Fiscal Impact Analysis and the Costs of Alternative Development Patterns
Abstract and Keywords
The growing recognition of the relationship between decisions about growth (how and where it occurs) and public sector budgets has contributed to the attention given in academia and in practice to fiscal impact analysis. This article explores the theory and practice of fiscal impact analysis with a focus on the relationship between government budgets and urban development patterns. Conventional wisdom states that higher-density development, or compact growth, produces more favorable fiscal impacts than low-density suburban sprawl. Furthermore, the article presents a framework for understanding fiscal impact analysis. It also presents a discussion of the methods of fiscal impact analysis and a discussion of the relationship between development patterns and fiscal impacts. Finally, it highlights some thoughts on the use of fiscal impact analysis for making development decisions.
Growth and local government budgets affect each other in complex ways. On the one hand, the fiscal health of a local government influences the decisions that it makes about growth (Thomas 2006; Edwards 2007). On the other hand, many studies have shown that different land-use types and patterns affect local government budgets (both costs and revenues) in different ways.
The growing recognition of the relationship between decisions about growth (how and where it occurs) and public sector budgets has contributed to the attention given in academia and in practice to fiscal impact analysis. Policy makers, and the public they represent, have many questions about what they term the “fiscal” impacts of growth and development.1 They have tried to use fiscal impact analysis to answer (p. 704) questions such as: Should we annex? Where should we grow? Is compact growth less costly than less compact growth? Who will pay for that growth, and how? At the same time, however, many researchers have cited concerns about the appropriateness of fiscal impact analysis in general, and about the specifics of the techniques employed.
In this chapter we explore the theory and practice of fiscal impact analysis with a focus on the relationship between government budgets and urban development patterns. Conventional wisdom is that higher-density development, or compact growth, produces more favorable fiscal impacts than low-density suburban sprawl. Because of the salience of this issue, we explore this issue in some detail.
The rest of this chapter presents (1) a framework for understanding fiscal impact analysis, (2) a discussion of the methods of fiscal impact analysis, (3) a discussion of the relationship between development patterns and fiscal impacts, and (4) our thoughts on the use of fiscal impact analysis for making development decisions.
Fiscal Impacts: Overview of the Theory and Methods
Urban growth and development provide benefits and impose costs. Built environments provide utility by providing not only places to live, play, and travel but also suitable places to produce, transport, and exchange goods and services. Built environments can change the costs not only of direct factors of production (e.g., land, labor, and capital) but also of natural resources and amenities that affect those costs less directly (e.g., loss of farmland, stream quality, and natural habitat). Some of the benefits and costs are easy to quantify because they are traded in markets; quantifying nonmarket benefits and costs is not so easy.
Figure 31.1 illustrates these points. It has two columns of impacts: one shows examples of costs, the other examples of benefits. It shows three categories of positive and negative impacts (i.e., of benefits and costs): private, public, and other. The full matrix of six boxes is what economists would refer to as a full benefit-cost analysis: it includes, in theory, all benefits and costs, public and private, internal and external, monetizable and nonmonetizable, quantifiable and qualitative. In concept, it includes everything, everywhere. The box in the top right corner emphasizes that point: a full evaluation would look at all types of impacts, on all people, over a long time period.2
Not illustrated in figure 31.1 is an analytical consideration that adds enormously to the scope of the evaluation: that these costs and benefits are not borne or enjoyed equally by all people in a relevant study area, but vary by location, socioeconomic (p. 705) characteristics, consumer preferences, and so on. In other words, figure 31.1 shows the structure for an aggregated evaluation of efficiency, but public policy often depends on a disaggregated analysis of the distribution of costs and benefits (equity).3
The middle two boxes show the subset of benefits and costs that is the focus of fiscal impact analyses: direct impacts on revenues and expenditures in the public sector. “Revenues” are a subset of “benefits,” and “expenditures” are a subset of “costs.”4 The single, highlighted box left of center is where most of the literature on (p. 706) the “costs of growth” (or “costs of sprawl”) tends to focus: on the costs to local governments of building and operating the public infrastructure that private development has come to rely on.5
Figure 31.1 also illustrates the importance of perspective. For example, from the private developer perspective, taxes and fees are a cost (box top left); from the public sector perspective, taxes and fees are a revenue (box middle right); from a social perspective, fees are largely a transfer, excluding administrative costs, which can be substantial.
Few, if any, studies have been able to adequately address an overarching question in the public debate about economic development and growth management: What are the costs and benefits of growth, per se? There are many reasons, but a key one is that growth has too many dimensions and relationships to measure. Growth is more than land development. Even if growth is defined that narrowly, there are many different kinds of land development (residential, commercial, and industrial) and many different ways that development can occur (by location, design, and density). A full benefit-cost analysis of some growth-related action (e.g., new public policy or new private development) would address all its impacts, on all parties, now and into the future. That is obviously an overwhelming task and is only a conceptual goal, not a practical analytical standard. Even the best benefit-cost analyses look only at a subset of potential impacts from a subset of potential perspectives. They do, however, try to distill the impacts and perspectives likely to be most significant and important (Gramlich 1990; Zerbe and Dively 1994; Moore and Thorsnes 2007).
A fiscal impact analysis has a more specific (and, in most cases, more manageable) scope. As figure 31.1 shows, fiscal impact analysis quantifies only those benefits and costs that accrue directly to local governments as a result of various policies about physical growth. Many critiques of fiscal impact analysis simply restate the methodological limitations of the approach and fault it for not being its more comprehensive relative, benefit-cost analysis.
Fiscal Impact Analysis in Theory
Because it focuses not broadly on social welfare (full benefits and costs) but narrowly on revenues and expenditures, a fiscal impact analysis can be framed much like an analysis of a multiproduct firm. That is, net public revenues are analogous to net profits (and profit maximization in the private sector) and can be expressed as the difference between revenues and expenditures: (p. 707)
NPR = ΣRi (population, property values, income, fee structure, other)
–ΣCj (population, density, service levels, other)
NPR = net public revenues;
Ri = revenues from revenue source i; and
Cj = costs of providing service j.
In this framework, a fiscal impact analysis requires the identification of those features of a development project that impact revenues and costs and an estimate of the changes in revenues and expenditures prompted by the development.
An important difference between a government entity and a multiproduct firm, however, is that for most governments, sources of revenues do not directly correspond with sources of costs. A multiproduct firm's costs of producing product i generally have directly corresponding revenues from the sales of product i. A government entity, however, may incur additional school, police, fire, and sewer service costs when accommodating residential growth but may or may not receive additional revenues in the form of property taxes, sales taxes, or service fees—at least those additional revenues are not likely to be directly related to the levels of service provision. Changes in government revenue depend on the types of instruments the government uses to collect revenues, the characteristics of the development project that affects those revenue sources, and the tax-base elasticity of the revenue source with respect to the pertinent characteristics.
