Editor's note: When the writer Barbara Ehrenreich spent some time
in the Twin Cities of Minneapolis-St. Paul to research her book
Nickel and Dimed: On (Not) Getting By in America, she had trouble
finding an inexpensive place to live. "There is one possibilityone
place in the entire Twin Cities that rents 'affordable' furnished apartments
on a weekly or monthly basis," she wrote, and had the presence
of mind to place quotation marks around the word affordable. She ended
up living somewhere else, and her quest for affordable housing became
a signature theme of her Twin Cities visit.
Ehrenreich certainly tapped the local zeitgeist on that one. The
Twin Cities media frequently report on affordable housing issues, and
policy forums on the topic are commonly held. The mayor of Minneapolis
made affordable housing a top priority of his campaign and first term
last year, promising to build a thermometer outside City Hall to measure
progress on the issue. Several months after election the thermometer
had yet to materialize, according to local press reports, because the
mayor's aides were unsure what the thermometer should actually measure.
That unbuilt thermometer is an apt metaphor for an issue that has
galvanized so many and yet is so hard to define. This is not only true
of the Twin Cities, but of cities and regions throughout the countryeven
many rural areaswhere housing affordability has become a top public
policy issue. The following article is an excerpt from a working paper
produced by the Minneapolis Fed that not only helps frame the debate,
but also offers a policy prescription. See the whole
paper, including appendices, notes, tables and additional graphs.
Many observers claim that we are in the midst of an "affordable
housing shortage" or, even worse, an "affordable housing crisis."
The primary concern is that too many households live in "unaffordable"
rental units. We hope to clarify the current debate by first measuring
the size of the problem, then diagnosing its underlying causes and,
finally, discussing treatments that policymakers should consider. While
our review is hardly exhaustive, we conclude that a shortage of income
is largely behind the housing affordability problem despite the current
focus on housing. Policymakers should recognize that government financing
of new housing units is unlikely to be a cost-effective response to
low household income.
Measuring the number of unaffordable housing units first requires a
definition of affordability. In this debate, a unit is considered unaffordable
if a household has to spend more than 30 percent of its income on it.
We use this standard to measure the size of the "shortage"
because of its public prominence, even though we recognize that such
a standard must be subjective. We also restrict our analysis to the
rental-housing market; the owner-occupied market remains affordable
by the commonly used standards. We find that the housing "crisis"
is heavily concentrated among one subset of the populationpoor
We then examine two of the most likely potential causes. First, low
incomes lead households to spend most of their income on necessities,
like housing. Second, government regulation, in part designed to improve
quality, can increase the cost of housing so that it is unaffordable.
The costs imposed by land-use regulation can be particularly pronounced
for the lowest-cost units.
After examining the data for the United States and for the Twin Cities,
a metropolitan area reputed to have a severe affordable housing shortage,
we find that low incomes are the primary reason why the poor live in
unaffordable rental units. Even if costs fell significantlyby
an amount roughly equal to estimates of the increase in cost due to
regulationthe vast majority of the poor living in the United States
and the Twin Cities would still live in rental units considered unaffordable.
Again, we note our review is limited in its scope and sophistication.
In some areas of the country, such as California and greater New York
City, regulation may play a large role in explaining high housing
cost-to-income ratios. However, others have come to similar conclusions
that these select areas are the exception rather than the rule (see
Glaeser and Gyourko 2002, for example).
Because of our diagnosis, we focus our discussion of policy responses
on the link between low incomes and high housing expenditures. If policymakers
believe that housing cost-to-income ratios are too high for low-income
households, they should provide low-income households with additional
cash or federal support, such as food stamps, for accessing other basic
necessities such as food and medicine. After all, concern about a high
housing cost-to-income ratio for low-income households only makes sense
if housing costs prevent those households from buying other basic necessities.
Concerns about affordability would assumedly disappear if households
could spend over 30 percent of their income on housing and still acquire
all the food, clothing and other necessities they need. Cash or federal
subsidies for nonhousing necessities offer a more cost-effective way
of increasing access to nonhousing necessities than subsidies that must
be used for housing. There are surely some households or individuals
for whom it makes sense to allocate public subsidies for housing (for
example, the homeless), but they constitute a small part of the population.
