Engineering Methods for Planning Affordable Housing and Sustainable Communities
MICHAEL P. JOHNSON
Carnegie Mellon University
Housing is a key component of the U.S. economy. In 2001, housing comprised more than one-third of the nation’s tangible assets, and, in the form of home building and remodeling, housing consumption and related spending represented more than 21 percent of the U.S. gross domestic product. Since 2001, home sales, prices, equity, and debt have all increased substantially, enabling millions of Americans to purchase goods and services (Joint Center for Housing Studies of Harvard University, 2006).
Decent, affordable housing (generally defined as housing that consumes less than 30 percent of a family’s income) often enables families to enjoy stability, good health, employment, education, and recreation. Decent, affordable housing also contributes to the physical, economic, environmental, and social health—the sustainability—of communities (Millennial Housing Commission, 2002). These impacts are especially important for lower income households and other underserved populations.
Despite the general strength of the U.S. housing market, the benefits of housing and stable, vibrant communities are not distributed equally. Examples of inequalities include: residential segregation, differences in homeownership rates by race, sprawl-type development patterns, and shortages of affordable housing. In the wake of Hurricane Katrina, for example, the challenges of securing basic shelter and rebuilding homes and communities have fallen disproportionately on minority and low-income populations (de Souza Briggs, 2006; Joint Center for Housing Studies of Harvard University, 2006; Millennial Housing Commission,
2002). These and similar circumstances justify social intervention by government and nongovernmental organizations.
The purpose of this paper is to highlight new, creative research in a variety of disciplines—especially decision sciences—that can help determine when, where, what type, and by what means affordable housing and sustainable communities might be built, redeveloped, and maintained. As a prelude to the subject, it is useful to link housing planning and supply chain management, the theme of this Frontiers of Engineering session.
A supply chain is a network of facilities and modes of transportation that uses production and logistics processes to transform inputs into finished goods and services, thereby integrating supply and demand management. A central feature of supply chain management is temporal planning—strategic, tactical, operational, and technical (e.g., the location of facilities at which operations are performed). Housing and community development (a social enterprise) are not literally examples of supply chain management. However, facility location— here, the location of housing—is central to both, and the temporal scope of housing and community development planning spans strategic, tactical, and operational time horizons. Finally, effective housing and community development planning, like supply chain management, is an attempt to match supply and demand for goods and services—in this case, affordable shelter and sustainable communities.
Initiatives to make affordable housing and sustainable communities more accessible must address the needs of stakeholders (e.g., employers, housing developers, citizens, government agencies); policy objectives (minimize housing costs and environmental impacts, “deconcentrating” poverty); and actions (the creation of new housing alternatives, protection of current alternatives, changes in attitudes and preferences) (cf. de Souza Briggs, 2005).
Engineering and related disciplines can influence all of these dimensions of housing policy. Civil, environmental, and mechanical engineering, for example, can generate methods of implementing housing initiatives with more efficient and effective construction. Urban and regional planning, especially land-use and transportation planning, in contrast, focus on social efficiency and equitable development outcomes, given current or best-practice construction technologies. Decision sciences (e.g., operations research and management science) represent a link between engineering and planning methods; they generate specific, actionable strategies for optimizing social efficiency, effectiveness, and equity. Decision sciences may take as given current or best practices in construction technologies or planning methods, or both, or neither.
The remainder of this paper is focused on research results in engineering construction methods and urban and regional planning methods related to the development of affordable housing and a discussion of the unique contributions of decision sciences. We also identify a number of promising areas for continued research.
ENGINEERING-BASED METHODS FOR HOUSING CONSTRUCTION
Traditional engineering is well suited to the efficient development of cost-effective housing. Improvements in construction technologies can result in increased affordability, energy efficiency, and structural integrity and decreased negative environmental impacts. Recent European research addressing “sustainable” development from an engineering perspective, focused mostly on minimizing negative environmental impacts, has shown that, even when construction techniques are modified to decrease the ecological impacts associated with “flows” of energy, construction materials, and water, the resulting innovations are often contradicted by increased resource usage by housing occupants and ineffective national policies (e.g., Priemus, 2005). Ultimately, Priemus argues, the policy with the greatest impact on sustainability may be a policy that discourages, or even decreases, the construction of new housing.
