Cities 72 (2018) 411–420

Contents lists available at ScienceDirect

Cities
journal homepage: www.elsevier.com/locate/cities

Why there? Developers' rationale for building social housing in the urban T
periphery in Latin America
Nora Ruth Libertun de Duren
Housing and Urban Development Senior Specialist, Inter-American Development Bank, 1300 New York Avenue, NW, Washington DC 20577, United States

A B S T R A C T

Inadequate housing has become endemic to Latin American cities for over six decades. All that has changed has
been who is going where. In the 1960s, the rural poor who came to the city solved their housing needs by
building their own informal settlements on peri-urban lands. Today, the urban poor relocate to peri-urban
housing complexes built by the private sector with state subsidies. Why have these new housing units for low-
income households been built in peri-urban areas? This paper examines some of the mechanisms behind the
location of the urban poor in cities, with a specific focus on the role developers have played in the construction of
affordable housing in peri-urban areas of Brazil and, Mexico. The paper explores these mechanisms through
interviews with affordable housing developers. We found that economies of scale – and not land prices – explain
developers' preference for building in peripheral areas. Initial savings that accrue to developers due to lower land
prices in the periphery are offset by the cost of having to build basic onsite infrastructure. Plus, large lots – which
are available almost exclusively in urban peripheries – enable developers to achieve significant cost savings
because these large lots make it possible for developers to build more than 500 units. In addition, weaker
municipal regulations and fewer bidders, both of which are typical for projects in difficult-to-access peripheral
locations, make for a shorter and easier approval process for these large housing projects.

1. Housing demand and housing policy
The dream of living in a safe and comfortable home eludes millions
of urban residents everywhere. In Latin America, more than a third of
the population still does not have adequate housing, and it has been
that way for more than six decades. This may be partly due to the way
the region urbanized. Cities grew very rapidly as a consequence of an
inflow of migrants who were often fleeing either poverty or violence.
In fact, the housing deficit has been one of the shaping forces of
today's Latin American city. Since 1950, the number of urban residents
in the region has grown sevenfold, and the urban population has now
reached 450 million (UN-DESA, 2013). But as the region's economic
growth has averaged only 3% annually over the last 60 years, in-
adequate housing has become endemic (Hardoy & Satterthwaite, 2014).
All that has changed has been who is going where. In the 1960s, the
rural poor who came to the city solved their housing needs by building
their own informal settlements on peri-urban lands (Aguilar,
Ward, & Smith, 2003). Today, the urban poor relocate to peri-urban
housing complexes built by the private sector with state subsidies.
Why have these new housing units for low-income households been
built in peri-urban areas? This paper examines some of the mechanisms
behind the location of the urban poor in cities, with a specific focus on
the role developers have played in the construction of affordable
housing in peri-urban areas of Latin America. By doing so, it con-
ceptualizes housing location as the outcome of a rational response to
the economic incentives policy frameworks create. In that sense, it
continues the scholarly tradition of analyzing urban form as the phy-
sical consequence of the search for economic surplus (Harvey, 1985). At
the same time as it emphasizes that the regulatory frameworks play a
significant role in determining the economic advantages of each specific
form of land occupation, which takes place on top of pre-existing im-
balances between the public and private powers regulating land uses.
These imbalances are not only between public and private actors, but
also between municipal jurisdictions within metropolitan areas.
The paper explores these mechanisms through a qualitative meth-
odology based on semi structured questionnaires and in depth open
ended interviews with developers of subsidized affordable housing in
Goiania, Brazil and in Puebla, Mexico. In total, ten developers re-
presenting four companies participated. The interviewees were evenly
divided between those who develop affordable housing in central lo-
cation and those who develop it in peripheral locations. Both Brazil and
Mexico have ambitious housing programs where most housing units
have been built in peripheral areas. Together, these countries account
for about half of the region's households. Both have demonstrated a

E-mail address: nlibertun@iadb.org.