Despite the reasoning showing that limited income is the primary problem
and that public subsidies that can be used outside the housing market
are generally the reasonable response, some policymakers will only support
government subsidies for housing. These policymakers will have to choose
between programs that build new units and those that provide households
with the equivalent of cash for housing (that is, vouchers). During
that decision process, policymakers must address the evidence that housing
vouchers appear able to put households into affordable housing units
at a significantly lower cost than production programs. Housing vouchers
might prove less effective than production programs where government
restrictions significantly impede the market supply of housing. In these
cases, policymakers should consider the removal of regulations, a challenging
task that will require balancing competing interests, including the
desire to increase housing quality.
We make no claim of addressing all concerns raised about housing affordability
in the United States in coming to these conclusions. We do not address
data on rural housing markets, for example. Nor do we claim to highlight
novel solutions. Instead, we hope this essay contributes to the more
modest goal of framing the issues and encouraging the use of economic
and policy analysis in future discussions. In that vein, we hope that
the continued debate on housing affordability stays focused on specific
failures in housing markets, the effects of existing government interventionparticularly
land-use regulationin housing market outcomes and the central
role that income appears to play in generating concerns about housing
The data on housing expenditures
Those calling current conditions an affordable housing crisis rely
on community or social standards to determine that a unit is "unaffordable"
even when a household has the resources to live in it. This approach
has been rightly criticized for several reasons, including its obvious
subjective nature (see Appendix 1 of the working paper). Despite these reservations, we use the
"30 percent of income" definition because it is the most frequently
cited standard and one used by the U.S. Department of Housing and Urban
Development (HUD). Specifically, this standard focuses on rental units
and defines a unit as unaffordable if a household would have to spend
more than 30 percent of its gross income on rent and utilities. Analysts
have used the
fixed-standard method in many ways, all of which suggest that the affordability
problem is confined to households with very limited financial resources.
(We focus on the rental market because the owner-occupied market, discussed
in working paper Appendix 2, seems affordable based on frequently used social standards.)
Before we review these data in detail, we have to define the geographic
boundaries for the housing market under examination. Much analysis and
data describe national trends in housing supply, demand and prices.
But, consumers and producers make decisions based on conditions in much
smaller local housing markets. To address both perspectives while keeping
our discussion focused, we review data for the United States and for
a metropolitan area reportedly suffering from one of the worst affordable
housing shortages in the country, the Minneapolis-St. Paul metropolitan
According to recent congressional testimony from a HUD economist, the
traditional approach to identifying a housing shortage compares the
number of units affordable to households whose income falls between
predefined amounts to the number of households that fall into that income
group (see Nelson 2001). Incomes are generally expressed as a percentage
of the median income for an area. So, for example, such analysis might
compare the number of units affordable to those households with incomes
at or below 80 percent of area median income to the number of households
that have that income. On the national level in 1999, the only group
of renters for whom a shortage existed were those considered by HUD
to have "extremely low incomes" (defined as at or below 30
percent of area median income). Households with extremely low incomes
typically fall below the official poverty line (see Nelson 2001). In
the Twin Cities, a shortage of rental units is also limited to those
households with extremely low incomes as of 1998. (Note that these 1999
and 1998 data from the American Housing Survey are the most current.
The detailed 2000 decennial census data have not been released.)
Some object to this approach because it does not differentiate between
existing affordable units and those affordable units actually available.
A household with an income at the median for an area could decide to
occupy a unit that is affordable to a household with an extremely low
income. While the low-cost unit is affordable to the lower-income household,
it is not actually available. To provide a more comprehensive view of
affordability, we also examine the total number of rental households
that spend more than 30 percent of their income on housing.
This second approach yields the same answer as the first: Virtually
all those living in unaffordable units have very low incomes. In the
Twin Cities, the most recent data show that 69 percent of all renting
households in unaffordable units had extremely low incomes (another
25 percent had incomes between 30 percent and 50 percent of the area
median). The national figures were generally comparable, although a
smaller percentage of renters living in unaffordable housing had extremely
Income as a potential underlying
cause of the housing crisis
The fact that households in unaffordable units have very low incomes
should naturally lead policymakers to consider low income as one of
the potential causes of the "shortage." Very low-income households
might have only enough money to pay for the basics. The most expensive
basics will consume a greater share of their income. Housing is typically
the largest expenditure for most households. Higher-income households
spend less of their income on housing, on average, than lower-income
households. As a result, housing ends up consuming a very large share
of expenditures for low-income households, especially relative to higher-income
households. (See graph for data on household expenditures and income
Looking to income as a source of a social concern is not a novel proposition
for policymakers. The current measure of poverty has its roots in the
amount of income a household needs to buy an adequate amount of food.
