Cover Image

Not for Sale



View/Hide Left Panel
Click for next page ( 78


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 77
42 Estimating Soft Costs for Major Public Transportation Fixed Guideway Projects LIGHT RAIL HEAVY RAIL LIGHT + HEAVY RAIL 100% Soft Costs (% of Construction) 100% Soft Costs (% of Construction) 100% Soft Costs (% of Construction) 90% 90% 90% 80% 80% 80% 70% 70% 70% R2 = 0.05 R2 = 0.00 R2 = 0.02 60% 60% 60% 50% 50% 50% 40% 40% 40% 30% 30% 30% 20% 20% 20% 10% 10% 10% 0% 0% 0% $- $2 $4 $6 $8 $- $5 $10 $15 $- $5 $10 $15 Soft Costs per LF (000) (2008$) Soft Costs per LF (000) (2008$) Soft Costs per LF (000) (2008$) R2 = 0.05 t-Stat = 1.07 R2 = 0.00 t-Stat = 0.09 R2 = 0.02 t-Stat = -0.81 Figure 24. Soft costs as a percentage of construction versus soft cost per linear foot of constructed guideway. It is unclear from these initial analyses which basic measurement of soft costs (percentage or dollar value terms) is most appropriate. Therefore this analysis shows results with both unless one measure appears more appropriate given the circumstances. 4.5. Relationships between Cost Drivers and Historical Soft Costs This section tests the relationship between various project characteristics such as mode, align- ment, and year (detailed above in Table 13) and actual soft cost expenditures. This research took two approaches to measuring how soft cost drivers have impacted actual soft costs: Univariate testing of soft cost drivers suggested in interviews and the questionnaire. First, this research began by creating a series of scatter diagrams comparing soft costs with the kinds of project characteristics that estimators currently use to choose higher or lower soft cost per- centages. This kind of analysis tests only whether one project characteristic alone influences soft costs. As the results below demonstrate, some of these tests showed that soft costs are cor- related with certain project characteristics, while other tests yielded less conclusive results. Many of the less conclusive results are presented in Appendix C. These single-variable results served to guide the research into the next phase described in Section 4.5.7. Multivariate testing of combinations of soft cost drivers. Second, this research tested a multitude of combinations of soft cost drivers and their effect on soft costs in a multivariate regression. Project characteristics were the independent variables, and soft costs as percent of construction costs acted as the dependent variable. After several hundred tests, a single multi- variate regression was developed that can explain approximately 60% of the differences in soft cost percentages by variations in project characteristics (R2 = 0.58), as will be described later. This kind of analysis tests the cumulative effect of how changes in a variety of project attri- butes have affected resulting soft costs. 4.5.1. Assembling Data on Soft Cost Drivers A set of characteristics was gathered for the projects to help identify cost relationships, including the following: Physical attributes, such as alignment length, profile (e.g., below grade, at grade, aerial), number of stations, or whether the project initiated new service or extended an existing line.

