NSPA-2003_Report.pdf

US DEPARTMENT OF

TRANSPORTATION

FEDERAL TRANSIT

ADMINISTRATION

PPRREEDDIICCTTEEDD AANNDD AACCTTUUAALL IIMMPPAACCTTSS OOFF

NNEEWW SSTTAARRTTSS PPRROOJJEECCTTSS

CCaappiittaall CCoosstt,, OOppeerraattiinngg CCoosstt aanndd RRiiddeerrsshhiipp DDaattaa

DRAFT

Office of Planning and Environment with support from

SG Associates, Inc

September 2003

Acknowledgements

This report was primarily authored by Mr. Frank Spielberg of SG Associates and Mr.

Steven Lewis-Workman of the Federal Transit Administration. Portions of this report were also written by Mr. Tom Buffkin and Ms. Sarah Clements of the Federal Transit Administration and Mr. Joe Swartz of SG Associates. Ms. Jennifer Terry and Mr. Jason Mumford of URS Corporation compiled cost data from project reports. The authors would like to thank all of the project sponsors and FTA regional office staff who took the time to review and ensure the accuracy of the information contained in this study.

Table of Contents

1. Overview

1.1. Methodology

1.2. Findings for Capital Costs

1.3. Findings for Operating Costs

1.4. Findings for Ridership

1.5. Organization of this Report

2. Capital Costs

2.1. Analysis Approach

2.2. Forecast and Actual Capital Cost

2.3. Changes Between Planning and Construction

2.3.1. Project Scope

2.4. Trends in New Starts Capital Cost Estimates

2.5. Results and Interpretation

3. Operating and Maintenance Costs

3.1. Analysis Approach

3.2. Predicted and Actual Operating and Maintenance Cost

3.3. Changes Between Planning and Operation

3.3.1. Service Levels

3.4. Trends in New Starts O&M Cost Estimates

3.5. Results and Interpretation

4. Ridership

4.1. Analysis Approach

4.2. Forecast and Actual Ridership – 19 New Starts Projects

4.2.1. Average Weekday Boardings

4.2.2. Average Weekday Boardings Adjusted to Forecast Year

4.3. Updated Ridership – Ten Projects from Urban Rail Transit (1990)

4.4. Results and Interpretation

Appendix 1: Project Profiles

Atlanta North Line Extension

Baltimore MTA Light Rail Extensions: BWI, Hunt Valley, and Penn Station

Baltimore Heavy Rail Section C Northeast Subway: Extension to Johns Hopkins Hospital

Chicago Southwest Transitway Project

San Francisco Colma BART Station

Dallas South Oak Cliff Light Rail Project

Denver – North I-25 Bus/HOV Lanes

Denver Southwest LRT Project

Houston Southwest Transitway

Jacksonville Automated Skyway Express

Los Angeles Red Line

Miami Metromover Omni and Brickell Extensions

Pittsburgh Airport Busway/Wabash HOV Facility Phase 1 (West Busway)

Portland Westside-Hillsboro LRT Extension

San Jose, CA – Tasman West LRT

San Jose Guadalupe Corridor

San Diego El Cajon Extension

Downtown Seattle Transit Project

Salt Lake City North/South Light Rail Transit Line

St. Louis Metrolink Light Rail Project

St. Clair County MetroLink Extension Project

Appendix 2: References

Federal Transit Administration Page 1 of 206 US Department of Transportation

1. OVERVIEW

The Federal Transit Administration’s Office of Planning and Environment has implemented a program of analysis of the predicted and actual impacts of major transit projects implemented in the past two decades through Full Funding Grant Agreements and using Section 5309 funds (formerly the Section 3 program). The analysis has three purposes:

To examine the adequacy of procedures and technical methods used to develop information for decision making by project sponsors and the FTA;

To provide an up-to-date assessment of the actual performance of projects compared to the forecasts made for these projects; and

To establish a frame of reference for the Before and After studies of future1 New Starts projects that will be prepared by grantees in response to requirements of the New Starts regulation.

This first phase of the effort was an inventory – from existing sources – of the forecasts prepared at various stages of the project planning process and of the actual results.

These inventory data have been used to identify differences between projections of capital cost and the actual costs of constructing the projects,2 of the operating costs and the actual costs of operating the services, and of projected ridership and the actual ridership achieved.

All data necessary to examine the cause of differences between forecast and actual values was not available during this inventory phase. While the data assembled during this phase and reported here provide some insight into trends in the forecasts associated with major transit investment projects many details unique to each project cannot be reflected in such a summary analysis. This phase focused on data and documentation that would identify topics that would be fruitful areas for further research in more detailed phase two case studies. Those detailed studies will strive to identify, with the assistance of project sponsors, the specific factors that contributed either to successful projections or particularly large errors.

The work in this first phase has been particularly useful in meeting the third objective cited above – establishing a frame of reference for the for the Before and After Studies of New Starts projects. The lessons learned in reconstructing project histories from available published documents have been reflected in the Before and After Study guidance developed by FTA.

1 New Starts project sponsors are now required to collect and analyze information on the impacts of the new start project and the accuracy of the forecasts prepared during development of the project. The effort includes collection of “before” data on the current transit system; documentation of the “predicted” scope, service levels, capital costs, operating costs, and ridership of the project; collection of “after” data on the transit system two years after opening of the new start project; and analysis of the consistency of “predicted” project impacts with the “after” data.

2 However, since the projects were built in different time periods and the costs are expressed in dollars adjusted to different base years, capital costs cannot be compared between projects.

Federal Transit Administration Page 2 of 206

1.1. Methodology

Data on project scope, service levels, costs and ridership were identified and documented, to the extent possible, at four milestones during project planning and implementation. These key milestones are:

Completion of Alternatives Analysis and Draft EIS;

Completion of a Final EIS;

Signing of the Full Funding Grant Agreement; and

Project completion, for capital cost data, or two years after opening, for operating and maintenance costs and ridership results.

