Draft Evaluation of the
Florida High Speed Rail Project
January 1998
Draft Ridership Study

Prepared by
Wendell Cox
Wendell Cox Consultancy

March 1998

Abstract

The Florida Department of Transportation and Florida Overland Express intend to build and operate a high speed rail line between Miami, Orlando and Tampa. Last year, the James Madison Institute published research indicating that the high speed rail line was likely to produce a substantial deficit, which would be the responsibility of Florida taxpayers. Recently, FDOT's ridership projections were increased by 28 percent. This report uses eight independent tests to evaluate the new ridership projections and finds that under each, a substantial deficit would occur. Two comprehensive tests are used to estimate the overall deficit from not only ridership, but also higher costs, which conclude that the eventual deficit could cost Florida taxpayers from $11 billion to $37 billion more than planned.

Executive Summary

A 1997 report by the James Madison Institute concluded that the proposed Miami-Orlando-Tampa high speed rail line would carry many fewer passengers than projected, cost much more than projected and expose the taxpayers of the state to enormous financial exposure. New ridership projections were provided in January 1998. The new 2010 ridership projection is up nearly 30 percent from the original FOX projection, while overall fare revenues are projected to be seven percent higher. Fare revenues were projected to rise at a lower rate than ridership because a lower average high speed rail fare was assumed.

Florida taxpayers would pay for deficits created by unreliable forecasts: Accurate ridership projections are crucial to the success of this project. In the final analysis, the taxpayers of Florida will take the entire risk for both over projection of revenues as well as under projection of capital and operating costs. The high speed rail developers will be liable only to the extent of their investment (less than 5 percent of project capital and debt service cost). Private bondholders will be protected by a state guaranteed "covenant to complete and operate," which will have the same effect as the full faith and credit of the state of Florida.

New ridership forecasts appear to be unreliably high: Considered in the light of recent market and research developments, the new ridership and revenue projections appear to be high and are likely to translate into substantial additional subsidy obligations for Florida taxpayers. Eight independent sensitivity analyses (tests) are presented. Each of the deficit amounts estimated assumes that the project would be built and operated within current cost projections. That is considered highly unlikely (see "Comprehensive Tests" below).

The USDOT Federal Railroad Administration forecasts are considerably lower than the new forecasts: Revenue projections by the Federal Railroad Administration (FRA) for virtually the same route are at least 35 percent lower than the new projections. This could require $6.5 billion to $8.5 billion in additional Florida taxpayer subsidies.

2. Proposed high speed rail fares will need to be reduced to compete with air fares: To compete for airline passengers, high speed rail will need to charge fares that are no higher than air fares. Currently air fares are 20 percent lower than projected high speed rail fares, and could fall to at least 32 percent below in response to high speed rail competition. Lowering fares to equal air fares would require $3.5 billion to $6 billion in additional taxpayer subsidies.

The higher airline passenger attraction is inconsistent with falling air fares: Since the original FOX projection, air fares have fallen 52 percent relative to proposed high speed rail fares. According to an FRA formula, this should have resulted in a reduction of 16 percent to 39 percent in high speed rail ridership, instead of a 28 percent increase. This discrepancy could produce an additional deficit of $3 billion to $7.5 billion to be paid by Florida taxpayers.

4. The new projections assume an unreasonably high attraction of airline passengers: At the more realistic air passenger attraction rate estimated by FRA, the Florida high speed rail project would sustain a $7 billion higher deficit, which would be the responsibility of Florida's taxpayers.

The connecting passenger airline market potential appears to be nil. The new projections supplied by French National Railways consulting unit, Systra, assume that nearly 1.5 million annual connecting ("air connect") passengers will be supplied by airlines operating in Florida markets. Systra assumes that airlines could reduce operating losses by using high speed rail for connecting passengers, instead of short distance airline flights. But it appears that airlines are earning profits on short distance flights. There does not appear to be any compelling reason for any airline to outsource its short distance Florida passengers to FOX. Further, Systra appears to overestimate the size of the air connect market. There seems to be no realistic potential for high speed rail to attract any of the "air connect" market, which would reduce high speed rail revenues by $7 billion, which would have to be made up by Florida taxpayers.

6. The short distance projections are unreasonably high: The new projections continue to assume unreasonably high ridership in the short distance markets in which high speed transportation (airlines and high speed rail) is incapable of competing with door to door automobile travel times (the Tampa to Orlando and Palm Beach to Miami corridors). This over-projection is likely to produce a deficit of $1.5 billion, which will be the responsibility of Florida taxpayers.

A higher service level is assumed, which increases costs: The new projections assume a 15 percent higher level of service, which will increase operating costs $1.5 billion.

8. Europe: Lower ridership and higher costs: The two London-Paris-Brussels high speed rail lines are on the brink of commercial collapse. There is a very real possibility that the private developers will either withdraw or have their franchise relinquished. Ridership projections for the two lines is only 10 percent higher than the new Florida projections for a single line, despite a market three times as large as the Florida market. Further, there are indications that new high speed rail lines in Europe will be poorly patronized and sustain heavy losses.