OCR for page 77
As-Built Soft Cost Analysis 43 Installation conditions, such as whether the project interacted with other active rail transit lines. Schedule information, including major milestones in the project lifecycle for a subset of projects in the dataset. While each project had a midyear of expenditure, only some projects had full schedule data available. Characteristics of the project sponsor, such as experience level, internal policies on capital costs, and use of outside contractors. The context of the project development process, such as the level of public involvement, delivery method, or whether a significant redesign was necessary. For these last two types of characteristics, the definition and determination of values required some judgment based on knowledge of those projects' development process. Many measures were derived from this primary dataset that were intended to act as a proxy to capture other project characteristics, such as project magnitude (e.g., construction costs per linear foot), complexity (e.g., percent of alignment below grade), unique circumstances (e.g., real estate acquisition costs, project occurred prior to certain federal requirements), and many others. 4.5.2. Soft Costs by Mode and Year Figure 25 shows the average soft costs as percentage of construction costs across modes and by decade. The amount spent on soft costs appears to vary little depending on mode, as indicated in the left pane. Light rail projects averaged 33.8%, heavy rail projects averaged 28.0%, and the combined database projects averaged 30.9% of soft cost percentage of construction. Soft costs have been rising over time since the 1970s. The right pane of Figure 25 shows that on average, soft costs for both heavy and light rail have recently amounted to approximately 34.6% of construction costs, and this figure is an increase from about 21.4% three decades ago. 4.5.3. Soft Costs by Project Delivery Method Project delivery method or procurement strategy also appears to affect expenditures on soft costs. Although most projects in the dataset were delivered via a DBB methodology, evidence for light rail projects indicates that DB projects have lower soft costs, as shown in Figure 26. With only nine designbuild projects and one construction management/general contractor (CM/GC) project out of all database projects, these findings need to be considered within the limitations caused by the small sample size. 50% 40% Soft Costs (% of Construction Costs) Soft Costs (% of Construction Costs) 33.0% 34.6% 45% 35% 40% 27.7% 33.8% 30% 35% 30.9% 21.4% 28.0% 25% 30% 25% 20% 20% 15% 15% 10% 10% 5% 5% 0% 0% Light Rail Heavy Rail All Modes 1970s 1980s 1990s 2000s Sample Size: 25 26 51 6 14 10 21 Figure 25. Average soft costs by mode and by decade.

OCR for page 77
44 Estimating Soft Costs for Major Public Transportation Fixed Guideway Projects LIGHT RAIL HEAVY RAIL LIGHT + HEAVY RAIL 40% 40% 40% Soft Costs (% of Construction) Soft Costs (% of Construction) Soft Costs (% of Construction) 36.3% 35% 35% 35% 29.9% 31.1% 29.9% 27.3% 30% 30% 30% 25.6% 25.6% 25% 25% 25% 20% 20% 20% 15% 15% 15% 10% 10% 10% 5% 5% 5% 0% 0% 0% DBB DB CM/GC DBB DB CM/GC DBB DB CM/GC Sample Size: 16 8 1 22 0 0 38 8 1 Figure 26. Soft costs as a percentage of construction versus project delivery method. Projects selected for designbuild delivery method may be chosen for their simplicity, how- ever, so care should be exercised when considering the above chart. An agency may choose to advertise for designbuild projects that would incur low soft costs regardless of delivery method. The Hudson-Bergen project, for example, was delivered with a designbuild contract, which may have contributed to lower soft costs. Alternatively, designbuild contractors may classify soft costs in different ways than a public agency (e.g., in the construction line item), which might make soft costs appear lower. One of the problems with these delivery methods is that they are not yet very common in the United States, and transit agencies may not fully understand them. Some transit agencies may award a designbuild or other alternative delivery contract but then continue to perform engi- neering work in a more traditional project delivery mode, unknowingly duplicating soft costs. 4.5.4. Soft Costs by Project Development Schedule Figure 27 shows the effect of pre-construction duration (from planning/DEIS to construction phases) on soft costs in dollar terms. Total soft costs are presented in the left pane, and engineer- LIGHT + HEAVY RAIL: ALL SOFT COSTS L IGHT + HEAVY RAIL: PE + FD COSTS ONLY $18,000 $7,000 PE + FD Costs per LF (2008$) Soft Costs per LF (2008$) $16,000 $6,000 $14,000 $5,000 $12,000 $10,000 $4,000 $8,000 $3,000 $6,000 R2 = 0.24 $2,000 2 $4,000 R = 0.22 $1,000 $2,000 $- $- - 2 4 6 8 10 12 - 2 4 6 8 10 12 Years Elapsed between Planning/DEIS and Years Elapsed between Planning/DEIS and Construction Construction Sample Size: 11 R2 = 0.24 t-Stat =1.69 Sample Size: 11 R2 = 0.22 t-Stat =1.58 Figure 27. Soft costs per linear foot versus years elapsed between completion of the draft environmental impact statement and construction.