Other engineering approaches have focused on best practices for reducing energy consumption through energy-conserving materials, such as windows, insulation, and appliances; alternative energy sources, such as solar power; improved construction methods for foundations and walls; and more efficient heating and air-conditioning systems (Steven Winter Associates Inc., 2001). Building-design strategies are based on advanced computer simulations comparing energy savings from novel designs with actual outcomes, as well as architectural choices, such as site selection and building orientation for maximum passive solar exposure, and compact floor plans. A specially designed house that incorporated these technologies used 46 percent less energy than the average U.S. house (Balcomb et al., 1999).
These technologies are also available for the rehabilitation of existing housing in low-income areas through retrofitting, improved gas metering, and increased cooperation between stakeholders. Estimated cost savings in energy for a low-income family are on the order of one month’s rent per year (Katrakis et al., 1994).
Engineering methods also influence construction processes. Examples include concurrent engineering to help meet customer requirements for industrialized housing (Armacost et al., 1994) and knowledge management to improve coordination between the owners, designers, and developers of affordable housing (Ibrahim and Nissen, 2003).
URBAN PLANNING FOR AFFORDABLE HOUSING AND COMMUNITY DEVELOPMENT
American planners and analysts have been dealing, with limited success, with the problems of affordable housing and community design for more than 80 years (von Hoffman, 1996). In central cities, planners in the 1930s and 1940s embraced the idea of vertical towers grouped in communities distinct from sur-
rounding neighborhoods. These enclaves often resulted in social dysfunction and physical decay, which have only been remedied in a substantial way in the past decade under the Federal HOPE VI Program. In contrast, post-World War II suburbs were designed to be affordable, accessible to central cities via freeways, and uniform in appearance.
In recent years, dense, transit-friendly, mixed-use developments in central cities or nearby suburbs, often on land previously used for residential or industrial purposes, have converged with the redevelopment of distressed inner-city neighborhoods into mixed-income, joint ventures (Bohl, 2000). Although U.S. consumers still overwhelmingly prefer the traditional suburban model of detached, single-family, owner-occupied housing, market demand is increasing for housing units and communities that appear to be more sustainable socially and environmentally (Myers and Gearin, 2001).
The impact of assisted housing development has been limited in recent years because of stagnant federal funding for subsidized and affordable housing. Planning researchers are turning to decision models and geographic information systems to generate alternative strategies for optimizing social objectives (Ayeni, 1997). However, very little work in this area, or in traditional urban planning, is being done on decision-support models designed specifically for planning affordable housing.
DECISION-SCIENCE METHODS FOR AFFORDABLE HOUSING POLICY AND PLANNING
Decision models can help planners improve access to affordable housing and sustainable communities by simultaneously, and explicitly, addressing space, opportunity, design, and choice alternatives. Space and opportunity are factors in decisions about the physical location of housing units and their proximity to community amenities, which are important to improved quality of life. Design decisions are important to the development of policies that enable families to participate in housing programs, as well as in establishing development priorities and configuring mixed land-use and mixed-housing communities. Choice decisions are essential to individuals choosing housing and neighborhood destinations that best meet their needs and preferences. In contrast to engineering construction and planning methods, decision models for housing development are quantitative, stylized, prescriptive, forward-looking, and multiobjective.
One type of strategic decision we address is choosing and evaluating housing and community development policies. A solution to this problem consists of program types (e.g., housing subsidies) and intensities (e.g., funding levels or number of program participants). Caulkins et al. (2005) developed a model to predict long-term population outcomes associated with stylized, large-scale programs in which low-income families use housing subsidies to relocate to low-poverty neighborhoods. The purpose of the model is to identify the circum-
stances under which a large-scale housing program might preserve the health of destination communities. The authors model changes in the stock of middle-class families in a typical region as a result of (1) normal demographic changes, (2) a large-scale housing mobility program resulting in low-income families that “assimilate” to the middle class, and (3) middle-class “flight” in response to in-movers. Figure 1 shows that, for base-case values of structural parameters, equilibrium would be maintained over the long term (near X = 1) in a generic metropolitan area with a low-intensity housing-mobility program; in the long term, the size of middle-class communities would decrease only slightly.