There are no consulting firms that have a record of success in projecting US high speed rail ridership: Despite claims of previous accuracy by the two consulting firms that produced the projections, the new projections contain significant inconsistencies compared to their own previous projections. There are also significant inconsistencies between the projections of the two firms. Moreover, no consulting firm can claim to have accurately projected high speed rail ridership in a nation that has near-market petroleum prices and a vibrant competitive airline industry, because that exists only the United States, where no high speed rail has ever operated.

Comprehensive tests show substantial deficits: Two sensitivity analyses are presented that take into consideration the ridership, revenue, operating cost, capital cost and debt service with respect to the proposed Florida high speed rail line.

National Research Council: Transportation Infrastructure Projections are Often Incorrect: A recent study published by the National Research Council notes that large transportation infrastructure costs are commonly 50 to 100 percent above projections, while usage (ridership) can be 20 to 60 percent lower than projections. If these figures are applied to the Florida high speed rail project, additional taxpayer financed subsidies of $11 billion to $31 billion could be required. The inaccuracy of such forecasts has led noted economist Charles A. Lave to propose that consulting firms be required to bond the accuracy of their projections.

James Madison Institute: Substantial Deficits Projected for FOX: The 1997 JMI Evaluation projected that under the most favorable conditions, the Florida high speed rail line would carry 55 percent fewer riders than forecast, generate 45 percent less revenue, with 15 percent higher capital costs and 29 percent higher operating costs. At these rates --- all below or within the range suggested by the National Academy of Sciences research, Florida taxpayers would be required to pay an additional $11 billion in subsidies. In the worst case, additional taxpayer subsidies of more than $37 billion could be required --- a range similar to that suggested by application of the National Academy of Sciences research findings.

The tab for Florida taxpayers: $11 billion to $37 billion more: These comprehensive tests suggest that Florida taxpayers could be required to provide from $11 billion to $37 billion in additional subsidies to high speed rail. Once the bonds are issued, the taxpayers will have little or no financial control, because of the state's covenant to complete and operate the high speed rail system.

1. Introduction

A high speed rail line has been proposed for the Miami-Orlando-Tampa corridor. Trains would operate at up to 200 miles per hour. The Florida Department of Transportation (FDOT) and the developer, Florida Overland Express (FOX) believe that the line will attract significant numbers of travelers from automobiles and airplanes. The state of Florida would provide $3 billion in subsidies, the federal government $300 million, Orlando International Airport $100 million and Miami International Airport $50 million. FOX would provide $350 million in equity, construct the line, build the trains and operate the system for 40 years. Infrastructure debt of $6.5 billion would be required (1995$).⁽¹⁾

The James Madison Institute published a detailed analysis of the high speed rail project in April 1997, Evaluation of the FDOT-FOX Miami-Orlando-Tampa High-Speed Rail Proposal (JMI Evaluation).⁽²⁾ This analysis was based upon FDOT, FOX, generally available planning and market data and high speed rail experience around the world. The JMI Evaluation found that high speed rail is likely to cost much more than projected, carry many fewer passengers than projected and expose the taxpayers of the state to enormous financial exposure.

Recently FDOT and FOX have published a new ridership study,⁽³⁾ which projects 28 percent more ridership than the original FOX proposal, at 8,256,000 for the year 2010. This new projection is the mid point of two separate projections by Systra, the "consulting arm"⁽⁴⁾ of the French Railways, SNCF,⁽⁵⁾ (8,504,000) and KPMG Peat Marwick (8,008,000).⁽⁶⁾

While the ridership projection has risen, the projected average fare per trip has been reduced by nearly 17 percent from the original FOX proposal (though costs per trip remain higher than automobile and airlines).⁽⁷⁾ Annual fare income is thus approximately seven percent (approximately $30 million annually) above the level in the original FOX proposal.⁽⁸⁾

This report applies eight independent tests to the new ridership projection and provides sensitivity analyses estimating financial implications. The report also applies two comprehensive tests to the ridership and cost projections, including the findings of National Research Council research on the experience in forecasting the performance of transportation infrastructure projects and the JMI Evaluation findings.

2. Why the Ridership Forecast is so Important

The "down-side" accuracy of ridership forecasts is crucial to the financial viability of the Florida high speed rail project. If ridership is lower than projected, revenues will also be lower and the result will be a larger capital and operating deficit, and a bailout of project bonded indebtedness could be required.

In this project, the private developers (FOX) risk only their capital --- a substantial amount, but very small in relation to project costs and the potential for a larger deficit. Moreover, FOX profits are paid before debt service payments are made. Promoters have claimed that the state will be provided with a "fixed price" under the contract with the developers. A fixed price contract, however, will provide the state no guarantees, since the private developers will not (and probably could not) supply financial guarantees to bond against capital cost, operating cost and subsidy increases. The FOX partners will limit their legal liability to the extent of their investment, which is less than five percent of the project cost. As would be the case with any rational business organization, FOX cannot be expected to shoulder the substantial losses that could arise from what have become typical capital and operating cost overruns, or substantial shortfalls in fare revenue arising out of ridership forecasts that were too high.