Revenues may also accrue at various points in time. For example, the capital costs for a new road to service a new development may be collected at the time of construction by systems development charges, but the ongoing maintenance of that road may be paid by general fund dollars, which the property taxes of the new development add to. Revenue impacts are difficult to forecast because it is difficult to (1) identify how a development will change the base of revenue sources, and (2) estimate future tax rates and fee schedules. Because most local governments are required to balance revenues and expenditures, tax rates and fee structures are often adjusted to meet this requirement. Further, because capital expenditures are often assumed to be covered by fees and systems development charges, some fiscal impact analyses focus only on the costs and revenues associated with the operation and maintenance of a facility. Thus, unlike a multiproduct firm, the revenue projection in a fiscal impact analysis is largely independent of the projection of additional costs.
The structure of costs is more directly similar to that of a multiproduct firm, although cost functions may differ substantially for, say, the provision of parks, safety, education, and sewer services. That is, the cost of each public service can be expressed as a function of the level of services provided. Like a multiproduct firm, the cost of each service will exhibit varying degrees of diminishing returns (or rising marginal cost), economies of scale and scope, and discontinuities that occur following lumpy capital investments (see, e.g., Knaap, Ding, and Hopkins 2001). An aggregate cost function reflecting the total cost of providing all services to each resident would consist of a weighted average of the individual cost functions where the (p. 708) weights represent the bundle of services provided. A fully detailed fiscal impact analysis would examine impacts on each individual cost function, but there has been little research on aggregate government cost functions, and less on the cost functions of any particular public service.
Fiscal Impact Analysis in Practice
The use of fiscal impact analysis increased markedly since the publication of The Fiscal Impact Handbook (Burchell and Listokin 1978, henceforth, The Handbook), and recent survey research shows that it continues to be a popular form of analysis among planners and city officials (Edwards and Huddleston 2009). Fiscal impact analysis is typically used in one of two ways (Edwards and Huddleston 2009):
• To determine the fiscal impacts of a site-specific development (e.g., Does a new residential development increase the demand for and costs of providing schools, parks, roads, and other public goods more or less than it increases revenues in taxes and fees?).
• To evaluate the cost of alternative development patterns (e.g., Are compact or sprawling developments more likely to pay for themselves?).
The Handbook standardized methods for projecting the costs and revenues associated with the service requirements of anticipated development. It defined fiscal impact as “a projection of the direct, current, public costs and revenues associated with residential or nonresidential growth to the local jurisdiction(s) in which this growth is taking place.” In this definition, costs include “operating expenses (salaries, and statutory and material expenses) and capital outlays.” Revenues include “municipal and school district own source (local) contributions (taxes, charges, and miscellaneous revenue) and state and federal intergovernmental transfers.” As described in The Handbook, this definition means that fiscal impact analysis is inherently limited:
• Only direct impacts are counted. Fiscal impact analysis projects only the “first round” of costs and revenues associated with a development. It counts the revenue from property taxes from a new residential development, but does not count the revenue from income taxes that might be earned by the residents of that new development. It counts the costs of providing a new school teacher in a district, but does not count the cost of providing public services to that school teacher.
• Only current expenditures and revenues are counted. Fiscal impact analysis evaluates the expenditures and revenues of a facility as if it were complete and operating today. It assumes that expenditures and revenues will both increase over time, but that the relative relationship between the two will change little.
• Only expenditures and revenues that accrue to the public sector are counted. Governmental expenditures are counted, but private revenues and costs are not.
(p. 709) • Only a subset of public-sector expenditures and revenues are counted. Analysis is done from a particular geographic perspective. Expenditures and revenues are often estimated for only the area in which the population or employment change is occurring, and often for a subset of the services and service districts in that area.
There are four basic steps in the evaluation of the impact of new (usually private sector) development on the (usually public) expenditures and revenues for the infrastructure and services that new development will require:
• Step 1: Forecasting growth. How much will the improvement or development increase population or employment?
• Step 2: Forecasting revenue. How much revenue will the new development generate in the form of fees, taxes, or transfers from other jurisdictions for the local government?
• Step 3: Forecasting expenditures. How much will it cost the local government to provide services (schools, roads, sewer, police and fire protection, etc.) to that increased population or the new jobs?
• Step 4: Comparing expenditures and revenues. The comparison gives an estimate of the net fiscal impact of the improvement or development.
The next sections describe in more detail the first three steps; the fourth step is just a comparison of the results of the work in steps 2 and 3.
Forecasting growth is often the easiest step: not because getting the forecast right is easy (it is not—there is a lot of uncertainty) but because the data and methods for doing an acceptable forecast are well documented and accepted. For the typical development project, the forecasting parameters—the independent variables such as number of housing units, square feet of building space, acres of land, and size of investment—are often well specified.
Forecasting gets more difficult and speculative as the geography, purposes, and time period grow. In a cost-of-growth study, such parameters are chosen as a critical part of the research design. To isolate the fiscal impacts of differences in density, it is preferable to change only density and to leave all other parameters constant. But this is often difficult. And the inability to change specific parameters, and hold others constant, is what limits the insights provided by many fiscal impact studies.
Moreover, growth may beget growth (the multiplier effect). The construction of a new road, for example, involves direct public expenditures on road construction, part of which is again spent by construction workers on goods and services in the community. This additional expenditure can stimulate additional growth. These demand-side multipliers are often addressed using economic base or input-output multipliers from models such as REMI or IMPLAN.6 Supply-side effects include the (p. 710) stimulus effects of new infrastructure. The construction of a new road, for example, can cause new residential or commercial growth even if the demand side effects are negligible. Because demand- and supply-side multiplier effects are difficult to estimate, they are often implicitly or explicitly ignored.
Future revenues can be reasonably estimated when other things remain constant. Table 31.1 shows some of the main local sources of revenues, and the factors that might influence their growth or decline over time.7
Table 31.1 illustrates that revenue sources are influenced not only by growth but also by forces that are outside of the direct influence of most local jurisdictions. The largest source of revenue for many municipalities is transfers from other governments (38 percent on average). These transfers are more influenced by policy factors than by market or other factors that have more stable underlying relationships with growth.
It is also difficult to predict, from year to year, how much revenue might come from any given source. The data in table 31.1 are measured as an average percentage of total revenue collected by local governments across the entire nation: measured this way, the percentages do not change significantly from year to year. At the local level, however, there can be much more variation. As an example, in the city of Portland in fiscal year 2007–2008, federal sources composed about 3 percent of the total revenue; in 2009–2010, federal sources composed 11 percent of the total, probably as a result of a federal economic stimulus package available to local municipalities (Multnomah County Tax Supervising and Conservation Commission 2009–2010).