http://dx.doi.org/10.1016/j.cities.2017.10.006
Received 1 June 2017; Received in revised form 29 August 2017; Accepted 7 October 2017
Available online 02 November 2017
0264-2751/ © 2017 Elsevier Ltd. All rights reserved.
N.R. Libertun de Duren
century-long commitment to supporting affordable housing, but they
have seen a shift from the state as the producer of affordable housing to
the state as the regulator of private production of such housing
(Bouillon, 2012).
The findings of these paper show that developers who built af-
fordable housing in the urban periphery are intrinsically indifferent to
project location. Their preference for building in those areas rest in
their ability to find larger plots of unused land with a lower number of
bidders. This is because these plots allow them for shortening the time
for municipal approvals, and for accommodating larger number of units
and hence achieving economies of scale that contribute to lower the
cost per housing unit built. The consequence on the urban form is that
the larger housing complexes get built in the less attractive areas of
peripheral municipalities.
With this finding this paper also contributes to the urban form
theory based on the seminal works of William Alonso, which poses that
households' economic resources determine where they choose to reside
and that the aggregate result of these decisions shapes residential
geography. A household decision is the outcome of a tradeoff between
housing amenities and housing location. In as much as the cost of
commuting remains low, it could choose to live in suburban areas
where they can afford larger plots (Alonso, 1964). But the dynamics is
different for a household who depends on a subsidy for buying a
housing unit. This subsidy does not allow him to participate in the
existing real estate market, but stimulate developers to produce ad-hoc
housing units at below market prices. Hence, and as households are a
captive market, it shifts the trade-off decision between amenities and
location from households to developers.
The next section provides background on housing demand and
housing policies in Brazil and Mexico. The profile of the cities studied,
the methodology and the main findings of the study are then presented.
The final section concludes with a critical analysis of current housing
policies.
The extent to which inadequate housing shapes urbanization today
in Latin America is striking. Of the more than 150 million Latin
American households suffering some housing deficit, the vast majority
are urban households that lack access to basic infrastructure. In terms of
the percentages of people affected, the housing deficit stands at 32% in
Brazil (79.6 million people) and 27% in Mexico (31.2 million people).
Two-thirds of these housing deficits are due to the quality of housing –
that is, people living in overcrowded units, lacking proper titles, or not
having access to basic services such as tap water and sewerage. The
other third suffers what are called quantitative deficits, which implies
that they either do not have a home, share one with another family, or
live in an improvised dwelling.
If the current urban population growth rate continues, it is expected
that 2 million out of the 3 million households that are added to Latin
American cities every year will suffer a housing deficit (Bouillon, 2012).
The magnitude and the persistency of this housing deficit is an in-
dication of the difficulties faced by national governments trying to close
this gap (Rolnik, 2013).
Both Brazil and Mexico have specific institutions responsible for
designing and supporting affordable housing (Murray & Clapham,
2015): the Ministry of Cities in Brazil; and the Secretariat of Territorial,
Agrarian, and Urban Development in Mexico (Secretaría de Desarrollo
Agrario, Territorial y Urbano – SEDATU) (Table 1). These ministries
have solid frameworks though which they channel funding for afford-
able housing programs. Brazil relies on the Federal Savings Bank, also
known as La Caixa, (La Caixa Econômica), a government-owned cor-
poration that provides banking and financial services. The Federal
Savings Bank is the fourth-largest bank in terms of assets in Brazil
(Caixa, 2015). Mexico's financing mechanism is similar, although more
complicated. The Federal Mortgage Society (Sociedad Hipotecaria
Federal - SHF), a second-tier bank with funding from the Ministry of
Finance, sells credit products to financial intermediaries, including the
National Housing Fund Institute for Workers (Instituto del Fondo
412
Cities 72 (2018) 411–420
Nacional de la Vivienda para los Trabajadores - INFONAVIT) for private
sector employees and the Social Security and Services Institute's
Housing Fund for State Employees (Fondo de la Vivienda del Instituto
de Seguridad y Servicios Sociales de los Trabajadores del Estado –
FOVISSSTE) for public sector workers. These two agencies collect
mandatory contributions from employers and employees for a fund to
finance affordable housing, which is then built by the private sector
(SEDATU, 2013). As the Table 1 shows, the aim and organizational
structure of both countries is somewhat similar, although the Mexican
case is more complex because it includes more than one financial in-
termediary.
In addition, the national budgets of these countries allocate a sub-
stantial amount of funding for housing programs that historically has
ranged between 1 and 2% of their respective gross domestic products
(GDP). By comparison, the United States allocates about 0.55% of its
GDP to support housing programs (OMB, 2015), and member countries
of the Organization for Economic Co-operation and Development
(OECD) allocate an average of less than 1% of GDP (OECD, 2012).
Currently, expenditure on social housing programs in Mexico and Brazil
amounts to about 1.7% of their respective GDPs.1 This is twice what
Mexico spent and four times what Brazil spent in the 1990s. The con-
tribution of these funds to increasing the housing supply is consider-
able. In Brazil, the flagship national program, My House My Life (Minha
Casa Minha Vida - MCMV), facilitates housing credit for low-income
households without access to mortgages, thus fostering the construction
of housing for these households. The program has been responsible for
about 750,000 housing units per year, which represents about 30% of
annual housing production by Brazil's formal housing sector
(Magalhaes, 2013). Determining the amount of housing production in
Mexico is particularly complex because 60% of it is informal (Hebert,
Belsky, & DuBroff, 2012). One approximation is that 70% of the
600,000 housing mortgages filed each year are linked to national pro-
grams for social housing (Monkkonen, 2011).
Table 2 shows the housing deficits in Brazil and Mexico, the
amounts of subsidies for affordable social housing, and the numbers of
beneficiaries. Brazil has a larger, more urbanized population and a
higher level of housing deficit than Mexico. In both countries, most the
deficit is of a qualitative nature, although in the Brazilian case one in
three households in housing deficit classifies as needing a new unit, and
in the Mexican case only one in twenty. Nonetheless, in both countries
most funding goes for building new units. In terms of size, Brazil's total
number of households with housing deficits more than doubles that of
Mexico, while its funding for programs amounts for more than six times
the Mexican funds.
The impressive scale of these national programs has aimed for a
large impact as measured by the number of units delivered. For ex-
ample, in 2013 the Mexican government pledged to deliver 1 million
social housing units in its first year of office, while in 2014 the Brazilian
government promised to deliver 2.75 million housing units by the end
of its administration. While these numerical targets did not imply any
specific urban focus, they certainly increased the pressure for quick
delivery of large-scale projects, the magnitude of which could only be
attained in urban peripheries.
Data on the budget allocations of national programs show a marked
preference for sponsoring programs that support new housing units
over those that improve existing ones (Tables 3 and 4). This signals a
pronounced mismatch between the nature of the housing deficit and the
type of housing support offered, as surveys indicate that the vast ma-
jority of households need to improve their current unit (qualitative
1
Expenditure on housing is particularly centralized in Latin American countries. In
2000, after the decentralization process was almost consolidated, housing accounted for
about 14% of total central government spending but less than 9% of local government
spending. In contrast the average share of housing expenditure among OECD countries
was less than 3% for national governments and more than 13% for local ones (De Mello,
2003)
N.R. Libertun de Duren Cities 72 (2018) 411–420

Table 1
Institutions responsible for designing and supporting affordable housing in Brazil and Mexico.

Brazil Mexico

Policy Ministry of Cities SEDATU (Secretariat of Territorial, Agrarian, and Urban Development)
CONAVI (National Housing Commission)
Funds (main sources) Growth Acceleration Program Federal Mortgage Society (SFH)
Financing and management (main La Caixa (Federal Savings Bank) INFONAVIT (National Housing Fund Institute for Workers)
players) FOVISSSTE (Housing Fund of the Social Security and Services Institute for
State Employees)
Territorial plan Municipal authorities Municipal authorities
Execution Private developers Private developers
Flagship program analyzed MCMV (My House, My Life) National Development Plan 2007–2012
Program aim New housing for first-time homeowners with household New housing for first-time homeowners with household incomes of one to
incomes of up to three minimum salaries four minimum salaries working in the formal sector
Annual target (number of units) 690,000 units 1,000,000 units