The project will be financed with Florida tax revenues and bonded indebtedness.⁽⁹⁾ FDOT and FOX are seeking federal guarantees for all or part of the bonded indebtedness. The bond holders, however, do not represent a resource for bailing out the project if revenue projections are not reached. Like FOX, the bondholders will have no risk. To make the bonds marketable will require that the state of Florida covenant to complete and operate the project. This represents a "de facto" pledge of the full faith and credit of the state of Florida. This leaves the risk for financial failure to the taxpayers of Florida.

The risk is genuine. Each one percent reduction in fare revenue translates into a nearly $200 million bill for state taxpayers.⁽¹⁰⁾ Each one percent increase in costs translates into more than $150 million in additional taxpayer subsidies. The record with respect to large infrastructure projects suggests that the Florida high speed rail ridership and cost projections will prove to be considerably less accurate (see "Unreliable Projections").

The financial risk is substantial. If, for example, revenue falls 20 percent short of forecast, revenues would be insufficient to fund the bonded indebtedness and Florida taxpayers would be required to bail the project out financially. Similarly, a 15 percent increase in projected costs would also leave the project unable to service its debt --- even if the ridership and revenue projections were met. The financial status of the project is tenuous. The project's success depends upon closely achieving the ridership, revenue and cost projections.

The risk of failure could be imminent. As was indicated in the 1998 JMI Evaluation, project revenues could be insufficient to service bonded indebtedness within five years (see "Developments in Europe: Losses and Higher Subsidies").

3: Independent Tests

A series of independent tests (sensitivity analyses) were applied to the new ridership and revenue projections. These tests assume that capital and operating expenses will be at the projected level --- an assumption considered highly unlikely (see "Comprehensive Tests" below).).

3.1: The New Projection: Higher than USDOT FRA Projection

The new ridership and revenue projections are much more optimistic than projections developed for the US Department of Transportation Federal Railroad Administration (FRA)⁽¹¹⁾ over virtually the same route. The Tampa-Orlando-Miami corridor was included in a review of 10 high speed rail corridors, none of which were projected to be profitable by 2020. FRA projected the 2020 Florida high speed rail revenue would be 45 percent lower than the original FOX projections.

3.2: The Projected Rail Fares are Higher than Airfares

More than 30 percent of the Systra-KPMG projected high speed rail ridership is expected to be diverted from the airlines.⁽¹²⁾ Attraction of air travelers will require, among other things, high speed rail fares that are competitive with (no higher than) air fares. Average air fares are now nearly 20 percent below high speed rail fares.⁽¹³⁾ The present air fare structure is not indicative of the ability of airlines to compete on price. At their lowest point in the last two years, Florida air fares were 32 percent below the new lower rail fares.

3.3: Ridership is Up Despite Significant Air Fare Drop

The new projections are questionable in light of the downward trend in air fares.

At the time of the original projections, FOX assumed that airfares would be approximately 65 percent above the projected high speed rail fares. Air fares are now 20 percent below projected high speed rail fares.

In an analysis of the Florida market, FRA estimated that a 30 percent reduction in air fares would translate into a 10 to 24 percent reduction in high speed rail ridership.⁽¹⁵⁾ Air fares have dropped 52 percent relative to the high speed rail fares as originally projected by FOX. At the FRA rate, this would translate into a reduction of 16 to 39 percent in diversion of airline passengers --- 300,000 to 800,000 passengers.

Instead, however, the new projections indicate an increase of 28 percent in high speed rail ridership (Chart: Relationship: Change in Air Fares & High Speed Rail Ridership). At least one of the two projections is faulty. In view of the inordinately high airline market share projected by Systra-KPMG and the unsupportable air connect passenger projection (below), the earlier projection would appear to be the least incorrect. (The JMI Evaluation projected the original FOX projections to be nearly 100 percent high.)

3.4: The Projected Air Passenger Attraction is Unreasonably High

The new ridership projections assume that the high speed rail line will attract 58 percent⁽¹⁶⁾ of the airline market. This figure appears to be well beyond the range of probability.

FRA projected a 21 percent high speed rail airline market share for virtually the same route. None of the 10 high speed rail corridors studied by FRA achieved ridership approaching the 58 percent FOX projection.

High speed rail has attracted only 30 percent of the air market in the London to Paris corridor, and 17 percent in the London to Brussels corridor.⁽¹⁷⁾

In the Washington-New York corridor, where Amtrak's near high speed rail trains offer point to point time savings relative to airline service, rail achieves only a 40 percent of the combined air-rail market share, despite first class rail fares that are at least one-third less costly than airline coach fares and more spacious and comfortable accommodations.⁽¹⁸⁾

The airline business is fluid and competitive. Airlines will respond to competition by FOX in the same manner that they have responded to airlines that charge lower fares --- by lowering fares. It is reasonable to assume that FOX would attract no more than 20 percent of airline passengers.

