Intercity Passenger Rail: Financial and Operating Conditions Threaten
Amtrak's Long-Term Viability (Chapter Report, 02/06/95, GAO/RCED-95-71).

Amtrak's financial condition, always precarious, has declined steadily
to the point that its ability to offer service over its existing 25,000
mile system is seriously threatened.  It is unlikely that Amtrak can
overcome its problems in financing, capital investments, and service
quality without significant increases in passenger revenues or subsidies
from federal, state, and local governments.  In recent years, Amtrak has
responded to deteriorating conditions by taking on debt, deferring
maintenance, and cutting staff.  Some of these measures have diminished
the quality and reliability of Amtrak service.  In December 1994, Amtrak
announced an aggressive plan to reduce expenses by $430 million annually
by eliminating routes, retiring its oldest cars, cutting staff, and
improving productivity.  Although this plan is an aggressive first step,
it will not solve the railroad's long-term problems.  Amtrak and the
federal government face hard choices.  GAO believes that continuing the
present course--maintaining the same funding level and route system,
even with proposed service cuts--is neither feasible nor realistic
because Amtrak will continue to deteriorate.  GAO discusses the
implications of several alternatives, ranging from privatization of the
railroad to limiting service to routes that carry the largest number of
passengers in the most cost-effective manner. GAO summarized this report
in testimony before Congress; see: Amtrak: Deteriorated Financial and
Operating Conditions Threaten Long-Term Viability, by Kenneth M. Mead,
Director of Transportation Issues, before the Subcommittee on Railroads,
House Committee on Transportation and Infrastructure. GAO/T-RCED-95-98,
Feb. 7, 1995 (15 pages).

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-95-71
     TITLE:  Intercity Passenger Rail: Financial and Operating 
             Conditions Threaten Amtrak's Long-Term Viability
      DATE:  02/06/95
   SUBJECT:  Railroad transportation operations
             Railroad industry
             Financial management
             Cost effectiveness analysis
             Future budget projections
             Differential subsidies
             Federal aid to railroads
             Mass transit operations
             Transportation industry
             Maintenance costs
IDENTIFIER:  Amtrak Northeast Corridor
             Amtrak Northeast Corridor Improvement Project
             Beech Grove (IN)
             Bear (DE)
             Wilmington (DE)
             
**************************************************************************
* This file contains an ASCII representation of the text of a GAO        *
* report.  Delineations within the text indicating chapter titles,       *
* headings, and bullets are preserved.  Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are           *
* identified by double and single lines.  The numbers on the right end   *
* of these lines indicate the position of each of the subsections in the *
* document outline.  These numbers do NOT correspond with the page       *
* numbers of the printed product.                                        *
*                                                                        *
* No attempt has been made to display graphic images, although figure    *
* captions are reproduced. Tables are included, but may not resemble     *
* those in the printed version.                                          *
*                                                                        *
* A printed copy of this report may be obtained from the GAO Document    *
* Distribution Facility by calling (202) 512-6000, by faxing your        *
* request to (301) 258-4066, or by writing to P.O. Box 6015,             *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time.                                    *
**************************************************************************


Cover
================================================================ COVER


Report to Congressional Committees

February 1995

INTERCITY PASSENGER RAIL -
FINANCIAL AND OPERATING CONDITIONS
THREATEN AMTRAK'S LONG-TERM
VIABILITY

GAO/RCED-95-71

Amtrak's Financial and Operating Conditions


Abbreviations
=============================================================== ABBREV

  DOD - Department of Defense
  DOT - Department of Transportation
  FRA - Federal Railroad Administration
  GAO - General Accounting Office
  ICC - Interstate Commerce Commission
  NECIP - Northeast Corridor Improvement Project

Letter
=============================================================== LETTER


B-259656

February 6, 1995

Congressional Recipients

This report assessing Amtrak's deteriorating financial and operating
conditions was conducted as part of our legislative responsibilities
under the Rail Passenger Service Act (P.L.  91-518, 84 Stat.  1327
(1970)).  The report addresses the likelihood that Amtrak can
overcome its financial and operating problems and presents
alternative actions that could be considered by the Congress in
deciding on Amtrak's future mission and on commitments to fund the
railroad.  On the basis of our review, we are making a recommendation
to the Congress and several recommendations to the President of
Amtrak. 

We are sending copies of the report to the Secretary of
Transportation, the President of Amtrak, and interested congressional
committees.  We will also make copies available to others upon
request. 

This work was done under the direction of Kenneth M.  Mead, Director,
Transportation Issues, who may be reached at (202) 512-2834 if you or
your staff have any questions.  Other major contributors to this
report are listed in appendix V. 

Sincerely yours,

Keith O.  Fultz

List of Recipients

The Honorable Larry Pressler
Chairman
The Honorable Ernest F.  Hollings
Ranking Minority Member
Committee on Commerce, Science,
 and Transportation
United States Senate

The Honorable Trent Lott
Chairman
The Honorable Daniel K.  Inouye
Ranking Minority Member
Subcommittee on Surface Transportation
 and Merchant Marine
Committee on Commerce, Science,
 and Transportation
United States Senate

The Honorable Mark O.  Hatfield
Chairman, Committee on Appropriations
United States Senate

The Honorable Robert C.  Byrd
Ranking Minority Member
Committee on Appropriations
United States Senate

The Honorable Frank R.  Lautenberg
Ranking Minority Member, Subcommittee
 on Transportation
Committee on Appropriations
United States Senate

The Honorable J.  James Exon

The Honorable Bud Shuster
Chairman
The Honorable Norman Y.  Mineta
Ranking Minority Member
Committee on Transportation and
 Infrastructure
House of Representatives

The Honorable Robert Livingston
Chairman
The Honorable David R.  Obey
Ranking Minority Member
Committee on Appropriations
House of Representatives

The Honorable Frank R.  Wolf
Chairman
The Honorable Ronald D.  Coleman
Ranking Minority Member
Subcommittee on Transportation
Committee on Appropriations
House of Representatives

The Honorable Susan Molinari
Chairwoman
The Honorable William O.  Lipinski
Ranking Minority Member
Subcommittee on Railroads
Committee on Transportation and
 Infrastructure
House of Representatives

The Honorable John D.  Dingell
House of Representatives


EXECUTIVE SUMMARY
============================================================ Chapter 0


   PURPOSE
---------------------------------------------------------- Chapter 0:1

Since 1971, the federal government has provided the National Railroad
Passenger Corporation (Amtrak) with over $13 billion to support
intercity passenger rail service.  In early 1994, GAO testified
before several congressional committees that Amtrak's financial and
overall operating conditions had deteriorated, seriously threatening
the corporation's ability to provide an acceptable level of service. 
GAO continued to assess (1) Amtrak's financial and operating
conditions, (2) the likelihood that Amtrak can overcome its financial
and operating problems, and (3) alternative actions that could be
considered in deciding on Amtrak's future mission and on commitments
to fund the railroad. 


   BACKGROUND
---------------------------------------------------------- Chapter 0:2

In 1970, the Congress created Amtrak as a for-profit corporation to
provide nationwide intercity passenger rail service.  Amtrak was
expected to help alleviate the overcrowding of airports and highways
and to offer the public a convenient and efficient transportation
alternative.  Until 1970, the railroads were required to provide
passenger service, but by that year, their combined annual losses for
passenger services had increased to about $1.7 billion in today's
dollars.  In return for eliminating the requirement to provide
passenger service and because of these losses, most railroads
provided personnel and equipment to Amtrak, and it began operations
in 1971. 

Like all major national intercity rail services in the world, Amtrak
operates at a loss, and it has always needed government funding.  In
1995, Amtrak will receive $972 million in operating and capital
grants, funds to improve the infrastructure that Amtrak owns in the
Northeast, and a payment for retirement and unemployment benefits. 


   RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3

Amtrak's financial and operating conditions have declined steadily
since 1990, and Amtrak's ability to provide nationwide service at the
present level is now seriously threatened.  Amtrak's federal support
has increased from $640 million in 1990 to almost $1 billion in 1995,
but this increase has not covered the widening gap between expenses
and revenues.  Also, requirements for capital investment continue to
grow, and unmet needs for new equipment and improvements to
facilities and track now total several billion dollars.  Over the
past several years, Amtrak has taken action to address this situation
by assuming debt, deferring maintenance, and reducing staffing. 
However, many of these actions, while necessary for day-to-day
survival, have simultaneously diminished the quality and reliability
of Amtrak's service.  Most recently, on December 14, 1994, Amtrak
announced an aggressive plan to improve its service quality and
productivity and to reduce its annual expenses by $430 million by
adjusting routes and service, retiring its oldest cars, and reducing
staff.  Amtrak expects that these actions will close the gap between
the operating deficit and federal grants for 1995.  However, the gap
will begin growing again in 1996, and the announced actions do not
resolve the problems with Amtrak's equipment and facilities. 

It is unlikely that Amtrak can overcome its problems in financing,
capital investments, and service quality--and continue to operate the
present nationwide system--without significant increases in passenger
revenues and/or subsidies from federal, state, and local governments. 
Amtrak's ability to overcome these problems is exacerbated by an
unfavorable operating environment.  For example, competition from
airlines in fares has increased, and the growth in Amtrak's revenues
from ridership has suffered as a result.  In addition, Amtrak
estimates that it needs over $4 billion to bring its equipment and
facilities systemwide and its track in the Northeast Corridor\1 into
a state of good repair.  Also, within the next 2 years, Amtrak must
negotiate new labor agreements and may confront substantial
additional costs for new agreements with freight railroads to use
their track. 

Amtrak and the federal government face difficult choices.  GAO
believes that continuing the present course--maintaining the same
funding level and route system, even with the proposed cuts in
service--is neither feasible nor realistic because Amtrak will
continue to deteriorate.  Another option--substantially increasing
funding--would permit Amtrak to make capital investments and improve
service quality so that it could retain current riders and attract
new ones.  This alternative would be costly and extremely difficult
to achieve in the current budget environment.  An option at the other
extreme would be to eliminate subsidies for Amtrak and privatize the
railroad.  This alternative may be difficult to achieve because few
private firms would be willing to assume the risks of providing
intercity passenger service, considering that no Amtrak route earns
sufficient revenues to cover all its costs.  Another option would be
to refocus Amtrak's efforts and greatly realign or reduce the current
route system, retaining service in locations where Amtrak can carry
the largest number of passengers in the most cost-effective manner. 
This option does not preclude retaining relatively unprofitable
routes or operating high-speed service outside the Northeast Corridor
if the states or other entities are willing to make the necessary
capital investments and cover any operating deficits. 


--------------------
\1 The Northeast Corridor is the area between Washington, D.C., and
Boston.  High-speed rail service is currently provided between
Washington, D.C., and New York. 


   PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4


      AMTRAK'S FINANCIAL CONDITION
      HAS REACHED A CRITICAL STAGE
-------------------------------------------------------- Chapter 0:4.1

Over the years, Amtrak has made numerous attempts to reduce expenses
and improve the efficiency of its operations.  To Amtrak's credit,
these actions have served to hold down the corporation's operating
deficit, but they have not arrested the long-term decline in finances
and operations.  Since 1990, Amtrak's problems have accelerated, and
the corporation is finding it increasingly difficult to provide
convenient and efficient nationwide intercity passenger service. 
From 1991 to 1994, revenues were lower than projected, while expenses
were higher than planned.  As a result, Amtrak's revenues and federal
operating subsidies did not cover operating deficits.  Amtrak
overestimated passenger revenues by $600 million from 1991 through
1994.  The projected revenues did not materialize for several
reasons, including declining service quality and increased fare
competition from airlines.  To help cover the gap, Amtrak drew down
its cash resources.  At the end of 1994, it had a negative balance in
working capital of $227 million.  Amtrak also deferred maintenance on
train equipment and reduced staffing levels and some services. 
Despite these efforts, the 1994 deficit exceeded the federal
operating grant by $76 million.  Amtrak had projected that this gap
would increase to almost $200 million in 1995. 

In an effort to address the cash shortfall, Amtrak's Board of
Directors adopted, in December 1994, a plan to cut expenses by
restructuring the railroad's route network and improving
productivity.  Amtrak plans to reduce the train miles of service it
provides by more than 20 percent, but it will accomplish the
reduction largely by cutting back on the frequency of service,
especially on its long-distance routes, rather than by eliminating
routes entirely.  Only 3 short-distance routes and segments of 10
others will be eliminated.  Amtrak plans to negotiate with the
affected states and could retain service where the states are willing
to subsidize the losses on the routes.  Amtrak estimates that these
actions will result in the termination of about 5,000 of its 25,000
employees and will reduce operating expenses by about $200 million in
1995, closing the gap between the projected operating deficit and the
corporation's federal operating grant for 1995.  According to Amtrak,
collective bargaining and/or legislative change may be required to
achieve about 26 percent of these savings.  However, even if fully
implemented, these actions will not solve Amtrak's longer-term
problems.  Revenues will continue to fall short of expenses on most
routes, and Amtrak projects that operating expenses will exceed
operating revenues and the federal subsidy by $1.3 billion from 1996
through the year 2000.  In addition, Amtrak will still need over $4
billion for capital investments. 


      AMTRAK'S REVENUES WILL NOT
      KEEP PACE WITH CAPITAL
      INVESTMENT NEEDS AND OTHER
      COSTS
-------------------------------------------------------- Chapter 0:4.2

The cost of replacing and modernizing Amtrak's physical
assets--maintenance facilities, train equipment, and support
assets--is a greater challenge to the corporation's financial
well-being than resolving the current shortfall in operating funds. 
To cope with funding shortages, in the late 1980s Amtrak started
reducing car maintenance.  By the end of 1993, costly heavy overhauls
were overdue for 40 percent of its nearly 1,900 cars.  Amtrak also
deferred renovating and modernizing its outdated maintenance
facilities, which has contributed to costly and inefficient
operations. 

Focusing exclusively on the shortfall in operating funds masks the
critical problem of Amtrak's capital needs.  Today, the average age
of Amtrak's cars is about 22 years--similar to what it was when
Amtrak first began operating.  Amtrak now estimates that it needs to
invest about $1.5 billion for equipment overhauls and new equipment,
primarily locomotives.  Over the past 10 years, Amtrak's equipment
and facilities have depreciated at the rate of $200 million per year,
while investment has averaged only $140 million.  Yet most of
Amtrak's annual capital grant is already committed to paying off
prior purchases and meeting legal mandates such as environmental
cleanup.  Also, because capital grants are subject to the annual
appropriations process, it is difficult for Amtrak to formulate and
implement long-term investment projects. 