The primary data recorded for each milestone were:

Project scope (Length, horizontal/vertical alignment, stations, parking, vehicles, ancillary facilities);

Service levels (headways, service hours, fares, vehicle miles and hours, if available);

Capital cost (construction, right-of-way, engineering, administration, vehicles, etc.);

Operating and maintenance (O&M) cost (vehicle operation, vehicle maintenance, guideway maintenance, station operation and maintenance, etc.); and

Ridership (daily and annual boardings on the project, system-wide boardings and/or linked trips).

The projections at each stage of planning and project development were taken from the various reports prepared by project sponsors and/or their consultants. These documents were collected from the FTA or from one of the repository libraries. FTA Staff obtained the reports, reviewed each document to identify relevant sections and provided copies of these sections to the project team. The FTA Office of Planning and Environment staff recorded the data for project scope, service levels, and ridership.

FTA staff also contacted each grantee to document the actual ridership experience for each project. The consultant team reviewed the published materials and abstracted the forecast and actual capital and operating cost data from the reports.

In most cases, the information in the published documents was not as detailed as desired. This lack of detailed information affects the ability to use the data to examine sources of error in the cost estimates or ridership forecasts but does not affect the assessment of overall project deviations.

Estimated capital and operating costs and the related “dollar years” were taken from AA/DEIS, FEIS and FFGA reports. As-built capital costs were derived from Project Management Oversight (PMO) reports. As-operated O&M costs were developed from the National Transit Database.

Federal Transit Administration Page 3 of 206

For all cost estimates, the base year of the estimates was also documented. This base year is the year from which the dollar values for unit costs applied in cost estimation were derived.3 For purposes of comparison of predicted and actual costs, all cost projections were expressed in midpoint-of-construction year dollars. While not a perfect comparison to actual costs, year-by-year expenditures were generally not available so the approximation of using the midpoint-of-construction year was adopted. This adjustment is different for each project.

The adjustment of Capital Costs used the Engineering News Record Construction Cost Index. The specific city value was used when applicable; when not applicable the twenty-city average index value was used. The adjustment of operating costs from projection year dollars to year of operation dollars used an index based on the cost per vehicle-hour for bus service in the specific city as reported in the National Transit Database. For selected projects that are extensions of pre-existing lines or a new line of an existing mode, the operating costs could not be derived directly from the NTD. In those cases, an estimate was made based on the approximate relation of vehicle miles of service on the segment of interest relative to total vehicle miles of service for that mode.

Ridership forecasts were derived from the AA/DEIS, FEIS and related supplemental reports while actual ridership for each project was obtained directly from the transit agencies. This study compares forecasts of average weekday boardings to actual average weekday boardings. FTA attempted to match the actual ridership data to the forecast year. In cases where the forecast year remains in the future, typically 2005, 2010 or 2015, the actual 2002 ridership was inflated by the average annual growth in transit system ridership between 1990 and 2002. This “adjusted” actual ridership was then compared to the forecasts to evaluate the performance of those forecasts.

FTA and the contractor team condensed the relevant information collected regarding each project into project profiles. The profiles contain a description of the project purpose, summarize the planning and development of each project, and document the predicted and actual scope, service levels, ridership, capital costs, and operating costs for each project. These profiles were then sent to FTA regional offices and to each grantee for review and validation. The information in the project profiles, including the forecasts and actual data used for this analysis, has been reviewed for accuracy and validated by each grantee.

No attempt was made to adjust costs to reflect changes in the scope of a project.

However, when changes in scope that would have a significant impact on capital or operating costs were identified, those changes were noted in the text of the Project Profiles. In cases where the actual project constructed was closer to an identifiable

3 Construction cost estimates are typically developed by applying unit costs to estimates of the number of units required. (e.g. six miles of track @ $8 million per mile). The unit costs are developed from the cost experience on similar projects built in the years preceding the estimates. If the base year for a cost estimate is 1988, it means that the unit costs applied in estimating reflected the cost of construction in 1988.

Federal Transit Administration Page 4 of 206 alternative included in the planning studies, actual costs and ridership were compared to forecasts from the alternative that most closely resembled the constructed project.

1.2. Findings for Capital Costs

Adjusted for inflation between the time of the estimate prepared for the AA/DEIS (generally concurrent with the selection of the Locally Preferred Alternative) the cost of the completed projects averages about 20 percent greater than the initial estimates.

The accuracy of early stage planning estimated for major transit projects has been improving over time.

Inflation is the largest component of the difference in absolute dollars between the estimated cost and the as-built expenditure.

The FFGA process has had a positive impact in controlling project costs.

1.3. Findings for Operating Costs

Projections of operating costs, adjusted for general inflationary increases in the local costs of transit operations, are being achieved.

Experienced project sponsors have better O&M cost estimates. This probably reflects not only the ability to use local experience in developing unit costs but also greater realism in the assessment of efficiencies that can be achieved.

1.4. Findings for Ridership

The quality of ridership forecasts has improved markedly since the 1980’s.4

While ridership forecasts have improved, several ridership forecasts reviewed in this study have proven to be optimistic.

1.5. Organization of this Report

The material presented in this report is based on the forecasts made and values achieved for twenty-one transit New Starts projects opened for full service between 1990 and 2002. These projects are listed in Table 1.

4 Pickrell, Don H., Urban Rail Transit Projects: Forecast Versus Actual Ridership and Costs, DOT-T-91- 04, Office of Grants Management, Urban Mass Transportation Administration, Washington DC, October 1990 found that ridership projected for major transit projects compared to ridership actually achieved, none of the ten projects examined had achieved, at the time of the analysis, ridership greater than 72% of the forecast. Nine of the ten projects had achieved less than 50% of the forecast.

