National Association of Home Builders

Get serious about the stats we use

(Reading time: 9 minutes)

The housing stock working group has been tasked with two assignments in advance of its May meeting. The first is to collect ideas on misconceptions about—well, that’s not entirely clear, but it has to do with housing. The second is to review a supposedly detailed study that analyzes the costs of home building regulations. This is the “study,” formulated by the National Association of Home Builders, which concluded that regulatory costs and fees add $94,000 to the price of a new home.

It seems, therefore, that our second assignment fortuitously satisfies the first. The idea that the NAHB study is either detailed or has any relevance to the SAW region is a glaring misconception that the working group should be wary of promoting in any way.

As I wrote earlier, the NAHB study is deeply flawed because its respondents were woefully few in number and were self-selecting; because it is completely opaque about where the respondents were located or what kinds of housing they had built; and because it is, in any case, using data that predates the pandemic and the subsequent run-ups in housing costs, wages and interest rates. Its only apparent virtue is that it provides an easy if meaningless talking point, which endows it with a zombie-like persistence.

The question of whether regulatory costs and fees are excessive and therefore contributing to the unaffordability of new housing is a legitimate one, deserving of serious consideration. Unfortunately, doing anything in a serious way usually takes work, and it’s really tempting to avoid that kind of exertion when a supposed authority or expert offers an answer already gift-wrapped and tied up in a bow. So let me counter with another study, one that will take some effort to peruse and which superficially, at least, has even less relevance to the SAW region than the NAHB effort. For all that, however, I promise it can teach us all something useful.

“The High Cost of Producing Multifamily Housing in California,” which can be downloaded at https://www.rand.org/pubs/research_reports/RRA3743-1.html, was published earlier this month by the RAND Corporation’s Center on Housing and Homelessness. Its mathematically dense goal is “to identify policy reforms than can lower production costs and increase housing affordability in California,” which as the authors note, had seven of the ten most expensive metro regions in the U.S. in 2021. It does this by looking at data from more than 100 multifamily housing projects in California, Colorado and Texas and—hold on, here comes the wonky stuff—uses a regression-based statistical model to account for differences in development costs according to building type, size, whether financing was private or public and other variables.

In other words, the RAND study takes great pains to describe its sample base and methodology in a way that the NAHB doesn’t even acknowledge, much less detail. And while its greatest relevance is to California policy makers, its conceptual framework and analysis are applicable to all housing markets. At the very least, therefore, the RAND study provides a model for how to interrogate our own housing needs.

WHAT DISTINGUISHES THE RAND effort from its anemic NAHB cousin is its understanding—and willingness to explain—the complex interplay of cost-drivers that result in a final price tag for a housing unit. How much does it cost to build a multifamily housing project? Well, there are construction costs, which include labor and materials. There is the cost of the land itself. And there are “soft” costs, which include architectural, engineering and legal fees, the costs incurred by filing required environmental reports, inspection fees and regulatory compliance costs, financing costs and, in many jurisdictions, development or impact fees. The RAND study finds that, on average, 70% of the cost of a multifamily housing project is tied up in construction, 10% in the land and 20% in soft costs.

It is this last category that many people think about when they refer to burdensome regulatory costs, and there is some basis for that. Inspection fees, regulatory compliance costs and impact fees are all obviously the result of policy decisions that can be increased, modified or erased, depending on community standards. Likewise, architectural and engineering fees are sensitive to a community’s design and permitting requirements.

But land and construction costs are also sensitive to community policies. Restrictive zoning can easily drive up the cost of land. Hard costs can be driven higher by building codes that require certain construction materials or installation of life safety systems, like alarms or sprinklers, as well as landscaping or parking requirements. And labor costs can vary considerably, depending on minimum wage laws and whether subsidized housing projects are required to meet prevailing wage rates.

The point of detailing all these factors is to underscore the difficulty of teasing out how much a project’s costs are the result of regulatory and other policy decisions, which are almost always unique to that project in that location at that point in time. That’s why the NAHB “study” is so pointless, and why there is nothing comparable on a state or county level. There are too many variables, and it’s a moving target to boot.

So does that mean we’ve hit a brick wall? Not exactly. Not if we collect a significant amount of data in a targeted area and sort it according to certain specific categories. That’s what RAND did, categorizing multifamily units into four groups by dwelling size across the three states it targeted, further divided between those built with private funding and those participating in the low-income housing tax credit program. The results were starkly disparate:

Total development costs (TDC) per net rentable square foot (NRSF) for market-rate developments in Texas were $167, compared with $531 in San Francisco (2019 dollars). TDC for low-income tax subsidized units was $236 in Texas, versus $731 in San Francisco.
Total development costs per apartment in Texas averaged $133,000 for the market-rate units, $96,000 for the subsidized ones, indicating that the latter were considerably smaller in size. The comparable figures for San Francisco were $485,000 and $487,000, again indicating a sharp reduction in size of the subsidized units.
Average predevelopment time in Texas was 13.1 months, average construction time was 13.9 months, for a total time to completion of 27 months. The California (not just San Francisco) equivalents were 27.9 months in predevelopment and 21 months in construction, for a total time of more than four years before a project is completed.