Labor costs are also a major factor in Amtrak's finances.  Beginning
in 1995, Amtrak will be negotiating changes to wages, benefits, and
work rules with the 14 unions that represent 90 percent of its
employees.  Labor costs account for about 52 percent of Amtrak's
operating costs.  Amtrak has done a good job of improving labor
productivity and plans to achieve further increases in productivity. 
However, Amtrak already pays train and engine crews less on average
than freight railroads pay for comparable jobs.  Continuing to hold
down labor costs will present a difficult challenge. 

Amtrak could also face increased costs for track leases and liability
coverage.  Freight railroads own about 97 percent of the track that
Amtrak operates on.  In 1971, Amtrak entered into 25-year agreements
with the freight railroads to compensate them for the use of their
track and for providing essential services such as maintaining track
and stations, dispatching trains, and making emergency repairs to
Amtrak equipment.  The agreements expire in April 1996.  The freight
railroads do not believe that Amtrak's payments, which total about
$90 million annually, are adequate compensation for their services,
and they will seek higher payments.  Freight railroads are also
concerned about their liability in accidents involving passenger
trains and will likely seek reductions in their own exposure or
increases in the amount of risk assumed by Amtrak. 

Passenger revenues are not likely to increase enough over the next
few years to reverse Amtrak's deteriorating condition.  None of
Amtrak's routes--including those in the Northeast Corridor--are
profitable when capital costs are taken into account.  Revenues in
the corridor cover about 65 percent of costs on the routes, compared
with about 50 percent for routes elsewhere.  Furthermore, passenger
revenues have declined about 14 percent in real terms--from over $1
billion in 1990 to about $880 million in 1994.  The decline resulted
from, among other things, a weak economy; intense price competition
from airlines in certain markets; Amtrak's old, unattractive, and
poorly maintained facilities and equipment; and accidents involving
Amtrak trains.  While the economy has recovered and the impact of
train accidents on ridership has begun to abate, the other factors
continue to inhibit growth in ridership. 

Amtrak's fastest growing source of revenues is contracts to operate
local commuter rail systems.  These contracts generated over $270
million in 1994.  Over the long term, Amtrak believes that high-speed
rail service will increase ridership and revenues.  While high-speed
service is now limited to the electrified portion of track between
Washington, D.C., and New York City, Amtrak is extending
electrification to Boston, improving the track, and purchasing new
trains that will allow high-speed service from Washington, D.C., to
Boston.  Amtrak expects its market share between New York City and
Boston to be similar to its 45-percent share between New York City
and Washington, D.C.  To realize this expectation, however, Amtrak
will continue to need funds to expand rights-of-way, rehabilitate
track and facilities, purchase new train equipment, and increase
coordination with freight and commuter operators that share or own
segments of the rights-of-way.  High-speed service beyond the
Northeast Corridor is unlikely without greatly increased federal and
state funding.  Private-sector efforts to sponsor high-speed rail
without substantial government funding have been unsuccessful. 


      REASSESSMENT IS NEEDED OF
      AMTRAK'S MISSION AND
      COMMITMENTS FOR FUNDING
-------------------------------------------------------- Chapter 0:4.3

The Congress needs to make important decisions about the quality and
extent of future intercity passenger rail service, including whether
to maintain the current route system.  To maintain the present
nationwide route network will require increased funding from federal,
state, and/or local governments.  Passenger rail service, however,
competes for limited transportation funds, and unlike aviation,
highways, and mass transit, it does not have access to a federal
trust fund.  State and local governments have some flexibility to
allocate federal transportation funds among different modes, but
their ability to support intercity passenger rail operations is very
limited. 

Increased funding, especially capital investment, would improve
service quality and encourage more riders.  Doubling Amtrak's capital
grant to $500 million annually--a difficult task in today's fiscal
environment--would allow Amtrak to improve its maintenance facilities
and rights-of-way and to purchase new equipment, primarily
locomotives.  But even if gains in efficiency and ridership resulted
from such improvements, GAO estimates that Amtrak would continue to
need more than $400 million in annual operating subsidies through the
year 2000.  If funding for Amtrak is reduced or maintained at its
current level, GAO believes that the route network will have to be
restructured and reduced beyond the recently announced changes so
that quality service can be provided within the available funding. 
By reducing its route network, Amtrak could eventually reduce its
requirements for federal funding.  However, the short-term savings
will be limited without changes in labor agreements and/or
legislation, which, among other things, provide for up to 6 years'
wages for workers who lose their jobs when routes are eliminated. 
Amtrak has identified other legislative changes--such as eliminating
its obligation to pay federal fuel taxes, providing statutory
limitations on punitive damages against Amtrak, and removing
restraints against Amtrak's ability to contract out for work--that it
believes would also reduce future expenses. 

If Amtrak's route network is to be restructured or substantially
reduced, the Congress could direct Amtrak or a temporary commission,
similar to the one established to close military bases, to define and
realign a basic network, and, as appropriate, to recommend routes to
be eliminated.  Options could be developed for routes commensurate
with various levels of federal, state, and local funding.  A basic
network could be defined by determining where Amtrak carries the most
passengers and has the greatest economic potential.  In this regard,
GAO found that 11 of Amtrak's 44 routes earn 68 percent of Amtrak's
revenues.  However, these routes also account for 61 percent of the
railroad's expenses.  Also, interconnections between routes or the
presence of important public benefits defined by the Congress, such
as helping alleviate congestion and pollution, would be relevant in
evaluating how best to restructure the route network.  The basic
network could be augmented by regional routes supported by those
states that were willing to contract with Amtrak to cover shortfalls
between revenues and the full cost of operations. 

If subsidies to Amtrak are eliminated and the railroad is privatized,
it is unlikely that a nationwide passenger rail system could be
preserved.  Under this option, intercity service would be reduced to
a few regional corridors, at most, because only a few well-traveled
routes could potentially generate sufficient revenues to cover
operating costs.  Even in these cases, substantial federal investment
in the infrastructure would likely be needed before the railroad was
privatized.  The experience with the Consolidated Rail Corporation
(Conrail) in 1976 may illustrate how a profitable business can be
established after substantial federal investment.  However, the
analogy with Conrail is limited because GAO found no evidence that
intercity rail passenger operations can be profitable.  In Europe and
Japan, where population densities and geographic conditions are more
conducive to rail travel than they are in the United States,
intercity passenger service requires substantial public funding. 


   RECOMMENDATION TO THE CONGRESS
---------------------------------------------------------- Chapter 0:5

In light of Amtrak's financial and operating problems, GAO recommends
that the Congress consider whether Amtrak's original mission of
providing nationwide intercity passenger rail service at the present
level is still appropriate.  If the Congress decides to reassess the
scope of Amtrak's mission, it could direct Amtrak or a temporary
commission, similar to the one established to close military bases,
to make recommendations and offer options defining and realigning
Amtrak's basic route network so that efficient and quality service
could be provided within the funding available from all sources.  The
Congress could then decide what Amtrak's future route network should
be. 


   RECOMMENDATIONS TO THE
   PRESIDENT OF AMTRAK
---------------------------------------------------------- Chapter 0:6

GAO is making several recommendations to the President of Amtrak to
improve the financial information the corporation provides to the
Congress, including a recommendation for Amtrak to provide the
Congress with proposed legislative changes, such as those noted in
its comments on a draft of this report, that Amtrak believes could
improve its long-term viability.  Amtrak should also estimate the
effect of these changes on its finances and other affected parties. 
These recommendations will facilitate congressional decision-making
on the scope of Amtrak's mission and funding. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:7

Amtrak said that GAO's draft report accurately presented the
financial and operating status of the railroad and correctly
portrayed its problems with capital investment.  However, Amtrak had
four principal points. 

First, Amtrak stated that the draft report understated the impact of
actions adopted by its Board of Directors in December 1994,
contending that the board had, in effect, implemented GAO's
recommendation to redefine and reduce the route system consistent
with the funding available.  GAO revised the report to specifically
highlight the board's actions (see ch.  2).  GAO believes the board's
new business plan is an aggressive first step.  The plan does not,
however, implement GAO's recommendation.  GAO recommends that the
Congress reassess Amtrak's mission and establish likely levels of
federal support for operations and capital investment so that either
Amtrak or a temporary commission could identify options for a route
network that is consistent with the funding available from all
sources.  Amtrak's proposed plan, even if fully implemented, still
leaves a cumulative operating deficit, after federal subsidies, of
$1.3 billion for fiscal years 1996-2000.  Moreover, Amtrak's plan
does not resolve the problem of unmet capital needs, which now total
$2.5 billion in the Northeast Corridor alone.  Furthermore, although
the board set a goal that could ultimately eliminate the federal
operating grant by 2002, doing so was made contingent on Amtrak's
receiving (1) "sufficient capital funding to achieve a
good-state-of-repair," (2) the current level of operating grants
until 2002, and (3) increased funding from states to cover operating
deficits for the services they receive.  These are very significant
funding assumptions that will require congressional endorsement as
part of the process of defining options for Amtrak's future route
network. 

Second, Amtrak raised concerns about the timing of GAO's
recommendation for a commission because the board's recent analysis
showed that reducing the frequency of service was more economical
than closing routes and because Amtrak is a commercially driven
corporation, which should not act like the Department of Defense in
making decisions about closures.  Amtrak's point on timing is well
taken if the railroad's requirements for capital and operating
funding for the national system, as presently constituted, are met. 
If they are not, reductions in service frequency and routes beyond
those announced by the board will be required.  As for Amtrak's
reservations about a commission offering options to the Congress, GAO
recognizes that Amtrak could offer options for defining the national
route network, including realigning and closing routes, and the
recommendation explicitly provides for this option.  However, GAO
realizes the difficulties inherent in deciding which of the 44 states
and more than 500 stations should continue to receive service.  GAO
believes a commission is an option that could provide the Congress
with an independent perspective on defining the national route system
and any necessary realignments or closings. 

Third, Amtrak thought the role of the states in intercity rail
passenger service and the restrictions on the states' access to the
transportation trust funds should be emphasized.  GAO revised the
report to reflect the fact that unlike aviation, highways, and mass
transit, intercity passenger rail has no trust fund.  In 1991, the
Intermodal Surface Transportation Efficiency Act provided the states
with some flexibility to use highway dollars for mass transit, but it
did not authorize the direct use of such funds for intercity
passenger rail service.  Although the states' access to federal
transportation trust funds is a key element of Amtrak's plan to have
the states cover significantly more of the railroad's costs, it is
clearly up to the Congress to decide whether and to what extent such
flexibility should be extended to intercity passenger rail. 

Finally, Amtrak observed that it needs to be freed from certain
legislative restraints to help it operate more as a competitive
commercial entity.  Amtrak envisions changes to labor laws to give
the corporation greater latitude in negotiating such matters as
severance pay, contracting, and work processes.  Also, Amtrak wants
exemptions from requirements to pay federal fuel taxes and wants the
authority to issue tax-exempt debt.  GAO added a recommendation to
its report that Amtrak provide its proposals to the Congress, along
with the estimated effect of each proposal on Amtrak's finances and
other affected parties.  This information will provide a vehicle for
congressional deliberation on the merits of each proposal. 

Amtrak also provided comments that clarified certain technical
information or statements made in a draft of this report.  GAO
incorporated these changes in the report where appropriate.  Amtrak's
written comments, including its legislative proposals, are presented
in appendix IV, and GAO's responses are discussed in chapter 5. 


-------------------------------------------------------- Chapter 0:7.1

Figure 1 shows some key indicators of Amtrak's financial and
operating conditions.  The first three charts illustrate trends in
Amtrak's revenues, and the next three charts show the age of Amtrak's
fleet, the railroad's operating deficit, and the decline in working
capital. 

   Figure 1:  Key Indicators of
   Amtrak's Deteriorating
   Financial and Operating
   Conditions

   (See figure in printed
   edition.)



   (See figure in printed
   edition.)


INTRODUCTION
============================================================ Chapter 1

Passenger trains, at one time the nation's primary means of intercity
transportation, began losing riders to automobiles in the 1920s and
to airplanes in the 1950s and 1960s.  Revenue losses on passenger
services mounted, and in 1970, the final year in which these services
were operated by private railroads, over $1.7 billion (in 1994
dollars) was lost on these services.  Railroads had stopped investing
in new passenger cars and facilities, and service declined, leading
to the near demise of passenger rail service.  To revitalize
intercity passenger rail service, in 1970 the Congress created the
National Railroad Passenger Corporation, commonly called Amtrak. 
Amtrak began operations on May 1, 1971, with the task of reversing
the decline in intercity rail ridership and making up for years
during which the railroads had neglected passenger operations. 


   BACKGROUND
---------------------------------------------------------- Chapter 1:1

Until World War I, intercity travel generally meant train travel, and
the railroads operated up to 20,000 passenger trains daily.  After
World War II, however, ridership declined dramatically, and by 1970
only about 300 intercity passenger trains were operated daily,
providing only about 0.5 percent of intercity travel.  In response to
declining revenues, most railroads did not invest in new passenger
equipment and allowed the service to deteriorate.  Trains frequently
failed to run on time, speeds were slow because of poor track and
roadbeds, and equipment was old and prone to breakdowns.  Intercity
train travel was rapidly becoming extinct in the United States. 

The nation's railroads, which focused primarily on freight service,
had to obtain permission from federal or state regulators to abandon
unprofitable passenger train services.  Before passage of the
Transportation Act of 1958, state regulatory commissions had the
authority to discontinue passenger trains.  Some state agencies
allowed the railroads to discontinue unprofitable passenger
operations, and between 1920 and 1958, service was discontinued for
more than one-half of the railroads' passenger miles.\1 Because of
the large decline in the number of trains, state commissions became
more reluctant to allow the railroads to discontinue more services. 
The 1958 act transferred effective control over decisions to
discontinue passenger services to the Interstate Commerce Commission
(ICC).  The ICC was concerned that continuing losses from passenger
train operations could threaten the financial health of the entire
railroad industry.  It therefore allowed railroads to eliminate
passenger operations that generated serious deficits. 


--------------------
\1 A passenger mile is one passenger traveling 1 mile. 


      THE CONGRESS CREATES AMTRAK
-------------------------------------------------------- Chapter 1:1.1

The Congress passed the Rail Passenger Service Act\2 in 1970,
creating Amtrak to operate and revitalize intercity passenger rail
service.  The act called for Amtrak to be

     a for-profit corporation to provide intercity and commuter rail
     passenger service, employing innovative operating and marketing
     concepts so as to fully develop the potential of modern rail
     service in meeting the Nation's intercity and commuter passenger
     transportation requirements.  The Corporation will not be an
     agency or establishment of the U.S.  Government. 