Federal Transit Administration Page 5 of 206

Table 1: New Starts Projects Included in This Study

City Project Mode Year of

AA/DEIS FEIS FFGA Open

Atlanta North Line Extension HR 1990 1991 1994 2000

Baltimore BWI, Hunt Valley, Penn Station Ext. LR 1991 1993 1994 1997

Baltimore Extension to Johns Hopkins HR 1984 1987 1988 1995

Chicago SW Transitway HR 1982 1985 1987 1993

Dallas South Oak Cliff LR 1990 1991 1993 1996

Denver North I-25 Bus/HOV 1988 1989 1989 1994

Denver Southwest LRT 1995 1996 1999 2000

Houston Southwest Transitway Bus/HOV 1985 1985 1987 1993

Jacksonville Skyway Express AGT 1982 1983 1994 2000

Los Angeles Red Line HR 1983 1983 1997 2002

Miami Omni and Brickell Extensions AGT 1987 1988 1989 1994

Pittsburgh Airport Busway/ Wabash Bus/HOV 1992 1994 1999 2000

Portland Westside - Hillsboro LRT 1982 1994 1994 1998

Salt Lake City I-15/State Street LRT 1990 1994 1995 1999

San Diego El Cajon Extension LRT 1985 1986 1986 1989

San Francisco Colma BART Station HR 1988 1990 1993 1996

San Jose Guadalupe Corridor LRT 1981 1983 1984 1991

San Jose Tasman West LRT 1991 1992 1996 1999

Seattle Downtown Project Bus 1984 1985 1987 1990

St. Clair Co. MetroLink Extension LRT 1995 1996 1996 2001

St. Louis MetroLink LRT 1984 1987 1995 1993

Planning for projects included in this analysis began as early as 1977 and as late as 1993. The time between the Alternatives Analysis report and the opening of service ranged from four years for the San Diego El Cajon Extension to nineteen years for the Los Angeles Red Line.

The body of this report addresses the findings that have been gleaned from the project data. The details of any specific project are referenced only to illustrate points of interest.

The remainder of this report addresses the variations identified between projects as planned and projects as constructed and operated. The sections discuss, in turn, capital costs, operating and maintenance costs, and ridership. Tables and charts reporting the project information and suggesting some possible interpretations are presented. Project Profiles in Appendix 1, one for each project, include more detailed information on the project history, the scope of the project as conceived and executed, and other information necessary to interpret the summary statistics.

Federal Transit Administration Page 7 of 206

2. CAPITAL COSTS

This chapter compares the AA/DEIS, FEIS, and FFGA capital cost estimates to the actual costs incurred to complete these projects. The capital costs are compared both in nominal terms and in real (inflation adjusted) terms in order to ascertain whether the cost estimates developed in the planning and implementation of New Starts projects have accurately predicted the actual costs of constructing them. The capital costs in this analysis include:

design and engineering, right-of-way acquisition, construction of all facilities and systems, and vehicles.

The purpose of this Phase 1 study is simply to identify the extent that the costs of New Starts projects have deviated from their cost estimates. The costs are not evaluated in enough detail to identify the specific reasons for any deviations from the cost estimates, though any obvious changes in scope or specific known construction difficulties are noted where appropriate. The detailed analysis of particular cost estimation problems is left to the future Phase 2 study.

2.1. Analysis Approach

Information about the nature of the project and the associated capital cost estimates were derived from a variety of available sources. The primary source documents for the planning stage estimates were the Alternatives Analysis documents, usually in the form of an AA/DEIS, and the FEIS. For some projects, particularly those with a longer planning history, these were supplemented with other documents, such as a Preferred Alternative Report, a local Feasibility Study, and/or a Supplemental DEIS/FEIS. The primary factors used to describe the scope of the project were the length of the line; the length proposed at-grade, below grade or above grade; and the number of stations.

When available other information affecting the capital cost (e.g., amount of parking, number of vehicles to be purchased) was also documented. A list of the documents from which data were derived is presented in Appendix 2.

Capital cost estimates at the stage of the Full Funding Grant Agreements (FFGAs) were taken directly from FFGA documents. The cost reported includes all project costs identified in the FFGA, both local and federal. When known, other capital costs items funded through different sources have also been identified.

In many cases, the project as constructed differs from the project that was identified in the earlier studies as the Locally Preferred Alternative (LPA). For some projects these differences are minor. For others they are more significant (e.g., additional stations, fewer vehicles, portions of the project not built). No attempt has been made to adjust

Federal Transit Administration Page 8 of 206 either the earlier estimates or the as-built costs to reflect these differences. Rather, the major differences are noted in the project profiles in Appendix 1.

The as-built costs of the projects are taken from reports prepared by the FTA Project Management Oversight (PMO) contractors assigned to each project. A final PMO report is prepared prior to closing out the FFGA. The cost reported in the final PMO report serves as the as-built cost for this analysis.

Capital cost estimates prepared at the AA/DEIS, FEIS and FFGA have been adjusted for inflation to the midpoint of construction year of the project. This adjustment attempts to approximate the year in which the as-built costs are expressed. These adjustments have been made based on the Engineering New Record Construction Cost Indices.

When available the specific city indices were used; when the city specific indices were not available the twenty-city average was used. As-built costs are as reported (dollars expended) and were not adjusted.

2.2. Forecast and Actual Capital Cost

The inflation adjusted predicted and actual capital costs for the twenty-one projects are summarized in Table 2. Since the costs are adjusted to reflect midpoint-of-construction-year dollars for each project, costs cannot be compared between projects.

The cost information provides an indication of how estimated costs changed as design became more definite over the project planning period and how the predicted costs compare to the actual costs of constructing the projects.

