Case Study:

Envision Utah

Methodology

Infrastructure Costs

Overview

The purpose of the infrastructure cost assessment was to provide reasonable approximations of the infrastructure costs associated with new residential development at the scale of the Greater Wasatch Area. The infrastructure cost analysis was conducted by the Utah Governor's Office of Planning and Budget, which developed the Infrastructure Cost Assessment Model in collaboration with Psomas Engineering and the Utah Division of Water Resources. Project sponsors note that there are various uncertainties associated with the cost estimates. The estimates provided by this tool are not considered to be exact predictions of infrastructure costs, but instead are reasonable estimates that provide a relative understanding of how land use effects public investment in infrastructure.

Three levels of infrastructure costs were identified:

1. Regional infrastructure, including regional roads, transit, and water supply facilities. These projects are planned by regional or state governments and financed by state and or federal funds.
2. Subregional (off-site) infrastructure, maintained by the municipality or service district. This includes water and waste water treatment facilities along with distribution lines, storm drain lines and basins, and minor arterial roads. These projects are financed by local governments through the sale of bonds, levying of impact fees, or use of tax revenues.
3. On-site infrastructure provided by developers. On-site infrastructure is classified into the categories of roads, water transmission lines, sewer transmission lines, dry utilities (telephone, electric, etc.), and storm drains. Private developers generally finance the bulk of on-site infrastructure and reclaim their money through the sale of improved lots.

Separate methods were developed to assess infrastructure costs at each of these levels. For off-site and on-site infrastructure, a two-step procedure was followed: first, cost estimates were produced on a per-unit basis by type and density of development; and second, total regional costs were estimated from per-unit costs, according to development characteristics under each scenario.

Regional Infrastructure

The road and transit networks and their associated costs under each scenario were developed by the Wasatch Front Regional Council and Mountainland Association of Governments. Transportation modeling results for the various scenarios indicated that much of the same road network will be required as for the baseline scenario, but total costs will be lower because some road expansion projects are not needed. Transit costs are greater under Scenarios C, D, and the Quality Growth Strategy, although the cost-effectiveness of this service is also greater since population is clustered around transit hubs.

The water supply facilities needed and the cost of these facilities under each scenario were developed by the Division of Water Resources, in cooperation with various municipal water agencies. Savings in water demand from the Quality Growth Strategy were substantial enough to delay some parts of the baseline regional water projects, thus reducing the regional component of water supply costs.

Off-Site Infrastructure Costs per Unit

Per unit estimates were prepared by Psomas Engineering through a collaborative effort with the Division of Water Resources and the Governor's Office of Planning and Budget. Experts working in local, regional and state government also provided input into the development of these estimates. The bulk of the information used in estimating municipal costs came from municipal impact fee studies of selected municipals and special districts, in conjunction with an analysis sponsored by the Utah League of Cities and Towns on the compliance of municipals with the Utah Impact Fee Act. Based on this information and on professional judgment, the median impact fees for roads, water, sewer and storm drains were estimated.

Infill and reuse developments benefit in that much of the needed off-site infrastructure in the form of roads and storm drains are already in place. Psomas Engineering found that municipal roads and storm drains make up 20 to 23 percent of the total impact fees levied. Based on these findings it was recommended that the median impact fees be reduced by 20 percent for infill and 15 percent for reuse. The end result of this process was the schedule of off-site cost by density and land use shown in Table 1. These costs were estimated from local data in the Greater Wasatch region of Utah and are not necessarily applicable to other parts of the country.

Table 1. Off-Site Cost Schedule in Envision Utah Study¹

¹Off-site costs include water and waste water treatment facilities, distribution lines, storm drain lines and basins, and minor arterial roads maintained by the municipality or service district.

Work is planned in 2000 to develop an engineering cost model for off-site costs, similar to the on-site cost model. The purpose is to develop a model that can estimate costs for specific development proposals and locations, rather than just considering average regional characteristics by development density and type.

