by Thomas Silverstein
There is a type of article that appears periodically and that predictably generates reactions ranging from handwringing, at best, to outrage, at worst. The article will generally look at a single affordable housing project and highlight the property’s high per-unit costs. The article may or may not provide the context that the highlighted project is a bit of an outlier and average per-unit costs are lower. If it does, it will still try to paint the average costs as being intolerably high. In doing so, it may compare those costs to the median value of existing homes or to the per-unit cost of producing unsubsidized housing, both of which will inevitably be lower than the cost of producing affordable housing. We – the readers – will be left to decide which of two forks in the road to take: should we modify the policy design of our affordable housing programs to reduce costs, or should we give up on the premise of developing affordable housing because the private market is more efficient? Asking these questions naturally invites us to draw connections to the abundance discourse and abundance proponent Ezra Klein’s critique of “everything bagel liberalism,” a putative cause of these elevated costs.
Poppy Seed Bagel Progressivism
If you are reading this article, you probably are not inclined to turn off the funding (or tax expenditure) spigot for the affordable housing programs that exist in the United States. Thus, we are left with the former option of figuring out how to reduce the per-unit cost of affordable housing development, and we should figure out how to do that. The thrust of this article, however, is that we need to undertake that task with extreme care because many of the policy design features that increase the cost of affordable housing development actually reduce costs that would be absorbed elsewhere within our social safety net, effectuate important civil rights goals, or both. Looking at the ways in which labor, environmental, and civil rights provisions incorporated into affordable housing programs contribute to project costs while mitigating what we might otherwise be externalized social costs is helpful for charting a path forward. That path forward, to stick with the bit, might be a poppy seed bagel progressivism that is more just and equitable and involves fewer trade-offs than the plain bagel with which a myopic focus on cost reduction would leave us but more scalable and therefore effective at ending the housing crisis than the everything bagel liberalism of the status quo.
Labor
There is no serious dispute that increasing labor costs are one of the drivers of high affordable housing development costs. Growing labor costs also affect market-rate residential and non-residential development but not to the same extent because affordable housing developments may be, depending on the project, subject to the requirements of the Davis-Bacon Act (which requires payment of local prevailing wages to construction workers performing work on federally funded construction projects). Affordable housing developers might also opt for what would otherwise be suboptimal (and cost increasing) configurations in order to avoid compliance with Davis-Bacon. Some states and localities may impose prevailing wage-like requirements on a broader set of affordable housing projects than those covered under federal law, as well. It is self-evident that there is some cost associated with complying with these requirements.
If the conversation is left there, the obvious answer may be to abandon prevailing wage requirements and to put affordable housing developers and other developers on an equal footing in terms of labor costs. Choosing that answer, however, requires us to ignore a lot. First, by simply comparing the hourly wages of construction workers employed under prevailing wage requirements with those who are not, we are likely to miss cost savings flowing from increased worker productivity and decreased staff turnover. Second, even if we ignore those savings, paying construction workers less creates social costs that have to be borne somewhere. As of May 2024, Bureau of Labor Statistics data reflected that the median hourly wage for construction laborers was $22.47 per hour. The 25th percentile hourly wage, by contrast, was $18.32 per hour. According to the National Low Income Housing Coalition’s 2025 Out of Reach report, there was not a single state in the country where a worker earning $18.32 per hour could afford a modest two-bedroom apartment while workers earning $22.47 per hour could afford one in 14 states. In other words, the average construction laborer is already teetering on the edge of housing precarity. Pushing down wages could make that economic tenuousness universal among construction laborers rather than merely common. The same point that applies with respect to construction laborers’ ability to afford housing extends to other life necessities, such as food, clothing, and health care, that may be pushed further out of reach if wages declined. This precarity results in costs that are borne societally, ironically including through affordable housing programs themselves, and not just by low-wage workers.
