Guidelines for Analysis of Investments in Bicycle Facilities (2006)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

CHAPTER 4 BENEFIT-COST ANALYSIS OF BICYCLE FACILITIES INTRODUCTION AND PURPOSE Based on the research conducted for this project, the team crafted a set of guidelines to be used by transportation pro- fessionals and government agencies to better integrate the planning of bicycle facilities into the transportation plan- ning process. The web-based guidelines will assist state DOTs and other state, regional, and local agencies in con- sidering bicycling in all transportation projects. Additionally, the guidelines will support local agencies’ review of bicycle projects as part of the transportation improvement plan. Transportation planners will be able to use the guidelines for the following purposes: • Estimating the likely cost of specific facilities based on type and on key characteristics, • Estimating how a facility will impact the overall bicycling environment in an area, and implicitly how it will affect the amount of riding based on characteristics of the facil- ity and of the surrounding area, • If information is available for calibration, estimating the usage of a facility (and the change in usage of comple- mentary and competing facilities), and • Estimating the specific types of benefits and their rela- tive sizes based on characteristics of the facility and of the surrounding area. TRANSLATING DEMAND AND BENEFITS RESEARCH INTO GUIDELINES Demand Estimating the use of a new facility rests on two main assumptions. First, all existing commuter bicyclists near a new facility will shift from some other facility to the new one. Second, the new facility will induce new bicyclists as a function of the number of existing bicyclists. Research for this project uncovered that people are more likely to ride a bicycle if they live within 1,600 m (1 mi) of a facility than if they live outside that distance (Appendix B). The likelihood of bicycling increases even more at 800 m and 400 m. The team therefore estimates existing and induced demand using 400-, 800-, and 1600-m buffers around a facility. 38 Estimates of existing bicycling demand are based on U.S. census journey to work mode shares. Establish the number of residents within 400-, 800-, and 1,600-m buffers of the facil- ity by multiplying the area of each buffer by a user-supplied population density. Multiply the number of residents in each buffer (R) by 0.4, assuming the national averages of 80% of residents are adults and 50% of adults are commuters, to cal- culate the number of daily commuters. Then multiply this number of commuters in each buffer by the region’s bicycle commute share (C). Use the bicycle commute share for the Metropolitan Statistical Area (MSA) as the default value; the user has the option to enter a commute share for the specific area if it is known. Adult commuters represent only a portion of adult bicy- clists. The team compared U.S. census commute shares to NHTS data and found that the total adult bicycling rate ranges from the census commute rate at the low end to 0.6% plus three times the commute rate at the high end (Appen- dix A). This allows the use of readily available census com- mute shares to extrapolate total adult bicycling rates (T). Multiply the estimated low, moderate, and high rates by the number of adults—estimated to be 80% of the population— in each buffer to arrive at the total number of daily adult cyclists. Additional research (Appendix B) found that people who live near a facility are more likely to bike than those that do not; multipliers were developed to describe these probabilities. Multiplying the numbers of both commuters and total adult cyclists by the likelihood multipliers found in this research for various buffers around the proposed facility provides an estimated number of induced cyclists in each group. Total daily existing adult cyclists = • •T Rj 0.8 T C T C T C high moderate low = + = + = 0 6 3 0 4 1 2 . . . Daily existing bicycle commuters = • •R C 0 4.