OCR for page 77
As-Built Soft Cost Analysis 45 ing costs (preliminary engineering and final design) are presented in the right pane. In the left pane the results are pronounced, from zero soft cost at 4 years to a maximum of about $16,000 per linear foot at about 15 years between the DEIS completion and construction. This relation- ship holds for engineering soft costs as well, as shown in the right pane. This finding seems to suggest that the duration of pre-construction phases should be consid- ered within the estimate of soft costs. However, the findings in Figure 27 may simply show that costly projects take longer to plan and design. The relatively small sample size (11) and the role of one relatively costly project in this chart should be recognized in a careful consideration of these findings. 4.5.5. Soft Costs by Project Complexity The remainder of the univariate analysis focuses on project characteristics that address com- plexity (such as percentage of guideway not at grade), number of stations, and other factors and the impact of these characteristics on soft costs. In general, indicators of complexity tend to cor- relate well with soft costs when measured in dollar terms per linear foot. Many of these relation- ships where soft costs are measured as a percentage of construction costs are presented in the appendices. The following figures compare soft cost percentages to the project's alignment profile and typify many of the other results addressing project complexity. The alignment profile of new rail construction can substantially influence the technical com- plexity of the project. In the proposed hypothesis, as the proportion of guideway that is not at grade (in tunnels, on aerial structures, etc.) increases, complexity increases, and soft costs may increase likewise. In the first part of this analysis, "not at grade" is defined as an aerial structure, built-up fill, underground cut and cover, underground tunnel, or retained cut or fill guideway. Figure 28 shows little correlation between the proportion of alignment not at grade and soft costs as a percentage of construction costs. The light rail soft cost percentage is flat at about 40%, while heavy rail shows an increasing trend in the soft cost percentage from 25% to about 35%. The combined project database is flat at about 38%. The statistical trend line for all three relation- ships shows a very weak correlation (R2 less than 0.04), and all relationships are statistically insignificant. The issue of project complexity can be examined in another way, by measuring soft costs in terms of dollar per linear foot. Figure 29 expresses soft costs in dollars per linear foot and shows that soft costs indeed rise as greater portions of the alignment are not at grade. LIGHT RAIL HEAVY RAIL LIGHT + HEAVY RAIL 60% 60% 60% Soft Costs (% of Construction) Soft Costs (% of Construction) Soft Costs (% of Construction) 50% 50% 50% 40% 40% 40% 30% 30% 30% 20% 20% 20% 10% 10% 10% 2 2 2 R = 0.04 R = 0.04 R = 0.01 0% 0% 0% 0% 25% 50% 75% 100% 0% 25% 50% 75% 100% 0% 25% 50% 75% 100% % of Guideway Not At-Grade % of Guideway Not At-Grade % of Guideway Not At-Grade 2 2 2 R = 0.04 t-Stat = 0.94 R = 0.04 t-Stat = 0.941 R = 0.01 t-Stat: -0.51 Figure 28. Soft costs as a percentage of construction versus percentage of guideway not at grade.