The Congress gave Amtrak specific goals to (1) provide modern,
efficient intercity railroad passenger service; (2) help alleviate
the overcrowding of airports, airways, and highways; and (3) give
Americans an alternative to private automobiles and airplanes to meet
their transportation needs.  The act required Amtrak to operate a
basic passenger rail system that considered the passenger service
that existed at the time the act was passed as well as opportunities
to provide additional rail service that would be faster, be more
convenient, and serve more population centers at a lower cost.  Fig. 
1.1 shows Amtrak's route network. 



(See figure in printed edition.)Figure 1.1:  Amtrak's Route Network

Source:  Amtrak. 

Amtrak's first full year of operation was 1972.  Although the route
system today is not much larger than the one operated in 1973, it now
carries about 30 percent more passengers, serves 22 percent more
stations, and operates about 29 percent more train miles.\3 (See
table 1.1.)



                          Table 1.1
           
              Comparison of Amtrak's System and
               Operations--1973, 1983, and 1993

-------------  -------------  -------------  ---------------
Route miles       22,000         24,000          25,000
Train miles         27             29              35
 (millions)
Stations            451            497             535
 served
Passengers         16.9           19.0            22.1
 (millions)
Revenue-           1,717          1,480           1,853
 producing
 cars
Operating           337            273             360
 locomotives
Systemwide          60%            82%            72%\a
 on-time
 performance
Employees           \b           18,505          24,037
 (person-
 years)
------------------------------------------------------------
\a Long-distance on-time performance for 1993 decreased 14 percent
from fiscal year 1992 as a result of extensive flooding in the
Midwest. 

\b Data were unavailable for 1973. 

Source:  Amtrak. 


--------------------
\2 P.L.  91-518, 84 Stat.  1327 (1970) as amended. 

\3 A train mile is one train traveling 1 mile. 


      AMTRAK'S RELATIONSHIP WITH
      FREIGHT RAILROADS
-------------------------------------------------------- Chapter 1:1.2

In 1971, Amtrak took over routes from all but three of the railroads
that were providing passenger service at the time.\4 In return for
relief from their obligation as common carriers to provide passenger
service, the railroads turned over their passenger cars and
locomotives to Amtrak.  The original fleet of over 1,500 cars
averaged more than 22 years in age.  In 1994, Amtrak's fleet of 1,900
cars still included about 435 of the original passenger cars. 

Amtrak eventually took over control of yards, stations, train service
employees, and reservation offices.  By 1976, most of Amtrak's
services, other than those provided by train and engine crews, were
provided by Amtrak's own employees.  Using $120 million provided by
the Congress, Amtrak purchased 103 miles of track between
Philadelphia and Harrisburg, Pennsylvania; 83 miles joining Kalamazoo
with Michigan City; 62 miles linking New Haven, Connecticut, and
Springfield, Massachusetts; and 12 miles near Albany, New York.  In
addition, Amtrak acquired about 400 miles of the Northeast
Corridor--between Washington, D.C., and Boston--from the estate of
the bankrupt Penn Central railroad.  The freight railroads, however,
own the rest of the track over which Amtrak operates.  The Rail
Passenger Service Act grants Amtrak the right to use freight
railroads' tracks, and Amtrak is expected to pay the incremental
costs that the freight railroads incur to maintain the tracks for
passenger service.\5 The freight railroads also provide dispatching
and emergency services for Amtrak trains operating on their systems. 
In recent years, Amtrak has paid about $90 million annually for these
services.  Amtrak compensates the freight railroads according to
individual agreements established in 1971.\6

These agreements expire April 30, 1996. 


--------------------
\4 The three remaining railroads providing intercity passenger rail
service eventually turned over their business to Amtrak or went out
of business altogether. 

\5 Track conditions are governed by federal safety regulations, which
define six classes of track.  Maximum speeds are defined for each
class--the higher the speed allowed, the stricter the standard. 
Thus, if freight traffic would normally operate at 40 mph (class 4)
but Amtrak operated at 79 mph (class 6), Amtrak would have to pay the
incremental cost of maintaining the track to class-6 standards
instead of class-4 standards. 

\6 Amtrak's agreement with Conrail (Consolidated Rail Corporation)
was established in 1976, when that railroad was formed. 


      AMTRAK'S INTERCITY RAIL
      SERVICE AND OTHER MARKETS
-------------------------------------------------------- Chapter 1:1.3

Amtrak's primary function is operating the 25,000-mile intercity
passenger rail route system that serves 44 states.  Amtrak operates
about 350 trains to serve 6 routes in the Northeast Corridor, 17
short-distance routes (under 500 miles), and 20 long-distance routes. 
Thirteen of these routes are partially funded by eight states under
section 403(b) of the Rail Passenger Service Act; three of these
routes are fully funded by two states under section 403(d) of the
act.\7

Amtrak generated $1.4 billion in revenues in fiscal year 1994 and had
$2.4 billion in expenses.  Its federal subsidy was $909 million. 
Passenger-related services generated 65 percent of Amtrak's revenues
in fiscal year 1994 (see fig.  1.2) but covered only about 38 percent
of the expenses.  Another 19 percent of Amtrak's revenues came from
commuter rail service, which Amtrak operates in seven metropolitan
areas under contract to state or local transportation authorities. 
Finally, Amtrak earns revenues from other activities, such as real
estate development (including 30th Street Station in Philadelphia). 
Other revenues come from handling specialized freight, including U.S. 
mail and certain express deliveries, and leasing rights-of-way along
the Northeast Corridor to telecommunications firms for data
transmission lines, among other things.  Over half of Amtrak's
operating expenses in fiscal year 1994 were for salaries, wages, and
benefits to employees.  (See fig.  1.3.)

   Figure 1.2:  Amtrak's Operating
   Revenues for Fiscal Year 1994

   (See figure in printed
   edition.)

Note:  Total operating revenues were about $1.4 billion in fiscal
year 1994. 

Source:  Amtrak. 

   Figure 1.3:  Amtrak's Operating
   Expenses for Fiscal Year 1994

   (See figure in printed
   edition.)

Note:  Total operating expenses were about $2.4 billion in fiscal
year 1994. 

Source:  Amtrak. 


--------------------
\7 Section 403(b) of the Rail Passenger Service Act allows Amtrak to
initiate and/or operate intercity rail services that are financially
supported by non-Amtrak agencies--such as a state, group of states,
or regional or local agency--or by an individual.  As of September
1994, Amtrak provided such services in Alabama, California, Illinois,
Michigan, Missouri, New York, North Carolina, and Wisconsin.  Section
403(d) allows Amtrak to operate services that are fully funded by
states.  As of September 1994, Amtrak provided such services in
Pennsylvania and Washington State. 


      FEDERAL SUBSIDIES SUPPLEMENT
      AMTRAK'S REVENUES
-------------------------------------------------------- Chapter 1:1.4

Since 1971, the federal government has provided Amtrak with over $13
billion to support intercity passenger rail service.  Figure 1.4
shows the amounts appropriated to Amtrak over the past 8 years.  In
fiscal year 1994, Amtrak received federal funds through an operating
and capital grant, a grant for the Northeast Corridor Improvement
Program (NECIP),\8 and a mandatory payment by the Federal Railroad
Administration (FRA) to fund certain retirement and unemployment
benefits.\9 The capital grant pays for purchasing cars and
locomotives; overhauling fully depreciated equipment (when the
overhaul increases the value of the equipment); modifying equipment
as required by law; upgrading facilities for maintenance, overhauls,
and other work; and servicing debt.  NECIP is a long-term capital
improvement project that includes electrification of track between
New Haven and Boston so that trains will be able to travel at speeds
up to 150 miles per hour by the end of the

   Figure 1.4:  Federal
   Appropriations for Amtrak,
   Fiscal Years 1988-95

   (See figure in printed
   edition.)

Notes:  Before fiscal year 1990, FRA's mandatory payments to the
Railroad Retirement Trust Fund were included in Amtrak's operating
grant; its payments to the fund for fiscal years 1988 through 1990
are estimated. 

The fiscal year 1993 appropriation includes $20 million in
supplemental funds for operations and $25 million for capital
requirements. 

In Amtrak's early years, the need for federal subsidies steadily
increased as expenses grew faster than revenues.  The railroad's net
loss grew from $153.4 million in fiscal year 1972 to $716.9 million
in fiscal year 1980.  Inflation was partially responsible, but even
in real terms, the net loss was growing.  Beginning in the early
1980s, the President's budget stopped requesting annual operating
support for Amtrak.  Each year the Congress provided funding, but
Amtrak's management knew it was under pressure to reduce operating
losses.  Amtrak combined an aggressive cost-cutting strategy with new
efforts at revenue enhancement to try to improve its
revenue-to-expense ratio. 


--------------------
\8 NECIP is an ongoing project established in 1976 for the
construction and upgrading of tracks, bridges, communications and
signals, and electric traction between Washington, D.C., and Boston. 

\9 Amtrak is required to participate in the railroad retirement and
unemployment systems.  Each participating railroad pays a portion of
the costs for all retirement and unemployment benefits in the
industry.  Since Amtrak's payments exceed the corporation's specific
retirement and unemployment costs, FRA has agreed to pay the excess
costs for Amtrak. 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
---------------------------------------------------------- Chapter 1:2

This report's objectives were to assess (1) Amtrak's financial and
operating conditions, (2) the likelihood that Amtrak can overcome its
financial and operating problems, and (3) alternative actions that
could be considered in deciding on Amtrak's future mission and on
commitments to fund the railroad.  In this report,

chapter 2 describes the extent to which Amtrak's financial condition
has deteriorated and why;

chapter 3 describes the increased costs facing Amtrak over the next
several years that will make recovery from the financial and
operating problems difficult;

chapter 4 assesses the extent to which revenues are likely to
increase, pay for these increased costs, and improve Amtrak's
financial condition; and

chapter 5 assesses the likelihood that Amtrak will be able to
continue to provide nationwide service at the current federal subsidy
level.  It evaluates alternative levels of service and funding and
the potential long-term results of these alternatives. 

We performed this work as part of our legislative responsibilities
under the Rail Passenger Service Act to conduct performance audits of
Amtrak's activities and transactions.  In addition, five
congressional committees expressed interest in this work.  The issues
raised by the committees, along with our objectives, scope, and
methodology, are discussed in detail in appendix I.  The
organizations that we contacted in the course of our review are
listed in appendix II.  We report dollars as constant 1994 dollars
unless otherwise noted. 

We provided Amtrak with a draft of our report for comment.  Amtrak's
principal observations and our responses to these observations are
provided in chapter 5, and its written comments appear in full in
appendix IV.  We performed our work between August 1993 and December
1994 in accordance with generally accepted government auditing
standards. 


AMTRAK'S FINANCIAL CONDITION IS AT
A CRITICAL STAGE
============================================================ Chapter 2

We testified in early 1994 that Amtrak's financial condition had
deteriorated steadily since 1990,\1 causing a decline in the quality
of service.  Since our testimony, Amtrak's condition has worsened. 
As service quality has deteriorated, Amtrak has had more difficulty
attracting and retaining riders, resulting in further revenue losses
and exacerbating the problem.  All of Amtrak's intercity train
service loses money--no route or portion of a route comes close to
breaking even when capital costs are considered. 

For several years, revenues have been substantially less than
forecast, resulting in larger operating deficits than budgeted.\2

The operating deficits have exceeded federal operating subsidies by a
total of about $175 million since 1990.  Amtrak's actions to reduce
operating expenses have not been sufficient to offset the shortfall
in revenues.  To cover the gap between the operating deficits and
federal operating subsidy, Amtrak has drawn down its cash resources. 
In 1987, its working capital had a positive $113 million balance; by
the end of fiscal year 1994, the balance was a negative $227
million.\3 Nevertheless, while its financial condition was
deteriorating, Amtrak reported that its revenue-to-expense ratio was
improving.  This reported ratio was a misleading indicator of
Amtrak's financial condition because it omitted many important
expenses and showed an improving trend when Amtrak's financial
condition was deteriorating. 

Amtrak's management believes that the railroad's financial condition
will continue to deteriorate and estimated a cash shortfall of about
$200 million by September 30, 1995.  The situation had become so
alarming by December 1994 that Amtrak's Board of Directors approved
large reductions in service and staffing.  These actions, along with
anticipated increases in productivity, are expected to eliminate the
1995 cash shortfall.  However, even if these proposed actions are
successful in bringing costs in line with revenues and grants for
1995, they will not solve Amtrak's longer-term problems, such as the
need for significant capital investment, which we discuss in chapter
3.  Finally, without increased funding, Amtrak expects the gap
between its deficit and the operating subsidy to reappear in 1996 and
to produce a $1.3 billion shortfall through the year 2000. 


--------------------
\1 See Amtrak:  Financial Condition Has Deteriorated and Future Costs
Make Recovery Difficult (GAO/T-RCED-94-155, Mar.  17, 1994); Amtrak: 
Deteriorated Financial Condition and Costly Future Challenges
(GAO/T-RCED-94-145, Mar.  23, 1994); and Amtrak:  Key Decisions Need
to Be Made in the Face of Deteriorating Financial Condition
(GAO/T-RCED-94-186, Apr.  13, 1994). 

\2 The operating deficit is the difference between the cash paid for
operating expenses and the cash received from operations. 

\3 Working capital is the difference between current assets and
current liabilities.  As such, it indicates a firm's ability to pay
current liabilities from current assets. 


   DETERIORATION IN AMTRAK'S
   FINANCIAL CONDITION HAS REDUCED
   CASH RESOURCES
---------------------------------------------------------- Chapter 2:1

Over the last several years, the revenues that Amtrak projected have
not materialized.  As a result, the operating deficits have been
larger than the federal operating subsidies.  Although Amtrak cut
planned spending and developed strategies to conserve cash, these
actions have not compensated for the shortfall in revenues. 
Moreover, Amtrak has not specifically budgeted for operating
contingencies that interrupt service, such as accidents and the flood
in the Midwest in 1993.  As a result of shortfalls in revenues and
losses stemming from these events, Amtrak has had to draw down cash
resources and take on additional short-term debt. 