Where L400m = 2.93 L800m = 2.11 L1600m = 1.39 Mobility Benefit This research, based on stated preference analysis, found that bicycle commuters are willing to spend, on average 20.38 extra minutes per trip to travel on an off-street bicycle trail when the alternative is riding on a street with parked cars (6). Commuters are willing to spend 18.02 min (M) for an on- street bicycle lane without parking and 15.83 min for a lane with parking. Assuming an hourly value of time (V) of $12, the per-trip benefit is $4.08, $3.60, and $3.17, respectively. Multiply the per-trip benefit for the appropriate facility by the number of daily existing and induced commuters, then dou- ble it to include trips both to and from work. This results in a daily mobility benefit. Multiplying the daily benefit by 50 weeks per year and 5 days per week results in the following annual benefit: It should be noted that this methodology assumes that no bicycle facility previously existed nearby, aside from streets with parking. In the this equation, V is divided by 60 because the M is in minutes and V is in hours; dividing V by 60 con- verts it to minutes so that the result can easily be multiplied by the minutes. Health Benefit An annual per-capita cost savings from physical activity of $128 is determined by taking the median value of 10 stud- ies (Appendix E). Then multiply $128 by the total number of new bicyclists to arrive at an annual health benefit. Recreation Benefit The “typical” day involves about 1 hr of total bicycling activity, which is valued at $10 (D), based on a wide variety Annual health benefit total new cyclists= • $128 Annual mobility event existing commut = • • M V 60 ers new commuters+( ) • • •50 5 2 New commuters Existing commuters= −( )( )∑  L d d 1 = = 400 800 1 600, , , New adult cyclists Existing adult cyclists  L d d −( )( ) = ∑ 1 400 800 1 60, , , 0 39 of studies of outdoor recreational activities (Appendix G). From both NHTS and Twin Cities TBI, the average adult cycling day includes about 40 min of cycling. This is the amount used, plus some preparation and cleanup time. Mul- tiply this by the number of new cyclists minus the number of new commuters. (The value of the facility to new commuters is counted in the mobility benefit.) Reduced Auto Use Benefits These benefits apply only to commuter and other utili- tarian travel, because it is assumed that recreational rid- ing does not replace auto travel. These include reduced congestion, reduced air pollution, and user cost savings. Multiply the total benefit per mile by the number of new commuters, multiplied by the average round trip length from NHTS (L). Then consider two offsetting adjustments that ultimately leave the total number unchanged. First, there are utilitar- ian riders in addition to commuters and some of these trips will replace auto trips. Second, not all new bike commuters and utilitarian riders would have made the trip by car; evi- dence from NHTS suggests that something less than half of bike commuters use driving as their secondary commuting mode. For simplicity, assume that these two factors offset each other, and thus the total amount of new bike commuter mileage is a reasonable number to use to represent the total amount of new bike riding substituting for driving. The benefit per mile of replacing auto travel with bicycle travel is a function of location and the time of day. There will be no congestion-reduction benefits in places or at times when there is no congestion. Pollution-reduction benefits will be higher in more densely populated areas and lower else- where. User cost savings will be higher during peak periods when stop-and-go traffic increases the cost of driving. Based on reasoning documented in Appendix G, conges- tion savings will be 0 to 5 cents per mile and pollution sav- ings from 1 to 5 cents per mile depending on conditions. Assume the high end of this range in central city areas, the middle range in suburban areas, and the low end in small towns and rural areas. For simplicity, assume that all com- muting and utilitarian trips are during congested periods. User cost savings were determined to be 3 cents per mile dur- ing congested peak periods and 0 otherwise; thus, these are scaled by location in the same way as congestion savings. Overall, the savings per mile (S) is 13 cents in urban areas, 8 cents in suburban areas, and 1 cent in small towns and rural areas. Reduced auto use benefit new commuters= • • •L S 50 5• Annual recreation benefit new cyclist = • • D 365 s new commuters−( )