3.5: The Air-Connect Market Potential Appears to be Nil

Perhaps the weakest element of the new projections is the finding that high speed rail would become a subcontractor to the airlines for providing service in Florida markets. This "air connect" market segment would constitute more than 15 percent of FOX ridership. This proposition is based upon three assumptions that are likely to be faulty:

That the airlines are losing money on shorter flights.⁽¹⁹⁾

That it is in the interest of the airlines to outsource service to FOX.

That the number of "air-connect" passengers on two airlines is sufficient to generate 1.5 million annual FOX riders.

In its only serious criticism of the Systra-KPMG reports, the "peer group" expressed concern about the size of the "air connect" attraction assumption.⁽²⁰⁾ It is notable that FOX has provided no expressions of interest from either of the two potential partners, American Airlines and Delta Airlines.

The air connect market as described: Most of the ridership attracted from airlines is projected to be "air connect" passengers (connecting passengers). Air connect passengers begin or end their trips outside the high speed rail corridor. For example:

A domestic air connect passenger could be traveling from Indianapolis to Miami, with an intermediate stop and usually a plane change in Orlando.

An international air connect passenger could be traveling from Rio de Janeiro to Orlando, with an intermediate stop and perhaps a plane change in Miami.

Systra assumes that FOX would establish a "code share" arrangement with a hub airline in Miami and a hub airline in Orlando. A code share arrangement involves the use of smaller regional airlines to provide service under the larger carrier's name and logo in corridors that have insufficient demand or are too short to operate full sized airline jet equipment. Many airlines have "code share" arrangements with smaller airlines to provide services from hub airports to airports to which demand is too low or the distance is too short for the airline to cost effectively employ its jet fleet.

Systra indicates that in 1997 1.7 million air connect passengers traveled in the high speed rail corridor.⁽²¹⁾ By 2010, Systra projects the air connect market to grow to 2.7 million annual passengers. Approximately 55 percent of air connect passengers are projected by Systra to transfer to high speed rail.

Short distances flights do not necessarily lose money: Systra claims that the two hub airlines could achieve "huge cost savings by transferring their local connecting passengers to the rail mode." However, short distance flights appear to be profitable, and it is unclear how the airlines would achieve "huge cost savings" by using FOX. For example, Comair, the Delta code share operator in the Orlando to Miami and Orlando to Fort Lauderdale markets earned a profit margin more than double that of Delta Airlines in 1996, and ten times that of American Airlines.⁽²²⁾ It seems unlikely that the stockholders of either Delta Airlines or American Airlines would permit those carriers to make such substantial investments in their own regional airline subsidiaries, if they produced losses. Southwest Airlines' aggressive entry into Orlando and Tampa to Southeast Florida markets would not be rational if profits were not anticipated.

If the airlines were losing money on code share flights, the problem could be immediately solved by discontinuing regional services. Since deregulation, US airlines have had the authority to discontinue unprofitable routes whenever they deem appropriate and they have done so. For example, American Airlines has closed entire hubs in Raleigh-Durham, San Jose and Nashville due to unprofitability.⁽²³⁾

Systra indicates that the cost of operating smaller aircraft is likely to increase in the future. It is not clear why this should be given the fact that airline unit costs have been falling in inflation adjusted terms.⁽²⁴⁾ It is expected that short distance airline flights will become more cost effective in the years to come. Much more efficient smaller, "regional" jets are being purchased by many airlines and assigned to shorter routes.⁽²⁵⁾ Some are already being operated in Florida markets, and many more will be in the future. This means that code sharing with FOX could be less attractive to the airlines in the future. Moreover, as airline and airspace capacities are increased, lower fare airlines such as Southwest Airlines, could begin to serve more airports, such as Miami and West Palm Beach.⁽²⁶⁾

The size of the potential air connect market has been over-estimated: This can be illustrated by an analysis of the present air connect market.

Systra indicates that 954,000 annual air connect passengers traveled in the Central Florida to Southeast Florida market in 1997 (Orlando to Miami and Orlando to Fort Lauderdale).

One airline has a hub in Miami --- American Airlines, and there is one hub airline in Orlando that provides service between Orlando and the Southeast Florida market (Delta Airlines).

If all of the seats in every American Airlines and Delta Airlines code share flights in the Central Florida - Southeast Florida market were filled with air connect passengers, there could be no more than 700,000 air connect passengers.

It is estimated that the American Airlines code share flights carry approximately 200,000 annual air connect passengers and that the Delta Airlines code share flights carry approximately 150,000 annual air connect passengers.⁽²⁷⁾

If FOX were able to successfully execute code share agreements with American Airlines and Delta Airlines, it is estimated that approximately 560,000 annual air connect passengers would be carried in 2010 by FOX in the Central Florida - Southeast Florida market --- 480,000 fewer (40 percent less) than Systra projects.

Airlines have no reason to "code-share" with FOX: Systra assumes that FOX will execute code sharing arrangements with airlines only at Orlando and Miami hubs,⁽²⁸⁾ and further indicates that FOX "can attract a large portion of these (air connect) passengers only by having an agreement with the airlines companies." It is unlikely that any such agreements will be executed.