Keeping in mind the earlier discussion about unique variables affecting every project, it should be noted that these comparisons are hardly apples to apples. Construction costs in San Francisco, to pick just one of the most obvious differences, are affected by seismic standards that you won’t find in Dallas. Real estate costs more in California, wages are higher, etc. etc.

Yet for all that, the disparities are too great to be explained away by such factors alone, as RAND’s regression analysis found, and their consequences are that average rental prices in California are nearly twice those in Texas. Why? Lower construction costs leave room for lower rents. And according to RAND, “Discretionary local impact fees are dramatically lower in the state [of Texas], at least in part because of strict oversight of these fees. But substantially lower levels of regulation overall and state policies that tightly constrain approval times likely play the most important role.”

WE CAN LEARN AT LEAST TWO things from all this.

First, the Texas and California statistics provide us with two extreme data points for total development costs, as well as extremes of production time. If local builders report development costs or production times closer to those found in Texas, there’s probably not a whole lot to be gained by second-guessing existing regulatory policies and procedures. Any gains from doing so would be marginal.

If, on the other hand, local production times and development costs are more comparable to California’s, clearly that opens up some ideas for addressing our affordable housing deficit. The RAND report’s recommendations, although explicitly intended to lower multifamily housing production costs in California, could have relevance to us, depending on which excessive regulatory cost-drivers are identified locally. Among the more intriguing:

Adoption of a policy, similar to state law in Texas, requiring local jurisdictions to approve or deny a proposal for a housing development within 30 days of submission. Failure to meet the 30-day deadline would result in automatic approval.
Promotion of policies to speed construction timelines, such as having synchronized rather than sequential inspections.
The returns on municipal impact and development fees, if any, should be weighed against potential gains from increased property tax revenue and other revenue and welfare gains from more new housing.
Adoption of large-scale upzoning to lower per-unit land prices and increase overall production. Even modest reforms, such as allowing duplexes or fourplexes where only single-family homes are allowed, can make a significant difference.

There doubtless are other conclusions that can be drawn from this study—it runs to 58 pages, and there’s a separate 48-page annex for anyone who really wants to get in the weeds—and there is abundant talent within our group to make that happen. But that will take work.

That work also would benefit from the input of developers and builders who looked to build in our area but have concluded they just couldn’t make their ideas pencil out—assuming, of course, that this has happened. If our group has had one continuing weak spot, it has been the lack of an industry voice to identify what could be done at the policy-setting level to encourage more affordable housing construction. That’s not entirely our fault—one of our members invited three different developers/builders “to join in the discussion about cost of regulations/what we can do for them/holdups in construction processes/etc.” but none responded—but it does create an information vacuum.

As a result, we don’t really know if a “regulatory burden” is an important reason why we don’t have developers breaking down our doors to build multifamily homes. The RAND study should at least help us formulate the right questions to ask, and also could help us decide that this is an area of inquiry that isn’t worth our efforts.

We have to know what we don’t know

(Reading time: 8 minutes)

One of the underlying issues pervading all aspects of the affordable housing discussion we’ve been having locally is the lack of reliable, timely data. There’s a lot we don’t know, and much of what we think we know is flawed.

There is, for example, the recently released regional housing study, which not only lacks a lot of needed information but is burdened by a significant load of outdated and incomplete statistics. The problem this poses is a false sense of authority. Statistics just look so damn definitive. They’re precise and official looking, and they do such a nice job of reducing complexity into the numeric equivalent of a soundbite that it’s hard to put them aside. They brook no argument—even when they send you off on a wild goose chase.

But the regional housing study is far from unique. A couple of more recent pronouncements about the housing situation illustrate how apparently authoritative sources can paint a picture that on closer examination is at least questionable. Yet because such statements fit so well into a broader understanding of our circumstances, they get adopted and repeated and eventually blend into the background narrative without a challenge. They become accepted wisdom, shutting down further discussion.

Consider, for instance, the question of how much government regulations add to the price of a newly built home. Developers may have to pay for environmental impact or traffic studies, as well as zoning, impact, utility hook-up and other fees, and have the additional costs of complying with OSHA regs and specific design standards. Builders must comply with building codes and architectural design standards, as well as foot the bill for permit or inspection fees. Could relaxing or amending some of this regulatory burden allow for cheaper housing to be built?

To ask the question is to answer it: of course it would. What that doesn’t tell us is how much of a cost-saving is possible, and whether that reduction would be significant enough to prompt the construction of more affordable housing. Are regulatory costs so high that they are a major disincentive for more housing development? Or are they relatively minor, in the overall scheme of things, and therefore unlikely to produce more than marginal gains if cut back, possibly at the price of lower quality?

We don’t actually know. All the meetings and conversations locally about “solving” the housing shortage have been remarkably unbalanced, in the sense that the people sitting around the table overwhelmingly are from the demand side of the housing equation. The supply side—developers, builders, lenders, underwriters, property managers—has been remarkably rare.