40 Costs / Demand / Benefits General Inputs Cost Worksheet Demand / Benefits Inputs Output: Capital cost Annual maintenance cost Output: # of New Cyclists Output: $ Benefits Figure 2. Flow chart showing structure of the guidelines. BENEFIT-COST ANALYSIS TOOL The guidelines, titled Benefit-Cost Analysis of Bicycle Facilities, can be found at http://www.bicyclinginfo.org/ bikecost/. These guidelines provide planners, policy officials, and decisionmakers with the ability to use a standard method to analyze the costs, benefits, and induced demand associated with a planned bike facility in their community. These guide- lines allow the user the ability to tailor the information to a particular project. Figure 2 presents a brief outline of the sequence of logic and process contained within the web-based guidelines. The guidelines include useful accompaniments such as “i” buttons and a “Bicyclopedia.” The “i” buttons help explain variables so that the user can better understand the information that is being requested (Figure 3). The Bicy- clopedia comprises a glossary of terms, a guide on how to use the guidelines, and a brief description of the method- ology that was used to develop the guidelines. The expla- nations associated with the “i” buttons and the glossary appear in separate popup windows so that data entered by the user are retained. General Inputs When determining the benefits and costs of a bicycle facil- ity, the guidelines allow the user to tailor the outputs to the specific project that is being proposed. The user first selects a metropolitan area. Second, the user specifies if the proposed facility will be located in the city or suburbs (Figure 4). Such options allow the user to tailor output to local conditions. The user is asked to specify characteristics of the project that is being proposed. Options include on-street bicycle lanes with or without parking, off-street bicycle trails, and bicycle-related equipment. During the general input portion of the tool, it prompts the user for information about the proposed facility. The infor- mation that the tool requires depends on the different char- acteristics of the facility that the tool is analyzing. The fol- lowing is the detailed logic tree that tool goes through for the general input portion of the tool. The user is prompted by the following general inputs: 1. Are you interested in: Costs, Demand, Benefits? Or a combination of the three? 2. In which metro area will the facility be located? Central city or Suburb? 3. Mid-year of project? 4. Type of facility? a. On-street with parking 1. Restripe 2. Overlay 3. Full Depth 4. Signed Route b. On-street without parking 1. Restripe 2. Overlay 3. Full Depth 4. Signed Route c. Off-street 1. Stone trail 2. Asphalt trail 3. Concrete Trail d. Bicycle-Related Equipment Costs The guidelines prompt the user to select various features of a trail such as dimensions, signals, landscaping, and materials used for the trail. The guidelines provide the user the ability to enter a cost for various features or to use default settings as outlined in Chapter 1 of this report (see Figure 5). Based on such inputs, the tool calculates the cost of the facility.

41 Figure 3. Information from “i” buttons explaining the input variables provided in separate popup windows. Figure 4. General inputs.

Demand The next process estimates the induced demand as a result of the facility. In doing so, the guidelines consider the exist- ing bicycle commute share, residential density at 400, 800, and 1,600 m from the facility, household size, and length of facility. The guidelines have default settings for each metro- politan area. However, the user has the ability to change these figures if better information about the area around the facility is available (Figure 6). Output The guidelines present the user with easy to read tables showing the costs of a new facility and the induced demand and benefits related to mobility, recreation, health, and reduced auto use (Figure 7). APPLICATION TO PEDESTRIAN FACILITIES Introduction Some of the difficulty in planning for bicycling and walk- ing is that the bulk of the literature and subsequent planning tools typically aggregate these two modes. For abstract or general purposes this may suffice; the two modes are almost always aggregated in transportation research. In terms of daily 42 use, facility planning, and community design, bicycling and walking differ substantially. The cost, demand, and benefit tool for this project were developed specifically with bicycle facilities in mind. Many elements of the tool, however, can be applied to pedestrian facilities. For exam- ple, cost information can readily be adapted for pedestrians to the existing cost model, given the constraints described in this section. On the other hand, demand and benefit cal- culations would need to be considerably modified to meet the unique characteristics of pedestrians. This section describes the manner in which the designed tool could be applied to matters of pedestrian planning and some of the issues involved. Important Issues to Consider It is helpful to draw attention to two points when consider- ing specific facilities. The first is that there are facilities devel- oped specifically for pedestrian use and most of the time only pedestrian use (e.g., sidewalks, stairs, and street crossing improvements). Other facilities suitable for bicycle use, how- ever, tend to be used by pedestrians as well. The second point is that pedestrians, in a strict definition, are meant to include people walking, versus variants such as running and skating. There is an important difference in terms of speed and typical distances covered that impact how demand and benefits will be impacted by facilities. Figure 5. Facility costs spreadsheet.