Federal Transit Administration Page 9 of 206

Table 2: Project Capital Costs Adjusted for Inflation

City Project Mode Capital Cost (in Millions)1

Adjusted for Inflation

AA/DEIS FEIS FFGA As-Built

Atlanta North Line Extension HR 439.5 389.7 352.0 472.7

Baltimore BWI, Hunt Valley, Penn Station Ext. LR 81.9 110.2 109.5 116.2

Baltimore Extension to Johns Hopkins HR 313.7 310.5 310.5 353.0

Chicago SW Transitway HR 604.0 532.3 438.4 522.0

Dallas South Oak Cliff LR 325.4 374.6 377.0 360.0

Denver North I-25 Bus/HOV 189.5 189.5 205.2 228.0

Denver Southwest LRT 149.6 158.3 176.9 177.7

Houston Southwest Transitway Bus/HOV 95.6 100.0 111.2 98.3

Jacksonville Skyway Express AGT 85.8 85.8 142.0 137.3

Los Angeles Red Line HR 3,031.3 3,181.3 3,505.6 4,469.7

Miami Omni and Brickell Extensions AGT 221.2 221.2 161.3 228.0

Pittsburgh Airport Busway/ Wabash Bus/HOV 274.4 338.6 361.7 321.8

Portland Westside - Hillsboro LRT 559.3 804.0 886.5 964.0

Salt Lake City I-15/State Street LRT 305.6 245.9 299.5 298.5

San Diego El Cajon Extension LRT 114.4 114.4 100.4 102.7

San Francisco Colma BART Station HR 112.5 130.1 171.6 179.9

San Jose Guadalupe Corridor LRT 257.7 321.5 395.2 380.3

San Jose Tasman West LRT 451.2 462.5 346.1 325.2

Seattle Downtown Project Trolley Bus 299.6 348.7 400.0 468.7

St. Clair Co. MetroLink Extension LRT 367.7 367.5 322.2 339.2

St. Louis MetroLink LRT 379.7 346.5 455.8 464.0

1. Cost estimates are reported in “midpoint of construction” dollars

Table 3, below, summarizes the relationship between the as-built capital cost of each project and the inflation-adjusted estimates prepared at different phases of project planning. In the table, a percent value less that 100 percent indicates that the as-built cost was less than the estimate while a value greater than 100 percent indicates that the as-built cost was greater than the estimate. For example, a value of 96.6 percent in the column headed AA/DEIS means that the actual cost was 3.4 percent less than the AA/DEIS cost estimate while a value of 120.7 percent in the column headed FFGA means that the actual cost was 20.7 percent greater than the FFGA cost estimate.

Federal Transit Administration Page 10 of 206

Table 3: As-built Project Costs as Percent of Estimate

City Project Mode As-Built Capital Cost as Percent of

Estimate, Adjusted for Inflation

AA/DEIS FEIS FFGA

Atlanta North Line Extension HR 107.5% 121.3% 134.3%

Baltimore BWI, Hunt Valley, Penn Station Ext. LR 141.9% 105.5% 106.2%

Baltimore Extension to Johns Hopkins HR 112.5% 113.7% 113.7%

Chicago SW Transitway HR 86.4% 98.1% 119.1%

Dallas South Oak Cliff LR 110.6% 96.1% 95.5%

Denver North I-25 Bus/HOV 120.3% 120.3% 111.1%

Denver Southwest LRT 118.8% 112.2% 100.5%

Houston Southwest Transitway Bus/HOV 102.8% 98.3% 88.4%

Jacksonville Skyway Express AGT 160.0% 160.0% 96.7%

Los Angeles Red Line HR 147.5% 140.5% 127.5%

Miami Omni and Brickell Extensions AGT 103.1% 103.1% 141.3%

Pittsburgh Airport Busway/ Wabash Bus/HOV 117.3% 95.0% 89.0%

Portland Westside – Hillsboro LRT 172.4% 119.9% 108.7%

Salt Lake City I-15/State Street LRT 97.7% 121.4% 99.7%

San Diego El Cajon Extension LRT 89.7% 89.7% 102.3%

San Francisco Colma BART Station HR 159.9% 138.2% 104.9%

San Jose Guadalupe Corridor LRT 147.6% 118.3% 96.2%

San Jose Tasman West LRT 72.1% 70.3% 94.0%

Seattle Downtown Project Trolley Bus 156.5% 134.4% 117.2%

St. Clair Co. MetroLink Extension LRT 92.3% 92.3% 105.3%

St. Louis MetroLink LRT 122.2% 133.9% 101.8%

Average over twenty-one projects 120.9% 113.5% 107.3%

The actual capital cost of New Starts projects has been on average, 20.9 percent greater that the inflation adjusted AA/DEIS estimate, 13.5 percent greater than the FEIS estimate and 7.3 percent greater than the FFGA limit. The AA/DEIS is the point in project planning when the locally preferred alternative is selected, and therefore, is the most critical decision point for the implementation of any project. The data reported here suggest that in terms of the real economic cost the information presented at the time of selection of the LPA has been a generally reliable estimate of the cost of construction. Eleven of the twenty-one projects had construction costs that deviated less than 20 percent from the inflation adjusted AA/DEIS estimate. One project significantly overestimated capital costs (though this was due to scope reduction), while nine projects significantly underestimated capital costs in their AA/DEIS.

Figure 1 arrays the AA/DEIS capital cost results along a 45 degree line. Projects that fall within the 20 percent range were considered reliable. The results of the 1990 UMTA study, Urban Rail Transit Projects: Forecast versus Actual Ridership and Costs are included for comparison purposes.

F e d e ra l T ra n s it A d m in is tra tio n

P a g e o f 2

U S

D e p a rtm e n t o f T ra n s p o rta

$0

$200

$400

$600

$800

$1,000

$1,200

$1,400

$0 $200 $400 $600 $800 $1,000 $1,200 $1,400

A c tu a l

C a p it a l

C o s t

M il li o n s

AA/DEIS Capital Cost Estimate - $Millions

CHI

PIT-P

MHR-P

BAL-P

POR

BUF-P

POR-P

DET-P

MIA-P

ATL

SJ-TSL-I

BAL-HDAL

SLC

PIT

MIADEN-B

DEN-L

SAC-P

SD

HOU

BAL-L

JAX SF

SEA

SL-M

SJ-G

ATL-P WAS-PLA

+20% error interval

-20% error interval

Forecast Error less than 20%

Forecast error close to 20%

Forecast error well over 20%

1990 Study Results

Key to Project Symbols ig u re 1

: A A

/D E

IS

C o st E stim a tes v

s. A ctu a l

Atlanta HR* (’90) NA Dallas South Oak Cliff DAL Miami HR (’90) MHR-P San Diego El Cajon SD