Table 2 developers and the remainder for self-construction (SHF, 2014). In

Housing Deficits and affordable housing subsidies in Brazil and Mexico. Brazil, the MCMV program supports two subprograms: the National
Rural Housing Program, which provides loans to improve rural homes;
Brazil Mexico
and the National Urban Housing Program, which provides access to
Unit Numbers Percent Numbers Percent credit that enables poor urban households to become first-time home-
owners of new housing units.2 Urban dwellers who want to improve
Population People 202,700,000 100 112,300,000 100 their current home need to be included within the Precarious Settle-
Urban People 172,300,000 85 88,700,000 79
population
ment Program (Programa Assentamentos Precários), which funds in-
Rural People 30,400,000 15 23,600,000 21 tegral upgrading of urban slums. In practice, these programs have a
Total housing People 79,600,000 32 31,200,000 27 virtual monopoly on formal housing choices for households that cannot
deficit afford a unit at nonsubsidized market prices,3 which makes this bias
Qualitative People 52,000,000 21 21,500,000 19
toward ownership relevant.
Quantitative People 27,600,000 11 9,700,000 8
Total Housing Homes 5,431,000 100 13,444,000 100 National housing programs rely on their financing structure to en-
unit deficit force quality standards for private real estate developers. If projects do
New Homes 1,757,000 32 744,000 6 not meet policy standards, they are not eligible for program funding.
(quantitative) Policy standards are quite specific in terms of architectural features –
Improved Homes 3,674,000 68 12,700,000 94
(qualitative)
they determine minimum area per housing unit, basic service require-
Total units Homes 986,781 100 248,363 100 ments, and the quality of building materials. The less costly archi-
subsidized tectural housing typology that meet these standards is a residential
New Homes 942,283 95 136,212 55 building with up to four stories, which do not require elevators. In that
Improved Homes 44,498 5 90,905 37
sense, even if building codes do not impose any location, they have an
Total subsidies US$ 6.628 billion 100 851 million 100
New US$ 5.7 billion 86 587 million 69 intrinsic incentive to locate where larger lots are available. All the more
Improved US$ 928 million 14 155 million 18 when there is a cap to how much a household would be able to pay for a
Average subsidy US$ 6717 3426 housing unit that benefits from a social housing subsidy. Unlike the case
per unit of market built housing in which developers can transfer extra costs to
New US$ 6049 4309
Improved US$ 20855* 1705
the price of the unit, in the affordable housing market that option is
Source and year of housing Ministry of Cities 2012 CONAVI 2010 severely restricted.
deficit data (most recent Housing programs are somewhat vague in terms of location stan-
official censuses dards for new housing units. While national laws contain language in
available)
support of appropriate locations for social housing, this does not
Source and year of housing Federal budget/MCMV, CONAVI, 2015;
deficit data (government 2013; Fundo de Garantia INFONAVIT, 2015; translate into clear enforcement mechanisms. For example, the
estimates) do Tempo de Serviço, Ministerio de Hacienda, Brazilian National Housing Plan for 2009–2023 (Plano Nacional de
2013 2015 Habitação) emphasizes that the housing issue is fundamentally an
Exchange rate used based on 2 Brazilian reals = 1 US 13.5 Mexican Pesos = 1US urban one. It states that “the extra cost of well-located land [for
average value at time of $ $ (2014)
building affordable housing units] is overly offset by its benefits”
the program (local (2013)
currency/U.S. dollars) (Magalhaes, 2013). Yet, decisions in terms of location are a municipal
responsibility that is separate from funding mechanisms. The Mexican
Notes National Housing Law (2010) links funding to access to services in
* Figure includes the infrastructure work to serve the home (i.e. extending sewerage and specific areas (urban polygons). To date, only a small share of housing
water services).
projects has been affected by this regulation (Paquette & Arnold, 2013).
Data on the distribution of funds imputed from previous years. Total includes subsidies
for existing units, rentals, and leases. The lack of clarity in regulating housing location has allowed real

deficit) and not move to a new one (quantitative deficit). Simply stated,
80% of the budgets go to building new houses, whereas 80% of
households need to improve their homes.
During 2007–2012 in Mexico, more than 80% of all funding from
the National Housing Commission (Comisión Nacional de Vivienda -
CONAVI) went toward acquiring housing units. Furthermore, 87% of
funding for acquisition went to new market units built by private
2
Entidades, a small offshoot of the national urban housing program, funds co-
operatives that include residents' input on housing design. Funding for this program
amounts to less than 8% of the MCMV budget.
3
The housing choices of households are further skewed when deductions are made
from paychecks through a mandatory contribution to a housing fund. These contributions
can only be used to buy a new unit. Such is the case for Mexico's formal sector employees:
about a quarter of low-income houses financed between 2006 and 2009 remain vacant
(Fuentes & Hernandez, 2014), proof that these houses did not match their owners' needs.

413
N.R. Libertun de Duren Cities 72 (2018) 411–420

Table 3
Population, size, and housing deficit in Goiania, Brazil and Puebla, Mexico.

Goiania, Goias, Brazil Puebla, Puebla, Mexico

City population 1,300,000 1,500,000

Metropolitan area population 2,200,000 2,700,000
City area (square miles) 285 206
Housing deficit 15% 48%
Qualitative housing deficit 60% 87%
Quantitative housing deficit 40% 13%
City Gini index 58 48
City Gini versus the national Gini +5 (City Gini is worse) − 3 (City Gini is better)
Sources IBGE, 2013; and UNDP et al. (2013) INEGI, 2010; CONAVI (2015); and Consejo Nacional de Evaluación de la Política de Desarrollo
(Gini) Social, 2010 (Gini).

Table 4
Social housing parameters that apply to developers.