43 Figure 6. Demand inputs. Figure 7. Output of guidelines.

The first distinction between facility types primarily affects cost estimates. Facilities that are intended for mixed use including bicycles can obviously be addressed using the cost tool in the guidelines. However, it may not be possible to esti- mate costs for facilities that are for pedestrian use only, unless the physical characteristics of the facility have a close parallel to the bicycle facilities included in the model. Demand and benefit estimates tend to be more influenced by the second point. Specific facilities and environments that may be very useful to walkers may not be as suitable for higher speed alternatives like running and skating. Conversely, the value of a continuous off-road facility may be greater for higher-speed users, which in turn will influence how far they will go out of their way to use these facilities, which will impact demand. A final matter is that walking is 10 times as common as bicy- cling. An estimated 70% of adults walk at least once per week (in the NHTS baseline sample) while about 7% bike. Because so many people walk already it seems unlikely that new facil- ities will have a significant impact on total demand, although they may influence where walking is done. The major impact on benefits will likely be the value of improved safety or gen- eral conditions for the already large number of existing pedes- trians, rather than the value of additional activity by new users, as was the case with many of the bicycling benefits. Costs Costs and benefits obtained from a self-contained project such as a multipurpose trail, striping of an existing roadway for bicycles, or construction of a sidewalk next to an existing road- way, are relatively easy to estimate by applying the guidelines. However, bicycle facility costs as an element of much larger road construction costs are more difficult to reliably estimate. In like manner, the guidelines can be more readily applied to self-contained pedestrian paths than sidewalks built as part of roadway projects, particularly in urban environments. Urban roadway projects will often include essential elements to create a quality pedestrian facility which include not only the sidewalk material but also street trees, landscaping, benches, trash bins, and public art—all of which are often important to creating inviting pedestrian environments. The cost model was designed to estimate bicycle facility construction and design cost for all types of on-street and off-street facilities, equipment such as racks, and real estate 44 costs. Bicycle facilities may exclude pedestrians or be shared with pedestrians and other non-motorized users. Table 8 identifies, based on location, shared facilities for cyclists and pedestrians as well as exclusive facilities. Having identified the correspondence between the bicycle and pedestrian facility type, the next step is to identify the construction elements required for exclusive pedestrian facil- ities. Updating the cost model to include construction ele- ments for sidewalks and trails would require minimal effort as most construction elements are already included in the cost model. Table 9 identifies construction elements that are exclu- sive to bicycles and pedestrians as well as those that are shared. Note that most elements are shared and included in the existing model. To use the tool to estimate the cost of the bicycle facility, the user inputs information (such as path width) for the applicable elements of the pedestrian facility type, thereby generating the appropriate cost estimate. Demand Demand in terms of new bicycle users is generated based on an extrapolation from existing mode share of bicycle com- muters to find the number of new users attributed to a given facility. This is based on two different calculations: first, esti- mating the total number of bicyclists in an area based on the commute share and second, estimating the number of new cyclists as a function of the current number. Both of these calculations are based on the results of the research done for this project. Development of an equivalent demand estimation method- ology for pedestrians should be considered carefully. There is no reason to believe a priori that observed relationships between recreational and commuter bicycling would hold for walking or other potential facility uses. Nor is there any reason to think that there would be a similar relationship between marginal improvements to facilities and the amount of walking. Indeed, there are good reasons to believe that these relationships would not hold. 1. Adequate facilities for walking are much more wide- spread (e.g., sidewalks). There may not be good facili- ties on specific roads, but most people have some place that is reasonable for walking. In general, because of the Location Bicycle Shared Pedestrian Street Bike Lane Paved Shoulder Shared Street Shared Lane - - Adjacent Street Side Path - Sidewalk Cycle Track - - Off-Street Multipurpose Path Walking Trail Stairs TABLE 8 Conveyance