      REVENUES HAVE BEEN LESS THAN
      FORECAST
-------------------------------------------------------- Chapter 2:1.1

Each year Amtrak forecasts ridership and revenues to develop its
request for federal operating grants and to set planned spending
levels.  From fiscal year 1991 through fiscal year 1994, Amtrak
overestimated passenger revenues by a total of $600 million.  (See
fig.  2.1.) In fiscal year 1994, actual passenger revenues were about
7 percent below those for fiscal year 1993 and about 22 percent less
than forecast. 

   Figure 2.1:  Amtrak's Actual
   and Estimated Passenger
   Revenues, Fiscal Years 1990-94

   (See figure in printed
   edition.)

Note:  Amounts are in current-year dollars. 

Source:  Amtrak. 

Over the past 4 years, Amtrak has consistently overestimated its
passenger revenues in terms of both ridership and yield (revenues per
passenger mile).\4 For its 1993 grant request, for example, Amtrak
estimated $1.092 billion in passenger revenues--overestimating actual
passenger revenues by $149 million.  For 1990-94, Amtrak estimated
that yield would increase with inflation.  In fact, yield declined
after adjusting for inflation.  Management also contributed to
overestimates of revenues by adjusting upward the results of a
ridership forecasting model because of a belief that the nation's
general economic improvement would lead to increased ridership. 
Finally, in adjusting the model results, Amtrak did not consider the
extent to which events--such as floods and accidents--might occur in
the upcoming year, depressing revenues.  In fact, during 1993 and
1994, Amtrak was involved in several major accidents and sometimes
had to cancel trains because of severe weather, which caused revenues
to decline.  In preparing its budget for fiscal year 1995, Amtrak did
not rely on its ridership forecasting model. 


--------------------
\4 To estimate revenues for the upcoming year, Amtrak first forecasts
passenger miles.  Amtrak then multiplies the number of passenger
miles by its forecast for yield to estimate passenger revenues.  The
revenue forecast is adjusted several times--by Amtrak's market
planning analysts and upper management--according to their informed
judgments and other factors.  Amtrak submits the adjusted revenue
forecast to the Congress as part of its federal grant request.  See
app.  III. 


      OPTIMISTIC FORECASTS HAVE
      AFFECTED REQUESTS FOR
      FEDERAL FUNDS
-------------------------------------------------------- Chapter 2:1.2

Because it overestimated revenues, Amtrak underestimated its
operating deficits and therefore requested smaller federal operating
subsidies than it needed.  In fiscal years 1990 through 1994,
Amtrak's operating deficit exceeded the railroad's federal operating
subsidy by an average of about $36 million per year.  (See fig. 
2.2.) This gap has forced Amtrak to reduce operations.  These
reductions have contributed to poorer service quality (fewer on-board
service personnel) and delayed or deferred maintenance of equipment. 
Ultimately, such actions threaten Amtrak's ability to provide high
quality rail service and to compete effectively for customers. 

   Figure 2.2:  Amtrak's Federal
   Operating Subsidy and Operating
   Deficit, Fiscal Years 1988-94

   (See figure in printed
   edition.)

Note:  Before fiscal year 1990, FRA's mandatory payments to the
retirement fund were included in Amtrak's operating deficit and
grant.  Amounts are in current-year dollars. 

Source:  GAO's analysis of Amtrak's data. 

As a result of the shortfall in revenues and increased expenses,
Amtrak's net loss in fiscal year 1994 grew to over $l billion, while
the railroad's operating deficit exceeded the operating subsidy by
$76 million.  These operating losses, combined with capital
expenditures, caused a cash shortfall of about $50 million as of
September 30, 1994.  For fiscal year 1995, Amtrak projects that the
gap between its operating subsidy and operating deficit will grow to
$193 million.  Amtrak officials believe this situation will result
from a 5-percent decrease in passenger revenues in 1995 and an $145
million increase in expenses.  As a result of the gap between the
operating subsidy and operating deficit along with planned capital
expenditures, in the fall of 1994 Amtrak estimated a cash shortfall
of about $200 million by September 30, 1995. 

In addition, Amtrak does not specifically budget for accidents and
weather-related emergencies.  The added costs of these problems
exacerbate Amtrak's deteriorating financial condition.  For example,
in September 1993 the Sunset Limited derailed in Saraland, Alabama,
and Amtrak incurred (1) additional costs to repair equipment and pay
liability judgments and (2) a shortfall in revenues as a result of
equipment shortages.  In addition, whenever service is disrupted
because of weather or accidents, Amtrak pays the costs its passengers
incur to get to their destinations by another mode of transportation. 
Amtrak has estimated that in 1993 and 1994, such problems cost about
a total of $95 million in lost revenues and $11 million in additional
expenses. 


      AMTRAK REDUCED EXPENSES IN
      RESPONSE TO SHORTFALL IN
      REVENUES
-------------------------------------------------------- Chapter 2:1.3

To address the shortfall in revenues, Amtrak cut back budgeted
expenses by reducing the number of management staff by 10 percent in
1991, by reducing train and station staffing levels in 1992, and by
decreasing heavy maintenance programs for cars and locomotives in
1993.\5 According to Amtrak, these cutbacks were expected to save
about $77 million in fiscal years 1991-93.  In 1993, Amtrak took
steps to conserve cash by reducing inventory, requiring advance
payment for work that Amtrak performed for others, and delaying
payments made to others by 15 days.  From the beginning of 1993 to
the end of 1994, Amtrak's investment in inventory declined from $147
million to $135 million.\6 Finally, Amtrak requested and received a
supplemental federal grant of $45 million ($20 million for operating
expenses and $25 million for capital expenses) in fiscal year 1993. 

To offset the continuing shortfall in passenger revenues in fiscal
year 1994, Amtrak again initiated actions to reduce budgeted expenses
by about $90 million.  These actions included reducing (1) the
frequency of service on some routes, (2) the number of passenger
support staff, and (3) general overhead costs.  However, Amtrak's
actual expenses in 1994 exceeded budgeted expenses by over $100
million.  By September 30, 1994, Amtrak had increased its short-term
lines of credit to $120 million and borrowed $60 million.  In
addition, by that date Amtrak's long-term debt for capital projects
exceeded $650 million. 

Despite Amtrak's actions, the gap between operating deficits and
federal operating subsidies continued to grow.  As a result, Amtrak
drew down its working capital, which fell from $113 million in fiscal
year 1987 to a $227 million deficit at the end of fiscal year 1994
(see fig.  2.3), a decline of $371 million in 1994 dollars. 
Continued reductions in working capital will jeopardize Amtrak's
ability to pay immediate expenses. 

   Figure 2.3:  Amtrak's Working
   Capital Surplus/Deficit, Fiscal
   Years 1987-94

   (See figure in printed
   edition.)

Notes:  Working capital is the difference between current assets and
current liabilities. 

Amounts are in current-year dollars.  In 1994 dollars, working
capital declined from $144 million in 1987 to a deficit of $227
million in 1994. 

Source:  GAO's analysis of Amtrak's data. 


--------------------
\5 Train staffing was reduced from one attendant for every two cars
to one attendant for every three cars. 

\6 Although reducing investment in inventory reduces the cash needed
to purchase replacement parts, continued declines in inventory may
delay equipment repairs and maintenance by limiting the timely
availability of parts.  Amtrak expects the value of its inventory to
increase over the next several years as it receives spare parts for
new cars and locomotives. 


      AMTRAK PLANS SERVICE
      REDUCTIONS AND OTHER CHANGES
      IN 1995 TO REDUCE CASH
      SHORTFALL
-------------------------------------------------------- Chapter 2:1.4

Amtrak's financial situation became so alarming that on December 14,
1994, Amtrak announced an aggressive plan to eliminate cash deficits
in fiscal year 1995 and attempt to put the corporation on a sound
financial footing.  Amtrak's management saw only two options--to
abandon the nationwide intercity rail system or to minimize losses
through improvements in productivity and reduction in routes and
service.  The Board of Directors directed that Amtrak eliminate 3
routes and segments of 10 others and reduce service frequencies (the
number of trains per week on a given route).  As a result, about 20
percent of Amtrak's nationwide service (in train miles) is to be
eliminated.  Amtrak plans to negotiate with the affected states and
could retain service where the states are willing to subsidize the
losses on the routes.  These reductions should allow Amtrak to
eliminate about 5,000 jobs and remove most of its oldest cars from
service.  Amtrak expects that these actions will improve the on-time
performance of its remaining trains and lower its operating costs. 
The plan also calls for generating additional revenues through
adjusting ticket prices, developing commuter operations and other
businesses, and selling or refinancing assets. 

These and other actions to enhance productivity are expected to
produce an annualized net savings to Amtrak of $364 million, by
reducing costs by about $430 million annually while forfeiting only
$66 million in revenues.  Amtrak expects these actions to result in
savings of about $200 million in 1995, eliminating the expected cash
deficit for 1995.  These actions, however, will not eliminate
Amtrak's need for federal operating and capital grants now or in the
future.  Revenues will continue to fall short of expenses on most
routes.  Furthermore, collective bargaining and/or legislative
changes might be required before approximately 26 percent of the
anticipated savings can be achieved, according to Amtrak.  In
addition, the plan does not account for implementation costs, such as
pay protection for employees whose jobs are eliminated as a result of
route closures and capital investment needs that continue to grow. 
Finally, Amtrak projects that after 1995, if federal grants remain at
current levels, the deficit will exceed operating grants by $1.3
billion through the year 2000.  These issues are also discussed in
chapters 3, 4, and 5. 


   REVENUE-TO-EXPENSE RATIO IS
   MISLEADING
---------------------------------------------------------- Chapter 2:2

Amtrak has calculated and reported a revenue-to-expense ratio to
demonstrate that its operating revenues were covering a larger share
of its operating expenses from year to year.  As figure 2.4 shows,
this ratio improved from fiscal year 1982 to fiscal year 1991 but has
since leveled off.  Amtrak reported that revenues covered about 53
percent of expenses in fiscal year 1982 and about 80 percent in
fiscal year 1993.  However, in calculating this ratio, Amtrak did not
include all relevant expenses.  Because Amtrak deferred certain
costs, including some maintenance expenses, the ratio gave a
misleading picture of Amtrak's financial health.  Amtrak officials
agreed with our assessment that this ratio should not be used\7 and
stopped reporting it in fiscal year 1994. 

   Figure 2.4:  Amtrak's
   Revenue-to-Expense Ratio,
   Fiscal Years 1982-93

   (See figure in printed
   edition.)

Source:  GAO's analysis of Amtrak's data. 

In calculating the revenue-to-expense ratio, Amtrak has excluded
certain expenses, including (1) depreciation, (2) mandatory payments
for railroad retirement and unemployment insurance made by FRA on
Amtrak's behalf,\8 (3) various federal and state taxes, (4) user fees
paid to FRA for track inspections and other activities, (5)
miscellaneous expenses relating to accident claims, (6) losses
incurred in providing 403(b) service to the states,\9 and (7)
disbursements for labor protection.  Amtrak is required by statute to
exclude losses under section 403(b).  It excluded other expenses for
various reasons, including a belief that some items (such as labor
protection payments) do not represent operating costs.  If all these
excluded expenses had been included, the ratio in 1993 would have
been 66 percent--14 percentage points less than Amtrak reported. 

The revenue-to-expense ratio is a misleading indicator of a firm's
financial condition when the firm defers or forgoes expenses to show
improved performance.  Amtrak has deferred maintenance on rolling
stock and equipment and reduced other expenses.  While these actions
have resulted in short-term improvements to the revenue-to-expense
ratio, they have long-term implications for Amtrak's viability.  This
ratio was also misleading because it showed an improving trend when
Amtrak's financial condition was deteriorating.  Furthermore, as
figure 2.5 illustrates, the overall gap between revenues and expenses
has been widening since 1988. 

   Figure 2.5:  Revenues and
   Expenses, Fiscal Years 1988-94

   (See figure in printed
   edition.)

Note:  Amounts are in current-year dollars. 

Source:  GAO's illustration of Amtrak's data. 


--------------------
\7 See Amtrak:  Financial Condition Has Deteriorated and Future Costs
Make Recovery Difficult (GAO/T-RCED-94-155, Mar.  17, 1994). 

\8 These payments have been excluded only since fiscal year 1991. 

\9 Amtrak provides intercity service that is partially funded by the
states served under sec.  403(b) of the Rail Passenger Service Act
(see ch.  4). 


   CONCLUSIONS
---------------------------------------------------------- Chapter 2:3

Amtrak's financial condition has deteriorated over the last several
years, and the problems accelerated in 1994.  Actual revenues have
been less than Amtrak projected and expenses have been higher than
expected.  As a result, federal subsidies have not fully covered
Amtrak's operating deficits.  This condition has adversely affected
Amtrak's supply of cash, which is needed to pay bills and provide
high quality service.  Amtrak's financial deterioration worsened in
1994 in the wake of accidents and lower-than-expected revenues. 
Amtrak had expected that the general economic recovery would improve
its financial situation, but this did not occur.  Although Amtrak
reduced its budgeted expenses, continued declines in revenues, cash,
and working capital threaten Amtrak's ability to provide high quality
intercity passenger rail service and compete effectively for
customers.  The service reductions and other actions that Amtrak
plans to take in 1995, if implemented, will bring Amtrak's costs in
line with revenues and grants for that year.  However, these planned
actions will not solve the corporation's longer-term problems, which
we discuss in chapter 3. 

The Congress needs better information on how Amtrak estimates
revenues from passenger service, not only to determine the amount of
federal subsidy, if any, to provide but also to make decisions about
Amtrak's future.  In chapter 5 we make recommendations to the
President of Amtrak to improve the information provided to the
Congress. 


CAPITAL AND OTHER COSTS WILL MAKE
FINANCIAL IMPROVEMENT DIFFICULT
============================================================ Chapter 3

The depletion of Amtrak's physical assets poses an even greater
threat to the railroad's financial well-being than the current
shortfalls in operating funds.  Operating a passenger railroad is an
inherently costly undertaking.  The advanced age and poor condition
of Amtrak's rolling stock (locomotives and cars) and overhaul
facilities make it expensive to maintain the fleet.  While purchases
of new equipment will ease the burden of maintenance and overhauls,
the funds to pay for this equipment must come from already-strained
capital budgets.  Amtrak may also face substantial additional costs
to comply with environmental laws and to pay freight railroads for
the use of their track.  Amtrak is unlikely to be able to address
these challenges under the current operating environment and at the
current funding level.  Reductions in routes announced in December
1994 should allow Amtrak to retire most of its oldest passenger cars. 
But, according to new estimates, the investments needed in
infrastructure will more than offset any savings. 