Goiania (Brazil) Puebla (Mexico)

Eligible beneficiaries
Scale (number of housing
units)
Density (housing units/ha)
Housing unit floor space (m2)
Cost of housing unit
Required infrastructure
Required social services
Specified location within city
First-time homeowners with household incomes of up to three First-time homeowners with household incomes of one to four minimum
minimum salaries (US$$969/month) salaries working in the formal sector (US$602/month)
No minimum identified. Maximum of 500 units per module, which No minimum or maximum limits identified
can be assembled into larger projects.
None identified None identified
32 square meters for houses; 37 square meters for apartments 38 square meters
US$36,190 (US$1130/m2) US$23,767 (US$625/m2)
Water, sewerage, and electricity (developer should donate 15% of the Water, sewerage, and electricity (developer should donate 20% of land
built housing units to the municipality to cover cost) value to Puebla municipality to cover cost)
Schools, health post, day care center, security post None
None None

estate developers to make this decision. Historically, low-income
households are concentrated in deteriorated urban neighborhoods or
slums close to job markets or transportation hubs (Pearlman, 2011).
However, privately built affordable housing has most often been lo-
cated in peripheral areas, based on the decisions of developers rather
than the preferences of future beneficiaries (Libertun de Duren, 2017).
In Brazil, most MCMV Program funds go to housing units in peripheral
urban areas. Over 2010–2013, in mid-size cities where most households
with deficits are located within 15 km of the city center,4 the median
distance of new social housing complexes from the city center was
25 km (Correia Lima Neto, Krause, & Alves Furtado, 2015). This pattern
resembles the Mexican scenario, where the average distance from
downtown to the new developments now exceeds 40 km, whereas it
was less than 15 km a decade ago (SEDESOL, 2011). About half of all
units built are beyond what are known as the first “urban contention
perimeters” (perimetros de contencion urbana) (Davis et al., 2016).
It should be noted that, in theory, it is not the national government
but the local municipality that is responsible for regulating land use.
Municipalities have a say regarding housing location via their land use
maps, which often locate affordable housing on the outskirts of urban
areas. In Brazil, municipalities need to designate some areas as Special
Zones of Social Interest (Areas Especiais de Interesse Social - AEIS) in
their city plans, and these are then the only areas where social housing
can be built. The MCMV Program has required that these special zones
be fully serviced, or likely to be fully serviced soon, and that land tenure
not be contested or problematic in other ways. Neither the National
Housing Plan nor the MCMV Program have specified where such zones
should be – that has been left up to the local municipal authority to
decide. In Mexico, the situation is similar, as cities have often desig-
nated the city center as a “first zone of restricted density” which means
that social housing can only be built in peripheral areas. All in all, the
potential locations for affordable housing programs are many and ex-
pansive, albeit not in the central part of the city. As previously
4
The study is based on an empirical analysis of Belem, Fortaleza, Belo Horizonte, Porto
Alegre, and Goiania.
mentioned, this is due in part to the fact that social housing policy has
had no specific mandate to determine the location of social housing
within cities, so private developers have made the final decision as to
where to locate such projects.
2. Who develops what and why
2.1. Cities profile
Since program regulations have left decisions on where to locate
social housing up to private developers, we turn to them as our subject
of inquiry. To understand how developers think about location deci-
sions, we interviewed developers of affordable housing complexes in
central and peripheral locations in two mid-size cities: Goiania in
Brazil5 and Puebla in Mexico.6 Goiania, Brazil and Puebla, Mexico are
capitals of their respective states and each has between 1.3 million and
1.5 million residents. Both are relatively affluent by their respective
national standards, even within the Latin American context, these cities
seem to have a high level of socio-spatial segregation. In recent years,
there has been an increase in the numbers of affordable housing units in
both cities built by private developers under national housing program
regulations.
Goiania is home to about 1.3 million residents and is the center of a
metropolitan area of about 2.3 million persons. The city is the capital of
a wealthy rural area with average per capita income that is higher than
the Brazilian average (Instituto Brasileiro de Geografía y Estadística
5
In 2010, Goiania Master Plan added 12 more AEIS, bringing the total to 74. These
cover an area of 56.8 km2, or 13% of the current built-up area of the city. If fully occupied
at the current urban population density, this area would allow for 167,000 more people to
live in social housing projects in Goiania, enough space to accommodate expected po-
pulation growth for the next decade. However, most of Goiania's AEIS locate more than
10 km from the city center and there is none allowed within a five-kilometer radius of the
city center.
6
In Puebla, large social housing projects can only be developed beyond the city plan's
“first zone of restricted densities,” which extends for a radius of some 6 to 10 km beyond
the city center.

414
N.R. Libertun de Duren Cities 72 (2018) 411–420

Table 5
Questionnaire structure and variables.

Theme Variable Type of Response Unit/parameter

Firm Size Open ended number of employees

Experience Open ended number of years since funded
Practice Social housing target Open ended Type of household by income
Social housing requirements Open ended Building requirements
Location of prior projects closed Center/periphery/other
Factors determining location closed Rules/costs/infrastructure/risks/others…
Smallest project built Open ended Number of housing units
Largest project built Open ended Number of housing units
Smallest unit built Open ended Square meters
Largest unit built Open ended Square meters
Lowest building cost Open ended As percentage of total project cost
Highest building cost Open ended As percentage of total project cost
Lowest land cost Open ended As percentage of total project cost
Highest land cost Open ended As percentage of total project cost
Lowest infrastructure cost Open ended As percentage of total project cost
Highest infrastructure cost Open ended As percentage of total project cost
Project Completion date Open ended Year
Site area Open ended Square meters
Building total floor area Open ended Square meters
Distance from city center Open ended kilometers
Project size Open ended Number of housing units
Unit size Open ended Average Square meters per housing unit
Land cost Open ended As percentage of total project cost
Infrastructure cost Open ended As percentage of total project cost
Site preparation cost Open ended As percentage of total project cost
Foundations cost Open ended As percentage of total project cost
Building cost Open ended As percentage of total project cost
Finishing cost Open ended As percentage of total project cost
Community equipment cost Open ended As percentage of total project cost
Other costs Open ended As percentage of total project cost