Atlanta North Line ATL Denver I-25 HOV NA Miami Omni/Brickell MIA San Jose Guadalupe SJ-G

Baltimore HR (’90) BAL-P Denver SW LRT DEN Pittsburgh LR (’90) PIT-P San Jose Tasman West SJ-T

Baltimore Johns Hopkins BAL-H Detroit DPM (’90) DET-P Pittsburgh West B'Way PIT Seattle Bus Tunnel NA

Baltimore LRT Ext. BAL-L Houston SW Transitway HOU Portland LR (’90) POR-P St. Louis Initial System STL-M

BART Colma SF Jacksonville ASE JAX Portland Westside-Hillsboro POR St. Louis St. Clair Ext. STL-I

Buffalo LR (’90) BUF-P LA Red Line LA Sacramento LR (’90) SAC-P Washington HR (’90) WAS-P

Chicago Orange Line CHI Miami DPM (’90) MIA-P Salt Lake South LRT SLC

Federal Transit Administration Page 12 of 206

2.3. Changes Between Planning and Construction

The projects studied, as constructed and operated often differ from the projects as contemplated in the AA/DEIS studies. In some cases, these changes are minor and would have little effect on the capital or operating costs. For other projects the changes have been significant and the effect is reflected in as-built or as-operated costs that differ markedly from early planning estimates. Changes in project scope primarily affect capital costs. Operating costs can also be affected by scope changes, but more often reflect changes in the service operating plan.

2.3.1. Project Scope

On average, the time between the publication of the Alternatives Analysis (AA) and/or DEIS reports, roughly the time at which the Locally Preferred Alternative (LPA) was selected and the revenue operations date, was eight years. The most common duration was six years, though the range was quite broad. The shortest periods were four years for the San Diego East Line extension (El Cajon) and five years for the Denver Southwest Corridor project and the Hillsboro section of the Portland Westside-Hillsboro Extension. Each of these were extensions of light rail lines existing or under construction at the time the LPA was selected. The longest periods were nineteen years for the Los Angeles Red Line and eighteen years for the Jacksonville Skyway.

Many changes in project concept and design can occur over extended project planning and development periods. Projects often change due to changes in development patterns, changes in community goals, changes in the technology of construction, changes in the cost of construction and changes in the funding available to complete the project. Some of these changes have relatively minor impact on the cost of a project while others are significant. To the extent possible from the documentation available, changes in project scope that would have had a significant effect on project cost are noted in Table 4 and described in greater detail in the Project Profiles (see Appendix 1).

In the analyses reported here, adjustments have been made for the effects of inflation in the cost of construction. When comparing project cost estimates at the several milestones with the actual as-built costs, the estimated costs have been adjusted to the midpoint-of-construction year.

d e ra l T ra n s it A d m in is tra g e o f 2

D e p a rtm e n t o f T ra n s p o rta

City Project Changes

As-built Cost vs. Estimate

Inflation Adjusted Nominal Dollars

AA/DEIS FFGA AA/DEIS FFGA

Atlanta North Line Extension

Built in two stages. FFGA included only Section 2- 1.9 miles, 2 stations from Dunwoody to North Springs.

Parking in structure rather than surface lot. +7.55% +34.29% +27.7% +44.4%

Baltimore BWI, Hunt Valley, Penn Station Extensions

Two stations added in FFGA but were not built. Double track planned for 4.6 miles but all built as single track.

+41.91% +6.15% +65.5% +9.3%

Baltimore Extension to Johns Hopkins

Significant delay due to leaking underground fuel storage tanks. +12.52% +13.68% +37.5% +9.7%

Chicago SW Transitway Elevated structure reduced from 4.3 to 2.7 miles.

Stations: DEIS: 9 ( 1 underground, 3 elevated) As-built: 8 (0 underground, 2 elevated). -13.58% +19.08% +15.2% +27.0%

Dallas South Oak Cliff DEIS planned for 33 vehicles, FFGA and as-built included 19 vehicles. Yard and shop area in AA/DEIS estimate deleted from project. +10.63% -4.52% +25.5% -2.7%

Denver North I-25 Length: DEIS 6.6 miles, As-built: 5.3 miles. +20.31% +11.09% +27.6% +13.3%

Denver Southwest 0.2 miles planned as single track built as double track. 2 additional LRVs purchased. +18.80% +0.45% +39.4% +11.6%

Houston Southwest Transitway

Surveillance, communications and control system planned and funded in FFGA not implemented. Project is

1.2 miles longer than AA/DEIS plan. +2.78% -11.57% +17.3% -5.0%

Jacksonville Skyway Express System retrofitted to new technology. Original vehicles sold and replaced with new systems. As-built has 8 stations rather than 9 as planned in AA/DEIS. +59.98% -3.31% +126.2% -3.3%

Los Angeles

Red Line Realignment due to underground natural gas fields.

Delays due to funding constraints. Length: DEIS 18 mi.;

FFGA 23 mi.; As-built 17mi. Stations: DEIS 16; FFGA 22;

As-built: 16 +47.45% +27.50% +90.0% +42.9%

T a b le 4

: C h a n g es in

P ro jec t S co p e a n d C a p ita l C o sts d e ra l T ra n s it A d m in is tra g e o f 2

D e p a rtm e n t o f T ra n s p o rta

City Project Changes

As-built Cost vs. Estimate

Inflation Adjusted Nominal Dollars

AA/DEIS FFGA AA/DEIS FFGA

Miami Omni and Brickell Extensions

Only minor changes.

+3.07% +41.32% +16.0% +48.1%

Pittsburgh Airport Busway/ Wabash

Length reduced from 8.1 mi. to 6.1 mi.

Bridge across river not constructed.