   COSTS FOR MAINTAINING AGING
   FLEET ARE GROWING
---------------------------------------------------------- Chapter 3:1

Today, Amtrak's fleet is about as old as the aged fleet the
corporation inherited from the freight railroads over two decades
ago.  (See fig.  3.1.) Aging equipment requires more extensive
repairs, and Amtrak spends over $400 million annually to repair,
maintain, and overhaul this equipment.  Nevertheless, that amount
understates Amtrak's true costs for maintaining the equipment because
Amtrak has routinely deferred equipment overhauls to cope with
funding shortages and performed less comprehensive work when
overhauling some cars.  Amtrak introduced a new maintenance/overhaul
program in 1994 to reduce this backlog and improve the fleet's
operating condition.  However, this program is already underfunded,
and new backlogs will inevitably result. 

   Figure 3.1:  Average Age of
   Amtrak's Passenger Cars and
   Locomotives, Fiscal Years
   1972-93

   (See figure in printed
   edition.)

Notes:  Fiscal year 1993 is the last year for which data are
available. 

The average age of locomotives reflects only the age of locomotives
used to operate trains. 

Source:  Amtrak. 


      CARS ARE GROWING OLDER AND
      MORE EXPENSIVE TO MAINTAIN
-------------------------------------------------------- Chapter 3:1.1

A significant portion of Amtrak's equipment--31 percent of the cars
and 54 percent of the locomotives--is beyond its useful life.  About
23 percent of Amtrak's 1,900-car fleet consists of Heritage passenger
cars that Amtrak obtained in 1971 from other railroads.  These cars
now average over 40 years old--much older than the 25- to 30-year
average expected useful life estimated by Amtrak's Chief Mechanical
Officer.  The remaining passenger cars (Amfleets, Horizons,
Superliners, Turboliners, and Viewliners) in total average 14.2
years--about half-way through their useful life span.  (See fig. 
3.2.)

   Figure 3.2:  Average Age of
   Amtrak's Car Fleet, June 1994

   (See figure in printed
   edition.)

Source:  GAO's analysis of Amtrak's data. 

As equipment ages, it breaks down more often and requires more
extensive repairs.  In an August 1994 study, Amtrak found that
failures of Heritage cars cause more train cancellations and longer
delays than failures of any other type of car.  Heritage cars that
fail are out of service awaiting parts about three times longer on
average than newer cars.  Needed parts for Heritage cars must often
be manufactured because they can no longer be purchased off the
shelf.  Moreover, because there are 27 different Heritage models,
Amtrak cannot produce replacement parts in economical quantities.  As
a result, the cost per seat to perform a limited "intermediate"
overhaul on a Heritage car can be 25 to 50 percent higher than the
cost to undertake a complete heavy overhaul on a Superliner.  Before
December 1994, Amtrak planned to replace some Heritage cars as 245
new Superliner and Viewliner cars were delivered by 1997.\1

About 200 Heritage cars would have remained in service.  Now,
however, Amtrak expects to retire all but a few "specialty" Heritage
cars (diners, cab cars) as it reduces service frequencies and
eliminates routes in fiscal year 1995.  According to Amtrak, this
will have a significant impact on the costs to maintain equipment and
out-of-service rates. 


--------------------
\1 Amtrak had only three prototype Viewliner cars in its fleet as of
December 1994. 


      BACKLOGS AND FUNDING
      SHORTAGES SPURRED CHANGE IN
      OVERHAUL PROGRAM
-------------------------------------------------------- Chapter 3:1.2

Before fiscal year 1994, Amtrak's goal was to maintain all its cars
through a program of periodic preventive maintenance and regular
heavy overhauls (every 3 to 4 years).  These overhauls can cost about
$300,000 for each car.  In comparison, a new car costs about $2
million.  However, to cope with its deteriorating financial
condition, Amtrak began deferring maintenance in the late 1980s. 
Amtrak also began using capital funds in 1992 to overhaul many older
cars (Heritage and Amfleet I models) and locomotives.  Even with the
infusion of capital funds to overhaul these cars, overhauls were past
due for nearly 40 percent of the fleet by September 1993.  (See fig. 
3.3.)

   Figure 3.3:  Backlog of Heavy
   Overhauls on Cars, Fiscal Years
   1989-93

   (See figure in printed
   edition.)

Note:  Amtrak implemented a new "progressive maintenance" program in
October 1993. 

Source:  GAO's analysis of Amtrak's data. 

As the backlog grew, Amtrak officials recognized that the railroad's
preventive maintenance/overhaul program was not adequate.  Cars
looked shabby and were breaking down with increasing regularity. 
When the fiscal year 1994 budget provided no increase in funding, the
Mechanical Department began to implement a new "progressive
maintenance" program in October 1993.\2 Under this program, Amtrak
performs heavy maintenance when a car breaks down and also a limited
overhaul each year on every car.  Every third year, the annual
overhaul will be more comprehensive but still less extensive than the
heavy overhaul performed before.  Only Amtrak's newer passenger cars
will be maintained under this program.  The remaining Heritage cars
and Turboliner coaches will continue to receive preventive
maintenance and/or limited overhauls as before.  The progressive
program is intended to keep more cars in service--thereby generating
more revenues--but is not expected to reduce overall maintenance
costs.  By the end of fiscal year 1994, all Amfleet, Horizon,
Superliner, and Viewliner cars were eligible for inclusion in the
progressive program.  However, Amtrak is already falling short of
reaching its goals under this program.  While 1994 was a transition
year, fewer cars than required received annual overhauls because of
budget reductions.  Also, maintenance officials have found that the
program is not appropriate for all these cars.  For example,
increased customer complaints about the condition of Amfleet cars in
the Northeast Corridor have prompted Amtrak to return these cars to a
scheduled maintenance program in fiscal year 1995, although they will
continue to receive overhauls under the progressive program. 
Furthermore, the progressive program was intended to use only
operating funds, but Amtrak has already found it necessary to
allocate 1995 capital funds for the heavier 3-year overhauls of the
Amfleet I cars. 


--------------------
\2 We reviewed the progressive maintenance program and the program it
replaced in Amtrak:  Financial Condition Has Deteriorated and Future
Costs Make Recovery Difficult (GAO/T-RCED-94-155, Mar.  17, 1994). 


      NEW PROGRAM BRINGS HIGHER
      WORKLOAD AT SOME FACILITIES
      AND HIGHER OVERALL COSTS
-------------------------------------------------------- Chapter 3:1.3

As Amtrak implements the progressive program, two of its three
overhaul facilities--located at Bear, Delaware, and Beech Grove,
Indiana--will need to handle significantly more cars than they did
under the previous heavy overhaul program.\3 Both facilities fell
behind in performing overhauls in the past because funds were not
available. 

The Bear facility is responsible for annual progressive overhauls on
Amtrak's 629 active Amfleet cars, which provide service primarily in
the Northeast Corridor.  In fiscal year 1993 (the last full year of
heavy overhauls), the facility overhauled 59 cars, but it had
overhauled as many as 152 cars per year in earlier years.  Officials
said that this facility has the physical capacity to perform the 629
overhauls required under the progressive program, but only if its
production staff is increased by 90 from its current level of 183. 

The Beech Grove facility is responsible for overhauling nearly 1,200
cars and 265 locomotives that operate outside the Northeast Corridor. 
In fiscal year 1993, this facility performed heavy overhauls of 117
cars and 50 locomotives--about 47 percent less than the number
required to meet the established overhaul cycles.  During 1995, it
would need to perform at least 166 heavy overhauls and 350
progressive overhauls of the cars while continuing to overhaul an
average of 66 locomotives to keep up with the established schedule. 
However, the current budget for Beech Grove will fund only about 63
percent of these overhauls.  Furthermore, Amtrak's new Superliners
and Viewliners will add to Beech Grove's workload as they become
eligible for annual overhauls starting in 1996 and 1997,
respectively.  While the per-unit cost of a progressive overhaul is
substantially less than the cost of a heavy overhaul, the total cost
to Amtrak to fully implement the new program will be greater because
so many more cars will be overhauled each year.  For fiscal year
1995, Amtrak needs an additional $31.2 million to do this work. 


--------------------
\3 A third facility at Wilmington, Delaware, overhauls mostly
electric locomotives.  Because the process for overhauling
locomotives has not changed, Wilmington has not been affected by the
new progressive program. 


   CAPITAL INVESTMENT NEEDED TO
   IMPROVE EFFICIENCY AND INCREASE
   PRODUCTIVITY
---------------------------------------------------------- Chapter 3:2

The depletion of Amtrak's physical assets--maintenance and overhaul
facilities, rolling stock, rights-of-way in the Northeast Corridor,
and other support assets--is perhaps a greater threat to the
railroad's financial well-being than the current shortfall in
operating funds.  Over the past 10 years, Amtrak's equipment and
facilities depreciated at the rate of $200 million per year, while
investment averaged only $140 million annually.  Amtrak currently
estimates that even with the reduced route system announced in
December 1994, it needs capital investment of over $4 billion to
purchase rolling stock and to bring the infrastructure into a state
of good repair.  (See fig.  3.4.) This amount does not include money
needed for equipment and facilities for high-speed service in the
Northeast Corridor. 

   Figure 3.4:  Amtrak's Estimate
   of Capital Investment Needs

   (See figure in printed
   edition.)

Notes:  The amount estimated for purchasing rolling stock and for
overhauls does not include funds needed to purchase high-speed train
sets. 

Maintenance of way costs include repairs needed on the Northeast
Corridor right-of-way but exclude funds needed for new facilities and
for completion of the electrification project to allow high-speed
service on the north end of the corridor. 

Payments on equipment purchased in 1991-93 reflect the amount needed
for principal and capital interest from 1995 to 2004.  Regular
interest (paid out of Amtrak's operating funds) is not included. 
Amtrak will need an additional $615 million for principal payments on
this equipment between 2005 and 2017 to retire the debt. 

These capital needs total $4.3 billion. 

Source:  Amtrak. 

Amtrak's capital subsidies have not been sufficient for the railroad
to attempt this level of investment.  Amtrak borrowed much of the
capital needed to replace some of its oldest cars and locomotives
with new equipment ordered in the early 1990s, and a significant
portion of its annual capital subsidy must be used to pay this debt. 
The limited remaining capital funds are generally committed to
short-term projects that enable Amtrak simply to keep its equipment
operating and complying with federal laws.  Very little is left over
to invest in projects that might increase revenues, improve the
efficiency of operations, or increase the capacity and productivity
of overhaul facilities. 


      FUTURE FEDERAL CAPITAL
      GRANTS HAVE ALREADY BEEN
      COMMITTED
-------------------------------------------------------- Chapter 3:2.1

The ability of Amtrak to improve its facilities, overhaul its oldest
cars and most of its locomotives, and purchase new equipment depends
wholly on its federal capital grant.  The demands on these funds far
exceed the grants provided over the past few years.  Amtrak received
$560 million in capital funding from 1992 to 1994.\4

According to Amtrak, it must now set aside a sizable portion of the
subsidy to pay for the rail equipment and computers it purchased and
facility improvements it made with borrowed funds in the past, as
well as for legally mandated equipment modifications and
environmental cleanup efforts.  Additionally, Amtrak uses an
increasing amount of its capital funds to overhaul cars and
locomotives, leaving only a small amount for all other capital
replacement needs.\5 (See fig.  3.5.)

   Figure 3.5:  Commitments of
   Amtrak's Federal Capital Funds,
   Fiscal Years 1989-95

   (See figure in printed
   edition.)

Notes:  Uncommitted funds are those remaining from the capital grant
that are not already committed for specific requirements at the
beginning of the fiscal year.  Amtrak receives additional capital
funds for NECIP that are not included in this figure. 

Capital fund categories for fiscal year 1995 are estimated. 

Source:  GAO's analysis of Amtrak's data. 

Amtrak allocated $14.6 million of its capital funding to overhauls in
fiscal year 1992, $55.7 million in fiscal year 1993, and $70 million
in fiscal year 1994.  As a result, less money was available for
capital investment in new equipment or new infrastructure.  In its
capital subsidy request for fiscal year 1995, Amtrak identified more
than $800 million in needed capital spending, of which $195 million
will be used to pay off debt, comply with federal laws on equipment
modifications, or fund capital overhauls.  However, only $230 million
in capital funds was appropriated.  In fiscal year 1995, Amtrak will
have only $35 million to invest in new capital projects like facility
renovations and major repairs to track and rights-of-way--down from
more than $100 million in 1993. 

Furthermore, Amtrak recently recognized that the need to reinvest in
the Northeast Corridor right-of-way (track, signals, and auxiliary
structures) is becoming critical.  It now estimates that at least
$2.5 billion will be needed to bring this infrastructure into a state
of good repair.  Much of this investment is needed on the south end
of the corridor, from Washington, D.C., to New York, where Amtrak
operates high-speed trains and has captured the largest share of the
transportation market.  Amtrak will use $115 million of its $200
million NECIP appropriation for fiscal year 1995 to improve track,
signals, structures (e.g., bridges), electric traction (catenary\6
and related power structures), maintenance-of-way equipment, and
tunnels in this part of the corridor.  According to Amtrak's Chief
Engineer, however, these improvements are only a small fraction of
what needs to be done.  The remaining $85 million of the NECIP
appropriation will be used for electrification and improving track
and facilities on the north end of the corridor, as well as for
purchasing high-speed train sets (see
ch.  4). 


--------------------
\4 Amtrak received an additional $634 million for improvements in the
Northeast Corridor, bringing its total capital funding to nearly $1.2
billion for fiscal years 1992-94. 

\5 In 1992, Amtrak reclassified some overhauls of equipment as
capital expenses because the cost exceeded 50 percent of the book
value of the car or locomotive. 

\6 Catenary is the overhead wire system that delivers electricity to
the locomotive for traction, or movement. 


      MAINTENANCE AND OVERHAUL
      FACILITIES NEED SUBSTANTIAL
      CAPITAL IMPROVEMENTS
-------------------------------------------------------- Chapter 3:2.2

Amtrak owns and/or operates 18 facilities where cars and locomotives
are maintained and overhauled.  Amtrak has developed a 10-year master
plan for improving these facilities so they can accommodate future
requirements, including the new progressive maintenance/overhaul
program.\7 The plan is estimated to cost $326 million.  Five
facilities need substantial renovation and/or modernization--the
three overhaul facilities at Beech Grove, Indiana, and Wilmington and
Bear, Delaware, and two divisional repair facilities in Los Angeles
and New York.  Together, these facilities will require $262 million
to renovate or replace structures, improve efficiencies in operation,
repair damage caused by earthquakes, and build new structures. 