–IBGE, 2008). Unfortunately, it also boasts one of the highest levels of
income inequality in Brazil. Housing deficits affects about 15% of
households. One of the particularities of Goiania is that it is a planned
city, with an initial master plan dating from 1938. Actual urban growth
has surpassed what was originally planned, resulting in a city marked
by affluent neighborhoods in the north and poor and informal settle-
ments in the south.
Puebla ranks fourth in terms of population in Mexico. Its me-
tropolitan area accounts for 2.7 million people, of whom 1.5 million
live within Puebla's municipal boundaries (Instituto Nacional de
Estadística y Geografía -INEGI 2010). Both the national and local
governments have been very proactive in supporting development of
auto industry, which accounts for about 70% of all industries of the
region (Bernal-Mendoza, Ramirez-Juarez, Perez-Aviles, & Morett-
Sanchez, 2010). Geography has played a major role in determining the
expansion of Puebla. The Malinche Volcano is to the north, the Iztac-
cíhuatl Volcano to the west, and Lake Valsequillo to the south, each of
them situated within nature reserves. Most informal settlements are in
these areas, which by 2000 accommodated almost half of Puebla's en-
tire population. (Valverde, 2005).
Table 3 shows some specific data on each of the cities. Goiania has a
relatively low housing deficit for Brazilian standards, but highly con-
centrated in poor households. Conversely, Puebla has a very high level
of housing deficit which a higher incidence in the low-income house-
holds. Table 4 shows that even if there are some disparities at the na-
tional and at the urban level in terms of the characteristics of the
housing deficits, housing policy impose comparable regulations on
developers producing social housing. That is, similar parameters apply
to their buildings and regulates the market demand they supply. For
example, no checks on housing density, housing location, and generous
or inexistent limits on maximum number of units per building. These
similarities on market regulations explain much of the similarities in
term of outcomes.
2.2. Methodology
The paper relies on a qualitative methodology based on semi
structured questionnaires and in depth open ended interviews with
developers of subsidized affordable housing in Goiania, Brazil and in
Puebla, Mexico. In total, ten developers representing four companies
participated. Half of the interviewees represented companies that de-
velop social housing within the urban core and half of them companies
that build in peripheral areas. Developers were first asked to complete
12 questions, and then to sit for an hour-long, open-ended interview.
Initially, we selected developers randomly among a list of those who
were active and have completed a residential building targeted to the
affordable-housing market within the last five years. However, ob-
taining responses from developers was particularly difficult as many
were reluctant to share specific answers on their business practices or
numbers In the end, we end up moving to another developer in our list
until we completed the total number of interviewees. All the developers
who answered are from well-established and quite successful formal
enterprises, which may not represent the majority of the developers in
their region but it does represent the majority of those active in the
social housing market built in peripheral urban areas. All The devel-
opers interviewed were operating within the framework of current af-
fordable housing policies (Table 6), and had built at least five afford-
able housing complexes in total and received funding and completed
construction under the programs analyzed: Brazil's MCMV Program
(2010 −2013) and the Mexican National Development Plan
(2007–2012).
The principal aim of both the questions and the interviews was to
understand how developers make decisions as to where to build af-
fordable housing. The surveys also provide specific information on the
developers' profile, including size and prior experience; the typical
characteristics of the affordable housing units they produce; and the
overall cost structure of the housing investment, with a focus on the
incidence of land cost vis-à-vis the total cost of the investment. The
questionnaires included multiple-choice questions on firm size (small,