Wabash Tunnel rehabilitation not yet completed. +17.26% -11.03% +42.4% -1.6%

Portland Westside - Hillsboro

DEIS assumed mostly surface construction. FEIS and as-built include 3 miles of tunnel and 1 underground station. +72.37% +8.75% +104.7% +17.3%

Salt Lake City

I-15/State Street One planned station not built. One-half mile access single rather than double track. -2.32% -0.35% +29.2% +16.5%

San Diego El Cajon Extension

DEIS: 11 miles single track.

As-built: 0.3 mi. single track; 10.8 mi. double track. -10.25% +2.26% -2.7% +4.9%

San Francisco

Colma BART Station

Underground: DEIS 0.13 mi.; As-built 0.19 mi.

Elevated; DEIS 0.0 mi; As-built 0.06 mi.

Vehicles: DEIS 3; As-built 0 +59.90% +4.86% +89.6% +5.7%

San Jose Guadalupe Corridor

Project delayed for almost two years due to lawsuit.

Single track: DEIS 0.0 mi.; As-built 1.6 mi.

Double track: DEIS 20.0 mi.; As-built 18.4 mi. +47.60% -3.78% +100.9% +2.3%

San Jose Tasman West Length: DEIS 12.2 mi.; As-built 7.6 mi.

Stations: DEIS 18; As-built 12 Parking spaces: DEIS 2380; As-built: 312 LRVs: DEIS 35; As-built 0 -27.92% -6.04% -9.0% +0.1%

Seattle Downtown Project Minor changes only. +56.47% +17.19% +62.6% +18.5%

St. Clair Co. MetroLink Extension

Length: DEIS 25 mi.; As-built 17.4 mi.

Stations: DEIS 12-13; As-built 8 Parking spaces: DEIS 1800-2700; As-built 4500 Yard & Shop: DEIS – expand existing; FFGA – New -7.74% +5.28% +0.6% +10.2%

St. Louis MetroLink Some alignment changes and delay due to need to adjust for Airport plans. +22.20% +1.80% +68.0% +0.0%

Federal Transit Administration Page 15 of 206

Specific changes found in one or more of the projects considered include:

Change in alignment

These include changes in the horizontal alignment, vertical alignment, or both.

Changes in vertical alignment tend to have greater impacts on capital costs since these can result in addition or elimination of structure or subsurface construction.

Change in number of stations

The projects reviewed included situations in which stations were added in response to new developments or community concerns and situations in which stations were omitted or delayed in order to reduce project costs.

Change in technology

In one project (Jacksonville ASE), the technology used for the guideway was changed from a rubber tire on concrete guideway to a monorail. This change required that the entire functioning system be shut down to accomplish substantial modifications to previously built segments.

Delay in constructing one or more project elements

Inflation is the single largest factor in considering the difference between the cost as estimated at the time of the LPA and the cost as-built (see Figure 2). The longer the period between the LPA and the time of construction, the greater the effect of inflation on the cost of a project in year-of-expenditure dollars. Lengthy delays tend to lead to greater differences between predicted and actual capital costs.

Change in regulation

Over the period of time between original planning and actual construction of projects considered in this report there were many changes in legislation and regulation that affected project requirements. Most notable were regulations implementing the National Environmental Policy Act (NEPA) and the of the Americans with Disabilities Act. (ADA). While the data assembled for these studies have not been in sufficient detail to isolate the effects to these and similar regulatory factors, there can be no doubt that costs were incurred, especially when projects had originally been planned prior to the implementation of significant policy changes.

Federal Transit Administration Page 16 of 206

Change in park-and-ride facilities

Addition of park-ride areas, changes in the number of spaces, or changes from surface lots to structures were noted in several projects. For one project equipment related to collection of parking fees was added.

One goal of this Phase 1 study is to assess the degree to which there may be problems with the processes applied for estimating capital costs of New Starts projects.

Differences between the estimates prepared early in project planning and the cost incurred when the project is built can arise from several sources including:

Difficulties constructing the project as designed;

Changes in project scope;

Uncertainties regarding unit costs; and

Unanticipated inflation.

The analyses presented in this report focus on the first three possible sources. These are the aspects of project development and cost estimation that are under the control of project sponsors and their consultants. The estimates produced relate to the “real” economic cost of the project and are the basis of local decisions related to selection of alternatives. Project estimators know that the costs of construction are constantly changing due to a variety of factors related to both local and national economic conditions. While estimators can, and often do, identify a cost component related to expected inflation, neither the estimators nor others associated with planning or implementing New Starts projects have control over the processes that influence inflation.

In the analyses reported in this study, adjustments have been made for the effects of inflation in the cost of construction. When comparing project cost estimates at the several milestones with the actual as-built costs, the estimated costs have been adjusted to the midpoint of construction year.

d e ra l T ra n s it A d m in is tra g e o f 2

D e p a rtm e n t o f T ra n s p o rta

$0 $200 $400 $600 $800 $1,000

Atlanta North Line Extension

Baltimore Light Rail

Baltimore Johns Hopkins

Chicago SW Transitw ay

Dallas South Oak Cliff

Denver North I-25

Denver Southw est

Houston Southw est Transitw ay

Jacksonville Skyw ay Express

Miami Omni and Brickell Extensions

Pittsburgh Airport Busw ay

Portland Westside - Hillsboro

Salt Lake City I-15/State Street

San Diego El Cajon Extension

San Francisco BART Colma

San Jose Guadalupe

San Jose Tasman West

Seattle Dow ntow n Project

St. Clair Co. MetroLink

St. Louis MetroLink

Capital Cost in Millions of Nominal Dollars

Cost in year of estimate dollars Inflation component

Actual Capital Cost ig u re 2

: C a p ita l C o st E stim a tes –

A A

/D E

IS

B a se E stim a te, In fla tio n

, a n d A ctu l

Federal Transit Administration Page 18 of 206

2.4. Trends in New Starts Capital Cost Estimates

The analysis of predicted and actual capital costs (Figure 3) show that, adjusted for inflation, the discrepancy between the AA/DEIS estimate and the as-built cost has been decreasing over time.