But when the number of households in poverty rises, policymakers and
the public do not typically consider it evidence of an "affordable
food crisis" that requires the planting of additional crops. There
is no formal connection between the amount of subsidies that farmers
receive and the number of households below the poverty level, nor do
policymakers typically scrutinize agricultural processors and grocery
stores to find a cause of high poverty rates.
Likewise, widely available data indicate that those with low incomes
allocate 8 percent of their expenditures to utilities, twice the percentage
of those with the highest incomes. Yet few have called on the government
to build new generators to address excessive utility expenditures by
the poor. Instead, policymakers typically view such data as evidence
that some households have too little income and respond by, for example,
providing more income to spend on energy
(for example, the Low-Income Home Energy Assistance Program and "lifeline"
programs). But, before we can surmise that the housing crisis has its
roots in income, we must examine an alternative explanation for the
Another potential underlying cause:
The private market can and does produce low-cost housing, usually through
a process called filtering, where existing housing units drop in cost
as their relative quality falls, rather than through construction of
new, lower-cost units. The most direct test of filtering suggests that
this process captures the workings of the real world. Malpezzi and Green
(1996, p. 1811) found that "high quality new construction is associated
with a growth in the low quality stock. ... In 1995 ... new units led
to an increase of nearly 2.5 percent in sub-standard rental units."
Although the private market appears capable of providing units that
would meet affordability standards for the very poor, this solution
has an important drawback. In housing markets, the actions of one household
or firm can strongly influence the quality of life of neighboring households.
For example, the owner of an apartment complex that lets the building
fall in quality also has a negative effect on the value of the units
nearby (a new building of very low quality could have the same effect).
The owner of the building whose value is depressed might have less incentive
to maintain its units, which, in turn, reduces the value of other units
in the vicinity. Owners do not bear all the costs imposed by their actions
and do not fully incorporate these costs into their actions. The result
is that a private market might produce housing units that are cheap
but which have too low a quality level.
Local government regulation, in theory, could address this failure.
Local governments regulate land use in many ways, including restricting
the types of housing built (e.g., apartments vs. single-family homes)
as well as dictating the attributes of that housing (e.g., number of
garages). While these rules could increase the quality of housing, they
can reduce the availability of low-cost housing by, for example, limiting
the total number of housing units produced in a community. The result
is less housing and higher prices. As McFate (1999, p. 172) summarized,
"We have regulated substandard housing out of existencethe
result is that the poor have safer but more expensive housing, and less
money available for other goods." (See working paper Appendix
3 for a discussion of quality trends in rental markets.)
Coming up with empirically sound estimates of the cost of local
land-use regulation is quite difficult. One estimate using a reasonable
approach found that a significant increase in regulationmore specifically,
raising regulatory intensity from relatively low to relatively high
levelswould increase rents by 17 percent. These results are generally
consistent with other research findings, such as those reviewed by Fischel
(1990). Moreover, the small number of detailed cost/benefit analyses
find that regulatory costs exceed the benefits they produce. (Working
paper Appendix 4 discusses
another slightly more complicated justification for regulation, namely,
to ensure the efficient provision of government service.)
Income is the cause
We have described two possible causes for the fact that most very poor
renters live in unaffordable units: too little income or too much regulation.
Because unaffordable units are almost all occupied by the poor, our
expectation is that income plays a larger role than regulation. This
conclusion is consistent with anecdotal information. As Anthony Downs
(1991, p. 1105), one of the most prominent critics of local regulation,
noted, "It is certainly true that eliminating all regulatory barriers
to housing affordability would not come close to ending the existing
housing affordability problems of America's low income households. Those
problems are caused more by poverty and low income than by high housing
costs. ... " Similar sentiments have been expressed in the Twin
Cities by a mayoral committee investigating the affordable housing crisis:
"It is important to stress that these cost reductions [from reduced
local government regulation] ... will never be sizeable enough to fully
eliminate the need for subsidies to produce affordable housing for all
income levels" (Mayor's Regional Housing Task Force 2000, p. 8).