415
N.R. Libertun de Duren
Table 6
Social housing profile based on survey findings, Brazil and Mexico.
Brazil Mexico
Periphery
Housing cost (lowest) US$490/m2 US$452/m2
Infrastructure cost 20%–30% of total 12%–15% of total cost
parameters cost
Land cost parameters 10%–15% of total 5%–7% of total cost
cost
Central
Housing cost (lowest) US$435/m2 US$500/m2
Infrastructure cost 5%–11% of total cost Land plus infrastructure:
parameters 18%–28% of total cost
Land cost parameters 20%–20% of total
cost
medium, or large), and the top reasons for choosing project locations
(urban planning rules, environmental rules, building costs, infra-
structure costs, land costs, available infrastructure, natural disaster risk,
and others) (see Table 5 and Appendix 1). In terms of assessing costs,
the questionnaires asked developers to allocate percentages of total
expenditures between construction (including contracting labor), land
and infrastructure, paperwork, and other costs. This format allowed for
comparisons and facilitated dialogue, as many developers were re-
luctant to share actual figures. The follow-up dialogue touched on these
same topics but in an open-ended fashion, thus allowing for further
explanation of the rationale behind the answers.
3. Main findings
The questionnaires and the interviews revealed interesting data in
terms of the profile of developers and of social housing projects and
their cost structure.
3.1. Profile of developers
In all cases, the developers surveyed were also developing other
properties for nonsubsidized households. Those that build properties in
urban peripheries are larger, more capitalized, and more experienced
companies that had been responsible for the construction of large-scale
buildings in the past. Those that operate in central city areas tend to be
local, smaller, and less experienced. Whether the difference in scale
between providers of peripheral social housing projects and providers
of central ones is part of a more general phenomenon cannot be con-
cluded from our data. Evidence against this notion came from the de-
velopers themselves, all of whom stated their willingness to work in
both areas of a city.
3.2. Profile of social housing and costs structure
Table 5 presents a summary the sizes and costs of social housing in
peripheral and central areas, as described in the sections that follow.
3.2.1. Typical characteristics of units
Developed units range between 40 and 46 square meters, which
meets but barely exceeds the minimum requirements of national pro-
grams, in all locations. Most of the social housing provided is multi-
family apartment blocks, often with four floors, which is the accepted
height limit of walk-up buildings without elevators. Amenities offered
within the units are also similar in both central and peripheral areas, as
access to funding requires compliance with these standards as well.
3.2.2. Building costs
Most developers reported that they could provide social housing
units at less than US$500 per square meter, without including taxes and
profits. Contrary to expectations, they provided no evidence of higher
416
Cities 72 (2018) 411–420
construction costs for housing in central areas than in the periphery.
Indeed, the developers explained that their building costs were similar
for both locations. The main cost items for construction, labor, and
materials (often brought from a great distance) did not vary sig-
nificantly between the two locations, which are normally not more than
10 km apart.
3.2.3. Land costs
In peripheral areas, where land is expected to be cheap, land costs
ranged between 5 and 7% of total costs in Mexico and between 10 and
15% in Brazil. In both cases, land costs in central areas could reach 20%
of total costs. However, accounting for land value is complicated,
particularly in cases in which developers had acquired land in years
prior to the actual start of the project. Reported land costs can also
diverge from market values for other reasons. In both Brazil and
Mexico, developers told the study team that during the planning stages
of a social housing project they negotiated land prices as well as land
transfers and swaps with local authorities. Similar negotiations took
place over the provision of infrastructure, its standards, and who would
have to pay for it. The results were often cost-sharing and arrangements
and contributions that obfuscate the actual price and value of the land
involved, not for lack of transparency but due to the complexity of the
negotiations.
3.2.4. Costs of infrastructure and services
These costs include the provision of mandatory infrastructure for
the units, including water, sewerage, electricity, and gas connections,
and may also include paving of streets or sidewalks. At peripheral sites,
these services often require investments in offsite areas, so there is a
high initial cost in terms of onsite infrastructure, as most basic services
are lacking. In Mexico, developers are responsible for funding and
building the extension of basic services (water, sewerage, and elec-
tricity) to their developments. Once the works are completed, muni-
cipal authorities are responsible for their operation. Similarly, devel-
opers are responsible for the construction of sidewalks and urban
services when they are lacking. In Brazil, developers donate 15% of the
built housing units to the municipality as compensation for the ex-
penses the municipality will incur when extending basic services to the
complex. The distance to existing urban infrastructure networks weighs
heavily in determining these costs. In general, in Mexico the infra-
structure costs of peripheral projects ranged between 18 and 28% of
total costs, a bit higher than in Brazil, where they rarely surpassed 18%.
Offsite investments are also required for central areas. Such is the case,
for example, if the project generates demand for urban services (e.g.,
water) that requires increasing system capacity.
In both countries, Internet and phone services are not mandatory
and are negotiated directly with private providers, which are typically
given exclusive rights over provision of these services in exchange for
paying the cost of installing the infrastructure. Health and education
services are a state rather than a municipal responsibility, and hence
they are not required of developers. These services are often of poorer
quality in peripheral developments and harder to access from those
locations. However, they are not part of the conversation between
government officials and developers.
3.2.5. Rationale for project location
Developers did not cite any preference for the location of their
projects other than one that “makes more sense, as it allows for com-
pleting the project on time and within the budget.” When asked for
clarification, they mentioned three variables: the time required by the
approval process, the cost of land and infrastructure, and the number of
units they can accommodate on the site. To be sure, their response does
not preclude the existence of other variables at play, such as the poli-
tical alliances or personal relationships that may weight on developers'
site selection. Yet, these three variables offer an explanation on how
developers rationalize their own locational choices within the
N.R. Libertun de Duren
constraints of existing policy and planning frameworks.
Time is often a measure of a developer's ability to have a dialogue
with the local authorities. As one developer told us, the location deci-
sion is a complex one involving many departments of the firm.
Decision-making then involves repeated negotiations with local au-
thorities, considering local urban planning and building regulations.
Municipalities that are perceived as easier to negotiate with, either
because they have “weaker enforcement” mechanisms or are friendlier
with developers, are likely to be preferred. It was difficult to get precise
information on how the relationship between developers and municipal
officer works; and to which extent weaker enforcement is the outcome
of a municipality with low institutional capabilities or of the familiarity
between municipal officers and developers. The politics of permitting
process tend to be obscure at municipal level, and developers are re-
luctant to discuss them. However, developers did acknowledge that the
less demand there is for a plot of land, the easier it is to discuss potential
uses for the land with municipal officers. In particular, if the munici-
pality has large percentage of its area undeveloped.
Incentives for building in peripheral areas also involve the spatial
configuration of those areas. These lots tend to be far away from
highway access and lack adequate services and infrastructure, which
makes them unappealing to investors who are targeting more affluent
households, as in the case of developers building gated communities
(Libertun de Duren, 2011). Hence, low-income housing developers face
little resistance when vying for these lots. Further, since local land use
regulations are the exclusive domain of municipal governments, those
governments can exert considerable influence on how those regulations
are implemented. Unlike urban areas where developers are rarely the
single or most affluent party interested in an area, developers have the
upper hand when negotiating with peripheral municipalities, which
translates into easier and faster approval processes for their projects.
Developers calculate the cost of land and infrastructure as a single
variable by establishing a “land-plus-infrastructure” cost parameter that
should not exceed a third of total housing costs (Fig. 1). Within that
parameter, the trade-off between land and infrastructure costs allows
for many different project locations to meet the standard. This ex-
pectation is the same for developers in central and peripheral locations.
So even if, as expected, land values differ drastically between the center
and the periphery, the lower cost of peripheral locations is offset by the
higher cost of the onsite infrastructure required. Land for central de-
velopments accounts on average for 17% of the investment, while in-
frastructure expenses average below 15%. Conversely, land in periph-
eral developments accounts on average for about 6% of the investment,
while infrastructure expenses average some 30%. Overall, however, the
data collected do not point clearly to any significant difference between
peripheral and central developments with respect to these cost para-
meters: 36% of the total project cost goes for land and infrastructure in
peripheral areas and 32% in central areas. A more important finding is
how developers use the inverse relationship between land and infra-
structure costs as social housing project locations move between the
417
Cities 72 (2018) 411–420
center and the periphery of cities. At the center, land costs are high and
infrastructure costs are low. Conversely, at the periphery, land costs fall
while infrastructure costs rise.
The number of housing units builders can accommodate is directly
related to the ability to reach economies of scale, and hence increase
the profitability of projects. Developers can save substantial costs on
materials and labor once they get to housing developments larger than
500 units. In Mexico, where it is not uncommon to find housing com-
plexes that accommodate more than 2500 units, builders can afford
onsite production of some basic construction components like bricks
and blocks, wages and contractor payments, and rental equipment, and
can even help amortize legal fees, all of which reduces the cost per
housing unit built. In addition, some of the initial costs, such as legal
and administrative expenditures, are not proportional to the size of the
development.
However, accommodating these huge complexes requires large
tracks of empty land. These are rarely available in central areas, but
they abound in the far peripheries. Moreover, if developers bring large
projects to municipalities, they tend to have more leverage in nego-
tiations. Having said that, it should be noted that developers make a
profit from social housing projects no matter where they locate them. In
Puebla, for example, subtracting the reported total production costs
(building, public spaces, land, and infrastructure) from the selling price
yields a gross profit margin of 42% on peripheral developments and
32% on central developments. In peripheral developments in Goiania,
the equivalent figure was 20%.
In sum, developers choose the periphery because it allows them
higher profits. But it is worth emphasizing that is not the land cost per
se that drives developers to the periphery, as this is offset by higher
infrastructure costs. Rather, it is the per unit profit they can achieve in
the periphery, as this allows for accommodating larger complexes and
reduces approval times. Because building larger projects requires larger
flows of capital, larger developers tend to be the only ones developing
in peripheral areas. If land values were the main factor, we would not
see a concentration of smaller developers in central areas, where land is
more expensive. Nor would we see larger developers gravitating to the
periphery when the land-plus-infrastructure cost parameter should
make them indifferent to a project's location. Economies of scale and
weak regulations make the financing of social housing in peripheral
areas a more profitable investment for those who can afford the initial
investments.
4. Conclusions
One of the aims of this paper has been to elucidate the mechanisms
that link national affordable housing programs with the location of new
housing units in peripheral urban areas. We hypothesize that because
programs leave location decisions to private developers, those decisions
respond to developers' business needs as benefited households are a
captive market. This hypothesis is in line with William Alonso's view of
Fig. 1. Land and Infrastructure Costs as Percent Shares of
Total Housing Costs.
N.R. Libertun de Duren
urban residential geography as determined by household tradeoffs be-
tween housing amenities and housing location; albeit with the con-
sideration that it is developers and not households who make the lo-
cational choices. We then hypothesize that developers reap more profits
from building affordable housing in the periphery than in the centers of
cities because land is cheaper in the periphery, which would follow a
straightforward understanding of how market economy determines
housing location.
While our first hypothesis proved to be true, the second did not.
Developers do profit more in the periphery but not due to its lower land
prices. Initial savings that accrue to developers due to lower land prices
in the periphery are offset by the cost of having to build basic onsite
infrastructure in those areas as mandated by law. Eventually, the sum of
the cost of land plus infrastructure is about a third of the total invest-
ments both in central and peripheral locations. Thus, economies of scale
– and not land prices – explain developers' preference for building in
peripheral areas. Large lots – which are available almost exclusively in
urban peripheries – enable developers to achieve significant cost sav-
ings because these large lots make it possible for developers to build
more than 500 units. In addition, weaker municipal regulations and
fewer bidders, both of which are typical for projects in difficult-to-ac-
cess peripheral locations, make for a shorter and easier approval pro-
cess for these large housing projects. This finding is important because
it shows that at each point, institutional constrains determine how
markets work, and one sided analyses would lead to incorrect conclu-
sions on policy actions.
Developers have simply acted within the constraints and opportu-
nities afforded by the housing programs that have been established.
National programs have failed to adequately regulate housing produc-
tion because they measured the success of housing policies by the
number of units built, and let developers choose where to locate these
complexes. National governments went even further by actively trans-
forming housing into a product suitable for mass-scale production,
which by definition required large lots that are only available in per-
ipheral areas. Then again, the direction of causality is a bit more
Appendix A. Appendix 1. Interview guide for developers
Details about interviewee:
Name of firm:
___________ (name)
Firm Size ___________ (number of employees)
Name and ___________ (name)
function of _________ (function)
inter-
viewee:
Details about the project we are researching:
Name of ___________ (name)
housing
project:
Year ___________ (year)
completed:
Total area of _____________ (m2)
site:
Total floor _____________ (m2)
space built:
Distance from ____________ (kms)
city center:
Questions
1 In your firm, who do you understand to be the main buyers of social housing?
2. What kind of infrastructure and community facilities does social housing have?
418
Cities 72 (2018) 411–420
nuanced, as developers are not insulated from the political process that
goes into establishing building standards. But the salient point is that
there is political economy rationale that links land prices, architectural
typologies, and municipal practices into an urban form that created an
incentive for the private sector to build massive affordable housing
buildings in peripheral areas.
In addition, national housing programs in Brazil and Mexico have
disproportionately funded new housing instead of housing upgrades,
leading to an increase in the supply of housing for low-income house-
holds in the periphery at the expense of improving the conditions of
inadequate urban dwellings. Programs that promote mortgages and
subsidies for building new housing units received more than five times
the funding than programs that supply services and titles to existing
inadequate housing – the two main reasons why urban poor houses are
considered inadequate. Only 20% of all households with deficits need a
new housing unit, yet 80% of all housing funds go toward subsidizing
the production of new houses.
From a practical standpoint, some measures are starting to be put in
place to reduce the development of huge low-income housing projects
in difficult-to-access peripheral areas. Although not clearly successful as
yet, Mexico's aim to link funding to certain urban areas is a good start.
Brazil has aligned incentives among different constituencies, such as
transportation providers and real estate developers. Even with these
improvements, however, serious concerns remain. Funding structures
should be simplified as much as possible in order to increase trans-
parency and reduce administrative costs. It is worrisome that national
governments have been more effective in modeling financing structures
for potential developers than for households. Beyond these, some les-
sons transcend the Brazilian and the Mexican case: Matching funding
amounts for programs with actual demand would go a long way toward
allowing residents to stay in the city instead of being removed from it.
Policies should never force low-income households to decide between
having access to either housing or urban opportunities. Good policies
should always give voice to the people whom they claim to serve.
City: ___________ (name)
Date of
interview: ________ (dd/mm/yy)
Experience ___________ (number of
year since funded)
______ ____________ (email)
(telephone)
______ (cell)
Number of ______ (total number)
apartments: ______ (m2 per unit)
Number of _______ total number
single family _______ m2 per unit
homes:
Number of _______ total number
others _______ m2 per unit
(specify):
N.R. Libertun de Duren Cities 72 (2018) 411–420