Most of the sources of revenue in table 31.1 have a positive correlation with growth and development. On the capital side, impact fees will increase as development increases. On the service and operation side, user fees and sales, income, and property taxes will increase with growth. Since these revenues do go up with new development and new population, it should be possible to model their future levels, in part, as a function of growth and development. Even here, though, unanticipated events that are not in direct control of municipalities can have large effects on the revenue stream. The timing and severity of economic cycles is hard to predict, and both can significantly affect revenue.
Policy decisions by higher levels of government can also cause variations from the patterns that might be predicted. Figure 31.2 gives an example of this volatility at a state level, using property tax collections in Oregon from fiscal year 1969–1970 (p. 711) (p. 712) to 2009–2010. Collections increased every year during that forty-year period except for a seven-year period between 1990–1991 and 1997–1998, a period marked by big changes in property tax law. Predicting the passage of such legislation would have been difficult enough, let alone predicting how these revenue changes at the state level might have translated into cuts to programs administered at the municipal level.
Table 31.1 Some Main Sources of Local Government Revenues
Percent of Local Municipal Revenues from Source (National Average, 2007)
New commercial development; economic cycles; policy changes at the state or local level
New development; real estate market economics (construction costs, impact fee rates, land prices, demographic influences); citizen initiatives and tax revolts; urban renewal policy; local and state policy changes to tax rates; population growth rates
National industry and sector growth trends; strength of national and regional economy; regional comparative and competitive advantages and economic development policy; local and state policy changes to tax rates; graduation and higher education rates
Impact fees (system development charges), hookup fees, and user fees (e.g., utility rates)
Amount and type of development occurring; local and state policy changes to fee rates, demand for services, costs of providing services
Federal grants (block grants, stimulus funds), intergovernmental and interjurisdictional transfers
Federal policy changes; revenues from other government agencies; sharing legislation
Note: Figures do not add to 100%; we have not included a small “other sources” category that includes a wide range of revenue sources.
Source: ECONorthwest (2010) with data from the Urban Institute and Brookings Institution Tax Policy Center.
The key point is that for all revenue sources, growth is just one factor that influences revenue projections. The correlation of revenues with growth is therefore weaker than the correlation of expenditures or costs with growth.
Two broad categories of approaches are used to estimate expenditures or costs in fiscal impact analyses: marginal-cost approaches and average-cost approaches. Average-cost approaches generally assume that the average cost of services remains constant so that future costs can be estimated by multiplying current average cost times the quantity of new services required. Marginal cost approaches do not assume that cost remains constant, because some forms of public infrastructure are fixed in the short run, and diminishing returns set in as variable inputs are combined with fixed inputs.
Average-cost methods estimate the cost of providing services to new development by multiplying the cost per unit of service (usually extrapolated from the cost per unit of service in the past or in similar communities) by the amount of growth (measured, for example, in people, households, students, employees) the development is expected to accommodate. They presume that the relationship between cost and growth is linear, and that the cost to a jurisdiction increases proportionately with people or space. For this reason, average-cost methods cannot account for existing excess capacity in a system; they miss the large jumps in costs that occur as capital expenditures for major systems updates (such as new sewage treatment plants or major arterials) are required (figure 31.3).
(p. 713) Figure 31.3 shows long-run, average cost (LAC) is constant, but marginal cost exhibits a sawtooth pattern, jumping from MC1 to MC2 at the point of capacity expansion. At some output levels marginal cost exceeds average costs; at other levels, marginal cost is less than average cost.
Marginal-cost methods can account for this “lumpiness” in public spending that occurs to support growth because they determine the incremental cost occasioned by new development. Marginal-cost methods often use interviews with local officials and case studies to incorporate specific knowledge of the community into the analysis. Estimating the cost of increases in congestion on a fixed transportation facility is an example of a marginal-cost approach. Both approaches have analytical strengths and weaknesses, as summarized in table 31.2.
The choice to use an average- or marginal-cost approach is influenced by both practical and methodological factors. Average-cost methods are typically faster and therefore cheaper, at least in part because data are often more readily available (Burchell and Listokin 1978; Edwards and Huddleston 2009). Marginal-cost methods are more tuned to local circumstances and usually require more quantitative and qualitative work, including interviews of service providers, developers, and other stakeholders. The advantage of interviews is improved data that account for system capacity and variability; the disadvantages are (1) cost, and (2) some interviewees may have actual or perceived conflicts as information sources because they stand to influence policy decisions based on the fiscal impact analysis (Burchell and Listokin 1978; Edwards and Huddleston 2009).
Timing is another factor. In theory, the two approaches should give similar results in the long run: for every public service, excess capacity and low marginal costs now will eventually become a requirement for a big capital investment and high marginal costs.8 Marginal costs may cycle up and down because of lumpy investments, but over time they are the same costs that average costs are estimated from. But in the short run (one to five or ten years), the two methods can give very different results (Burchell and Listokin 1978; Edwards and Huddleston 2009).
A chief reason the results of average cost and marginal costs differ is excess capacity. Large capital facilities (highways, sewage treatment plants, schools, libraries) are built to last and to accommodate increased demand (growth). In other words, these facilities typically have substantial excess capacity when they are opened for operation. They also have substantial fixed costs (e.g., debt service on plant and equipment). Adding new users, at the margin, generates average revenues per user (they pay, on average, the same taxes and rates as existing users), but the marginal costs may be insignificant (e.g., adding the sewage of a new house has no measurable affect on the cost of sewage treatment). Another example: new development may be substantially more fire resistant than old development, and new development will have a higher average value per square foot than old development, with the result that new development may pay more in taxes and fees than (p. 714) (p. 715) a risk- or cost-based analysis would suggest. In other words, in the short run, for some services, the marginal revenues from new development may exceed its marginal costs. In that case, existing users can lower their rates (or, at least, reduce their increases) by bringing in new users to share the fixed costs.
Table 31.2 Overview of Approaches to Cost Estimation in Fiscal Impact Methods
Strengths of Approach
Weaknesses of Approach
Cost estimates based on a per capita or per household estimates in historical trends, extrapolated to new development
Data are more readily available
Assumes that the new investment does not change patterns of demand for services
Per Capita Multiplier; Service Standard; Proportional Valuation
Best used for small projects, when the demand for services likely to grow linearly with population, for long-run evaluations of costs of growth
Relatively inexpensive and fast
Cannot capture jumps in spending or revenue constraints
Cost estimates based on interviews with local officials and experts
Incorporates local knowledge
Can be expensive and time-consuming
Case Study; Comparable City; Employment Anticipation
Best used for large, complex projects, when demand for services is known to be nonlinear, when the system has unused capacity, for short-run evaluations tied to budgets
Accounts for existing capacity in the system
Interviewees may have vested interest in the outcome of the analysis
Source: ECONorthwest (2010).