Given support, in a strategic sense, for a particular housing policy, a tactical decision must be made about the amount and type(s) of housing to be provided in a specific region over a specific period of time. Addressing this decision requires specifying program locations (municipalities, neighborhoods, or land parcels) and configurations (different numbers of different-sized rental- or owner-occupied housing units). Gabriel et al. (2006) developed a multiobjective optimization model for identifying land parcels for development that balances the needs of planners, developers, environmentalists, and government.
Johnson (2006) solves two complementary optimization models specifically for affordable housing: (1) a longer range model for identifying regional investment levels that maximize social benefits and (2) a shorter range model for identifying specific locations and development sizes that balances social benefits and equity. Figure 2 shows Pareto frontiers associated with solutions to the multiobjective optimization problem for owner-occupied and renter-occupied housing using data for Allegheny County, Pennsylvania. The curves show that a range of policy alternatives can support a “most-preferred” solution.
The last decision problem considered here, operational in scope, is a client’s choice of a most-preferred housing program, neighborhood, or housing unit, within defined, affordable, housing-policy priorities. Solving this problem requires specifying detailed characteristics (attributes) of housing units and neighborhoods, decision models by which participants can rank potential destinations (alternatives), and information systems to help standardize and automate the process (decision support systems).
Johnson (2005) developed a prototype spatial decision-support system (SDSS) for tenant-based subsidized housing that addresses qualitative concerns (which attributes of housing units and neighborhoods are important to the client) and quantitative concerns (how a client can rank a “short list” of alternatives to
maximize satisfaction and minimize the burden of the housing search). The SDSS uses geographic information systems to illustrate neighborhood characteristics, a relational database to store information on specific housing units, and a multi-criteria decision model to help clients make relocation decisions. Figure 3 illustrates the spatial-data interface with fair housing data for Allegheny County, Pennsylvania.
RESEARCH NOW AND IN THE FUTURE
A number of analytical methods can be used to make affordable housing and sustainable communities more accessible. In one stream of current research, civil, environmental, and mechanical engineering methods are being used to design housing units that improve on current practices in terms of energy efficiency, cost, structural quality, and efficiency of construction processes. In another stream of current research, urban and regional planning are being used to help stakeholders define development strategies that reflect best knowledge of social science-based program evaluation, land-use and transportation planning standards, and community-level partnerships. Decision sciences can provide opportunities to design housing- and community-development policies that improve on current practices in construction-oriented engineering and planning in terms of social outcomes, multistakeholder negotiations, and housing program client choices.
Because affordable housing and sustainable community development are not currently top priorities for market-rate housing providers, government support for the engineering of residential housing may be necessary to increase environmental sustainability and reduce user costs. However, housing policies that optimize various social criteria must also address technological aspects of housing and be based on best practices in urban and regional planning to be considered sustainable and affordable.
The decision-sciences research described in this paper suggests a number of promising areas for future research. The most important is to provide evidence that implementation of the decision models described above result in improved outcomes for communities and individuals. Other areas for research include: (1) choices of housing design and construction strategies that balance housing-unit-and community-level sustainability measures; (2) the development of dynamic models for designing strategic housing policies to address place-based housing strategies (i.e., new construction and rehabilitation of existing housing units); and (3) the design of realistic and tractable decision models to guide developers of affordable housing who must routinely choose a handful of sites to develop from many alternatives, with limited funding, to maximize the probability of neighborhood revitalization.
As long as urban sprawl, environmental degradation, and geographical barriers to affordable housing and opportunity remain policy problems, researchers have an opportunity to devise novel and creative solutions at the nexus of engineering, planning, and decision sciences.