So when attendees at a recent SAW housing group meeting heard that $92,000 of a new home’s price tag is attributable to regulatory costs, it might have seemed that a significant information void had been filled. Moreover, given that this statistic was generated by the National Association of Home Builders (NAHB), it certainly sounded authoritative. And the implications are seemingly huge: with a new home in the SAW region going on the market for upwards of $326,000, as much as 28% of a new home’s selling price might be trimmed solely by government fiddling with the various requirements it imposes on builders and developers.

A closer look at that NAHB calculation, however, suggests otherwise. The actual regulatory cost calculated by the association was $93,870, of which it attributed $41,330 to developer’s regulatory costs and $52,540 to regulation during construction. The study was conducted four years ago. The developers’ costs were based on a survey sent nationwide to 2,071 NAHB members—with a scant 57 providing “complete and useable responses.” The association provides no information about which markets were represented in the responses, nor how widely they were distributed. The $41,330 number, in other words, is a wild-ass guess that has little to no relevance to the SAW region in 2025, despite its apparent precision.

Meanwhile, the survey based its conclusions about builders’ costs on 280 “complete and useable responses,” which sounds better than the developers’ stats but with no indication of whether this was a higher rate of return, since the NAHB doesn’t say how many builders were polled. And, again, the study provides no information about which markets were represented or what kinds of homes the builders were erecting. Nearly half of the increased costs those builders attributed to regulations were due to “changes to building codes over the past 10 years,” so there’s no applicability to localities with few or no changes to their building codes over that period. In addition, a substantial chunk of increased regulatory costs was attributed to “architectural design standards motivated by aesthetics, or possibly even, in some cases, a desire to price less affluent residents out of particular neighborhoods.” Is that relevant to the SAW region?

In short, the applicability of this extremely limited “study” to any particular housing market is less than zero— “less than” because using a misleading statistic can create a false sense of comprehension. But with the NAHB distilling a complex issue into a seemingly authoritative data point, there’s the temptation to think there’s no need to research the issue any further. The regulatory cost burden on new housing construction? Asked and answered.

A different kind of false certainty based on an apparently authoritative source was seen at the Staunton city council’s March 13 work session, at which the planning staff was asked what role developers play in terms of the city’s housing strategy. As just indicated, an accurate answer would have been “little to none.” The staff response, however, was to assert that “most of the land for larger developments has already been purchased, so they are looking at more of the low hanging fruit of the single lots and other smaller cottage-type developments.”

That answer not only was unresponsive, but highly questionable. Even a cursory look at a map of Staunton will disclose an abundance of undeveloped and open land, especially in the city’s northern reaches. The city’s 2018 comprehensive plan, currently being updated, noted that of Staunton’s 12,800 acres, nearly 3,000 acres was vacant land zoned for residential use. Some of that land undoubtedly has been developed in the past seven years, and a significant chunk of it is undevelopable because of steep terrain or flooding hazards, but even with that there’s clearly a lot of room within city limits for more housing.

But land availability isn’t subject merely to physical constraints. Land use ultimately is subject to political choices, notably over zoning. Those 3,000 vacant acres are apportioned among four zoning classifications, with the lowest-density classification claiming nearly a third of the total. Medium-density zoning, meanwhile, had 415 vacant acres, while high-density zoning weighed in with 325.8 vacant acres—all more than enough, one would presume, for at least some significant housing developments. But if that’s not enough, how much more housing could be built through upzoning? Is that something that should be at least examined, without a prior dismissal of the possibility?

Other political choices are reflected in the city’s decision to set aside considerable acreage for an ag-forestal district, “intended to support the growth of active farm, forestal, nursey and related enterprise.” Given Staunton’s location within a heavily agricultural county, it’s not unreasonable to ask whether preserving still more farmland within the city’s boundaries is the most appropriate use of such property, especially if doing so penalizes development of sufficient affordable housing. How many hundreds of acres of the ag-forestal district could be carved out for other uses while still preserving its most attractive natural features, such as the Bells Lane corridor?

Then there are the decisions that went into designing Staunton Crossing, the city’s premier economic development effort. Early on, planners for the project contemplated housing as part of its development mix, presumably in recognition of the need for new businesses to have adequate housing for their employees. But then, for reasons never made explicit, the housing idea got dropped, even as plans for an AI data center shrugged aside criticism that such centers provide only modest employment gains—the ostensible rationale for building Staunton Crossing in the first place. Meanwhile, in the seven or so years since those choices were made, new data centers have grown exponentially in size and become omnivorous consumers of water and electricity, raising the question of how well suited such an industrial application is for a region that has had frequent drought scares. Should that part of the project be reexamined to assess its suitability for housing?

None of this is to say that the city should be upzoning any particular area, that the ag-forestal district should be trimmed or rezoned entirely, or that Staunton Crossing should stop trying to recruit data center providers. But it does point to the fact that these and other land-use decisions are inherently political, made at a specific time for reasons that may change or become obsolete, and that new priorities—such as the growing need for affordable housing—may take on greater importance. To dismiss a question about new housing developments by saying, in effect, that there’s no room for big projects is therefore untrue. It also is needlessly self-limiting, forestalling fresh thinking that could open new possibilities.