The facility at Beech Grove needs not only the increased funds to
perform the progressive overhauls discussed above but also upgrades
to its physical plant if it is to meet Amtrak's future overhaul
requirements.  Beech Grove is responsible for overhauls of and
repairs to 61 percent of Amtrak's total fleet.  The facility is
nearly 100 years old and is in very poor condition.  Much of the
on-site track was installed in the early 1900s and has deteriorated,
resulting in frequent derailments.  The facility was not designed for
production-line overhauls of cars.  The buildings are run-down, and
some cannot accommodate the work for which they are used.  For
example, Amtrak's newest diesel locomotives are too large to fit
inside the locomotive shop building, and the shop's cranes are not
large enough to lift the locomotives so that wheel sets can be
removed. 

In 1990, Amtrak initiated a five-phase modernization plan to correct
some of Beech Grove's problems.  By September 1993, about $12 million
of the total cost of $47 million had been spent on such projects as
combining the truck and forge shops, improving a coach production
line, constructing employee welfare facilities (lunchrooms, rest
rooms, and locker rooms), replacing roofs, and replacing or
rehabilitating overhead cranes.  In August 1994, Amtrak committed an
additional $1.9 million to repair Beech Grove's transfer table\8 and
repair or replace some of the track most critical to the facility's
operation.  These projects will help Beech Grove perform its required
overhauls and reduce derailments




(See figure in printed edition.)Figure 3.6:  Facility at Beech Grove

Deteriorated track



(See figure in printed edition.)

Cars derailed at the facility



(See figure in printed edition.)

Rerailing equipment



(See figure in printed edition.)

No indoor storage for inventory of seats

Source:  Amtrak. 

However, many renovations and modernization projects remain unfunded,
including a new warehouse and distribution system for material,
modifications to the locomotive shop, additional replacement and
rehabilitation of track, and a new wheel shop.  With the introduction
of the progressive overhaul program, which presents even greater
challenges to the facility's operations, Beech Grove officials have
drafted a new modernization plan to accomplish these projects and
other work.  First-year costs of $9.8 million have been budgeted for
fiscal year 1995, but $28.4 million more is needed to complete the
work. 

Of all Amtrak's repair facilities, Sunnyside Yard in New York City is
most in need of improvement.  This facility has almost no maintenance
structures.  Virtually all work is performed outside, so that
equipment and personnel are exposed to the elements.  The single
service building for performing minor maintenance (called "running
repairs") can accommodate only 8 cars at a time and is thus barely
adequate for the 12 to 15 cars needing repair each day.  The only
locomotive repair building at Sunnyside was recently condemned
because of chemical contamination; it will cost $550,000 to remove
this building and the related contamination.  In addition, because
Sunnyside lacked auxiliary power hook-ups to allow car heaters to
operate while the cars were being serviced in the yard, the plumbing
systems on over 50 cars froze and broke during the winter of 1993-94. 
Repairs to these cars cost about $1.8 million.  Amtrak estimates
costs of more than $100 million for improvements at Sunnyside,
including (1) constructing a new service and inspection building and
a car-cleaning facility, (2) realigning track and constructing new
storage track, and (3) completing necessary environmental cleanup. 


--------------------
\7 According to Amtrak officials, the strategic plan adopted in
December 1994 supersedes the 10-year plan and calls for consolidating
Amtrak's overhaul and repair facilities.  However, no decision has
yet been made on how this consolidation will be implemented. 

\8 A transfer table is a moveable piece of track that moves rolling
stock to various workstations. 


      CAPITAL EQUIPMENT
      REQUIREMENTS ARE ALSO
      SIGNIFICANT
-------------------------------------------------------- Chapter 3:2.3

As stated previously, maintenance and repair costs remain high
because of the large number of aging Heritage cars.  Between 1991 and
1993, Amtrak purchased 245 Superliner and Viewliner cars and 72
locomotives for $743 million and $181 million, respectively.\9 These
cars should be cheaper to maintain than the Heritage cars they will
replace because they have standardized parts and modular components
to allow for easier repair.  Amtrak estimates that it will save $341
annually in maintenance costs for every seat in a Heritage car
replaced by a seat in a Superliner.  Also, most of the cars are
designed with many more seats per car than the cars they are
replacing, potentially adding passenger revenues without adding more
cars to the train.  Amtrak believes that passengers will be more
likely to travel by rail if they can ride in newer, more modern
Amtrak cars.  The new locomotives will provide greater power, will
increase fuel efficiency, and should contribute to better on-time
performance. 

However, in addition to the new equipment that is now being
delivered, Amtrak has estimated it will need over 700 more new cars
and locomotives in the future.  Its current long-term equipment
acquisition plan, approved in November 1992, includes estimates for
replacing all remaining Heritage cars with Viewliners and replacing
some of the nonpassenger cars and locomotives that are in poor
condition or are nearing the end of their useful lives.  Amtrak
estimates that it will need to purchase 299 locomotives and 416 cars
at a cost of about $1.5 billion from 1994 to 2002.  While these needs
will be reassessed following the service reductions in 1995, Amtrak
officials believe that a large portion of the locomotive fleet will
still need to be replaced within the next 5 years. 

According to Amtrak's President, it is very difficult to commit to
long-term capital projects because of uncertainty about future
funding.  As a result, Amtrak tends to focus on short-term operations
and is less able to invest in projects that would create long-term
operating efficiencies.  Capital grants are appropriated annually and
often fluctuate from year to year.  (See fig.  3.7.) As stated above,
a significant amount of money is needed just to repay debt and pay
for capital overhauls and equipment modifications.  Also, the
Committees on Appropriations have, in recent years, placed
restrictions on when Amtrak is allowed to withdraw its money from the
Treasury, usually mandating that all withdrawals be delayed until the
fourth quarter of the fiscal year. 

   Figure 3.7:  Amtrak's Capital
   Grant, Fiscal Years 1981-95

   (See figure in printed
   edition.)

Source:  GAO's illustration of Amtrak's data. 


--------------------
\9 One of the new locomotives was destroyed in the accident in
Saraland, Alabama, in September 1993, less than a month after being
delivered. 


   LABOR COSTS COULD INCREASE
---------------------------------------------------------- Chapter 3:3

Labor costs could increase as Amtrak begins to renegotiate contracts
in 1995 with the 14 labor unions that represent about 90 percent of
Amtrak's 25,000 employees.  Amtrak estimated that wages increased
between $120 and $140 million from 1991 to 1995 as a result of the
last round of collective bargaining.  Similar increases may result
from the upcoming negotiations.  For example, if wages and benefits
increase by 4 percent per year to keep pace with inflation, Amtrak's
costs could increase by about $40 million per year, or about $200
million (in current-year dollars) over a 5-year period.\10 The
current contracts provide for most union employees to receive annual
wage increases of about 4 percent between 1991 and 1995. 

Amtrak has reduced labor costs and increased productivity when
possible.  Its employees receive less compensation on average than
freight railroad employees performing comparable jobs.  For example,
in 1992 Amtrak compensated train and engine crews about $42,900 per
employee, while other Class I freight railroads compensated similar
employees about $54,800 per employee.\11 Also, since 1983 Amtrak has
increased productivity by (1) adopting an 8-hour basis of pay for
train and engine crews (before this change, crews earned a full day's
pay on the basis of the number of miles traveled), (2) eliminating
the requirement for a fireman on trips of less than 4 hours, and (3)
establishing a 5-year graduated entry wage for certain newly hired
employees.  Amtrak estimates that these changes have saved about $50
million per year. 

Amtrak's costs in 1994 for salaries, wages, and benefits accounted
for about 52 percent ($1.3 billion) of the railroad's total operating
expenses.  Federal laws unique to the rail industry keep Amtrak's
labor costs higher than they would be otherwise.  For example, in
1992 Amtrak paid about 26 percent of its payroll in retirement taxes
to meet requirements of the Railroad Retirement Act of 1937; this
percentage is similar to the rate paid by other railroads.\12 In
comparison, other industries paid only about 6 percent of their
payroll in retirement costs and retirement-related savings plans. 
Also, in 1992 Amtrak paid about $0.67 per employee-hour worked for
accident and injury claims under the Federal Employers Liability Act. 
This amount compares with about $0.36 per hour worked paid by private
industry under state workers compensation systems.\13

Labor costs could also increase if Amtrak eliminates entire routes or
reduces service below three round-trips per week.  Amtrak and freight
railroad employees are covered by the Rail Passenger Service Act,
which requires adoption of a labor protection agreement for employees
affected when intercity rail passenger service is discontinued. 
Under the agreement, known as Appendix C-2 and adopted in 1973,
employees who are dismissed may be eligible for payment of their
average monthly compensation for up to 6 years or, at their option,
may receive a separation allowance of up to 12 months' pay.\14
According to Amtrak, if the entire route system were shut down, the
railroad would incur labor protection expenses of between $2.1
billion and $5.2 billion. 


--------------------
\10 This estimate does not include the potential gains in
productivity that Amtrak could achieve as a result of such things as
changes in work rules.  Such gains could offset some or all of the
cost increases. 

\11 Class I freight railroads are those that earned $251.4 million or
more in annual revenues in 1992. 

\12 This percentage includes retirement costs for both management and
nonmanagement employees.  According to Amtrak, management employees
participate in a defined contribution retirement plan and receive
certain other retirement benefits that are not available to
nonmanagement employees. 

\13 See Railroad Competitiveness:  Federal Laws and Policies Affect
Railroad Competitiveness (GAO/RCED-92-16, Nov.  5, 1991). 

\14 Employees who are able to exercise their seniority rights and
find other jobs within Amtrak may be eligible for displacement pay
(the difference between their old and new average monthly
compensation) for up to 6 years. 


   NEW OPERATING AGREEMENTS WITH
   FREIGHT RAILROADS COULD
   INCREASE COSTS
---------------------------------------------------------- Chapter 3:4

Freight railroads own about 97 percent of the track over which Amtrak
operates and provide such essential services as dispatching trains,
making emergency repairs to Amtrak trains, and maintaining stations. 
These services are provided under operating agreements that Amtrak
maintains with 18 railroads.  On April 30, 1996, most of Amtrak's
initial 25-year operating agreements with freight railroads will
expire, and new agreements must be negotiated.  On the basis of our
discussions with freight railroad officials, it is likely that
Amtrak's costs under the new agreements will increase substantially. 
Amtrak currently pays about $90 million per year in both base
payments and incentive payments to freight railroads.  (Incentive
payments are additional amounts that railroads can earn when Amtrak
trains operate on time.) (See fig.  3.8.)

   Figure 3.8:  Amtrak's Payments
   to Other Railroads, Fiscal
   Years 1989-94

   (See figure in printed
   edition.)

Source:  GAO's illustration of Amtrak's data. 

Freight railroad officials told us that compensation will be a key
issue in negotiations with Amtrak.  They believe that their companies
are not adequately compensated for the services they currently
provide to Amtrak.  They offered several reasons.  First, the
methodology used to determine reimbursements for the incremental
costs of maintaining track--that is, the extra costs of wear and tear
on the track--does not adequately measure the costs resulting from
Amtrak's use of the track.  Of the $90 million that Amtrak pays
annually to freight railroads, about $20 million is for the
incremental cost of maintaining the track, and Amtrak estimates that
its costs for track maintenance could double if another methodology
is used.  Second, incentive payments do not consider delays caused by
Amtrak's trains--when an Amtrak locomotive fails, for example--in
calculating on-time performance.  Rather, these delays are held
against the freight railroads and consequently limit the amount of
incentives that they earn. 

Freight railroads may also seek higher compensation for
"level-of-utility" requirements that expire with the operating
agreements.  Amtrak trains generally travel at higher speeds than
freight trains.  As a result, the track must be maintained to higher
safety standards.  This increased standard is referred to as a higher
level of utility.  In addition, according to freight railroad
officials, Amtrak does not fully reimburse railroads for clearing
freight traffic and interrupting maintenance-of-way work to allow
Amtrak's trains to proceed on time. 

Freight railroad officials also said that liability arrangements
would be a key issue in negotiations.  They are concerned about their
liability in settling high-cost claims that result from
passenger-train accidents occurring on their track.  This concern
arose after a Conrail train collided with an Amtrak train near Chase,
Maryland, in January 1987; 16 people died and more than 350 people
filed injury claims.  In this case, Conrail paid about $94 million in
personal injury and death claims.  Freight railroad officials fear
that similar situations could develop on their property.  Under
current operating agreements, Amtrak and the freight railroads use a
"no-fault" liability arrangement, in which each party is responsible
for paying for its own equipment and personnel.  In addition, Amtrak
pays freight railroads between $0.0367 and $0.0734 per train mile for
the liability coverage that they must maintain in connection with
providing Amtrak service.  In fiscal year 1993, Amtrak paid about
$2.8 million for purchased insurance.\15 Freight railroad officials
told us that they would like to see this arrangement changed.  They
suggested that either Amtrak assume full responsibility for accidents
in which it is involved, regardless of who is at fault, and/or
liability claims be capped and handled through a pooled insurance
fund.  In either case, Amtrak's costs for liability protection could
increase substantially. 


--------------------
\15 Amtrak maintains about $200 million in liability insurance, of
which the first $25 million per occurrence is self-insurance.  The
balance is purchased from commercial insurance companies. 


   COSTS MAY INCREASE FOR EMPLOYEE
   BENEFITS AND ENVIRONMENTAL
   CLEANUP
---------------------------------------------------------- Chapter 3:5

Although health care and postretirement benefits are currently a
small part of Amtrak's budget, they could become more significant in
the near future.  In January 1995, the health care premiums that
Amtrak pays for its union employees are projected to increase by
about $61 per employee per month.  Amtrak estimates that this
increase could boost operating expenses by about $25 million per
year.  In addition, Amtrak provides postretirement health care and
life insurance benefits to salaried employees.  For fiscal years 1992
and 1993, Amtrak paid postretirement benefits of $1.4 and $1.5
million, respectively.  Amtrak estimates that cash outlays for these
benefits will grow from $3 million in fiscal year 1994 to around $20
million by the year 2010. 

Amtrak can also expect higher costs associated with environmental
cleanup.  Amtrak's current expenditures for environmental cleanup
projects have ranged between $2 million and $8 million per year. 
However, according to Amtrak, this amount represents only a fraction
of its known environmental problems.  For example, Amtrak may have to
spend between $17 and $69 million to bring the 69 fueling sites along
its routes into compliance with the Clean Water Act, according to
Amtrak officials.  Amtrak also estimated that an additional $1
million per year over a 5-year period will be needed to eliminate
asbestos at its stations and facilities.  Other environmental
projects are likely to increase costs further.  Amtrak's policy is to
limit spending for environmental cleanup to high-priority projects
that pose immediate dangers to the environment.  At the end of fiscal
year 1994, Amtrak recognized a $33 million liability for future costs
for environmental cleanup.  Amtrak acknowledged that costs would
increase if it fully complied with environmental standards and
requirements and implemented prevention and control measures. 