wealth of existing pedestrian facilities, new pedestrian facilities are less likely to create major changes to the overall opportunity set as they often do for cycling. They seem unlikely to have a major impact on the total amount of walking, although they may impact where it is done and the benefits that it provides (e.g., providing a side- walk connecting two formerly auto-oriented centers). 2. In more developed areas, pedestrians are generally sep- arated from traffic; therefore, off-road facilities do not create the same kind of unique advantages that they do for bicycles, at least in the case of walking (this would be different in rural areas). However, for run- ners and skaters, their higher speeds and longer dis- tances could mean that facilities without frequent intersections could be advantageous. 3. The relevant travelshed for walking is smaller than for cycling. Most walking trips are quite short; unless a facility is extremely close to a person’s home or work, they are unlikely to use it much. There could still be drive-in traffic for recreational walking, but only if the facility is special in some way (e.g., scenic). 4. Cycling also tends to be more influenced by attitudes and facilities; rates of commuting are an indicator of these attitudes and facilities that can also be used to predict recreational cycling. Commuting and recre- ational cycling to some extent benefit from the same kinds of conditions. Walking does not seem subject to the same cultural and facility limitations. Furthermore, commuting by walking is strongly constrained by local land use density due to the short distances involved, whereas walking for recreation is not. So there is no rea- son to believe that there would be a strong relationship between commuting and total walking at a neighbor- hood or even an urban area level, as there appears to be for cycling. The conclusion that follows from these points is that not only is the bicycling demand model in the guidelines not applicable to pedestrians; in fact, it cannot even be adapted because of the significant differences in the underlying circumstances for the two user types. In two respects, estimating pedestrian demand is simpler and more reliable than estimating bicycling demand. First, because walking is so much more common, there will not be such large variations across different cities or between differ- ent parts of the same city. Because it is more common, it would be much easier to complete representative counts in a given area to estimate local demand. Furthermore, local variations are smaller; it is therefore appropriate to estimate demand on a new or improved facility based on known demand on a similar facility at a different location. The second reason estimating pedestrian demand is more straightforward stems from the fact that walking facilities are also more common. For any facility with given charac- teristics—a sidewalk in a suburban commercial area—it is likely that one or more very similar facilities exist in the same urban area, and if not, then almost certainly in some other similar city. Thus estimating demand by comparing existing facilities is more feasible for walking facilities than it is for bicycling facilities, and probably more accurate as well. This is likely true of mixed use as well as pedestrian-only facilities, although some basic research to confirm these hypotheses would be valuable. Benefits Three bicycling benefits included in the guidelines rely in part on demand estimates. The number of new bicyclists is multiplied by a per-person dollar amount to calculate health, recreation, and mobility benefits. Of these, health and 45 Bicycle Shared Pedestrian Pavement Markings Earthwork Benches Bicycle Parking Pavement Bus Racks Drainage Landscaping Bridges Underpasses Signs Traffic Signals Barriers Lighting Security Real Estate Operations Costs Maintenance Costs TABLE 9 Construction elements

recreation could directly apply to pedestrian facilities with- out further study, assuming an established methodology for estimating induced pedestrian demand. It could be assumed that pedestrian facilities would generate $128 per year in health benefits to new pedestrians who did not formerly engage in physical activity. The same assumption would hold true for recreation; new pedestrians would value their recreational time at the same rate as their cycling counter- parts. The mobility benefit also relies on an estimate of demand, but cannot be applied directly to pedestrian facilities. The methodology employed to assign value to bicycle facilities did not consider pedestrian facilities. To determine the mobil- ity benefit, the research team used an adaptive stated prefer- 46 ence survey to assign value to five types of bicycle facilities. A similar survey could be used for pedestrian facilities, but would require a new framework for considering different types of facilities. The externalities benefit assumes that new bicycle com- muters induced by a new facility will generate benefits in the form of less congestion and air pollution. The former relies on the fact that bicycle facilities typically separate cyclists from autos. Applying this logic to pedestrian facilities presents the problem that increased pedestrian traffic, even on sidewalks, can actually increase automobile traffic congestion at inter- sections. For this reason, the methodology employed to cal- culate the externalities benefit would need to be substantially modified to apply it to pedestrian facilities.