Just as taking the problem of affordable housing costs seriously does not mean that we must scrap prevailing wage standards, taking the precarity of construction laborers seriously does not mean that we should view affordable housing program design as the most strategic intervention point for enhancing their economic well-being. Instead, a better policy mix, with some components geared towards protecting workers and some towards lowering costs, is possible. On the worker protection side, public policy should protect all workers, regardless of their sector, through substantially higher minimum wages and strengthened collective bargaining rights. The outcome of such shifts may – or, indeed, is supposed to be – higher wages, but it would be higher wages arrived at without the gymnastics that often accompany affordable housing developers’ choices about whether or not to comply with Davis-Bacon. Additionally, we should increase the long-term wage growth potential of construction laborers by investing in union apprenticeship programs that help workers transition into skilled trades.
There are also gains to be made through the more widespread utilization of modular construction, and, although that technological innovation would theoretically reduce demand for workers and therefore their bargaining power, it would not do so as significantly if accompanied by the scale of affordable (or ideally social) housing production that is needed to make up for our housing shortage. Immigration policy – albeit not during the Trump Administration – also holds potential for reducing labor costs. Both enabling more immigrants to enter the country and providing secure documented status and work authorization to those who are here already would make it easier for affordable housing developers to hire and retain workers, including without fear that their employees will be subject to removal proceedings during the middle of construction. Although, with a higher minimum wage or continuing prevailing wage requirements, the wages for immigrant construction laborers might be high (a good thing, to be clear), having a larger pool of potential workers would still reduce costs by helping developers avoid delays caused by staffing shortages. Lastly, if and only if we are able to significantly increase minimum wages and strengthen bargaining rights, we may reach a point where prevailing wage requirements that are baked into affordable housing programs are no longer needed to serve their laudable goal. In that scenario, taking a step back from prevailing wage requirements should not actually reduce construction laborers’ wages, but it should reduce affordable housing developers’ compliance costs.
It is impossible to ignore the fact that there can be tension between the ways in which we have historically attempted to boost construction laborers’ wages and the goal of increasing the volume of affordable housing unit production. However, it should be possible to harmonize these interests by strengthening generally applicable labor and employment laws, making career ladders to the skilled trades more available to construction laborers, embracing new technology, elevating the ambition of our affordable housing production goals, and reforming our immigration laws. That is the path forward for ensuring that attempts to reduce labor costs associated with affordable housing development do not result in significant externalities.
Environment
There are two analytically distinct sets of environmental considerations that can drive up the cost of affordable housing, and it is worth dealing with each in turn. The first relates to green building requirements, and the second relates to the environmental impact of development at particular sites.
With respect to green building requirements, the explanation for why compliance has greater upfront costs is obvious. If it cost the same amount or less to have thicker windows, more insulation, and energy-efficient appliances in comparison to thinner windows, less insulation, and energy-intensive appliances, requirements likely would not be needed in any case, and developers would default to green building. Moreover, for some green building criteria, it is also obvious why the cost would be higher: using more of the same type of insulation is going to cost more than using less of it.
The justifications for absorbing these costs when building affordable housing, however, are substantial. Many of these requirements pay for themselves through decreased utility bills over time. If the tenants would bear those increased utility costs, then avoiding that outcome prevents heaping more precarity on already low-income households. If the owner would bear those costs, then avoiding that outcome frees up resources for routine maintenance and important capital projects. Given the possibility that the climate crisis will motivate policymakers to force green retrofits in the future, it is also almost certainly more cost-effective to build green from the start. Lastly, if there are requirements that will not pay for themselves over time and that are unlikely to be forced on building owners down the line, it is plausible that some of those are means of avoiding the externalization of social costs that would be borne societally, whether through less clean air or disasters exacerbated by climate change.