If it is critical that the analysis be accurate in the short run for local government budget decisions, a marginal approach is the more appropriate choice: it is more likely to reflect actual expenditures over a one- or two-year budget cycle, or even over a six-year capital-improvement cycle. For studies about the long-run costs and revenues associated with various development types in various locations, an average-cost approach makes more sense.
The Fiscal Impacts of Alternative Development Patterns
In most applications, fiscal impact analysis is applied primarily to specific developments that are small in relation to the total amount of development in a region. But the cost of different development patterns is at the heart of debates about growth management, smart growth, and sustainable development. Thus it was inevitable that fiscal impact analysis would be enlisted to try to estimate the net fiscal impacts of growth and the different development patterns in which it might occur.
Table 31.3 summarizes the findings of some of the major studies that evaluated whether compact development provides more favorable fiscal impacts than sprawling development. Each of these studies has different scope and assumptions, different definitions of “compact” versus “sprawling,” different methodological approaches, and (potentially) different institutional and research perspectives. In many cases, the unqualified “yes” or “no” in the table is a simplification that leaves out important caveats about findings. The categories in the final column are described in more detail in the following.
• Changes in local land use. Fiscal impact analyses of this type focus primarily on the fiscal contributions of alternative land-use types and explore the fiscal impacts of land-use change—for example, what are the fiscal impacts of rezoning land from industrial to commercial use or from approving the residential development of farmland? Examples of this type of study we examine “cost of services analyses” that have been done frequently by the American Farmland Trust.
• Changes in local development patterns. Fiscal impact analyses of this type focus not on particular land uses but on the fiscal impacts of developments that differ in densities and regional location. There are two general approaches in this category: prototype development studies and comparative community studies.
(p. 716) • Prototype studies posit developments that feature alternative densities or land configuration, then estimate the capital and operating costs of providing services to those communities, typically based on engineering cost estimates.
• Comparative community studies compare the public services costs of two or more communities that differ in development densities or other features and try to identify differences in the capital and operating costs of providing public services to those communities.
• Changes in regional or national development patterns. Fiscal impact analyses of this type address differences in development patterns at the regional or national scale. We classify this type into two subcategories: community classification studies and econometric studies.
• Community classification studies use of cost and revenue data from multiple types of communities and then redistribute growth across those communities. That is, these types of studies examine how aggregate costs and revenues might change if growth occurred in certain communities and not in others.
• Econometric studies. In contrast to the standard fiscal analysis techniques, which tend to be conducted in spreadsheets as linear manipulations of tabular data (e.g., budget data, engineering costs for certain development or facility types), econometric studies use regional, and sometimes national, data sets to try to control statistically for alternative explanations of changes in costs or revenues.
Earlier sections of this chapter make it clear why it is so hard to compare studies: they have different purposes, audiences, definitions, data, time periods, geographies, services, service levels, and evaluation techniques, and they differ in their attempts to recognize those differences. Moreover, many of these studies have been conducted or sponsored by advocacy groups, which does not necessarily mean that they are biased, but that potential bias is something that must be considered.
Changes in Local Land Use
Changes in local land use are often explored using what is called cost of service (COS) studies. These types of studies are one example of efforts to estimate the effects of local land-use change on local government budgets. The American Farmland Trust (AFT 2002) is a common practitioner of this method. COS studies typically partition land uses into three classes: residential, commercial/industrial, and agricultural/open space. Expenditures and revenues from the municipal budget are then allocated to the three different land-use categories. Although the specific method for fiscal allocations differs among COS studies, the final result is usually a ratio of expenditures over revenues for each of the three land-use classes. For example, a residential ratio of 1.4 means that for every $1.00 of revenue raised from residential land uses, $1.40 is spent providing services to those residential uses. (p. 717) (p. 718) (p. 719) (p. 720) Studies then compare ratios across land-use categories to draw conclusions about their relative fiscal impacts. For AFT, a common conclusion is that land in agriculture or open space has a better (lower) ratio than many other land uses: it may not pay as much in taxes and fees, but that is more than offset by its lower use of facilities and services.
Table 31.3 Overview of Research Regarding the Fiscal Impacts of Compact versus Sprawling Development, Organized by Study Date
Is Compact Less Expensive?
Kotchen and Schulte (2009)
Quantitative meta-analysis of 125 cost of community service studies
For residential development: “Although many planning decisions tend to focus on density and factors that affect home values, the results here suggest that these planning dimensions may have relatively little effect on the balance of expenditures and revenues…”
a. Change in land use study
American Farmland Trust (2002, 1992, 1986)
Various cost of service analyses focused on development on farmland and urban development
Denser patterns are preferable because they cost less and also save farmland. One study estimates off-site capital costs to be around $3,500 to $5,000 per household in 1986 dollars, with denser patterns leading to lower cost.
a. Change in land use study
Real Estate Research Corporation (1974)
Comprehensive study examining the impact of urban form on cost in six hypothetical communities
Low-density development results in higher cost in energy and infrastructure capital/operating cost.
b. Prototype study of change in development patterns
James Duncan Associates (1989)
Detailed study of differences in cost of providing services to five local governments in Florida
Public capital and operating costs for close-in, compact development lower than they were for fringe, scattered, linear, and satellite development.
b. Comparative community study of differences in development patterns
Burchell et al. (2002)
Defines “sprawl” and synthesizes literature on impacts and costs of sprawl
Denser development produces fewer impacts to land, transportation, and natural systems and lower costs, overall.
c. Community classification study of national development patterns
Comprehensive, state-level analysis of New Jersey; examines road, water/sewer, and school costs
Compact growth per household costs less costly than sprawl growth: about 75% for local roads, 80% for utilities, and 95% for schools.
c. Community classification study of regional land use study
Econometric analysis to examine 1985 government expenditure data from 247 large counties
Except within a range of very low densities, public service costs increase with higher densities. Ladd finds a U-shaped relationship between the rate of growth and growth in local government per capita spending. According to regression analysis, spending decreases as density increases to around 1,750 persons per square mile, then begins to increase along with further density increase. This implies that there are diseconomies of scale as well as economies of scale.
c. Econometric study of national land-use patterns
Carruthers and Ulfarsson (2008)
Econometric analysis of Census of Governments data for twelve categories (including capital facilities, roadways, sewerage, trash collection, police, fire, parks, and education) used to model expenditures from 283 metropolitan counties nationwide
Per capita cost declines with density and increases with spatial extent. Low-density, spatially extensive development is more expensive to support. Compact development results in the largest savings for (in order): roadways, parks, education, and police protection.
c. Econometric study or national development patterns
Hortas-Rico and Solé-Ollé (2010)
Econometric analysis of 2,500 Spanish municipalities’ expenditures on basic infrastructure and transport, community facilities, local police, housing and community development, culture and sports, and general administration
Low-density developments led to greater provision costs in all the spending categories considered, with the exception of housing and basic infrastructure and transportation.
c. Econometric study or national development patterns
A meta-analysis of COS studies (Kotchen and Schulte 2009) reports that commercial/industrial and agricultural/open-space ratios usually produce expenditure-to-revenue ratios of less than one while residential ratios are greater than one. This is often interpreted to mean that commercial/industrial and agricultural/open-space land uses “pay their own way” while residential land uses do not. Burchell and Listokin (1993) identified a similar fiscal impact hierarchy. AFT and other land conservation advocates use these ratios to argue against the common perception that residential development will decrease the property tax burden for current residents. The results are also used to argue that open lands provide fiscal benefits and that current use valuation, rather than potential development value, is justified for tax purposes.
