Do Single-Family Dwellings Pay for Themselves?

This Advisor column was originally published in April 2010.

There was that assertion again: “… everyone knows that single-family houses do not support themselves.” This time I heard it over coffee from an economist who had just completed a fiscal analysis of a proposed development.

The perception that single-family houses do not generate sufficient revenue to pay for the local governmental services needed to serve them has become a basic assumption throughout local government. Until recent years, this assumption was an academic question that really did not affect much. However, as local governments struggle to keep fiscally afloat, many cities are acting on this assumption — encouraging commercial development and discouraging residential development. This perception is now exerting a significant influence over how cities are growing and developing.

My forty years in community development and finance tells me that this assumption is, at best, overly simplistic. Whether particular uses support themselves or not depends more on the location and circumstances of those uses. Like most planners, I ardently believe that sprawling single family dwellings do indeed not support themselves, but am persuaded to believe that well-sited (such as infill) housing may more than “pay for itself.” Similarly, poorly-sited commercial development can be very costly for local governments.

The perception that single-family uses do not support themselves has been generated by years and years of fiscal analysis that tends to overstate the costs of serving residential development and underestimates the revenue generated by such uses. Since the space available for this discussion cannot fully explore all aspects of this complex issue, this discussion will only touch on some of the major reasons that the fiscal benefits of single-family residences tend to be underestimated relative to commercial uses.

Artificiality of the Question

It is first important to note that the whole question is somewhat artificial. Our revenue system was not designed to balance revenue and expenditures between particular uses. Revenue systems were designed to generate revenue from a variety of sources to finance services for the entire community, not just portions of it.

For example, it is basic tax policy in America that education should be financed by the entire community and not just the parents of children. So, obviously, single-family dwellings, where those children live, will not generate enough revenue to support schools. If school costs are included in the analysis of single-family dwellings then, of course, single-family dwellings will not support themselves and the analysis has little meaning beyond the obvious implication that schools are supported the entire community, not just by houses with students.¹ Consequently, this discussion will only focus on whether single-family uses support themselves in terms of city revenues and costs, excluding schools.

If the expectation that different types of uses are to support themselves is carried to its logical extreme, wealthy residential areas should receive more governmental services than poorer ones because the wealthier areas pay more in property taxes.

Another element of artificiality is associated with estimating the future costs of supporting various types of uses, e.g., the cost of providing police services to serve new houses. In reality, the actual costs are going to be determined by budget decisions made well after the analysis, e.g., the city council decides to cut or increase its police services. These budget decisions are more likely to be driven by real operational considerations (such as a drop in revenues, actual calls for services in different areas, labor contracts, changes in state law, etc.) rather than by the abstract factors usually used in fiscal studies.

Faulty Revenue Analysis and Assumptions

One of the more significant factors that contributes to the perception that residential uses do not support themselves is the way most fiscal analysis studies treat revenue generated by sales taxes. Almost all studies of the fiscal impacts of various uses attribute retail sales tax² revenues to where the sales take place, i.e., to commercial uses. Rarely do these studies attribute any³ retail sales tax revenues to residential uses.

However, most retail sales are driven by the residential sector — businesses only take advantage of the increase in their market created by residential growth. More houses mean more sales tax revenues. What complicates this picture is that the residences that generate the revenues are often separated from where those revenues are collected (i.e., in retail stores) by jurisdictional boundaries. Although this separation may mean that the jurisdiction where the homes are located may not collect all of the revenue generated from its residences, almost all communities (except for a few exclusively residential communities) will experience some growth in sales tax revenues from the addition of new homes in the community.

Generally, the more developed a community’s retail sector, the more the community will capture of this revenue. The City of Lynnwood, for example, would capture most of the additional retail activity generated by new residential growth because it has a full, range of stores in the community where the local residents can shop, while the City of Sultan may capture only a small portion, because it only has a few stores available.

While in the short-term, communities with small commercial sectors may only capture a small amount of sales tax revenue from residential growth, the long-term benefits of residential growth can nonetheless be very significant. Residential growth in small communities increases the market available for all types of commercial activity. As these markets grow, more and more businesses will be attracted to locate in the community, increasing the amount of revenue captured from all residences in the community.

Although the amount of retail sales tax revenue captured by local governments from their residents varies from community to community, good fiscal analysis should show at least some additional retail sales tax revenues from new houses.

It is important to note that while most fiscal analysis only shows an increase in sales tax revenue if additional retail sales floor area is added, the increase in retail sales tax revenue resulting from new residential development will occur even if there is no increase in retail sales floor area, as the additional people in the new homes spend more in existing stores.

Most fiscal analysis studies also tend to overstate the revenue generated by new commercial development. Most studies calculate the amount of sales that is generated from commercial development by multiplying the amount of commercial floor area being added by an assumed amount of sales that is expected to be generated per square foot of commercial floor space.