Recognizing that the costs are probably worth it does not mean that we should not consider how we might reduce them. First, if there are green building requirements that, frankly, are not green and therefore neither reduce long-term utility costs nor internalize pollution- or climate-related externalities, then we should feel okay about discarding them. To be clear, discerning which requirements are worthwhile and which are not is likely to be difficult in practice. Second, government could change the cost of what is needed to comply with some green building requirements through subsidies or industrial policy. That could look like government paying producers of energy-efficient appliances to sell refrigerators to affordable housing developers at below market prices, and it could look like government providing the financing for the development of additional appliance manufacturing plants, theoretically leading to a greater supply of energy-efficient refrigerators and lower prices for those items. Lastly, if green building requirements are critical for addressing the climate crisis (among other objectives), there is a strong argument for holding market-rate developers to the same standards. Doing so may not directly reduce the per-unit cost of affordable housing, but it would close the gap in per-unit costs between affordable and market-rate development. Additionally, it actually could lower costs in absolute terms by establishing a larger market for, for example, energy-efficient refrigerators and incentivizing manufacturers to scale up their activities.
With respect to environmental requirements relating to project siting, a few considerations are noteworthy. First, environmental review, where it applies, is often a requirement of state laws like the California Environmental Quality Act that apply to development generally rather than affordable housing development specifically. Thus, while environmental review may increase costs for affordable housing development, it may not do so in ways that distinguish affordable housing development from market-rate development. Second, just as with green building requirements that reduce utility payments or avert the need for costly retrofits, some environmental constraints on siting may reduce or prevent long-term costs stemming from, for example, flood damage to properties. Third, it should still be fairly obvious that infill multifamily development that is not contributing to sprawl and that is not located on particularly inapt sites such as those that are in flood zones does not merit extensive environmental review. That is the case because such development is not actually causing externalities. Accordingly, streamlining environmental review requirements, whether under state environmental laws or under affordable housing programmatic criteria, does make sense.
In sum, we should have some tolerance for environmental requirements, whether to install an energy-efficient refrigerator or to pick a site that does not require filling wetlands, that increase the per-unit cost of building affordable housing, and a recognition that compliance with those requirements can reduce some long-term costs that still have to be borne by someone is critical to understanding why we should do that. At the same time, we can still attempt to reduce costs by making necessary goods less expensive to affordable housing developers through subsidies and industrial policy, eliminating requirements that are not grounded in evidence, and streamlining review for infill projects that are not on environmentally sensitive sites.
Civil Rights
Historically, affordable housing developments in the United States have been heavily concentrated in low-income communities of color and, in metropolitan areas with significant Black population, in poor and working class Black neighborhoods in particular. Intentional racial discrimination by both governmental and private sector actors has played a role in creating this dynamic, and factors ranging from land costs to zoning to the political capital of white communities to resist affordable housing development have reinforced it over time. With the Low-Income Housing Tax Credit (LIHTC) program having grown into the nation’s largest affordable housing production vehicle over the last four decades, efforts to achieve a more equitable distribution of affordable housing across neighborhoods and to foster residential racial integration have naturally focused on that program. Through both litigation and engagement in Qualified Allocation Plan (QAP) stakeholder engagement processes, fair housing advocates have pursued approaches like incorporating incentive points into QAPs, which guide which properties get credits, for properties located in low-poverty neighborhoods or having separate set-aside pools of tax credits available for properties in such neighborhoods. In metropolitan regions where central city neighborhoods are experiencing significant gentrification and displacement risk, some of these interventions have evolved to address both the need for more affordable housing in low-poverty neighborhoods and for the preservation of affordable housing in neighborhoods that are in the process of becoming higher income.
It is often but not always the case that a major cost component – land – will be higher for affordable housing developments in low-poverty neighborhoods than in higher poverty neighborhoods. Indeed, some of the prominent examples of high per-unit development costs that have received media attention are in such areas, including a project in the District of Columbia’s affluent and centrally located Adams Morgan neighborhood. It should not be surprising that land is often more expensive in higher income, lower poverty neighborhoods. To be clear, this is not always the case, especially in relatively low-density suburban contexts though the lower density limits for multifamily housing than in cities may erode the theoretical promise of lower land costs in those areas. Thus, we are left with a similar decision point in relation to the siting of affordable housing in low-poverty neighborhoods as were with respect to labor and environmental considerations: do we decide that the costs are simply worth bearing and that we should plow forward, do we decide that the costs are unsustainable and we should scrap the incentive points and set-asides that state housing finance agencies have devised in recent years, or do we chart a more nuanced path?