Under the FOX assumptions, airlines would outsource (contract) service to air connect passengers to FOX. But much of the code-sharing service in Florida is not out-sourced. It is provided by the airlines themselves through wholly owned subsidiaries. All American Airlines code-share service is provided by its "American Eagle" subsidiaries, while some Delta Airlines code-share service is provided by its "Delta Express" subsidiary. With the airlines having determined that it is not in their interest to outsource shorter flights to smaller regional airlines, it seems unlikely that they would outsource the same service to a high speed rail operator (FOX).

Cost reduction does not appear to be the driving consideration in airline management decisions on code sharing arrangements. For example, an argument could be made that in the Miami - Orlando market, American Airlines and Delta Airlines could reduce costs by combining their code sharing flights, operating larger, more cost effective aircraft. They avoid this arrangement not only in Florida markets, but in virtually all other markets. In some shorter distance markets, luxury bus service could be used to reduce costs relative to smaller airplanes, while not seriously retarding travel times. Yet there are no significant bus code-share arrangements among the major airlines.

There are practical problems with combining the code sharing flights of different airlines. Currently, code share flights are scheduled to meet the incoming and outgoing "banks" of flights at hub airports. Use of high speed rail as a code share provider could make this problem worse. Passengers would be inconvenienced by longer walks between gates, a second security clearance, potentially longer connection times and longer travel times. This would probably be unacceptable to airline marketing managers.⁽²⁹⁾

Airlines do not share code sharing flights as a matter of marketing practice. Each hub airline with code share arrangements requires that its logo and not that of any competitor appear on all aircraft. From a marketing perspective, it is unlikely that both American Airlines and Delta Airlines would have a code sharing arrangement with FOX under which the logos of both airlines would appear on rail cars.

According to Systra, 45 percent of air passengers are local passengers who travel within Florida markets, rather than air connect passengers. FOX proposes that the airlines would cancel flights and route its air connect passengers instead by high speed rail. Cancellation of these flights would leave local air passengers without service. It seems highly unlikely that the airlines would abandon their local passenger market.

It would therefore appear that the most likely scenario is that FOX would have no code sharing arrangements with airlines. FOX has provided no evidence that either American Airlines or Delta Airlines intends to cancel their own services and instead transfer their own passengers to high speed rail.

3.6: Short Distance Ridership is Still Grossly Over-Estimated

Originally, FOX projected more than 1.7 million annual passengers would be attracted from the Interstate 4 corridor between Tampa and Orlando --- more than five percent of the automobile market. The JMI Evaluation considered this figure exceedingly high because high speed travel modes are comparatively uncompetitive in time and cost in corridors under 100 miles.⁽³⁰⁾

In a short corridor, an automobile averaging 45 to 60 miles per hour can provide a faster door to door trip than an airplane capable of 600 miles per hour or a train capable of 200 miles per hour. If Tampa-Orlando were a lucrative high speed market, significant numbers of people would be using airlines today --- but they are not. Systra projects that high speed rail would attract 100 times as many automobile trips in the Tampa-Orlando corridor as there are local air trips today.

The new projections are lower for the Tampa-Orlando corridor. The reduction, however, is canceled by a new higher projection for the Miami-Palm Beach corridor. It is estimated that the short distance ridership projection could be 1,500 percent high.⁽³¹⁾

3.7: Higher Service Frequencies are Assumed

The new projections are based upon service frequencies that are approximately 20 percent higher than was originally proposed by FOX.⁽³²⁾ For example, the original FOX proposal anticipated 25 trains per day in each direction between Orlando and Miami. The new Systra projections anticipate an average of 32 trains daily in each direction between Orlando and Miami. This additional train service will increase operating costs and could increase infrastructure costs.

3.8: Developments in Europe: Losses and Higher Subsidies

There are already early indications in Europe of the type of financial difficulties the JMI Evaluation has predicted for the Florida high speed rail project.

A consortium of private developers were selected in 1996 by the British government to operate high speed rail service from London to Paris and Brussels (Eurostar service) and to build a high speed link from the Channel Tunnel to St. Pancras Station in London. Ridership, at six million annually, is running well behind the original SNCF (French National Railways --- parent of Systra) projections of 15 million and a revised projection of nine million. The Systra report attributes the failure to reach the nine million ridership projection to service interruptions in the Channel Tunnel. In the latest year, revenue was 40 percent below projections, and a loss of $240 million was sustained --- larger than its gross revenues of $220 million.⁽³³⁾ However, even if the nine million ridership projection had been reached, the service would have remained below revenue projections. It is now projected that the project will sustain losses of $5 billion over the next 50 years.⁽³⁴⁾

In January with capital costs rising 20 percent above projections made only two years before, the consortium approached the government to seek nearly $2 billion more in subsidies to complete the project. Their request was denied and the private developers have been given until the end of March to propose financing alternatives for the project. If such arrangements cannot be made, the private developers will relinquish their franchise. This is not a negative reflection on the private developers --- they have a fiduciary responsibility to their stockholders to abandon any endeavor that looks likely to produce intolerable losses.