OCR for page 77
46 Estimating Soft Costs for Major Public Transportation Fixed Guideway Projects LIGHT RAIL HEAVY RAIL LIGHT + HEAVY RAIL $12 Soft Costs (000) per Linear Foot $12 $12 Soft Costs (000) per Linear Foot Soft Costs (000) per Linear Foot 2 2 2 R = 0.35 R = 0.01 R = 0.36 $10 $10 $10 $8 $8 $8 $6 $6 $6 $4 $4 $4 $2 $2 $2 $- $- $- 50% 100% 0% 25% 50% 75% 100% 0% 25% 50% 75% 100% 0% 25% 75% % Guideway Not At-Grade % Guideway Not At-Grade % Guideway Not At-Grade 2 2 2 R = 0.35 t-Stat = 3.45 R = 0.01 t-Stat = 0.52 R = 0.36 t-Stat: 4.88 Figure 29. Soft costs per linear foot versus percentage of guideway not at grade. As Figure 29 shows, the relationship between the percent of guideway not at grade and soft costs per linear foot is statistically significant for light rail, and for both modes combined, but not for heavy rail alone. Indeed, the R2 value for light rail indicates that the proportion of guide- way not at grade can explain about half of the variation in soft costs per linear foot for heavy rail projects. As the not-at-grade percentage of the projects increase, the soft costs as measured in dollar value terms per linear foot increase. So while the dollar value of soft costs does measurably increase with project complexity as shown in Figure 29, the pattern is not significant enough to increase soft costs in percentage terms, as demonstrated in Figure 28. Although it is tempting to measure soft costs in dollar value terms because this measure pro- duces more correlation with complexity variables, it is worth exploring the measure further. One benefit of measuring soft costs in percentage terms is that the measure controls for variations in unit costs. Soft cost requirements of more expensive projects can be consistently compared to inexpensive projects in percentage terms. Measuring soft costs in dollars-per-linear-foot terms risks autocorrelation between unit costs--high soft costs could be correlated with higher other costs. In general, Figure 30 tends to confirm this hypothesis: in dollar terms, soft costs increase proportionately to construction costs. The correlations shown are strong and statistically signif- LIGHT RAIL HEAVY RAIL LIGHT + HEAVY RAIL $100,000 $100,000 $100,000 Soft Costs per Linear Foot Soft Costs per Linear Foot Soft Costs per Linear Foot 2 2 2 R = 0.51 R = 0.68 R = 0.56 $10,000 $10,000 $10,000 $1,000 $1,000 $1,000 $100 $100 $100 $1,000 $10,000 $100,000 $1,000 $10,000 $100,000 $1,000 $10,000 $100,000 Construction Cost (2008$) per Construction Cost (2008$) per Construction Cost (2008$) per Linear Foot Linear Foot Linear Foot 2 2 2 R = 0.56 t-Stat = 5.31 R = 0.51 t-Stat = 4.487 R = 0.68 t-Stat: 9.48 Figure 30. Soft costs per linear foot versus construction costs per linear foot on a logarithmic scale.

OCR for page 77
As-Built Soft Cost Analysis 47 icant for both modes and the combined database. This trend may help explain why soft costs measured in percentage terms appear unrelated to many other variables like alignment profile-- these other variables may simply drive up construction costs at the same rate. 4.5.6. Soft Costs by Other Characteristics As the questionnaire responses and interviews with cost estimators indicated, other impor- tant determinants of soft costs are the characteristics of the sponsor agency, and the political, operational, or other circumstances under which the project is being developed. Figure 31 shows the correlation between soft costs and the experience level of the sponsor agency (in the left pane), and the installation conditions of the project (in the right pane). The left pane shows a rough spectrum of experience levels across the x-axis, from inexperienced on the left to fairly experienced with both mode and delivery/procurement method at right. The experience level of the project sponsor has a mixed correlation with soft costs. The right pane of figure 31 shows how the level of a project's interaction with existing transit service can affect soft costs. Specifically, the more a project must coordinate with and work around other services, the more soft costs tend to increase in percentage terms. A project to con- struct a new, stand-alone transit line that is not adjacent to any previous service seems to require less design costs than projects to extend or expand an existing rail line. When a construction proj- ect interacts with existing transit service in any way, more engineering and design work has typ- ically been required in the final design phase. Working on or near an active rail right-of-way poses additional logistical challenges that must be planned for, and may also trigger additional safety requirements. Extending a rail line will mean integrating the new track and station(s) into the older infrastructure, and additional work is usually required to ensure that signal, power, safety, and other systems operate compatibly. Figure 32 summarizes the relationship between soft costs and three other project characteris- tics: whether the project required a direct interface with existing service, whether political or pub- lic influence was unusually high, and whether public involvement or opposition was significant. As Figure 32 shows, a project that requires a direct connection or interface with existing revenue service, such as a line extension, a new branch intersecting an existing line, or the rehabilitation of an existing line, tends to show somewhat higher soft costs. Projects where political influence Experience Level of Sponsor Agency at the Time Installation Under Active Revenue Service Soft Costs (% of Construction) 40% 45% 34.1% 34.5% 32.2% 41.5% Soft Costs (% of Construction) 35% 27.3% 40% 30% 25.8% 35% 31.3% 25% 28.1% 30% 20% 25% 15% 20% 10% 15% 5% 10% 0% No experience No recent Experience Recent Recent 5% with Mode or experience with mode, experience experience 0% Procurement with mode or new with mode, with mode and tunnel, other procurement not with procurement No Active Service Adjacent Active Rebuild Under rail or past procurement Service Operation experience Sample Size: 14 9 4 3 21 17 31 3 Figure 31. Soft costs versus sponsor experience level and installation conditions.