These kinds of studies have many limitations: they generally use average cost, not marginal cost, methods; they assume the cost of services can be directly attributed to particular land uses; and they assume that costs and revenues can be projected indefinitely. But two limitations are most salient. First, COS studies do not hold population constant. They do not presume that some amount of population or employment growth must be accommodated somewhere. An implication is that growth or development costs more than it pays, so that growth should not occur. Second, COS studies presume that the decision to convert land from agricultural to urban uses can be done independently for each urban land-use type. That is, they presume that it is possible to choose particular land uses for a community and to exclude other land uses.
The primary reason most COS studies find that agricultural and commercial land uses have favorable impacts is because nobody lives there. Since it is people who demand education, public health, and most other public services, and since agricultural and commercial uses house relatively few (if any) people, the cost of providing services to these land uses is very low. And for commercial uses, with high property values, the potential revenues are very high. Favorable fiscal impact ratios for agricultural and commercial land uses, and unfavorable ratios for residential uses are the inevitable result.9
A frequent use of COS studies is to counter perceptions that development, especially residential development, provides net fiscal benefits. But just as it is inappropriate to promote development for its fiscal impacts, it is inappropriate to limit development for the same reason. Fiscal impact analysis alone should not be used (p. 721) as the primary basis for such decisions. Because healthy communities require a balance of land uses, COS studies are poorly suited to drive such decisions.
Changes in Local Development Patterns
As opposed to local land-use studies that focus on particular development types, studies that examine changes in local development patterns evaluate developments that include several land-use types but differ in other important dimensions such as density, location, or design. Analysts have taken two general approaches to examine these differences: prototype development analyses and comparative community analysis.
Prototype studies are perhaps the most common type of fiscal impact analysis. They focus primarily on capital costs and revenues, often related primarily or exclusively to publicly provided infrastructure (roads, sewer, and the like). Like COS studies, prototype studies generally follow the steps outlined in The Fiscal Impact Handbook. Their history can be traced, however, to a seminal but much criticized study by the Real Estate Research Corporation (RERC 1974). The Costs of Sprawl found that capital costs per unit were higher in the “low-diversity sprawl” and “sprawl mix” neighborhood types than they were in the “planned mix” or “high-density planned mix” neighborhoods. When used to analyze the fiscal impacts of alternative development patterns, the procedure generally involves following the standard steps for at least two development alternatives: sprawl or smart growth, where the smart growth alternative involves more mixed uses and higher densities.
The findings of these studies are also quite consistent. Denser, mixed, and more compact development has more favorable fiscal impacts than lower-density (sprawl) development, and that cost increases with distance from centralized public services (Parsons-Brinkerhoff and ECONorthwest 1998; Speir and Stephenson 2002). Cost savings are attributable to capital costs, primarily roads and streets, sewer, and other utilities.
The criticisms of the RERC study (e.g., Windsor 1979) remain valid for prototype studies that are being conducted today. First, at the highest level, they often ignore operating and maintenance costs, leaving key expenses out of the equation. This is an especially important critique when considering cost incidence, as many municipalities have developer and other fees in place that partially or entirely recoup the major capital costs that are counted in these studies.
Second, most new development of any type takes advantage of some existing capacity in infrastructure systems. From this perspective, the higher costs for infrastructure to support sprawl could potentially be mitigated with better planning: sizing the pipes and roads to accommodate possible future development, even in a sprawl pattern, could create economies of scale that affect infrastructure costs measured on a per-unit basis (Altshuler and Gomez-Ibanez 1993). Further, in situations in which higher costs are incurred by leapfrog development patterns, the higher infrastructure costs attributed to this pattern could be recovered as areas that were initially passed over are later developed with little additional public cost.
(p. 722) A final set of concerns about prototype studies is that they do not consider the variability in the quality of what the public sector is providing in exchange for the costs it incurs. In essence, it assumes that high-density housing units and sprawl housing units are comparable, which may or may not be the case. Cost savings in higher-density developments might accrue from less open space, smaller houses, and potentially more congestion, all of which might not be desirable outcomes for the occupants of the units in question or for the local government tasked with servicing them. On the whole: prototype studies are useful for helping local governments estimate the costs of adding new roads or pipelines, but they must be considered against community preferences to meaningfully guide growth policy.
Comparative community studies are similar to prototype studies, but they draw data from real communities. James Duncan Associates (1989) is an example. Using case studies of several Florida communities, Duncan Associates examined not only differences in density but also the broader “regional” costs of different scenarios. These case studies included five different development patterns (compact, contiguous, satellite, linear, and scattered). They found that public capital and operating costs for close-in, compact development were lower than they were for fringe, scattered, linear, and satellite development. Costs per dwelling unit ranged from $9,252 for downtown Orlando (1989 dollars) to a high of $23,960 in Wellington, a low-density fringe development. In addition, the study concluded that planned growth could save significantly on the cost of providing roads and utilities, but only modestly on schools.
Comparative community studies, like all others, have advantages and disadvantages. Because they use actual data from real communities, they lack the hypothetical nature of prototype studies. The cost and revenue data are real, the development patterns are real, and the densities, levels of service, and taxpayers are real. The real-world features of the development patterns are especially important for examining operating costs, which can be particularly difficult to estimate using prototype studies.
Comparative community studies also share many of the limitations of prototype studies. That is, it is difficult for such studies to control for existing infrastructure capacity. Also, as in prototype studies, a number of development qualities are not held constant. While there is perhaps some assurance that the levels of public services are relatively comparable across developments with different densities, housing size, lot sizes, and public open spaces are not directly comparable across communities. Residents may be willing to pay more for these attributes and, perhaps, the higher public service costs they require.
Changes in Regional and National Development Patterns
A third type of fiscal impact study looks at impacts from a regional or national level, using data from a sample of regional or national communities to estimate the fiscal impacts of alternative development patterns at the regional or national scale. There (p. 723) are two common approaches taken to estimate the fiscal impacts of alternative development patterns at the regional or national scale: community classification approaches and econometric (regression) approaches.