There is, however, an important difference between the Channel Tunnel link and the Florida High Speed Rail project. In Britain, without the artificial market mechanisms of public bond or financial guarantees, the project could be abandoned even before construction begins. In contrast, the de facto full faith and credit of the state of Florida could permit that project to proceed to construction, with state officials being called upon to provide additional subsidies as costs escalate during the construction process. This has occurred in other recent large infrastructure projects, such as Denver International Airport, the Channel Tunnel itself, and the Boston Central Artery. Indeed, these projects tended to escalate in cost by approximately $1 billion annually during construction --- a figure nearly matched even before construction started in the case of the Channel Tunnel link. In the United Kingdom, the risk for cost overruns belongs to the developer. In Florida, the risk for cost overruns belongs to the taxpayers.

Further, the London-Paris-Brussels market is considerably larger than the Miami-Orlando-Tampa market. The Florida high speed rail project would be a single line serving a population of 8.4 million. The London-Paris-Brussels market has a population of nearly 25 million and is served by two high speed rail lines (London to Paris and London to Brussels). High speed rail has a considerable cost advantage over automobiles partly because of costly ferry and Channel Tunnel shuttle fares. High speed rail also has a greater travel time advantage over the automobile because of the time necessary to transition from freeways to Channel Tunnel shuttles. Yet the revised annual projection of nine million annual riders for the two London-Paris-Brussels lines is little more than the new 8.3 million annual projection for a single line in the Florida market barely a third the size.

But the problems with high speed rail costs and ridership are even broader. French government railway official Claude Martinand is reported to have told a recent Paris transportation seminar that "with few exceptions, France's and Europe's high-speed rail projects would have very low passenger volumes and he predicted that one of the much-touted EU priority projects - the Paris-Brussels-Amsterdam-Cologne TGV (high speed rail) link - will be a financial disaster because the level of users will never pay back the ... cost of building it, nor the ongoing costs of running it."⁽³⁵⁾

3.9: Summary of Independent Tests

The eight independent tests project deficits ranging from $1.5 billion to $12.5 (the average is $5.5 billion).

The comprehensive tests described later result in much higher additional Florida taxpayer obligations, because they take into consideration the operating and capital cost overruns that are probable as the project proceeds.

4: Unreliable Projections

There are also problems with the projection process itself.

The projecting firms: no relevant US experience: In relation to the FDOT financed ridership study, both Systra/SNCF and KPMG provide detailed information on the accuracy of their projections in previous projects.

Systra points to projection successes with respect to high speed rail lines in France and the Channel Tunnel link. As was noted above, Channel Tunnel high speed rail ridership is running well below the original forecasts (6 million annual riders compared to the 15 million annual rider projection).

The French experience is simply not transferable to the United States.

The cost of driving is much more than high speed rail fares in France. Exorbitant taxes raise the price of gasoline in France to $4.50, approximately four times that in the US. Intercity freeways are subject to excessively high tolls.

The French government has required air fares on competing routes to be higher than high speed rail fares, to encourage rail ridership.⁽³⁶⁾

France's artificial market manipulations have advantaged high speed rail compared to the automobile and airlines. High speed rail will not be the recipient of such preference in the US. In contrast to France, the US is characterized by much nearer market gasoline prices and the cost of driving will be lower than high speed rail fares. The US, unlike France, has a mature, competitive, deregulated air market, in which carriers have established their own fares, without government interference, for nearly two decades. The market price of air fares in Florida markets is lower, not higher than proposed high speed rail fares. In short, Systra has no experience that would demonstrate its projection accuracy or inaccuracy with respect to high speed rail in the much less regulated US market --- because no high speed rail line has been built in the US. The French experience is simply irrelevant to the much different US market conditions.⁽³⁷⁾

KPMG cites a projection success with respect to near high speed rail service in the US northeast corridor --- the Northeast Corridor Improvement Project. This assignment did not involve projecting ridership on a new route that was to be built. It was rather a projection of ridership after improvements were made to an existing route --- a considerably less risky undertaking.

No consulting firm has a history of successful high speed rail ridership projections in the US or any similar market, because high speed rail has not been built in such a market. Any such forecasts must therefore be highly speculative and subject to significant error.

Forecasting Discrepancies between Firms: There are considerable discrepancies between the new Systra and KPMG projections.

KPMG projects 500,000 more passenger diversions from the automobile than Systra.

Systra projects more than 100 percent more induced ridership (riders not diverted from airlines or automobiles) than KPMG.

Forecasting Discrepancies within Firms: There are substantial discrepancies between the ridership projections of each consulting firm.

Systra now projects annual ridership at 2 million more than its original 1995 FOX projection --- an increase of 32 percent (Chart: Comparison: 1995 & 1997 Projections).

KPMG projects annual ridership at 2.5 million more than its 1995 projection on the FOX proposal prepared for FDOT --- an increase of 45 percent (Chart: Comparison: 1995 & 1997 Projections).

Systra estimates the number of short distance riders between Southeast Florida and Palm Beach at 250 percent higher than its own original FOX projection.