OCR for page 77
48 Estimating Soft Costs for Major Public Transportation Fixed Guideway Projects Direct Interface w/Existing Service Unusually High Political or Public Influence Significant Public Involvement or Opposition 40% Soft Costs (% of Construction) 35% Soft Costs (% of Construction) 40% Soft Costs (% of Construction) 31.9% 33.4% 29.0% 35% 35.1% 30% 35% 29.2% 29.8% 30% 30% 25% 25% 25% 20% 20% 20% 15% 15% 15% 10% 10% 10% 5% 5% 5% 0% 0% 0% True False True False True False Sample Size: 32 19 20 31 10 41 Figure 32. Soft costs versus installation conditions, political influence, and public involvement. is unusually high or where public involvement or opposition is significant also tend to be correlated with higher soft cost percentages. Figure 33 compares soft cost percentages to the sponsor agency's tendency to use outside con- tractors to varying degrees, to whether the project was ever required to be redesigned for any rea- son, and to whether the project's planning phase was unusually long. As the left pane shows, sponsors that make more extensive use of outside contractors in early project development phases to design and plan tend to incur somewhat higher soft cost expenditures. However, spon- sors who use contractors in both the development and construction phases do not typically see significant differences in soft cost percentages. The middle pane of Figure 33 shows that the two projects in the dataset that had to undergo significant redesign do not show significantly different soft cost percentages. The right pane of Figure 33 demonstrates that when projects remain in development stages for an unusually long period of time (beyond approximately five to seven years), their soft cost per- centages tend to increase. A significant component of engineering and design soft cost is simply the salary and benefit costs of planners working on the project. When the planning phases for a project take an unusually long time, these costs tend to continue to be charged to the project, increasing Significant Use of Contractors Significant Redesign Required Abnormally Lengthy Project Development Soft Costs (% of Construction) 45% 40% 40% 39.7% Soft Costs (% of Construction) 40% Soft Costs (% of Construction) 41.3% 35% 35% 35% 30.6% 30.9% 30.1% 29.5% 29.4% 30% 30% 30% 25% 25% 25% 20% 15% 20% 20% 10% 15% 15% 5% 0% 10% 10% Little roles More Extensive outside of extensive use in project 5% 5% traditional use in project development development and 0% 0% construction True False True False Sample Size: 38 4 9 2 49 7 44 Figure 33. Soft costs versus use of contractors, redesign required, and lengthy project development phase.

OCR for page 77
As-Built Soft Cost Analysis 49 overall soft costs. When a significant amount of time elapses between entering preliminary engi- neering and the beginning of construction, projects incur higher soft cost percentages. 4.5.7. Soft Costs by Multiple Project Characteristics So far, this analysis has focused on testing the cost relationship between soft cost percentages and a project's characteristics one variable at a time. The next step of this analysis tests the ability of a number of variables in combination to predict the variability in soft costs between projects using multivariate regression techniques. To do this, various combinations of variables were tested, including those variables that did not show particularly strong correlations in the univariate analysis. Soft costs as a percentage of construction costs was the dependent variable, and different combinations of project characteris- tics were the independent variables. Variables that described broadly similar project characteristics were grouped, and the relative contribution of each variable to the overall predictive power (R2) of the regression was measured. In an iterative fashion, one or several variables for each broad facet of the project were retained while many other indicators were left out. The following describes the variables tested and the resulting decision. Project Magnitude The variables with the best ability to predict soft cost percentages were alignment length (in linear feet) and construction costs, adjusted to 2008 dollars. It may seem counterintuitive that alignment length and construction cost in combination pro- duced opposite signs since both measures broadly describe the magnitude of the project. However, these two measures in tandem are good predictors of soft costs and produce better statistical re- sults together than either of them alone, one divided by another, or other measures of project magnitude such as number of stations or station density. The two variables together capture the special cases where short, expensive projects (such as a tunnel project) or long, less- expensive projects (such as service on existing right-of-way or in less developed areas) may tend to demonstrate differing soft costs. Several other variables describing the magnitude of a project were tested but were eliminated since they contributed relatively less to the regression analysis: Construction costs per linear foot, ROW costs as a percentage of construction, Vehicle costs as a percentage of construction, and Number of stations. Project Complexity Of many measures of project complexity, its mode, an indicator of installation conditions (i.e., whether the project is a new standalone line with no active adjacent service or not), and an indicator of an unusually lengthy project development phase were the best predictors of soft cost percentages. Heavy rail projects tend to incur somewhat higher soft costs than light rail, other things being equal, perhaps due to their relative complexity. This finding contrasts somewhat with that of Figure 25 because this multivariate regression controls for other factors influencing soft costs. Heavy rail projects can typically involve constructing guideway and systems that have been designed to more rigorous engineering standards that support more complex systems, move higher passenger volumes, and operate at higher speeds relative to light rail. This finding in the multivariate analysis confirms the results of the industry questionnaire and the interviews with cost estimators.