The community classification approach is used extensively by Robert Burchell and his colleagues. Burchell and others (1998); Burchell, Dolphin, and Galley (2000) examined the fiscal impact of alternative development patterns in several states in the 1990s. These comparisons of development-as-usual (“trend”) and more compact (“planned”) development attempted to estimate the savings in road and water/sewer cost over a twenty-year period. In each case, these studies projected savings from modestly increased densities and shifting growth closer to population centers. For example, Burchell, Dolphin, and Galley projected that shifting from sprawl to planned growth could reduce total road-building expenditures 12 percent in South Carolina, 12 percent in Michigan, and 26 percent in New Jersey (Muro and Puentes 2004). For water and sewer infrastructure, savings varied from 8 percent in New Jersey to 13 percent in South Carolina to 14 percent in Michigan. In a 2000 update of the 1998 study, Burchell et al. estimated that New Jersey could save $2.32 billion, or 15 percent, of its total road and water/sewer infrastructure costs between 2000 and 2020. He calculated that more than half ($1.46 billion) of the savings would result from a 13 percent reduction in water/sewer expenditures due to more efficient clustering, more use of existing infrastructure, and more attached and multifamily housing.
Costs of Sprawl—2000 (Burchell and others 2002) estimates the public service savings and fiscal benefits from controlled growth. By classifying communities on a continuum from urban to suburban to rural, and using per capita service-cost estimates, Burchell et al. estimate that local governments could reduce their public service costs by $4.2 billion a year, or 3.7 percent, after twenty-five years if the country were to embrace controlled growth nationwide. According to the report: “The decrease in costs is possible because, under controlled growth development, more development will take place in developed areas where public service costs may be more expensive, but public-service demand can be absorbed more readily due to the excess capacity found there.” (13)
The community classification approach has limitations like all others. Perhaps most fundamentally, the approach of classifying communities according to a sprawl continuum and then redistributing growth among those communities only indirectly addresses differences in development patterns. It is quite possible, for example, that new development in a particular rural county is more compact than new development in an urban county. Further, not all communities in the same place on the urban-rural continuum are likely to have the same cost structure. Finally, many of the regional and national studies are based on data from the Census of Governments. These data reveal general differences in costs and revenues but do not have the specificity of data obtained directly from local governments.
Several studies have addressed the impact of development patterns using regression analysis. Key studies in this category include Ladd (1992, 1994), (p. 724) Carruthers and Ulfarsson (2003, 2008), Solé-Ollé (2006), and Hortas-Rico and Solé-Ollé (2010). These studies use as dependent variables cross-sectional data on government expenditures in various spending categories, and as independent variables demand and cost factors, including density and urban area size. Although these studies include revenue-related variables, their focus is on costs and often a variation of the question: Does sprawl increase the cost of providing public services? The studies by Ladd and Carruthers/Urfarsson use U.S. counties as units of analysis; the studies by Solé-Ollé use data from municipalities in Spain. For the most part, the methods used are similar. All use ordinary least-squares (OLS) regression, and the later studies by Carruthers/Ulfarsson and Solé-Ollé control for spatial spillovers. The latter studies also use measures of developed land instead of total land area to get better measures of urban sprawl.
The results of all these studies are remarkably similar. All find per capita expenditures decrease with development density, at least over some range, for selected services and for total public expenditures. In all the studies, roads, water, and wastewater services were found most sensitive to measures of urban sprawl. The studies by Ladd and Solé-Ollé use piecewise regression techniques to estimate the effects of density in specific ranges of density. Ladd finds per capita cost functions for many public facilities in urban areas are U-shaped with respect to the amount of population: per capita costs initially decline because of economies of scale in service provision but eventually increase as the public sector tries to provide more facilities and services to mitigate the “harshness of the environment” that accompanies metropolitan size and density. Solé-Ollé finds that per capita costs fall with density over all density ranges but do so at a decreasing rate.
Unlike COS and comparative community studies, regression studies have the advantage of drawing inferences from many real communities: not just one and not a hypothetical one. But they also have limitations. Census of Governments data, used by Ladd and Carruthers/Ulfarsson, are notoriously imprecise and reflect survey results, not actual budget data. Moreover, they include in each category of expenditure both operating and capital cost, which can vary widely from year to year. The measures of density have improved over time from Ladd to Carruthers/Ulfarsson to Solé-Ollé, but there is no way when using data from large numbers of communities to capture engineering and development details the way comparative community and prototype studies can.
A limitation of the econometric studies to date is that they use reduced-form equations and rely on expenditure data. This makes it impossible to isolate supply-side effects from demand-side effects. That is, although each of the studies goes to considerable lengths to control for factors that affect the demand for public services, they cannot rule out the possibility that higher levels of spending in low-density communities reflect higher levels of public service quantity or quality instead of higher levels of cost holding the level of services constant.
(p. 725) Conclusion
Our review and analysis of the theory and practice of fiscal impact analysis suggests that the practice is alive and well and is being extended to address a variety of issues, including the cost of alternative development patterns. The practice of fiscal impact analysis has changed little over the last thirty years, although new and better econometric studies have been published in recent years. For the most part, however, the practice must contend with the same limitations of data and methods that have been recognized for years. Moreover, the literature on fiscal impact analysis has not merged gracefully with research on the cost functions of public facilities or the elasticity of local government tax base. This presents a major opportunity for advancement through collaboration between academic economists and planners.
The studies we reviewed about the fiscal impacts or costs of different land-use and development patterns support three general conclusions:
• Different land uses have different fiscal impacts; residential land uses are generally the most costly to service and generally do not generate sufficient revenues to cover costs (though high-value residential is often an exception, especially if it is in an area with relatively low service costs, which is often the case in exurban development).
• Capital costs per person are lower under higher-density development than lower-density development; this is particularly clear for roads, sewer, and water infrastructure.
• Expenditures per capita are lower in higher-density counties at least over some range.
Beyond that, definitive conclusions are difficult, in part because every study shares similar limitations. None of the studies we reviewed rigorously controls for the quality of services provided. High-density development in urban centers can reduce the cost per dwelling unit relative to lower-density or suburban development, but the costs per square foot for both land and buildings will be lower in suburban locations (lower land costs, and the use of wood instead of steel, surface instead of structured parking, etc.). The problem is that if all urban growth is accommodated at suburban densities, other costs (e.g., of service extensions, environmental externalities) may get higher.
Studies that do find denser development to be less costly are often based on an underlying and intuitive logic that has not changed since the original Cost of Sprawl study was completed in 1974: leapfrog development requires expensive road and utility connections between neighborhoods that can be reduced through more compact development. But most of the studies are comparative static studies, not dynamic ones. How do development patterns play out over time? In particular, do less dense patterns now set the stage for denser ones later? Some academic research has made the case for the efficiency of leapfrog development, arguing that (p. 726) subsequent infill occurs at higher densities than would have been observed if denser build-out have been attempted at the start (Ohls and Pines 1975; Peiser 1989).