Figure 3: Trend in Accuracy of AA/DEIS Capital Cost Estimates

40%

60%

80%

100%

120%

140%

160%

180%

1980 1982 1984 1986 1988 1990 1992 1994 1996

Year of AA/DEIS Study

A s

-B u il t

C o s t a s

P e rc e n t o f in fl a ti

A d ju s te d E s ti m te

The Full Funding Grant Agreement process was developed in the early 1980s by FTA’s predecessor agency, the Urban Mass Transportation Administration. The FFGA process was formalized in a model FFGA drafted by UMTA in 1990 and by the Congress in the Federal Transit Act Amendments of 1991.

The Full Funding Grant Agreement process appears to be an effective tool for controlling cost escalation. For New Starts projects included in this analyses with AA/DEIS reports since 1982 the as-built cost, after appropriate adjustments for inflation, has with three exceptions been within 20 percent of the grant amount. Figure 4 shows that the difference between the as-built cost is close, and getting closer, to the inflation adjusted FFGA estimate. On average, the actual project cost has exceeded the inflation adjusted FFGA estimate by about 7 percent. Even without adjusting for inflation the as-built costs have been within 13 percent of the FFGA, on average.

Federal Transit Administration Page 19 of 206

Figure 4: Trend in Accuracy of FFGA Capital Cost Estimates

40%

60%

80%

100%

120%

140%

160%

180%

1980 1982 1984 1986 1988 1990 1992 1994 1996

Year of AA/DEIS Study

A s

-B u ild C o s t a s P e rc e n t o f F

F G

A E s ti m

Inflation, a factor outside the control of planners, is the biggest single factor in the difference between predicted and actual capital costs. Inflation makes up 41 percent of the difference in the nominal dollar AA/DEIS cost estimate and the actual project costs.

For the FEIS cost estimate, inflation makes up 44 percent of the difference.

Even after adjusting for the effects of construction cost inflation, there is a tendency for capital costs to increase as the time from original plan to implementation increases.

Figure 5 shows two groupings. For projects developed over the typical span of four to eleven years there is no clear pattern – half of the projects had costs greater than the inflation adjusted estimate and half had a cost lower. All of the three projects with a development span exceeding fifteen years had costs greater than estimated by more than thirty-five percent. Each of these projects also had major design changes.

Federal Transit Administration Page 20 of 206

Figure 5: Increase in Project Capital Cost vs. Time Between Planning and Opening

60%

80%

100%

120%

140%

160%

180%

0 5 10 15 20

Span from AA/DEIS to Opening (years)

A s

-B u il t

C o s t a s

P e rc e n ta g e o f

In fl a ti

A ju s te d

A A

/D E

IS

C o s t E s ti m a te

Cost less than estimate

Cost greater than estimate

The length of time to bring a project to fruition is getting shorter. Figure 6 shows that the time from AA/DEIS to opening for service was about 12 to 14 years for projects planned in the early 1980s. There has been a continuing trend to shorter durations with projects initially planned after 1988 having a development span of about 8 years.

Reducing the development span reduces the effect of inflation on the cost of construction as well as minimizes the likelihood of significant design changes between planning and construction.

Figure 7 shows that the scope of the projects, as measured by the cost of construction expressed in 1996 dollars, has declined slightly over the period for the projects considered. Reductions in the scale of the projects may have contributed, in part, to the reduction in the average time required to implement major transit projects.

Federal Transit Administration Page 21 of 206

Figure 6: Project Development Duration Trend

1980 1982 1984 1986 1988 1990 1992 1994 1996

Year of AA/DEIS

Y e a rs f ro m A

A /D

E

IS

t o

P ro je c t

O p e n in g

Figure 7: AA/DEIS Capital Cost Estimates (LA excluded due to scale)

1980 1982 1984 1986 1988 1990 1992 1994 1996

Year of AA/DEIS

P ro je c t C o s t E s ti m a te

M il li o n s o f

LA Not Included

Federal Transit Administration Page 22 of 206

2.5. Results and Interpretation

The forecasts of capital cost developed for New Starts Projects included in this study, after accounting for the effects of local inflation in construction costs, are showing steady improvement. For the eight projects with DEIS reports published in 1990 or later, the actual cost was within 20 percent of the AA/DEIS estimate for all but one project, while 8 out of 13 projects with AA/DEIS estimates published prior to 1990 had errors exceeding 20 percent. The causes for this improvement are not known with certainty and are outside the scope of this Phase 1 study. Clearly the experience gained in previous projects has some bearing. Some of the key findings from this study are summarized below.

1. Capital costs tend to increase between planning and construction, but less than commonly believed.

Between the time of the estimate prepared for the AA/DEIS (generally concurrent with the selection of the Locally Preferred Alternative) the cost of the completed projects averaged about 20 percent greater than the initial estimates after adjusting for inflation.

2. Delays in planning, project development and implementation strongly influence capital costs.

Less time from the selection of the Locally Preferred Alternative (LPA) to the completion of the project tends to reduce cost estimation errors, even adjusting for inflation. By placing greater emphasis on early planning activities, including the necessity of specifying a project in some detail in order to prepare satisfactory environmental documentation, many issues can be resolved before selection of the LPA. Reducing the number of changes during project development is one major factor in controlling capital costs.

3. The accuracy of early planning estimates has been improving over time.

Projects that were the subject of Alternatives Analysis studies in the early 1980s often resulted in as-built costs that were 40 to 60 percent greater than the estimates, even after adjusting for the inflation in construction costs. For projects with LPAs in the 1990s, the as-built costs are generally within 20 percent of the estimates.

4. Inflation is the largest component of the difference in absolute dollars between the estimated cost and the as-built expenditure.

During the period studied the costs of construction resulted in increases of about 6 percent per year in project cost. Inflation accounts for approximately 40 percent of the difference in nominal dollar cost estimates and the actual capital costs.