OCR for page 77
50 Estimating Soft Costs for Major Public Transportation Fixed Guideway Projects A project to construct a new, stand-alone transit line that is not adjacent to any previous ser- vice will usually require less design costs than projects to extend or expand an existing rail line. When a construction project interacts with existing transit service in any way, more engineering and design work has typically been required in the final design phase. Working on or near an active rail right-of-way poses additional logistical challenges that must be planned for and may also trigger additional safety requirements. Extending a rail line will mean integrating the new track and station(s) into the older infrastructure, and additional work is usually required to ensure that signal, power, safety, and other systems operate compatibly. Note that this variable is not statistically significant to a high degree of certainty (t-statistic of -1.25). A significant component of engineering and design cost is simply the salary and benefit costs of planners working on the project. When the planning phases for a project take an unusually long time, beyond approximately five to seven years, these costs tend to continue to be charged to the project, increasing overall soft costs. Other variables were eliminated due to their relatively low contribution to the regression analysis' predictive power: Station density (number of stations per mile of guideway constructed), Percentage of guideway below grade, Percentage of guideway not at grade, Rebuild or rehabilitation under operation (dummy variable), Project type (new service, extension of existing service, or rehabilitation of existing service), and Direct interface with existing revenue service required. Delivery Method A dummy variable indicating whether the project sponsor chose an alternative project delivery method (i.e., a method that is not the traditional designbidbuild) contributed the most to the regression analysis in a statistically significant way. Including the specific effects of a certain kind of alternative delivery method did not strengthen the regression analysis, primarily due to the small sample size of such projects. When sponsors choose to procure their project through an alternative delivery mechanism such as designbuild, designbuildownmaintain, or construction manager/general contractor, these projects have historically incurred lower soft costs. In addition, these alternative delivery methods tend to frontload more design and planning costs in preliminary engineering. However, the lower soft costs of projects implemented with alternative delivery methods may be partially the result of differences in measurement rather than a real reduction in cost. Con- tractors may simply categorize their costs in different ways than transit agencies (in the construc- tion line item, for example), which makes that project's soft costs as a percent of construction appear low. Sponsor Agency Characteristics An indicator of whether the sponsor agency tended to minimize capital charges contributed the most to the regression. A dummy variable indicating if the sponsor agency tended to rely heavily on outside contractors during project development phases did not demonstrate significant power to predict soft cost percentages, and was excluded. When a transit agency sponsors a construction project, it usually contributes some of its own labor and even materials. Agency employees often inspect construction activities, monitor safety, administer the contract, acquire property, manage the project, and perform many other tasks. As opening day approaches, agency staff contribute time coordinating testing, training, safety inspec- tions, and shared tasks with other agencies. The agency chooses whether to charge these expendi-