This calculation seldom takes into account the competition between retail establishments that frequently occurs when new commercial stores are built. For example, did the Kohl’s built in Lynnwood a few years ago add new revenues, or did the Kohl’s just pick up the business that was there after a Mervyn’s closed, or alternatively did the Kohl’s just attract business from the nearby Macy’s, Nordstrom’s etc.? Unless the additional floor area is accompanied by additional residential growth somewhere in the market area (or growth in the personal income of people residing in residential uses) no significant amount of additional retail sales growth is likely to occur no matter how much floor space is added.

Faulty Expenditure Analysis and Assumptions

Many fiscal analysis studies tend to overstate the costs of serving residential uses. The prime example of this is that studies often estimate many costs on the basis of population. For example, a common way to estimate police and fire costs (a very large part of any city’s budget) is on the basis of per-capita multipliers. Commonly, studies divide the current police and fire budgets by the population of the city. The studies then multiply this factor by anticipated resident population growth to forecast the future police and fire costs needed to serve the community.

Since this calculation is based only on growth in population, this methodology allocates ALL of any additional police and fire costs to new houses that provide the population and no future costs are allocated to growth in other uses. Consequently, when comparing the costs of serving new residential growth with the costs of serving new commercial development, all of the costs are attributed to the residential uses as if there were no police or fire costs associated with new businesses.

This is obviously a simplistic representation of cost relationships. Additional businesses do generate police and fire costs in the form of shoplifters, various forms of property crimes, problems associated with drinking establishments⁴, increased medical aid calls⁵ etc. To the extent that these crimes and calls for services are due to people who may not live in the community but shop there, they are not attributable to residential uses.⁶ Emergency aid costs are driven by the location of people, where they work, where they shop, where they play, and where they travel — not just where they reside.

Even more significant, police and fire costs are driven by diverse activities which are not directly associated with specific uses. The amount of traffic in a community, for example, drives a substantial amount of the costs of police and fire services, and that traffic may include a large amount of through-traffic or traffic originating outside the area to commercial centers, unrelated to the residential uses in the community.

Some studies do recognize the problem of relying on per capita multipliers and have employed other methods to allocate costs between uses. A common alternative is the use of “calls for service,” i.e., forecasting future costs on the basis of the calls for service from police and fire. While this, potentially, is a much better basis for allocating costs between uses⁷, this approach must include a forecast of calls for service by use, which in turn is based on a forecast of the amount of each use that will be developed. I have reviewed several studies that forecast calls for service on the basis of population growth; however, relating calls for service only to population growth produces the same results as forecasting costs based on population growth alone.

Excluding Capital Costs

It is common practice to prepare studies of the fiscal impacts of various uses on the basis of ongoing costs, excluding capital costs. While the capital costs associated with new development are very sensitive to the location of the development, commercial uses (on a per acre basis) tend to have more cost impacts on capital facility needs, by often requiring more capital facilities than residential uses.

For example, commercial development attracts traffic to the community. This traffic, when combined with existing and future traffic, frequently results in significant costs for expanding the capacity of the street system. These costs are difficult to anticipate and capture by means of impact fees or mitigation assessments and can be very costly. IF the traffic impacts of new commercial development are not carefully managed and mitigated, the costs of making the related traffic improvements can dwarf the potential sales tax revenue that may be generated by the commercial development.

Lack of Considering Location Factors

The most significant weakness of almost every fiscal impact study⁸ is not including the location of the development being studied, even though the location of new development is probably the most significant indicator of potential costs to serve any new development. While the location of new development has a very significant potential impact on serving the new development, it does not tend to have a significant impact on revenues; revenues tend to be much the same, irrespective where a development might be located within a city.⁹

In general, sprawling development of any type will tend to be more costly to serve than compact growth patterns, Not only do streets and utility lines have to be extended over greater distances, but patrol times, response times, and other distance related costs increase to serve sprawl.

While this is true for all types of uses, sprawling commercial uses are probably more costly (per unit or per acre of development) than residential uses since the commercial uses must attract people over broader areas, intensifying traffic beyond what would be needed to serve a residential pattern of uses alone. This traffic will, sooner or later, lead to expensive capital expenditures to increase the capacity of streets over a wider area.

Locational cost factors can be very significant in some types of services, such as fire protection. Fire protection costs, as distinct from emergency medical services, are very difficult to allocate between growth in different types of uses because the major cost of fire services (as separate from emergency aid costs) is a function of “readiness.” Responding to actual calls for service tends to be a secondary cost driver.¹⁰ The major driver for fire service costs is in being able to respond in a timely fashion to fire calls within a service area by constructing, equipping and staffing a fire station to provide that service. Although, at some level of growth, increased calls for service in existing stations reach the capacity of the fire station and additional trucks, equipment, and personnel may need to be added to a station, the location of new development outside the desired response area of existing fire stations tends to drive potential fire costs protection more dramatically.