To test this theory, we use the most detailed recent data on housing
costs in the Twin Cities to simulate a reduction in rents. We choose
a 15 percent reduction in rents because it approximates the amount by
which rents are estimated to increase by shifting from a lightly regulated
to a heavily regulated environment (as noted earlier, this estimate
was 17 percent). This reduction in rent drops the number of all renters
living in unaffordable housing from 44 percent of all renters to 34
percent of all renters. This reduction in rent drops the number of renters
with extremely low incomes living in unaffordable units from 80 percent
of all such renting households to 70 percent. National data show similar
results. The percentage of renters living in unaffordable units falls
from 44 percent to 35 percent, while the percentage of extremely low-income
households living in unaffordable units falls from 85 percent to 78
Certainly, this analytical exercise is more illustrative than definitive.
Using another methodology, analysts have identified regulation as having
particularly deleterious effects on housing prices in certain parts
of the country, especially California and New York City. Moreover, the
declines in the number of renters in unaffordable units produced by
the decrease in rents was not trivial. Nonetheless, this exercise suggests
that even a significant reduction in cost would not shift the vast majority
of renting households currently in unaffordable units to affordable
units. As a result, the simulation suggests that low incomes play the
dominant role in explaining current affordability concerns in the country
as a whole and in an area reputed to suffer from a severe affordability
Since we view low income as the primary cause of our concerns about
affordability, we focus our discussion on policies that address it.
Issues policymakers face in reforming regulation in housing markets
are discussed in Appendix 4 of the working paper. As a general rule, we believe that policymakers
should respond to the high housing cost-to-income ratios by facilitating the ability of low-income households
to acquire nonhousing necessities like food and medicine. Assumedly,
we only care about high housing costs because they prevent households
from acquiring these other goods and services. The provision of cash
or cashlike subsidies is the most cost-effective way of assisting low-income
households in acquiring the necessities that high housing costs crowd
out. If policymakers provide subsidies that can only be used in the
housing market, then they must address the fact that building new housing
units appears to be a relatively expensive method of providing low-cost
housing compared to the alternatives.
Taken literally, a high housing cost-to-income ratio justifies increased
nonhousing expenditures by the government. Observers care about high
spending on housing by lower-income households because it leads to less
spending on other items that observers consider necessary to thrive.
The Minneapolis Affordable Housing Task Force (1999, p. 3) starts its
report arguing that "children are going without enough food because
their families cannot afford both food and rent." The notion that
a high housing cost-to-income ratio justifies increased government support
for nonhousing items may still seem counterintuitive and abstract. But
simply consider the case of very high-income households paying more
than 30 percent of their income on housing that can still afford large
amounts of food, clothing, transportation and so on. We cannot believe
that policymakers would support legislation to fix this "affordability"
problem. This suggests that a high housing cost-to-income ratio in the
absence of other nonhousing deprivations cannot be the source of policy
If the problem is too little consumption of nonhousing goods, the government
should increase its support for low-income households' acquisition of
food, utilities, health care and the other items that high housing costs
may limit (see Olsen 2001 for a discussion of this point). Subsidizing
housing does not seem to be a cost-effective method for increasing a
household's access to nonhousing goods and services. Programs that subsidize
housing naturally seek to increase consumption of housing and thus drain
away resources from the ultimate goal of increasing the delivery of
nonhousing items. To increase a household's consumption of nonhousing
goods by a dollar would require the government to spend more than a
dollar, maybe much more, if the subsidy is restricted in its use to
So a logical response to a high housing cost-to-income ratio is to directly
increase the income of low-income households. The Earned Income Tax
Credit is an example of a program that provides a cash transfer. Cash
has the attractive feature of allowing the recipient to determine the
highest-value use of the money. Alternatively, the government could
provide these low-income households with additional food stamps, health
insurance and similar benefits that make nonhousing items more attainable.
We emphasize that our argument for a nonhousing response to a supposed
housing problem would not apply to households whose resources are so
limited that they have no shelter. Food could not replace housing for
these households. However, the number of households that would fall
into these categories is a very small part of the population.