OCR for page 77
As-Built Soft Cost Analysis 51 tures to the capital project (either directly or as an overhead-type allocation) or to absorb them into the operating budget, and project sponsors each have different internal policies for this. External Factors Of many indicators of the broader circumstances in which a project is developed, two variables stood out: economic conditions and unusual political influence. The overall health of the economy, as well as the level of construction activity, can affect the construction bids a transit project sponsor can expect to receive. If the construction sector or economy at large is in a downturn when a project sponsor accepts bids, contractors may reduce their bids due to economic forces. In this case, soft costs computed as a percentage of the engineered construction cost estimate might look relatively higher simply because the bid construction cost is lower. Historically, some change in soft costs can be attributed to the rate of gross domestic product (GDP) growth when construction contracts are bid, after accounting for other variables. Although GDP growth rises and falls with the economy, it has historically risen an average of 2.5% to 3.0% per year. However, it is difficult to use this driver to estimate soft costs for a project years away from construction since future GDP growth is difficult to predict. The Guidebook therefore recommends using this cost relationship only when a cost estimator can be reasonably sure the project is to be bid within one year. When public involvement or political pressures are high, such as in a contentious design and planning process, soft costs tend to rise relative to construction costs. When, for example, mul- tiple planning boards, citizen advisory councils, and officials must approve the design and could even call for a redesign, these external factors were shown to increase soft costs. Other measures of project context and external circumstances contributed less to the regression analysis and were excluded: Unusually high public involvement and/or opposition (dummy variable), Major project redesign required (dummy variable), and Decade. The multivariate regression also used midyear of expenditures as an independent variable. As Figure 25 showed earlier, soft costs have been rising over time, so including this variable controls for the effect of the historic rise in soft costs. However, in estimating soft costs for a given project, the Guidebook does not recommend increasing soft cost percentages for future projects. Extension regression analysis yielded a 10-variable equation that can explain approximately 60% of the difference in soft cost percentages by variations in the projects' characteristics (R2 = 0.58). Table 15 shows the resulting coefficients from this regression, whose dependent variable is total soft costs as percent of construction costs. Table 15. Multivariate regression results on soft costs as a percentage of construction costs. Variable Name Unit Coefficient t-Stat Guideway alignment length 10,000 linear feet 1.4% 2.69 Construction costs Billions, 2008$ -5.9% -2.49 Mode Dummy, heavy rail = 1 6.0% 1.64 Installation conditions Dummy, no active service = 1 -3.8% -1.25 Delivery method Dummy, non-DBB = 1 -7.2% -2.10 Economic conditions GDP % annual growth -1.4% -2.34 Unusually long project development phase Dummy, yes = 1 7.1% 2.08 Unusual political influence Dummy, yes = 1 6.6% 2.22 Agency tendency to minimize capital charges Dummy, yes = 1 -6.0% -1.65 Years from 2008 Years -0.4% 2.22

OCR for page 77
52 Estimating Soft Costs for Major Public Transportation Fixed Guideway Projects Using the projects contained in this FTA capital cost database, the strongest correlation that could be produced is the regression described above. After testing many combinations of explanatory independent variables, these 10 could best predict the relationship between soft and hard costs. Although the strength of this correlation is not ideal (the R2 and t-statistics are rela- tively small), the relationship does highlight the importance of judgment in cost estimation. In addition, as more projects are included in this cost database, it may be possible to perform analysis with stronger cost relationships. 4.5.8. Preparing Multivariate Results for Use in Guidebook Alternative multivariate regressions were examined using different actual soft cost components (rather than total soft costs) as the dependent variable. The coefficient from the overall soft cost analysis was distributed to the soft cost components that correlated to the project characteristics in a statistically significant way. For example, alignment length showed an overall coefficient of around 1.4% per 10,000 linear feet regressed against overall soft costs, and this relationship was strongest when regressed against project management and other soft costs, so the Guidebook rec- ommends adjusting the percentage estimate for those two components to a total of 1.4% per 10,000 linear feet. Finally, the starting points and recommended percentage adjustments were validated against the original projects to gauge the potential error in the Guidebook methodology. Some minor adjustments to the coefficients were made to minimize the sum of each component's root mean square error for all projects.