one thing, Charlottesville is quite small, with well under 10,000 owner-occupied units in the whole city. That amounts to turnover of only 700 single-family homes a year. At any given point in time, the market will not provide expansive choice (even outside of times of very tight inventory, days-on-market tends to be under 30, so we’d typically see about 50-70 units on market at a time). Second, the characteristics of the housing stock may not line up well with contemporary preferences. The median size (by finished living space) of single-family units in the city is 1,300 square feet, a size that would have been small even by 1970s standards (in 1973, the median new single-family home size in the USA was 1,660 square feet). By current standards, the difference looks even more extreme — 2019’s average was 2,300 square feet. In the county, the corresponding median is 1,900 square feet. When people shop for housing, they don’t care about statistical and abstract measures of affordability of a theoretical and fungible “housing unit”. They care about whether they can find a house that meets their criteria within their price range. We get the sense, therefore, that some frustration is about finding a “house we like” at an affordable price rather than a house tout court. What the NAHB methodology shows is that there is indeed “a house” to fit the typical budget. It may not be a house that delights the modern sensibility; that said, given that the average multifamily condo price in Charlottesville (according to Zillow) trades at just over half the price of the average single family house, there is little evidence that the unmet desire is for such a housing form. Yet this is precisely the housing form the new zoning aims to supply!

High affordability for small houses, lower affordability for larger houses
High affordability at lower grades, lower affordability at high grades

We calculated the HOI for each census tract of the city, using data from 2019-2Q2022 in order to have a sufficient sample size for these smaller areas. We used the city-wide AMI, rather than calculating a separate AMI for each census tract, in order to see how affordable different parts of the city are to the typical city resident (as opposed to tract resident). Again, the rhetoric of upzoning would strongly suggest that large parts of the city are extremely unaffordable relative to others. We find, however, that the claims of “exclusion” are overblown and the accusation tends to be misplaced geographically to the extent is has any validity. The map below shows tract-level HOI, with the scale running from dark-green (highest HOI) to red (lowest HOI). You can click on each tract to see the HOI, the number of sales considered, median house size and mean assessor’s grade (A grade = 4, D grade =1). A neighborhood-level map is available here.

Most of the city scored between 45 and 65. The three “low affordability outliers” consisted of two of the densest tracts, with already-low R-1 fractions (the downtown area and the JPA area) plus the area directly to the north of UVA (Lewis Mountain and parts of Venable and Barracks-Rugby). When we look, however, at the median size and average assessor’s grade in each tract, we find that size and grade differences of houses sold explain virtually all of the variation in HOI (87% at the tract level and 80% when we drill down to the neighborhood level).

Quality/Size metric explains 78,5% of HOI variance at neigborhood level

This is an important difference from other jurisdictions where we often find that a quantitatively and qualitatively similar house in one neighborhood trades at a large discount to its analog in a more “exclusive” neighborhood (often driven by school attendance zones more than zoning).

Obviously, we understand that NAHB’s methodology necessarily simplifies matters. Not every household at or below 100% of AMI has a good enough credit score to qualify for a 90% LTV mortgage. Not every one of these households has enough capital available for a 10% downpayment. But by the same token, many households may have access to low-downpayment VA or FHA loans. And the “rule of 28” (the rule of thumb that says mortgage plus taxes should not exceed 28% of household income) may be overly conservative. In point of fact, many households in the US “violate” this rule of thumb without serious adverse consequences. In 2021, fully 27% of all households in owner-occupied homes with a mortgage had a housing cost-to-income metric above 30%; in 2015 the figure was even higher, at 29%. NAHB, as an institution, is heavily invested in the idea that the USA is “underbuilt” with respect to homes — it is a homebuilder organization after all. NAHB has tended to cheerlead the loosening of zoning rules, building codes, and land use restrictions generally. We should therefore bear in mind that the organization’s metrics may well tilt toward overestimating affordability challenges. But even so, we find that NAHB’s methodology shows most houses sold in Charlottesville have sold at levels affordable to the “middle of the market” in terms of household income.

We are left wondering: do the upzoning proponents play so loudly to emotion in order to drown out their audience’s intellect? A dispassionate look at almost any of the high-quality numbers available for the measure of housing affordability will find Charlottesville doing relatively well.

Appendix: A Look At the Highest-Affordability Markets

Some readers have asked us about those markets that score the highest on the NAHB’s HOI. Charlottesville may score in line with the US average, but a handful of MSAs have HOI scores as high as the low 90s. Why should Charlottesville not aim for that level of affordability? The answer is that having 90% of home sales taking place at a level affordable to the median income turns out to be a reliable indicator not of a highly functional housing market with robust new housing production (what most people would consider a desirable endpoint, and what upzoning proponents assert will be reliably generated by loosening zoning standards), but rather of sustained population decline. An interesting feature of housing markets (and markets for long-lived/hard-to-destroy durables generally) is that when demand tips into absolute decrease, asset prices quickly collapse to very low levels — levels far below cost of production. In areas where population is falling, many houses can change hands at near-zero prices. We find that the highest-scoring MSAs on the NAHB HOI metric all have falling populations, extremely low new housing production, and median home prices far below cost-of-production.

Extremely high HOI signals population collapse, not housing production

In fact, among the top 20 MSAs by latest (2h022) HOI (out of 230 for which an HOI is published) only two had statistically significant positive population growth from 2010 to 2020 according to the Census. Of these, one, Elkhart, IN, had population growth below the USA average (5.1%) and a housing production rate of only half of the USA average. The other, Indianapolis, IN, had population growth and housing production slightly above the USA average. However, neither exemplify dense, urban growth. Single-family detached homes represent around 70% of total housing units in these MSAs, well above the USA average of 61%.

While we are on the topic of extremes of the HOI index, we point out that the ten worst MSAs have HOIs ranging from 3.6(!!) to 11.1. Every single one is in California. Each of these California MSAs has its own set of zoning schemes, but all of them operate under the California Environmental Quality Act (CEQA). CEQA arguably imposes much heavier burdens on housing production than does zoning, illustrating how “land use regulatory restrictiveness” should not be conflated with “zoning” alone. It is also interesting to note that the HOI for every one of these MSAs is worse than the Cville HOI for families at 60% of AMI!

October 2022