Unsurprisingly, the more nuanced path is the one most conducive to positive policy outcomes. While there may be some instances in which it makes sense to invest housing subsidy in locations with the highest land costs, it still makes sense to prioritize subsidy for locations that simultaneously have low poverty rates and low land costs. Intriguingly, at least in some contexts, these places may have higher performing schools than do the places with the very highest land costs. In the context of LIHTC, it should be possible to continue to prioritize affordable housing development in low-poverty neighborhoods by utilizing QAP provisions like separate set-aside pools for projects in low-poverty neighborhoods while incentivizing lower land costs within the pools for low-poverty areas by assigning scoring points for lower land costs. It is worth noting that, while this approach should work in theory, its potential in practice will be influenced by the need for zoning and land use reform given that places with low land costs and low poverty rates often have relatively restrictive zoning.
It should be noted that pursuing this approach with respect to how we prioritize housing subsidies like LIHTC need not mean that we let areas with low poverty rates and high land costs off the hook when it comes to developing their fair share of affordable housing. That is because it is precisely in those areas where inclusionary zoning has the greatest potential to produce significant numbers of affordable housing units because market-rate demand and therefore the potential for cross-subsidization is the greatest. An approach that combines subsidies like LIHTC in places with low-poverty levels and relatively low land values with inclusionary zoning in places that also have low-poverty rates but that have high land values should help us achieve a better balance of affordable housing units across high- and low-poverty neighborhoods in metropolitan regions. Accordingly, we should be able to mitigate the impact of high land costs on overall affordable housing development costs across the country without perpetuating segregation.
Before concluding this discussion of how civil rights-oriented policy implicates cost concerns in affordable housing development programs, it is worth addressing the issue of family-sized units. Some of the developments that have become poster children for high per-unit costs featured substantial portions of units – such as three-bedroom units – that are capable of accommodating families with children, a protected class under the Fair Housing Act. It is absolutely true that the per-unit cost of producing three-bedroom units is higher than that of producing studio or one-bedroom units; however, there is an emphatic rejoinder to that framing: the per-occupant cost of producing larger units, if it is higher at all, is not nearly so disproportionately high. Ultimately, people – as opposed to households – motivate the imperative of building affordable housing.
Tackling Inefficiencies That Do Not Implicate Social Justice Tradeoffs
Of course, just because some of the drivers of elevated affordable housing development costs implicate social justice concerns, that does not mean that they all do. Policymakers at all levels of government should be especially mindful of the opportunities to trim costs by tackling these other drivers. Going into these in depth could be the subject of an entirely separate article, but some notable ones include: complicated program design that requires the intensive use of lawyers, accountants, and consultants; exclusionary zoning, including restrictions on housing type and density as well as aesthetically-oriented design criteria; and unduly slow approval processes. Prioritizing the reduction of costs through tackling these contributing factors is preferable to doing so by reducing protections for workers, eliminating green building requirements, and concentrating affordable housing in high-poverty neighborhoods. That does not mean that doing so will be easy, as powerful political constituencies will predictably fight to preserve the status quo. But no one ever promised that reducing affordable housing development costs without sacrificing socially important interests would be easy.
Conclusion
If nothing else, this article should be taken as a call to avoid binary thinking as we decide how to try to reduce affordable housing development costs. Avoiding binary thinking will mean that we acknowledge that some cost-inducing policies reduce costs that would be borne elsewhere. It will mean that we look for ways of solving the problems that cost-inducing policies seek to address through policy frameworks that are outside of affordable housing program design. It will mean that we look for ways of reducing costs that do not involve trade-offs with social justice implications. At the end of the day, rather than being left to choose between an everything bagel or a plain bagel, we can have a poppyseed bagel (or sesame or garlic, if one prefers) that is nearly as satisfying as an everything bagel and nearly as efficient as a plain bagel. n
Thomas Silverstein is the Executive Director of the Poverty & Race Research Action Council and a Visiting Clinical Lecturer in Law at Yale Law School. He is a national leader in multiple approaches to advancing fair housing and community development.