Neither firm provides a detailed explanation of the differences between these projections. It is clear, however, that if high speed rail is built in Florida, each firm will have produced at least one set of incorrect projections.

Forecasting: An Inexact Science: There are valid reasons why ridership and revenue projections are often high and cost projections are low. The planners and administrators who oversaw each of the projects above can supply a litany of reasons why forecasts were not met. Unforeseen circumstances, such as additional environmental mitigation requirements, changes to project scope and construction delays can add to costs. Usage projections can be high because projected demographic trends or market conditions do not materialize. But there are additional reasons for the unreliability of forecasts. Infrastructure decisions are often made without regard to the historic inaccuracy of forecasts. Forecasts can also be influenced by political factors.

... forecasts that underscore a priority which is out of political favor are likely to be ignored, whereas forecasts that support politically favorable positions are likely to be embraced.⁽³⁸⁾

Major infrastructure projects can take on a "life of their own." The experience demonstrates that, once authorized, even cost escalation that double or triples the cost of a project will not result in its cancellation. There will always be detailed explanations for cost escalation and failure to attract projected ridership and revenue --- some are more valid than others. But with regard to subsidized projects, the effect is the same --- the excess cost is paid by the taxpayers.

It is not unusual for different ridership modeling approaches to produce different results. It is simply an indication of the volatility --- and possible inaccuracy --- of ridership projections. It is simply not known which of the projections is most correct or whether any of the projections would be remotely close to the actual ridership that would be carried.

4: Comprehensive Tests

Two sensitivity analyses are presented that take into consideration the comprehensive impact of all factors (ridership, revenue, operating cost, capital cost and debt service) with respect to the proposed Florida high speed rail line.

4.1 National Research Council Study

The potential inaccuracy of projections is not limited to ridership. The JMI Evaluation summarized the experience with recent large Florida, U.S. and international transportation infrastructure projects, demonstrating that usage and cost projections have been very inaccurate, which has translated into higher user and taxpayer costs.⁽³⁹⁾

This finding has been repeated in a recent international study published by the National Research Council's Transportation Research Board, which is perhaps the most highly respected professional transportation organization in the United States:⁽⁴⁰⁾

... cost overruns of 50 to 100 percent are common and ... overruns of more than 100 percent are not uncommon. Traffic forecasts that are off by 20 to 60 percent when compared with actual development are frequent in large transportation projects.

Dr. Charles Lave, chair of the Economics Department at the University of California, Irvine has suggested that the solution to forecasting inaccuracies would be to require consulting firms to bond their forecasts.⁽⁴¹⁾ Nothing short of such a solution would guarantee that the taxpayers will not be forced to finance a financial bailout of the Florida high speed rail project.

4.2: James Madison Institute Study

The JMI Evaluation projected that, under the most optimistic circumstances, the Florida high speed rail line would carry 55 percent fewer riders than projected, and generate 45 percent less revenue. Capital costs were projected at 15 percent higher than projections, while operating costs were forecast at 29 percent higher than projection.

5: Florida Taxpayer Tab: Another $11 Billion to $37 Billion

The two comprehensive tests produce similar results and suggest a range of additional taxpayer obligation of $11 billion to $37 billion (Table: Comprehensive Test Comparisons)

There is nothing in the new ridership projections to suggest moderation of the previous JMI Evaluation projections. Once the bonds are issued, there will be no turning back. The taxpayers of the state of Florida will be required to provide the additional subsidies --- regardless of the cost --- because Florida's covenant to complete and operate will have, in effect, pledged the state's full faith and credit.

1. All financial amounts are in 1997$ unless otherwise indicated.

2. Wendell Cox, Evaluation of the FDOT-FOX Miami-Orlando-Tampa High-Speed Rail Proposal, James Madison Institute, April 1997.

3. Systra and KPMG Peat Marwick, Preliminary Summary Report: Florida High Speed Rail Project: Ridership Study, January 1998.

4. Systra and KPMG Peat Marwiick, Preliminary Summary Report: High Speed Rail Project: Ridership Study, January 1998, p. 1.

5. Systra, Florida Overland Express Ridership and Revenue Study: Draft Final Report, January 1998.

6. KPMG Peak Marwick LLP, Florida Overland Express High Speed Rail Study: Draft Ridership and Revenue Report, January 1998.

7. The 1995 FOX projected average passenger fare was $59.85 ($62.50 in 1997$). The 1997 projected average passenger fare is $51.83, 17 percent lower. Calculated from FOX, Systra and KPMG data.

8. All financial data in this report is in 1997 dollars (inflation adjustment has been performed).

9. FDOT and FOX are also seeking a $300 million grant from the federal government --- this would represent approximately five percent of project capital costs at current projections.

10. Calculated from original FOX/FDOT data.

11. US Department of Transportation Federal Railroad Administration, High-Speed Ground Transportation for America, September 1997.

12. This is the percentage of riders on high speed rail that would have been on airplanes if high speed rail were not available. In contrast, the new projections indicate that 58 percent of the airline passengers in the high speed rail corridor would switch to high speed rail.