Federal Transit Administration Page 23 of 206

5. The FFGA process seems to have helped control project costs.

As-built expenditures differ from FFGA allocations by about 7 percent, on average.

However, this is not purely a function of better cost estimation. Reductions in project scope or other changes to control costs have been necessary for some projects.

Federal Transit Administration Page 25 of 206

3. OPERATING AND MAINTENANCE COSTS

3.1. Analysis Approach

Projected operating and maintenance costs for New Starts transit projects were derived primarily from the AA/DEIS and FEIS reports. There are no operating cost estimates presented in the FFGA documents. The operating and maintenance (O&M) cost estimates presented in the AA/DEIS and FEIS are for the full system in the design year

– typically twenty to twenty-five years after the planning studies. The as-built O&M costs used for comparison are a best estimate of the O&M costs for the first full fiscal year after opening.

The O&M cost estimates in the AA/DEIS and FEIS documents were adjusted to reflect the change in transit operating costs in each specific metropolitan area between the year used as the basis for the cost estimate and the year following the opening of the project. The measure used for this adjustment was the change in cost per vehicle hour of bus service derived from the National Transit Database.

Actual operating costs were estimated from information in FTA’s National Transit Database. For some projects, primarily rail lines that were the only application of light or heavy rail in a metropolitan area, the operating cost was taken as the cost for light rail or heavy rail operations reported in the National Transit Database. When a rail project was not the first application of that mode in an area, the operating and maintenance cost for the line of interest was estimated based on the observed change in rail operating and maintenance costs reported in the National Transit Database. For projects that were extensions of rail lines, publicly available data were used to estimate the proportion of rail-car miles in the region operated on the segment of interest. This proportion was then applied to the total operating and maintenance costs reported in the National Transit Database.

Estimating the actual operating and maintenance costs for bus projects presents a greater challenge. Each project required special treatment. For instance, the Denver I- 25 North planning documents provided system-level O&M estimates, which was compared to operating costs in the National Transit Database. For the Pittsburgh West Busway Airport an estimate was prepared based on the proportion of the peak bus fleet using the Busway. For the Houston Southwest Transitway no operating and maintenance cost estimates were developed in the planning documents. For the Seattle Bus Tunnel the planning documents included an estimate of operating and maintenance cost savings, but there was no way of determining whether or not these savings had been realized.

Federal Transit Administration Page 26 of 206

3.2. Predicted and Actual Operating and Maintenance Cost

Table 5 summarizes the projected O&M costs from the planning studies and the estimated O&M costs for the first full fiscal year after opening.

Table 5: Predicted and Actual Operating Costs (Adjusted for Inflation)

City Project Mode Operating Costs (in Millions)

Adjusted for Inflation

AA/DEIS FEIS As-Built

Atlanta North Line Extension HR 6.1 7.7 5.4

Baltimore BWI, Hunt Valley, Penn Station Ext. LR 4.5 3.5 2.7

Baltimore Extension to Johns Hopkins HR 5.2 3.2 0.1

Chicago SW Transitway HR 24.0 24.8 16.7

Dallas South Oak Cliff1 LR 18.8 18.8 23.1

Denver North I-252 Bus/HOV 169.4 169.4 129.9

Denver Southwest LRT 5.3 5.4 3.2

Jacksonville Skyway Express AGT 1.8 1.8 3.2

Los Angeles Red Line HR 78.6 84.7 48.5

Miami Omni and Brickell Extensions AGT 4.4 4.5 3.8

Pittsburgh Airport Busway/ Wabash Bus/HOV 14.1 14.1 17.1

Portland Westside - Hillsboro LRT 11.3 10.5 12.2

Salt Lake City I-15/State Street LRT 7.6 9.7 7.4

San Diego El Cajon Extension LRT 4.4 4.4 4.2

San Francisco Colma BART Station3 HR 4.9 5.2 6.8

San Jose Guadalupe Corridor LRT 10.9 13.7 19.2

San Jose Tasman West LRT 12.2 8.2 10.2

St. Clair Co. MetroLink Extension4 LRT 29.8 14.8 10.2

St. Louis MetroLink LRT 11.3 10.0 11.5

1. DEIS and FEIS estimates include North Central and South Oak Cliff ; As-built includes West Oak Cliff

2. Costs are for entire bus system

3. As-built cost based on estimated proportion of car-miles

4. As-built cost based on estimate of 45 percent of LRT revenue vehicle-hours

Table 6 presents the relationship between the achieved O&M cost and the estimates.

The values shown are the as-built O&M cost as a percent of the estimates presented in the AA/DEIS or FEIS. The table confirms that, on average, predicted and actual operating costs are quite close, though close inspection of the data show that many projects either over or underestimated operating and maintenance expenses in their planning studies.

Federal Transit Administration Page 27 of 206

Table 6: Actual O&M Costs as Percent of Planning Estimates

City Project Mode Actual O&M Costs as Percent of Estimates (Adjusted for Inflation)

AA/DEIS FEIS

Atlanta North Line Extension HR 88.5% 70.1%

Baltimore BWI, Hunt Valley, Penn Station Ext. LR 60.0% 77.1%

Baltimore Extension to Johns Hopkins1 HR 1.3% 2.2%

Chicago SW Transitway HR 69.6% 67.3%

Dallas South Oak Cliff LR 122.9% 122.9%

Denver North I-25 Bus/HOV 76.7% 76.7%

Denver Southwest LRT 60.4% 59.3%

Jacksonville Skyway Express AGT 177.8% 177.8%

Los Angeles Red Line HR 61.7% 57.3%

Miami Omni and Brickell Extensions AGT 86.4% 84.4%

Pittsburgh Airport Busway/ Wabash Bus/HOV 121.3% 121.3%

Portland Westside – Hillsboro LRT 108.0% 116.2%

Salt Lake City I-15/State Street LRT 97.4% 76.3%

Sa…