Some might object to policies that address the income aspect of the
housing crisis through cash or other nonhousing transfers. We now turn
to policy responses that would seek to subsidize the consumption of
Housing-related responses: vouchers vs. production programs
Some policymakers would only support policies that direct government
funds to expenditures on housing. A particularly important question
for policymakers is if they want to subsidize specific new units or
tenants through a housing voucher.
A summary of available research on housing assistance aimed at
low-income households, albeit often limited and dated, offers several
reasons to support vouchers. The most important in our view is that
vouchers provide affordable units at a much lower cost than new production
programs by relying on older, already existing housing units (the kind
of housing that nearly all households live in). The General Accounting
Office (2001, p. 2) recently found that after controlling for unit size
and general location, "total per-unit costs for housing production
programs are from 32 to 59 percent greater than for housing vouchers
in the first year and from 12 to 27 percent greater over 30 years."
Because they are more cost-effective, vouchers could serve more households
than production programs.
More generally, a shift to vouchers is also consistent with the view
that the "housing" problems for low-income households represent
"income" problems. The voucher system relies on private markets
to provide low-cost housing. Vouchers increase housing consumption by
increasing the income of households. Private markets can provide lower-cost
housing more cost effectively by relying on existing units than can
new production programs. As the case of filtering revealed and Quigley
(1999, p. 48) highlights, "It is far more expensive to create low-income
housing through new construction than through depreciation." A
wide range of economists argue that increased income for poor households
represents the most effective policy solution for an affordability problem.
As long as private markets are not overly constrained by regulation,
the increase in consumption catalyzed by the voucher program should
signal private markets to increase supply. Thus, vouchers should be
effective even in areas with low vacancy rates over the long run. However,
in cases where regulation binds the market's supply response, the increase
in consumption induced by vouchers may only end up raising prices. In
these cases, policymakers should consider attempts to reduce regulation,
although, as discussed in working paper Appendix
4, making such changes will be a challenge.
Advisory Commission on Regulatory Barriers to Affordable Housing. 1991.
Not in my back yard. Washington, D.C.
Downs, Anthony. 1991. The Advisory Commission on Regulatory Barriers
to Affordable Housing: Its behavior and accomplishments. Housing
Policy Debate 2 (4): 1095-1137.
Fischel, William. 1990. Do growth controls matter? A review of empirical
evidence on the effectiveness and efficiency of local government land
use regulation. Cambridge, Mass.: Lincoln Institute of Land Policy.
General Accounting Office. 2001. Federal housing programs: What they
cost and what they provide. July 18.
Glaeser, Edward and Gyourko, Joseph. 2002. The impact of zoning on housing
affordability. Working Paper 8835. National Bureau of Economic Research.
Linneman, Peter and Megbolugbe, Isaac. F. 1992. Housing affordability:
Myth or reality? Urban Studies 29 (3/4): 369-392.
Malpezzi, Stephen and Green, Richard. K. 1996. What has happened to
the bottom of the U.S. housing market? Urban Studies 33 (10):
Mayors' Regional Housing Task Force. 2000. Affordable housing for the
region: Strategies for building stronger communities.
McFate, Katherine. 1999. General commentary. Federal Reserve Bank
of New York Economic Policy Review 5 (3): 171-177.
Minneapolis Affordable Housing Task Force Report, July 15, 1999.
Nelson, Kathryn P. U.S. Department of Housing and Urban Development.
Testimony before the House Committee on Financial Services, Subcommittee
of Housing and Community Opportunity, May 3, 2001.
Olsen, Edgar. O. 2001. Housing programs for low-income households. Working
Paper 8208. National Bureau of Economic Research.
Orr, James A. and Peach, Richard W. 1999. Housing outcomes: An assessment
of long-term trends. Federal Reserve Bank of New York Economic Policy
Review 5 (3): 1-11.
Pozdena, Randall. J. 1988. The modern economics of housing. New
Quigley, John. M. 1999. A decent home: Housing policy in perspective.
Working Paper W99-007. Institute of Business and Economic Research,
University of California-Berkeley.
State of Minnesota, Office of the Legislative Auditor. 2001. Affordable
Susin, Scott. 2002. Rent vouchers and the price of low-income housing.
Journal of Public Economics 83 (1): 109-152.
Turner, Margery Austin. 1998. Moving out of poverty: Expanding mobility
and choice through tenant-based housing assistance. Housing Policy
Debate 9 (2): 373-394.