13. Comparison of weighted average of high speed rail fares and air fares in the Central Florida to Southeast Florida and Tampa Bay to Southeast Florida markets. Calculated from data in Systra report and Domestic Airline Fares Consumer Report (US Department of Transportation, 1996: 3^rd & 4^th quarter, 1997: 1^st & 2^nd quarter).

14. The lower figure assumes air fares at their current level. The higher figure assumes air fares at their lowest level within the past two years.

15. Federal Railroad Administration.

16. This is the percentage of air passengers projected to switch to high speed rail. These passengers would comprise 30 percent of the high speed rail ridership.

17. Calculated from data in Systra report.

18. Cox, p. 26.

19. High Speed Rail Peer Review Board, Draft Preliminary Peer Review Report of Ridership and Revenue Forecasts for the Florida Overland Express, February 3, 1998.

20. FOX and FDOT have established a "peer group" of transportation and financial experts to review projections. In their report (High Speed Rail Peer Review Board, Draft Preliminary Peer Review Report of Ridership and Revenue Forecasts for the Florida Overland Express, February 3, 1998), the peer group referred to the projections on air connect passengers as "the most problematic aspect of the forecasts" and recommended that letters from airlines be submitted to document the assumptions. It should also be noted that peer group analysis does not necessarily reduce the uncertainty of projections and has been widely used in major infrastructure projects, including projects that fallen far short of their ridership projections.

21. The Systra report does not indicate the source of this estimate and JMI was unable to obtain sufficient documentation from FOX before this report's deadline.

22. Calculated from "The World Airline Report," Air Transport World, July 1997.

23. Perry Flint and Danna K. Henderson, "American at Bay," Air Transport World, March 1997.

24. Cox, p. 23.

25. A unit of FOX partner Bombardier is one of the world's leading suppliers of the new regional jets and other airliners designed for regional markets.

26. Airport capacity is being increased in Florida and additional airspace capacity is being created by improved computer flight technology. See Cox, p. 36-37.

27. Assumes load factors from the Systra report and a distribution of direct and air connect passengers equal to the overall ratio in the Central Florida - Southeast Florida market.

28. Systra report, page 5-20.

29. Projected FOX travel times have increased since the original FOX projections. For example, travel time from Orlando ro Miami has increased from 1:25 to 1:32.

30. U.S. Department of Transportation data indicates that the overwhelming majority of intercity trips of less than 100 miles are by automobile. Less than 0.5 percent of such trips are by airplane. Air travel tends to be slower than auto travel times because their "overhead" time --- traveling to the airport, checking in and waiting to depart. As a result, approximately 99 percent of travel in such short travel markets is by auto. High speed rail would face virtually the same competitive disadvantages (Source: calculated from Nationwide Personal Transportation Survey).

31. 1.8 million annual riders are anticipated in the short distance Tampa-Orlando and Miami-Palm Beach Markets. Using Nationwide Transportation Survey data, it is estimated that high speed rail demand in these corridors would be approximately 100,000 annually.

32. Based upon an estimate of train miles calculated from the original FOX proposal and the Systra report.

33. "Channel Tunnel: Prescott Vows to Ensure Link is Built," The Financial Times, February 25, 1998.

34. "Channel Tunnel" Last Gasp for Tunnel Rail Project," The Financial Times, January 30, 1998.

35. "Senior Railway Official Admits Folly of European Union Fail Policies," World Highways, January 1998 (quotation is from the article, which summarized Mr. Martinand's remarks).

36. Cox, p. 9.

37. There are a number of additional factors that render the French experience valueless in US projections, such as higher urban densities, more effective transit systems, higher automobile prices and lower incomes (purchasing power parity basis).

38. William Ahser quoted in Edward A. Mierzejewski, "Recognizing Uncertainly in the Transportation Planning Process: A Strategic Planning Approach," paper presented to the 76^th Annual Meeting of the Transportation Research Board (Washington, DC, January 1997).

39. Cox, p. 14-17.

40. Mette K. Skamris and Bent Flyvbjerg, "Accuracy of Traffic Forecasts and Cost Estimates on Large Transportation Projects," Transportation Research Record (Washington, DC: Transportation Research Board, National Research Council), 1996.

41. Charles A. Lave, "Playing the Rail Forecasting Game," TR News 156, September-October 1991.

42. The lower estimate assumes costs at 50 percent higher and ridership 20 percent lower than projections. The higher estimate assumes costs at 100 percent higher and ridership at 60 percent lower than projected. If cost overruns were 200 percent --- less than Denver International Airport --- an additional $45 billion would be required from Florida taxpayers.

43. Cox, p. 32. Converted to 1997$.

WENDELL COX CONSULTANCY
Contact Us by E-Mail
The Public Purpose | Demographic Briefs | Government Cost Review | Intercity Transport Fact Book
Labor Market Reporter | Realities | School Transport Fact Book | Transport Fact Book
Urban Transport Fact Book | Competitive Tendering Website
International Competition & Ownership Conference | Publications | New Items | To Subscribe (Free)