On-Site Infrastructure Costs per Unit

On-site infrastructure costs proved to be the most difficult and complex level of costs to estimate on a per-unit basis. The complexity and difficulty is due to the multiple ways subdivisions can be designed based on size, available land, and the negotiating that takes place between developers and municipalities. To simplify these and other variables, a simulation model was designed by Psomas Engineering to predict a mean cost estimate per unit. The model has the ability to create estimates by density and land type.

The simulation model was prepared based on actual estimates and sketch designs. These estimates and sketches are composed of detailed information of the infrastructure that is necessary to prepare a parcel of land for residential development based on conformity to Salt Lake City building code. The simulation model uses input data for parcel size and density, and then calculates a standardized lot size (net lot size, depth, and width), the relationship of the lots within the parcel (rows, tiers, block length, and street width), and the quantity of materials required (square feet of roads and linear feet of water, sanitary, storm drain, dry utilities). Cost coefficients per quantity of material are then applied. The on-site cost model was developed for the most stringent local development standards in the region (Salt Lake City), but was found to be 98 percent accurate when applied to other site plans.

Because this model is built out of mathematical relationships, it is possible to insert a variety of densities into the model and receive varied per unit costs by density. Four models in total were developed for different land uses. An index model serves as the base model; a raw land model accounts for peripheral roads that are general to raw land development; the reuse model accounts for demolition costs and a parcel size of 10 acres; and the infill model is based on a parcel size of five acres. The resulting on-site cost schedule is shown in Table 2. Once again, these costs were estimated from local data in the Greater Wasatch region of Utah and are not necessarily applicable to other parts of the country.

Table 2. On-Site Cost Schedule in Envision Utah Study¹

¹On-site costs include roads, water transmission lines, sewer transmission lines, dry utilities, and storm drains provided by developer.

Off-Site and On-Site Infrastructure Costs by Scenario

To estimate total off-site and on-site infrastructure costs for each scenario, the GIS land use data contained in the 50.3-meter grid cell format were utilized. Each cell contains a value that indicates the number of units located within that cell. The baseline was prepared by taking current land use and populating cells to represent the baseline population controls. Average household size was used to match population with the number of new units.

The Quality Growth Strategy represents development as occurring through eight broad development types. Each type is defined by specific land use characteristics such as residential area, density, jobs, and population. GIS cells received a code representing one of these eight development types. The cost model used the GIS to count up cells based on development and type, to derive the number of new housing units by density and type of land use (raw, infill, reuse). Reuse is defined as cells with new population in 2020 in addition to existing population as of 1997, while infill development is defined as cells surrounded by cells that were heavily developed as of 1997. Raw land is all development occurring on the peripheral of existing 1997 population.

The GIS exercise resulted in around 30,000 cell aggregations of like cells next to one another for the baseline and about 50,000 for the Quality Growth Strategy. Piecewise log linear mathematical functions were fit to the on- and off-site cost schedules. Clusters were then sorted by land use and by density, and were assigned to functions based on density and type.

The on-site infrastructure cost model continues to be improved. It was refined between the analysis of the four scenarios and the analysis of the Quality Growth Strategy, and will continue to be refined in future work.

Level of Effort

The development of the off-site and on-site infrastructure cost functions involved roughly two-thirds of a staff person's time for over a year, as well as roughly one person-week each for a water quality and wastewater engineer. Costs for an outside engineering firm included roughly $10,000 for the municipal (off-site) analysis and $20,000 for development of the on-site cost analysis models. The on-site model went through a number of iterations before staff were satisfied with its accuracy.

To apply the cost models to the land use data for each scenario, two approaches were possible. The first was to code the cost functions in GIS, which took roughly two to three person-days to write, check, and run the script and analyze results. The second approach was to use GIS to do a quick aggregation of the land use data (e.g., land development by type and density), and then apply cost functions in an Excel spreadsheet. This approach turned out to be quicker, taking roughly a half day.

[TOP]