By definition, fiscal impact analyses address public revenues and expenditures, not private benefits and costs. Some studies treat the issue of development patterns as if it were a cost-minimization problem, not a benefit-cost optimization problem. The less land, built space, and service provided, the less the cost of growth. But households are trying to optimize on a complex bundle of desired goods and services. Thus, even if it were theoretically true that a high-density pattern of development is cheapest to consumers and to the public sector,10 consumers (perhaps a large percentage) may be willing to pay for something different, and perhaps much different.
Over time and across jurisdictions, studies may also be capturing differences in policy rather than actual cost differences. Bunnell (1997) notes that “the local fiscal impact of development depends in large measure on the system of government finance that is in place. Indeed fiscal impact studies that now find that development is fiscally unbeneficial to local governments probably say more about changes in the structure of public finance that have occurred in the last 20 years…than they say about the desirability or undesirability of development.”(149) We agree.
Our review suggests that basic fiscal impact techniques are not infinitely scalable. They work best on a small scale (a single jurisdiction) and over a short period (one to five years). Within those boundaries an analyst can consider local budget data (costs and revenues), and assume that new development will not be large enough relative to the base of existing development to change demographics, preferences, service standards, or tax or fee structures. As one moves to long-run, regional development patterns, those options are no longer available. A region may comprise dozens of municipalities and taxing districts, each with its own fee structures and budgeting processes. Thus, regional fiscal studies tend to use different methods and data sets, which create some of the problems we described.
Given our findings, we believe that the techniques of fiscal impact analysis as typically practiced are not up to the task of making definitive findings about the desirability of different development patterns. The proper question, in the context of concerns about sustainable development, is one that is more suited to benefit-cost analysis (see figure 31.1): What are the full benefits and costs of alternative development patterns? Stating the question that way makes it bigger and harder to answer. It requires looking not just at average annual service costs but also at capital costs; not just at capital costs but at life-cycle operation and maintenance costs; not just at public costs but at private costs; not just at direct costs but at indirect and external costs; not just at the amount of cost but at the distribution of those costs across locations and groups; not just at cost but at benefits; not just at least cost but at net benefits.
(p. 727) Because local governments must balance revenues and costs, and because developments have varying impacts on revenues and costs, a carefully crafted fiscal impact analysis can provide information pertinent to a development decision—it is important to know what things cost. But a development decision must consider more than the fiscal impacts alone. We noted earlier the evidence that a local government might reduce its net fiscal costs by preserving farmland instead of building places where people live and demand services, or by building neighborhoods that use fewer roads, sewer and water pipes, and smaller school, police, and fire facilities. But it cannot, by itself, tell policy makers whether those are preferred development choices. Further, without a reduction in fertility or immigration, growth must occur somewhere. The economic question is where that growth can be accommodated most efficiently, not whether growth should occur.
A point fundamental to economists that planners often miss: finding optimal growth patterns is not a cost-minimization problem. While it may cost less to develop at higher densities, lower-density development may provide benefits that residents value. Put more simply: smaller houses are generally cheaper than large houses, but that does not mean that building only small houses maximizes social welfare. To economists, whether to build large or small houses should depend on relative costs and benefits. In the housing market, markets provide some assurances that both costs and benefits are entered into consumer decision making. But because the public sector cannot (or, at least, does not) similarly price all public services at marginal cost—though impact fees move in that direction—the cost-benefit evaluation must be left to decision makers.
Despite the obstacles, local evaluations of the costs (and benefits) of development will continue to be important in the inevitable debates about whether, how much, where, and how to grow. The professional literature may provide local governments with a sense of possible and likely outcomes, but it will not give them definitive answers about their specific situation. Local governments will ultimately require local data tied to local conditions to inform the discussion about the costs and benefits of growth.
There are, of course, plenty of opportunities for future research. Research certainly would be helped if local governments maintained and distributed information about service levels, revenues, and expenditures in consistent ways. This would, for example, permit the analysis that could separate cost increases due to increases in levels of services from other increases that occur holding service levels constant.
More broadly, the problems for addressing the costs of development patterns are of both data and technique. What would be helpful is something our review of the literature did not uncover: a set of cost curves with means and standard deviations, based on cross-sectional studies, that clearly describe the facilities and services being evaluated, and show the capital and life-cycle operations and maintenance costs change as a function of density and development patterns, controlled for type of service, quality of service, topography, distance from central facilities and services, and so on. Given what it would take to do the study that would generate those curves, we are not surprised not to find it. Studies of public (p. 728) service cost functions have been conducted for specific public services. But the central question in most of those studies is not the impacts of growth or development density. Given the importance of these variables to public policy as drivers of land use, transportation, economic development, the environment, and climate change, it would be useful to introduce such variables into these kinds of cost and fiscal impact analyses.