Once a fire station is constructed and staffed, it can usually serve a wide range of population within its response area. In most developed single-family residential areas that are adequately¹¹ served with established fire stations, there is usually capacity to accommodate more infill without significantly needing to add more equipment or staff. Additional residential growth within such areas would then create very little need for additional fire protection costs.¹²

In contrast, the location of any new use can be a very significant cost driver if that development is outside the desired response area of the existing stations; creating or adding to the need for a new station. Similarly, very little additional costs may be required for parks, branch libraries, police substations, etc. when that development is located within the established service area for existing facilities. However, when such development is located outside the service area for a facility, it creates or adds to the demand for a new facility. Once a facility is built, it must be staffed and maintained, adding to the costs of ongoing operations.

Conclusion

The “great recession” illustrates the hazard of placing too much reliance on new commercial development to relieve fiscal pressures. In good economic times, sales tax revenues from new commercial uses often do provide ample revenues to support growing needs for governmental services. However, those revenues are fickle and can quickly decline in the face of economic adversity. Other revenue sources, such as property taxes, utility taxes, state shared revenues etc., that are spread over all uses (residential, commercial, and industrial) tend to not decline as rapidly or as severely. Some of these sources may not decline at all. Communities that have become very reliant on the sales taxes relative to other revenues have been harder hit and have had a more difficult time adjusting to the current recession. Those communities that have a balance of revenues from a diversity of sources tend to fare better.¹³

The location of new development, especially in relation to existing facilities and services, is probably a more significant consideration than the type of use when assessing the potential costs of serving new development. It is important to note that since revenues do not tend to vary with location and costs do, significant benefits can accrue by reducing the costs of serving new development by the more efficient siting of new development.

The desired balance between the types and location of uses for a healthy fiscal capacity is probably the same as it is for creating a healthy community in general. A compact, diverse, and balanced pattern of uses (where no particular use dominates relative to the others) tends to provide more fiscal stability and the potential to generate revenues to support services, compared to a sprawling disconnected pattern of uses or where particular types of uses tend to dominate.

End Note

While the analysis techniques discussed can lead to erroneous conclusion about the fiscal impacts of various uses, they nonetheless can be useful for many types of tasks. Fiscal analysis and forecasting is very complex, and it would be very expensive, if not impossible, to develop analytical models that can precisely determine fiscal relationships. Consequently, analysts have developed various simplified methods, to estimate factors which drive costs and revenues.

Each of the methods critiqued above may be an appropriate tool to apply to certain circumstances.¹⁴ However, the use of any technique must be consistent with the assumptions embedded in each technique which create limitations on how their results should be interpreted. These assumptions must be consistent with the objectives of the analysis. For example, the key assumption embedded in the using per-capita multipliers to forecast police costs is that the balance between the amount of each type of use will be consistent in future with the current balance. This assumption may be appropriate in a general forecast of the fiscal capacity of a city. The problem arises when these generalized techniques are used to determine whether particular uses are more fiscally beneficial, or worse, to set policy regarding whether particular uses should be encouraged or discouraged. These objectives demand a much more rigorous analysis that focuses on the fiscal characteristics of each use as critiqued above.

Footnotes

• 1 In Washington, this actually applies only to special levies. Since “basic education” is financed by the state on a largely per-student basis, single-family dwellings at the local level do “support themselves” for basic education; more students bring more revenue to pay for the basic education of those students.
• 2 This discussion focuses on retail sales tax revenue. Retail sales are only a portion of total sales; consisting primarily of sales taxes associated with consumer oriented businesses. For simplicity sake, the effects internet sales and the Streamlined Sales Tax legislation is not included in this discussion.
• 3 Most studies now tend to recognize that the construction of single-family residences do generate revenue during their construction from the sales tax on new construction.
• 4 Early in my career when doing a fiscal impact study, I asked a small city police chief how he expected the anticipated population growth would affect his department. He said something like “Son, don�t tell me how many people there will be, tell me how many taverns there will be.”
• 5 Cost drivers for fire protection services are discussed below.
• 6 A high proportion of the people charged with crimes in most cities with large commercial centers are people who do not reside in the community.
• 7 This approach, however, usually has a problem in effectively handling the costs associated with traffic “calls for service,” which cannot be easily allocated by use. Anecdotally, I recently encountered a large accident on a freeway running through a city. There were a large number of police, fire, and emergency vehicles responding to the incident. While all of these vehicles were from the city that the freeway ran through the odds would be very greatly against any of the victims in the incident residing there.
• 8 Alas, including even those I prepare.
• 9 However, revenues do vary between jurisdictions.
• 10 Fire protection is like insurance, you need it even if you never use it.
• 11 Adequacy is usually defined by response time.
• 12 Unfortunately, the costs relationship for fire services is often not clear because fire protection in most cities is integrated with emergency aid services, which is directly affected by population growth.
• 13 A common rule of thumb among finance directors is that the revenues from sales taxes, property taxes and utility taxes should be roughly equal, creating a “three-legged stool” of revenue as a fiscal base.
• 14 Indeed I have used all of them and will probably continue to do so.