   CONCLUSIONS
---------------------------------------------------------- Chapter 3:6

In the coming years, Amtrak faces increased costs that it can no
longer defer or avoid.  These costs include keeping its equipment in
operating condition, purchasing new equipment, meeting its debt
obligations, negotiating and paying for new labor agreements, paying
reasonable fees to use other railroads' track, and paying for
increased employee benefits and environmental cleanup at Amtrak
sites.  Its past efforts to reduce costs, while necessary, have
resulted in deterioration of equipment and some facilities. 

The need for improvements is becoming critical, and these
improvements will not come cheaply.  The new progressive
maintenance/overhaul program, designed to keep more cars in service
at a lower unit cost, will still cost over $30 million more to
implement in fiscal year 1995 than was spent in 1994 and will
eventually cost nearly $60 million more as new equipment enters the
program.  However, without this commitment, Amtrak risks growing
backlogs of equipment in need of overhaul, depleting the number of
cars available for service.  The new cars and locomotives currently
being delivered should ease the pressures, in both the short and long
term, as older cars that are costly to maintain are replaced and more
revenues per car are generated.  By the turn of the century, however,
this equipment will need to be overhauled each year, increasing the
burden on Amtrak's aging overhaul facilities. 

We believe that capital investment is important and will lower
operating expenses in the long term by increasing productivity and
improving efficiency.  The Beech Grove overhaul facility could be
made much more economical by replacing, repairing, and modernizing
the track and structures that now impede efficient work there.  Such
investments could enable Amtrak to increase efficiencies and provide
comfortable, modern, on-time service to its passengers.  If the
problems that have been driving customers away from Amtrak are
reduced, demand for service may well increase.  In chapter 4, we
discuss the prospect that Amtrak can generate sufficient additional
revenues to offset these increased costs. 


REVENUE INCREASES ARE NOT LIKELY
TO RESULT IN MAJOR IMPROVEMENTS IN
AMTRAK'S CONDITION
============================================================ Chapter 4

Amtrak's revenues have never covered all the costs for providing
intercity rail passenger service on any route.  This gap, after
narrowing during the 1980s, has again widened.  Even as the economy
has recovered, Amtrak's ridership and revenues have not improved
correspondingly.  There are several opportunities for Amtrak to earn
additional revenues, including completing a computerized system to
maximize its revenues per seat, increasing the amount of compensation
the states pay Amtrak for losses on services it provides in the
states under section 403(b) of the Rail Passenger Service Act,
introducing high-speed rail in selected corridors, and expanding
commuter rail operations.  However, none of these actions is likely
to eliminate Amtrak's operating deficit. 


   REVENUES WILL CONTINUE TO BE
   INFLUENCED BY SERVICE QUALITY
   AND COMPETITION FROM OTHER
   MODES OF TRANSPORTATION
---------------------------------------------------------- Chapter 4:1

Since 1990, Amtrak's passenger revenues have fallen about 14 percent
in real terms (see fig.  4.1)--from over $1 billion in 1990 to about
$880 million in 1994--and riders have been experiencing more problems
on their trips as Amtrak continues to operate aged equipment and to
defer maintenance.  Amtrak has cited a weak economy and intense price
competition from the airlines as some of the key reasons for its poor
performance.  But deteriorating service quality and a spate of recent
accidents have also contributed to declining revenues. 

   Figure 4.1:  Amtrak's Passenger
   Revenues in 1994 Dollars,
   Fiscal Years 1989-94

   (See figure in printed
   edition.)

Source:  GAO's analysis of Amtrak's data. 


      REVENUES HAVE DECLINED AS
      SERVICE QUALITY HAS
      DETERIORATED
-------------------------------------------------------- Chapter 4:1.1

Although Amtrak's ridership generally increased during the 1970s,
financial losses persisted, and in recent years Amtrak's annual
deficit has risen steadily.  Since the 1980s, Amtrak's ridership has
remained relatively steady, fluctuating between 19 million and 22
million passengers annually.  (See fig.  4.2.) To deal with the
widening gap between revenues and expenses, Amtrak has had to rely
heavily on cost reductions, because substantial revenue increases
could not be expected with ridership stagnant and fares constrained
by falling air fares and gasoline prices (when adjusted for
inflation).  As Amtrak cut expenses, the quality of service
deteriorated, making increases in ridership even less likely. 

   Figure 4.2:  Amtrak's Intercity
   Ridership, Fiscal Years 1980-94

   (See figure in printed
   edition.)

Note:  Ridership data for 1994 are preliminary. 

Source:  GAO's illustration of Amtrak's data. 

According to a June 1994 survey of Amtrak's passengers, about 61
percent had experienced at least one problem during their trip.  Both
the proportion of passengers experiencing problems and the number of
problems per passenger increased with the distance traveled.  About
74 percent of passengers on western long-distance trains and about 67
percent of those on eastern long-distance trains experienced
problems.\1 However, even riders on Amtrak's high-speed Metroliners
encountered problems--roughly 44 percent of passengers reported
problems. 

The most common problems cited in the customer survey concerned
on-time performance.  Late arrivals and departures accounted for
about 21 percent of the problems that passengers cited. 
Dissatisfaction with the cleanliness of facilities was also common,
accounting for about 10 percent of the problems mentioned.  Between
1989 and 1992, overall on-time performance improved from 75 to 77
percent; performance on certain routes and for certain trains,
however, remains poor.  In 1993, Amtrak's systemwide on-time
performance declined to 72 percent.  On one route--the Empire
Builder, which runs between Chicago and Seattle--trains arrived on
time only 4 percent of the time, according to Amtrak.  Amtrak
estimates that late arrivals and departures alone result in about $80
million in lost revenues annually.  Altogether, Amtrak believes that
it is losing over $300 million annually because of problems affecting
customer satisfaction.\2


--------------------
\1 Amtrak defines long-distance trains as those traveling 500 miles
or more. 

\2 These numbers are based on estimates of the number of potentially
lost customers--both those experiencing problems and those lost for
other reasons, such as word of mouth--multiplied by an expected
annual revenue of $514 per passenger. 


      RAIL ACCIDENTS HAVE CAUSED
      TEMPORARY DECLINES IN
      RIDERSHIP
-------------------------------------------------------- Chapter 4:1.2

The series of accidents involving Amtrak trains--one near Mobile,
Alabama, in September 1993; others in Boise, Idaho; Kissimmee,
Florida; and Gary, Indiana, during November and December 1993; a
derailment at Selma, North Carolina, in May 1994; and an accident
near Batavia, New York, in August 1994--appear to have resulted in
temporary declines in ridership and thus revenue losses.  After the
September 1993 accident, Amtrak's ridership nationwide fell by about
4 percent, affecting revenues for about 2 months.  After the
subsequent accidents, revenues dropped by as much as 15 percent. 
Revenue recovery took increasingly longer and occurred only after
aggressive marketing campaigns. 


      ECONOMIC CONDITIONS AND
      COMPETITION HAVE STRAINED
      AMTRAK'S REVENUES AND
      PASSENGER YIELDS
-------------------------------------------------------- Chapter 4:1.3

Competitive pressures have limited Amtrak's ability to increase
revenues by raising fares.  From 1990 to 1993, Amtrak's overall
yield--revenue per passenger mile--fell by about 10 percent, after
adjusting for inflation.  The declines were larger, about 12 percent,
for routes outside the Northeast Corridor.\3

Yields on traffic in the Northeast Corridor--where Amtrak derives
about one-half of its ridership--fell by about 6 percent.  These
falling yields represent fare reductions made in response to, among
other things, lower fares on airlines and buses.  Meanwhile,
passenger yields on domestic airlines fell by about 12 percent from
1989 to 1993 and are now below Amtrak's.  Yields on intercity buses
fell about 10 percent from 1989 to 1992 and are well below the yields
of Amtrak and the airlines. 

Most intercity trips are made by private vehicles.  Automobiles and
other private vehicles account for about 80 percent of total
passengers miles of intercity travel, while Amtrak represents only
about 0.3 percent.  Falling real gasoline prices continue to
encourage people to drive.  Since 1990, the real price of regular
unleaded gasoline has declined by about 10 percent.  This price
decrease, combined with a general increase in the average fuel
efficiency of new cars, continues to make driving a relatively
inexpensive option, especially for leisure travel by families.  Many
people make the decision to drive on the basis of only the
out-of-pocket cost of their trip.  Many related costs, such as
insurance and depreciation, are perceived as largely fixed and do not
enter into the traveler's choice of mode. 

Thus, lower fares for air and bus travel and lower gasoline prices
combined to improve the relative attractiveness of the alternatives
to taking the train, while the quality of train service declined. 
These trends have continued for some time, and Amtrak will find it
difficult to overcome them. 


--------------------
\3 The numbers cited are passenger yields, which includes revenues
from both tickets and food and beverage services. 


      YIELD MANAGEMENT SYSTEM MAY
      HAVE BENEFITS, BUT REVENUE
      GAINS MAY BE MARGINAL
-------------------------------------------------------- Chapter 4:1.4

To increase revenues and improve yields, Amtrak is computerizing its
yield management system, which allocates seats among several service
classes.  This system is similar to those used by the airlines to
control seat inventories and maximize revenues.  Amtrak does not want
to sell a seat to a passenger traveling only a short distance on a
long-distance train if by doing so it is unable to sell the seat to a
long-distance passenger.  Still, because most Amtrak passengers
travel only on a portion of a route, Amtrak cannot reserve all the
seats for long-distance travelers and must optimally manage its seat
inventory.  Moreover, because of the number of stops a train makes
and the number of passengers boarding and deboarding at each stop,
programming a yield management system for passenger trains is much
more complicated than computerizing a system for airlines.  As of
July 1994, 100 long-distance trains were included in the computerized
system, and Amtrak plans eventually to have 220 of its reserved-seat
trains in the system.  Because of budget constraints, the system is
being implemented in phases.  It was introduced in 1991 and is not
expected to be completed until the end of fiscal year 1995. 

This system should help Amtrak maximize revenues, especially on
heavily patronized long-distance trains during the peak travel
seasons.  However, it is not likely that the added revenues resulting
from better yield management will make a substantial impact on
Amtrak's deficit because most trains are far from overbooked,\4 and
Amtrak has used a manual yield management system for some time. 


--------------------
\4 The load factor--the number of passenger miles per seat mile--for
all Amtrak trains was 52 percent in fiscal year 1993.  In 1994, the
load factor dropped to 49 percent. 


   CURRENT INTERCITY AND
   STATE-SUPPORTED ROUTES LOSE
   MONEY
---------------------------------------------------------- Chapter 4:2

Amtrak as a whole loses money.  However, to judge the financial
performance of individual Amtrak routes, it is necessary to determine
the revenues and costs associated with each route.  Doing so can be
difficult.  When all costs, including administrative and capital
costs, are fully allocated, all Amtrak routes--even the heavily
traveled Northeast Corridor--generate sizable deficits.  Amtrak also
operates intercity passenger rail services that receive financial
assistance from the states where they operate.  The support from the
states does not fully compensate Amtrak for the cost of operating
these trains.  Under its December 1994 business strategy, Amtrak will
seek reimbursement from the states for all the losses incurred by
these trains. 


      ALL OF AMTRAK'S INTERCITY
      ROUTES LOSE MONEY
-------------------------------------------------------- Chapter 4:2.1

As noted above, none of Amtrak's routes are profitable if the costs
are fully allocated, and only services in the Northeast Corridor and
on a few special trains\5 generate revenues that exceed avoidable
costs.  Some costs would cease if a route were discontinued or,
conversely, start if a new service were introduced.  These costs
include short-term avoidable costs, such as those for train and
engine crews and fuel.  Other costs would not end immediately if a
route were eliminated.  These include not only long-term avoidable
costs, such as expenses for heavy equipment maintenance and training,
but also the short-term costs directly attributable to a route. 

Losses on individual routes vary depending on what costs Amtrak
considers in the calculation.  If costs are fully allocated,
passenger revenues covered only about 54 percent of the costs in
fiscal year 1993.  In calculating fully allocated costs, Amtrak
excludes certain expenses, including general and administrative
overhead, interest, non-intercity passenger operations, and
adjustments from previous periods.  For short routes--those less than
500 miles--revenues covered 83 percent of the long-term avoidable
costs, while revenues from long-distance service covered 75 percent
of such costs.  The Northeast Corridor made a positive contribution
on the basis of long-term avoidable costs.  If only short-term
avoidable costs are considered, then both short- and long-distance
routes outside the Northeast Corridor come close to covering their
costs.  (See table 4.1.)



                          Table 4.1
           
             Costs and Revenue-to-Cost Ratios for
           Routes, Calculated Using Different Cost
            Allocation Methodologies, Fiscal Year
                             1993

                    (Dollars in millions)

                                      Corr
                                      idor
                                rout  rout
                    routes        es    es  passenger routes
--------------  --------------  ----  ----  ----------------
Train                $482       $160  $375  $1,019
revenues\a

Fully                $961       $348  $579  $1,891
allocated
costs

Long-term
avoidable            $639       $193  $231  $1,064
costs

Short-term
avoidable            $532       $163  $190  $885
costs

Ratio of
revenues to          0.50       0.46  0.65  0.54
fully
allocated
costs

Ratio of
revenues to          0.75       0.83  1.62  0.96
long-term
avoidable
costs

Ratio of
revenues to
short-term           0.91       0.98  1.97  1.15
avoidable
costs
------------------------------------------------------------
Note:  Totals do not add because of rounding. 

\a Revenues exclude non-intercity passenger sources (e.g., commuter
operations). 

Source:  GAO's analysis of Amtrak's data. 

Part of the reason that services in the Northeast Corridor appear
more profitable than services in other parts of the system is that
Amtrak treats a significant portion (60 percent) of the costs to
maintain track in the Northeast Corridor as fixed costs and therefore
excludes them from the measures of avoidable costs.  Elsewhere,
Amtrak considers track maintenance costs as avoidable costs because
they represent a contractual obligation between Amtrak and the
freight railroads that own the track.  Furthermore, the Northeast
Corridor includes both conventional trains and high-speed Metroliner
trains.  While the Metroliners recover 90 percent of their fully
allocated costs, the conventional trains do not perform as well.  The
Metroliners skew the overall cost recovery ratio of the Northeast
Corridor so that it performs much better than other routes.\6 (See
table 4.2.)