My thanks to Jeannie Kim and Vincent Chiou for assisting in this research and to Julie Swann and Jennifer Ryan for encouraging me to participate in the 2006 Frontiers of Engineering Symposium.
Armacost, R.L., J. Paul, M.A. Mullens, and W.W. Swart. 1994. An AHP framework for prioritizing customer requirements in QFD: an industrialized housing application. IIE Transactions 26(4): 72–80.
Ayeni, B. 1997. The Design of Spatial Decision Support Systems in Urban and Regional Planning. Pp. 3–22 in Decision Support Systems in Urban Planning, edited by H. Timmermans. London: E & F N Spon.
Balcomb, J.D., C.E. Hancock, and G. Barker. 1999. Design, Construction, and Performance of the Grand Canyon House. National Renewable Energy Laboratory, U.S. Department of Energy. Available online at: http://www.nrel.gov/docsfy00osti/24767.pdf.
Bohl, C.C. 2000. New urbanism and the city: potential applications and implications for distressed inner-city neighborhoods. Housing Policy Debate 11(4): 761–801.
Caulkins, J.P., G. Feichtinger, D. Grass, M.P. Johnson, G. Tragler, and Y. Yegorov. 2005. Placing the poor while keeping the rich in their place: separating strategies for optimally managing residential mobility and assimilation. Demographic Research 13(1): 1–34. Available online at: http://www.demographic-research.org/Volumes/Vol13/1/default.htm.
de Souza Briggs, X. 2005. Politics and Policy: Changing the Geography of Opportunity. Pp. 310– 341 in The Geography of Opportunity: Race and Housing Choice in Metropolitan America, edited by X. de Souza Briggs. Washington, D.C.: Brookings Institution Press.
Gabriel, S.A., J.A. Faria, and G.E. Moglen. 2006. A multiobjective optimization approach to smart growth in land development. Socio-Economic Planning Sciences 40: 212–248.
Ibrahim, R., and M. Nissen. 2003. Emerging Technology to Model Dynamic Knowledge Creation and Flow Among Construction Industry Stakeholders During the Critical Feasibility-Entitlements Phase. In Information Technology 2003: Towards a Vision for Information Technology in Civil Engineering, edited by I. Flood. Reston, Va.: American Society of Civil Engineers. Available on CD-ROM.
Johnson, M.P. 2005. Spatial decision support for assisted housing mobility counseling. Decision Support Systems 41(1): 296–312.
Johnson, M.P. 2006. Planning Models for Affordable Housing Development. Forthcoming in Environment and Planning B: Planning and Design.
Joint Center for Housing Studies of Harvard University. 2006. The State of the Nation’s Housing 2006. Available online at: http://www.jchs.harvard.edu/publications/markets/son2006/son2006.pdf.
Katrakis, J.T., P.A. Knight, and J.D. Cavallo. 1994. Energy-Efficient Rehabilitation of Multifamily Buildings in the Midwest. Argonne National Laboratory, Decision and Information Sciences Division. Available online at: http://www.eere.energy.gov/buildings/info/documents/pdfs/multi-gu.pdf.
Millenial Housing Commission. 2002. Meeting Our Nation’s Housing Challenges: A Report of the Bipartisan Millenial Housing Commission Appointed by the Congress of the United States. Available online at: http://govinfo.library.unt.edu/mhc/MHCReport.pdf.
Myers, D., and E. Gearin. 2001. Current preferences and future demand for denser residential environments. Housing Policy Debate 12(4): 633–659.
Priemus, H. 2005. How to make housing sustainable?: the Dutch experience. Environment and Planning B 32(1): 5–19.
Steven Winter Associates Inc. 2001. Building America Field Project: Results for the Consortium for Advanced Residential Buildings (CARB), January to October 2001. National Renewable Energy Laboratory, U.S. Department of Energy. Available online at: http://www.nrel.gov/docs/fy03osti/31380.pdf.
von Hoffman, A. 1996. High ambitions: the past and future of American low-income housing policy. Housing Policy Debate 7(3): 423–446.