Altshuler, A. A., and J. A. Gomez-Ibanez. 1993. Regulation for Revenue: The Political Economy of Land Use Exactions. Washington, DC: Brookings Institution.Find this resource:
American Farmland Trust. 1986. Density-Related Public Costs. Washington, DC: AFT.Find this resource:
———. 1992. Does Farmland Protection Pay? The Cost of Community Services in Three Massachusetts Towns. Washington, DC: AFT.Find this resource:
———. 2002. Cost of Community Services Studies: Making the Case for Conservation. Washington, DC: AFT.Find this resource:
Bunnell, G. (1997). “Fiscal Impact Studies as Advocacy and Story Telling.” Journal of Planning Literature 12:136–151.Find this resource:
Burchell, Robert, William R. Dolphin, and Catherine Galley. 2000. The Costs and Benefits of Alternative Growth Patterns: The Impact Assessment of the New Jersey State Plan. Trenton: New Jersey Office of State Planning. http://www.nj.gov/dca/osg/plan/impact.html.Find this resource:
Burchell, Robert, et al. others. 1998. The Costs of Sprawl—Revisited. Washington, DC: National Academy Press.Find this resource:
Burchell, Robert, and others. 2002. Costs of Sprawl—2000. Washington, DC: National Academy Press.Find this resource:
Burchell, R. W., and David Listokin. 1978. The Fiscal Impact Handbook: Estimating Local Costs and Revenues of Land Development. New Brunswick, NJ: Center for Urban Policy Research.Find this resource:
———. (1993) “Fiscal Impact Procedures and State of the Art: The Subset Question of the Costs and Revenues of Open Space and Agricultural Lands.” In Lincoln Institute of Land Policy, Does Land Conservation Pay? Determining the Fiscal Implications of Preserving Open Land, 1–32. Cambridge, MA: Lincoln Institute of Land Policy.Find this resource:
Carruthers, J. I., and G. F. Ulfarsson. 2003. “Urban Sprawl and the Cost of Public Services. Environment and Planning B: Planning and Design 30 (4):503–522.Find this resource:
———. (2008). Does “Smart Growth” Matter to Public Finance?” Urban Studies 45:1791–1823.Find this resource:
James Duncan and Associates. The Search for Efficient Urban Growth Patterns: A Study of the Fiscal Impacts of Development in Florida. Presented to the Governor's Task Force on Urban Growth Patterns and the Florida Department of Community Affairs, July 1989.Find this resource:
Edwards, Mary. 2007. “Public Finance in Planning Education and Practice.” Journal of Planning Education and Research 27:217–227.Find this resource:
Edwards, Mary M., and Jack R. Huddleston. 2009. “Prospects and Perils of Fiscal Impact Analysis.” Journal of the American Planning Association 76:25–41.Find this resource:
(p. 729) Gramlich, Edward. 1990. A Guide to Benefit Cost Analysis. Englewood Cliffs, NJ: Prentice Hall.Find this resource:
Hortas-Rico, M., and A. Solé-Ollé. 2010. “Does Urban Sprawl Increase the Costs of Providing Local Public Services? Evidence from Spanish Municipalities.” Urban Studies 47:1513–1540.Find this resource:
Knaap, Gerrit, Chengri Ding, and Lewis Hopkins. 2001. “Managing Urban Growth for the Efficient Use of Public Infrastructure: Toward a Theory of Concurrency.” International Regional Science Review 24:328–343.Find this resource:
Kotchen, Matthew J., and Stacey L. Schulte. 2009. “A Meta-analysis of Cost of Community Service Studies.” International Regional Science Review 32:376–399.Find this resource:
Ladd, H. F. (1992). “Population Growth, Density and the Costs of Providing Public Services.” Urban Studies 29:273–295.Find this resource:
———. 1994. “Fiscal Impacts of Local Population Growth: A Conceptual and Empirical Analysis.” Regional Science and Urban Economics 24:661–686.Find this resource:
Moore, Terry, and Paul Thorsnes. 2007. The Transportation/Land Use Connection. American Planning Association Planning Advisory Service Report 546/547.Find this resource:
Multnomah County Tax Supervising and Conservation Commission. 2010. Tax Supervising and Conservation Commission Annual Report: Fiscal Year 2009–10.Find this resource:
Muro, M., and R. Puentes. 2004. “Investing in a Better Future: A Review of the Fiscal and Competitive Advantages of Smarter Growth Development Patterns.” Brookings Institution Discussion Paper. Washington, DC: Brookings Institution.Find this resource:
National Association of Home Builders. 2009. The Local Impact of Home Building in a Typical Metro Area: Comparing Costs to Revenue for Local Governments. Washington, DC: NAHB.Find this resource:
Ohls, J., and D. Pines. 1975. “Discontinuous Urban Development and Economic Efficiency.” Land Economics 3:224–234.Find this resource:
Oregon Department of Revenue. 2010. Oregon Property Tax Statistics, Fiscal Year 2009–10.Find this resource:
Parsons-Brinkerhoff and ECONorthwest. 1998. The Full Social Costs of Alternative Land Use Patterns: Theory, Data, Methods and Recommendations. Prepared for the Federal Highway Administration. Washington, DC: FHWA.Find this resource:
Peiser, R. B. 1989. “Density and Urban Sprawl.” Land Economics 65:193–204.Find this resource:
Real Estate Research Corporation (RERC). 1974. The Costs of Sprawl: Environmental and Economic Costs of Alternative Residential Development Patterns at the Urban Fringe. Washington, DC: U.S. Government Printing Office.Find this resource:
Solé-Ollé, A. 2006. “Expenditure Spillovers and Fiscal Interactions: Empirical Evidence from Local Governments in Spain.” Journal of Urban Economics 59:32–53.Find this resource:
Speir, C., and Stephenson, K. 2002. “Does Sprawl Cost Us All? Isolating the Effects of Housing Patterns on Public Water and Sewer Costs.” Journal of the American Planning Association 68:56–70.Find this resource:
Thomas, John V. 2006. “Dividing Lines and Bottom Lines: The Forces Shaping Local Development Patterns.” Journal of Planning Education and Research 25:275–293.Find this resource:
Windsor, D. 1979. “A Critique of The Costs of Sprawl.” Journal of the American Planning Association 45:279–292.Find this resource:
Zerbe, R., and D. Dively. 1994. Benefit Cost Analysis in Theory and in Practice. New York: HarperCollins.Find this resource:
(1.) As this chapter will discuss, the concerns they might lump under the heading of fiscal impacts may go well beyond the more precise definitions used in the professional literature.
(2.) In concept, a full analysis looks at those costs over a long time period (not just current costs, but future costs) and fairly incorporates those future costs and benefits into the analysis by bringing them back to a present value at an appropriate discount rate.
(3.) As defined here, benefit-cost analysis is a comprehensive framework for evaluating all impacts on all people over all time periods, not a specific technique for calculating net monetizable benefits in the aggregate. In concept, it addresses the current call for evaluating the “sustainability” of private and public activities: for considering the “triple bottom line” and external impacts (i.e., all impacts, not just some), impacts on future generations, and the distribution of impacts (e.g., on developing versus developed countries, on lower-income versus higher-income groups).
(4.) Throughout this chapter we try to use the term cost when we mean the full costs (internal and external, whether monetizable or quantifiable or not) of some action, and expenditure when we mean the direct, monetizable cost that a public entity incurs to support its actions (which means it is a budget item). That distinction gets fuzzy in some of the literature on “the costs of growth,” which sometimes addresses the broader idea of costs and sometimes the narrower idea of public expenditures (i.e., fiscal impacts).
(5.) The definition of fiscal impacts that figure 31.1 illustrates includes both capital and operating expenditures. Some studies of fiscal impact limit themselves to operating expenditures, perhaps because for many developments many capital costs are covered by developers. Obviously, the question about “costs of growth” requires a consideration of capital costs and the sources of revenue to pay for them, so we use the broader definition here.
(6.) REMI and IMPLAN are regional input-output models.
(7.) Another way to pay for facilities and services that is not strictly a revenue source is to require the private sector (e.g., developers) to provide them. It is common, for example, for developers to provide sewer, roads, and other infrastructure in a residential subdivision and to then convey those facilities to a municipality or service district when the development is occupied. The capital costs are a real cost of development, but in the context of a fiscal impact analysis for a local service provider, both the capital costs and revenues to pay those costs are off the books and not considered.
(8.) In the long run under constant returns to scale, long-run average cost must equal long-run marginal cost.
(9.) One recent exception is provided by the National Association of Home builders (2009), which finds positive impacts of residential development. The reason the study finds positive impacts is because it includes multiplier effects of housing construction. Subsequent increases in commercial uses produce a net positive impact.