3 Where are social housing projects located within cities? (In cities of 1–2 million population)
☐ less than 5 km from center
☐ more than 10 km from center
☐ other (specify) __________
4. In your experience, what are the three most important factors that determine where ☐ urban planning rules
social housing projects are located—near or far from city centers? ☐ environmental rules
☐ building costs
☐ infrastructure costs
☐ land costs
☐ available infrastructure
☐ natural disaster risk
☐ other (specify) ______________________
5. Among the social housing projects you have built in the past five Your smallest
years, please describe the following: project: ________ (number of units)
Your largest
project: ________ (number of units)
Your smallest
housing unit: ___________ (m2 of unit)
Your largest
housing unit: ___________ (m2 of unit)
Your building
cost - lowest: _______ (as percent of unit cost)
Your building
cost - highest: _______ (as percent of unit cost)
Your land cost –
lowest: _______ (as percent of unit cost)
Your land cost –
highest: _______ (as percent of unit cost)
Your
infrastructure _______ (as percent of unit cost)
cost – lowest:
Your
infrastructure _______ (as percent of unit cost)
cost – highest:
6. For this _______________ (name) social housing project that we are Land: _______ (as percent of unit cost)
currently researching, can you please provide information about your Infrastructure _______ (as percent of unit cost)
total investment costs in dollars in each of the following? (specify):
Site _______ (as percent of unit cost)
preparation:
Foundations: _______ (as percent of unit cost)
Building: _______ (as percent of unit cost)
Finishing: _______ (as percent of unit cost)
Community _______ (as percent of unit cost)
equipment:
Other (specify): _______ (as percent of unit cost)

References Graduate School of Design.

Aguilar, A. G., Ward, P. M., & Smith, C. B., Sr. (2003). Globalization, regional develop-
ment, and mega-city expansion in Latin America: Analyzing Mexico City's peri-urban
hinterland. Cities, 20(1), 3–21.
Alonso, W. (1964). Location and land use: Toward a general theory of land rent. Cambridge:
Harvard University Press.
Bernal-Mendoza, H., Ramirez-Juarez, J., Perez-Aviles, R., & Morett-Sanchez, J. (2010).
Importancia de los territorios rurales en el proceso de reestructuración territorial: el caso de
la región metropolitana de la ciudad de Puebla.
Bouillon, C. P. (2012). Room for development: Housing markets in Latin America and the
Caribbean. Palgrave Macmillan.
Caixa. (2015). Um pouco da história da poupança na Caixa Econômica Federal. Retrieved
from http://www.caixa.gov.br/Downloads/poupanca-historia/HISTPOUP.pdf.
Correia Lima Neto, V., Krause, C., & Alves Furtado, B. (2015). O déficit habitacional intra-
metropolitano e a localização de empreendimentos do programa minha casa minha vida:
mensurando possibilidades de atendimento. Brasília: Instituto de Pesquisa Econômica
Aplicada.
Davis, D., Escobedo, N., Granados, F., Lara, F., Schoen, D., & Scott, M. (2016). Building
better cities with strategic investments in social housing. Case study compendium:
Understanding the barriers and enablers to densification at the metropolitan level.
Qualitative evidence from seven Mexican cities. Cambridge, MA: Harvard University
De Mello, L. (2003). Fiscal decentralization and subnational expenditure policy.
International Monetary Fund. (Unpublished) manuscript.
Fuentes, C. M., & Hernandez, V. (2014). Housing finance reform in Mexico: The impact of
housing vacancy on property crime. International Journal of Housing Policy, 14(4),
368–388.
Hardoy, J. E., & Satterthwaite, D. (2014). Squatter citizen: Life in the urban third world.
Routledge.
Harvey, D. (1985). The urban process under capitalism: A framework for analysis. The
urbanization of capital (pp. 1–31). Baltimore: Johns Hopkins University Press.
Hebert, C., Belsky, E., & DuBroff, N. (2012). The state of Mexico's housing: Recent progress
and continued challenges (working paper 12–8). Cambridge, MA: Joint Center for
Housing Studies, Harvard University.
Instituto Brasileiro de Geografia e Estatıstica (IBGE) (2008). Regiones de influencia das
cidades 2007. Rio de Janeiro: IBGE.
Libertun de Duren, N. L. (2011). The national embeddedness of urbanization trajectories.
City & Community, 10(4), 424–430.
Libertun de Duren, N. R. (2017). The social housing burden: Comparing households at the
periphery and the centre of cities in Brazil, Colombia, and Mexico. International
Journal of Housing Policy, 1–27.
Magalhaes, I. (2013). Desafíos actuales y nuevas herramientas para enfrentar el déficit
cuantitativo y cualitativo. Cumbre América Latina y China de políticas y conocimiento.
Retrieved from http://www.iadb.org/en/topics/urban-development/latin-america-

419
N.R. Libertun de Duren
and-china-policy-and-knowledge-summit,17728.html.
Monkkonen, P. (2011). The housing transition in Mexico: Expanding access to housing
finance. Urban Affairs Review, 47(5), 672–695.
Murray, C., & Clapham, D. (2015). Housing policies in Latin America: Overview of the
four largest economies. International Journal of Housing Policy, 15(3), 347–364.
Office of Management and Budget (OMB) (2015). The budget fiscal year 2015:
Department of Housing and Urban Development. Retrieved from https://www.
whitehouse.gov/sites/default/files/omb/budget/fy2015/assets/housing.pdf.
Organisation for Economic Co-operation and Development (OECD) (2012). Economic
surveys: Chile.
Paquette, C., & Arnold, P. (2013). Las consecuencias de la producción masiva de vivienda
social. Universidad Autonoma Metropolitana. Retrieved from http://hdl.handle.net/
11191/1211.
Pearlman, J. (2011). Favela: Four decades of living on the edge in Rio de Janeiro. Oxford, UK:
Oxford University Press.
Rolnik, R. (2013). Late neoliberalism: The financialization of homeownership and
housing rights. International Journal of Urban and Regional Research, 37(3),
420
Cities 72 (2018) 411–420
1058–1066.
Secretaría de Desarrollo Agrario, Territorial y Urbano (SEDATU) (2013). Programas na-
cionales de desarrollo urbano y vivienda 2013–2018. Ciudad de México, DF: Comisión
Nacional de Vivienda.
Secretaría de Desarrollo Social (SEDESOL) (2011). La expansión de las ciudades mexicanas
1980–2010.
Sociedad Hipotecaria Federal (SHF) (2014). Estado actual de la vivienda en Mexico. Ciudad
de Mexico: Centro de investigación y documentación de la CASA AC. Fundacion
CIDOC.
United Nations Department for Economic and Social Affairs (UN-DESA) (2013). World
urbanization prospects. New York: UN-DESA.
United Nations Development Programme (UNDP) (2013). Institute of applied economic
research, & João Pinheiro Foundation. Atlas of human development in Brazil (pp.
2013). . Retrieved from http://www.atlasbrasil.org.br/2013/en/o_atlas/o_atlas_/.
Valverde, F. (2005). El plan del ayuntamiento para el desarrollo de la Ciudad de Puebla.
México: Universidad Iberoamericana de Puebla.