                          Table 4.2
           
           Comparison of Revenues and Expenses for
           Trains in the Northeast Corridor, Fiscal
                          Year 1993

                    (Dollars in millions)

                                                  Convention
                                      Metroliner          al
                                               s    trains\a
------------------------------------  ----------  ----------
Train revenues                            $127.9      $233.5
Fully allocated costs                     $142.1      $392.7
Ratio of revenues to fully allocated        0.90        0.59
 costs
------------------------------------------------------------
\a Excludes Philadelphia-Harrisburg and Atlantic City service. 

Source:  GAO's analysis of Amtrak's data. 


--------------------
\5 Special trains provide unscheduled, intermittent service at a
profit.  These trains earn less than 0.2 percent of Amtrak's total
revenues and contribute 0.1 percent to total expenses. 

\6 Special trains, which account for 0.2 percent of Amtrak's
revenues, covered 88 percent of their fully allocated costs in fiscal
year 1993. 


      STATE-SUPPORTED SECTION
      403(B) SERVICE OPERATES
      AT A LOSS
-------------------------------------------------------- Chapter 4:2.2

The Rail Passenger Service Act allows Amtrak to initiate and/or
operate intercity rail service, known as 403(b) service, that is
financially supported by the states.  In fiscal year 1994, Amtrak had
contracts with eight states to operate such service over 13 routes.\7
This service accounted for about 14 percent of Amtrak's ridership. 

Under the provisions of the act, the states pay 45 percent of
operating losses for such service in the first year of operation and
65 percent of losses in subsequent years.  For service that began
before 1989, states reimburse Amtrak for short-term avoidable losses,
while for service that began after 1989, states reimburse Amtrak for
long-term avoidable losses.  In fiscal year 1994, three of the eight
states--Missouri, New York, and Michigan--reimbursed Amtrak for
short-term avoidable losses; three states--Alabama, North Carolina,
and Wisconsin--reimbursed Amtrak for long-term avoidable losses; and
two states--California and Illinois--reimbursed on both long- and
short-term bases, depending on the route.  States also pay 50 percent
of the capital costs (a calculation based on depreciation and
interest) associated with the equipment used for this service.  Any
losses (capital or operating) not paid by the states are absorbed by
Amtrak.  For the most part, Amtrak uses its own equipment to provide
this service. 

In fiscal year 1993 (the last year for which financial data on 403(b)
service are available), Amtrak absorbed about $82 million in losses
on section 403(b) services.  This amount included about $78 million
in operating costs and $4 million in capital costs.  Amtrak absorbed
such costs as heavy maintenance and overhaul of cars and locomotives,
repairs following accidents, and an allocated portion of fixed costs
(e.g., expenses to operate yards and stations and various overhead
costs).  The states paid about $26 million.  Amtrak absorbs other
costs from the service as well.  For example, Amtrak's use of
equipment for section 403(b) service precludes its use on other
intercity routes where equipment shortages could occur.  Amtrak is
not reimbursed for these lost opportunity costs.  In 1992, Amtrak
adopted a policy whereby no new 403(b) service will be initiated
unless a state purchases and provides the cars and locomotives.  Few
states have been willing to make this type of investment. 

As part of the new business strategy adopted in December 1994, Amtrak
will seek to gradually eliminate the "deeply discounted" service
provided to the states under section 403(b).  To accomplish this
goal, it plans to renegotiate the reimbursement terms of all 403(b)
service over the next several years so that the participating states
subsidize all costs not covered by revenues.  Doing so may require
changes to the Rail Passenger Service Act, according to Amtrak, but
the corporation has not yet decided what legislative changes it may
seek. 


--------------------
\7 These states were Alabama, California, Illinois, Michigan,
Missouri, New York, North Carolina, and Wisconsin. 


   HIGH-SPEED RAIL SERVICE COULD
   GENERATE NEW REVENUES BUT
   REQUIRE SIGNIFICANT NONFEDERAL
   SUPPORT
---------------------------------------------------------- Chapter 4:3

Many supporters of intercity rail passenger service believe that more
people would ride the trains and that routes could be more profitable
if high-speed services were introduced.  Amtrak continues to work on
the Northeast Corridor Improvement Program (NECIP) and plans to
extend high-speed service from New York to Boston.  Amtrak still
requires about $1.5 billion to complete this modernization. 
High-speed trains have been proposed for other corridors around the
nation, but attempts to build them have encountered serious
obstacles.\8 Significant capital costs will be incurred for
improvements to the infrastructure and for new equipment regardless
of where these systems are built.  The High-Speed Ground
Transportation Development Act of 1994 authorizes $169 million to
assist in planning these systems, but no funds have been
appropriated.  Moreover, other questions, such as who will operate
these systems and who will pay for them, remain unanswered. 


--------------------
\8 See See High-Speed Ground Transportation:  Issues Affecting
Development in the United States (GAO/RCED-94-29, Nov.  17, 1993). 
See also In Pursuit of Speed:  New Options for Intercity Passenger
Transport (Special Report 233), Transportation Research Board
(Washington D.C.:  1991). 


      FINANCIAL BENEFITS FROM
      NORTHEAST CORRIDOR
      HIGH-SPEED SERVICE MAY BE
      DIFFICULT TO ATTAIN
-------------------------------------------------------- Chapter 4:3.1

Extending electrification from New Haven, Connecticut, to Boston
under NECIP will allow train speeds of up to 150 miles per hour, will
cut travel times between New York and Boston from 4 hours to under 3
hours, and should generate increases in both ridership and revenues. 
Amtrak estimates that annual revenues from this project will exceed
long-term avoidable costs by about $36 million (in 1993 dollars) in
the year 2010.  Although this amount will not cover capital costs,
any revenues received in excess of long-term avoidable costs will
help reduce Amtrak's need for federal operating subsidies.  These
increased revenues are predicated on Amtrak's capturing about 45
percent of the rail/air travel market between New York and Boston, a
share equal to that currently held by Amtrak between Washington,
D.C., and New York, where trains achieve speeds of 125 miles per
hour.  As of September 30, 1994, Amtrak estimated that the major
improvements in infrastructure required to provide 3-hour service
between New York and Boston would be completed in 1999. 

Since 1976, about $3.3 billion (in current-year dollars) has been
appropriated for NECIP.  Amtrak estimates that about $1.5 billion
more is needed to complete the project.  FRA estimates that an
additional $582 million (in constant 1993 dollars) will be needed to
address problems with track capacity.  Counting prior expenditures on
the Washington-New York segment, the corridor will cost about $5
billion.  The track between New York and Boston is currently shared
by Amtrak and freight and commuter railroads--all of which plan to
increase their use of the track in future years.\9

The existing track cannot accommodate all of these plans.  Therefore,
either some train operations will have to be shifted to off-peak
hours or additional capacity will have to be constructed to ensure
that all parties' needs are met.  Since Amtrak does not own about 95
miles of the rights-of-way between New York and Boston, it may have
difficulty negotiating shifts in train schedules and/or costs to gain
additional capacity.  Consequently, Amtrak may either have to absorb
more of the additional costs than it expects or delay its planned
increases in train schedules until capacity problems can be resolved. 
Either action could significantly affect Amtrak's ability to achieve
timely revenue gains as a result of NECIP. 


--------------------
\9 For example, Amtrak plans to more than double the number of trains
on most sections of the route by the year 2010. 


      HIGH-SPEED SERVICE OUTSIDE
      THE NORTHEAST CORRIDOR FACES
      SIGNIFICANT OBSTACLES
-------------------------------------------------------- Chapter 4:3.2

Providing high-speed rail service outside the Northeast Corridor will
be influenced by the recently enacted High-Speed Ground
Transportation Development Act of 1994.  This legislation authorizes
$169 million over a 3-year period through fiscal year 1997 for
planning assistance for high-speed rail corridors.  To date, no funds
have been appropriated and no construction funds have been
authorized.  A number of important questions have yet to be answered,
including who will be designated to operate future high-speed rail
service.  There is no guarantee that Amtrak will be selected to
operate these services.  Nor is it clear how much federal assistance
might be provided to build high-speed rail systems.  The
responsibility for funding these projects is likely to fall on the
states, localities, or private investors. 

The impact of high-speed rail systems on Amtrak's need for federal
subsidies is also unclear.  If Amtrak's role is limited to operating
such systems under contract to others and federal operating subsidies
are limited, Amtrak and the federal government could largely be
shielded from losing money on these operations.  If high-speed rail
service is operated as part of Amtrak's national intercity route
system, then federal operating subsidies will rise or fall depending
on whether revenues exceed long-term avoidable costs.  Any federal
capital assistance would presumably be provided through separate
appropriations.  Amtrak estimates that nine high-speed train sets, at
a total cost of about $170 million, would be needed to provide nine
daily round-trips on a 300-mile corridor. 


   ANCILLARY ACTIVITIES ARE A
   GROWING SOURCE OF REVENUE, BUT
   FINANCIAL CONTRIBUTION IS
   UNCLEAR
---------------------------------------------------------- Chapter 4:4

Amtrak also earns revenues by operating commuter rail services under
contract to state and local governments and regional transit
agencies, developing its real estate holdings, and providing mail and
express service.  Amtrak was unable to provide us with an accounting
of the costs associated with these activities until December 1994;
therefore, we have not assessed their profitability.  In fiscal year
1993, these activities accounted for about $435 million--or about 30
percent--of Amtrak's $1.4 billion in revenues.  In general, ancillary
activities have been a growth area for Amtrak and, overall, appear to
have produced a modest profit.  In particular, commuter rail
services, which generated revenues of about $270 million in fiscal
year 1994, were Amtrak's second largest source of operating revenue. 
Amtrak operates eight commuter rail systems in five states under
competitively awarded cost-plus contracts.  Amtrak primarily operates
these services and does not use its own equipment.  Since 1983, when
Amtrak entered into its first contract with the Maryland Rail
Commuter System, its commuter revenues and ridership have grown as
Amtrak has entered into new contracts.  (See figs.  4.3 and 4.4.)
Amtrak began operating two major new systems in 1993--the Virginia
Railway Express in northern Virginia and Metrolink in Los Angeles--to
boost its commuter ridership from 20 million in 1992 to 29 million in
1993.  Amtrak now carries more passengers on its commuter services
than on its intercity operations. 

   Figure 4.3:  Amtrak's Revenues
   From Commuter Railroads, Fiscal
   Years 1989-94

   (See figure in printed
   edition.)

Source:  GAO's analysis of Amtrak's data. 

   Figure 4.4:  Amtrak's Ridership
   on Commuter Railroads, Fiscal
   Years 1989-93

   (See figure in printed
   edition.)

Source:  Amtrak. 

The financial contribution from Amtrak's ancillary activities is
largely unknown.  During our review, Amtrak had difficulties in
identifying the costs of these activities and, for the most part, was
unable to provide us with financial statements for them in a timely
manner.  This was particularly true for commuter rail activities. 
Data on revenues and ridership were available but data on expenses
were not.  We also had difficulty identifying general and
administrative expenses for these activities and the way these costs
are allocated to specific lines of business or specific contracts. 
According to Amtrak, these costs are not accounted for separately but
are instead allocated according to standard corporatewide formulas. 
We did not audit these formulas.  As a result, it is not clear what
the financial contribution of Amtrak's ancillary activities might be,
nor whether costs are being allocated correctly. 


   CONCLUSIONS
---------------------------------------------------------- Chapter 4:5

Since 1990, Amtrak's ridership has been relatively stagnant, revenues
have declined, and the quality of service has deteriorated.  Amtrak's
problems can be attributed partly to uncertain economic conditions
and to competition from the airlines and from buses.  But they are
also the result of continued reliance on old, unattractive equipment
that is prone to breakdowns and delays.  Lacking the resources to
purchase the new equipment that would increase the quality of service
and constrained to match operating costs with federal subsidies,
Amtrak has been forced to cut costs, delay maintenance of equipment,
and generally let the quality and attractiveness of train travel
deteriorate further.  The prospects for recovery of ridership and
revenues are poor. 

Increasing the amount of compensation the states pay for services
provided under section 403(b), contracting to provide more commuter
rail operation, and introducing high-speed trains outside the
Northeast Corridor could all enhance revenues, but none of these
initiatives can be expected to close the current deficit.  In chapter
5, we make recommendations to Amtrak to determine the profitability
of its ancillary activities and to provide further information to the
Congress on the corporation's potential involvement in high-speed
service outside the Northeast Corridor.  Chapter 5 also discusses
alternatives for matching the country's needs for viable intercity
rail passenger service with the realities of limited federal
resources. 


REASSESSMENT OF AMTRAK'S MISSION
IS NEEDED
============================================================ Chapter 5

Amtrak and the federal government need to make important decisions
about the future of intercity passenger rail service and the
government's commitment to subsidize such operations.  Amtrak's
condition will only get worse if it continues to operate the current
system--even with the reductions in service planned for 1995--at the
current level of state and federal funding.  High quality nationwide
passenger service of the present scope that might attract and retain
passengers would require substantially higher levels of additional
support, particularly for capital investment.  The key question is
whether the federal government or the individual state governments
are willing to make the required investments, given the competing
demands on their resources.  A substantially reduced passenger rail
system, while more feasible from a fiscal viewpoint, requires that
difficult decisions be made on the type, quality, and location of the
remaining services. 


   CONTINUED OPERATION OF EXISTING
   SERVICE IS NOT FEASIBLE AT
   CURRENT SUBSIDY LEVEL
---------------------------------------------------------- Chapter 5:1

We do not believe that the current situation--the present nationwide
passenger rail system at the current subsidy level--represents a
viable option.  Amtrak's financial condition will continue to
deteriorate, and the railroad's ability to provide nationwide service
at the present level will be seriously threatened.  To maintain the
current nationwide system will require significantly increased
resources if Amtrak is to offer quality service.  Without additional
funds from federal, state, and local governments, Amtrak will have to
cut expenses significantly by eliminating some routes and reducing
the frequency of service on others.  In either case, ridership could
fall as the level of service declines. 

In September 1994, Amtrak officials announced their decision to
reduce Amtrak's management force from 2,700 to 2,100 people by the
end of 1994 through voluntary or forced separation.  The staffing cut
is part of a plan to shift control over pricing, marketing, and
service to local offices and train crews by reducing headquarters
staff in Washington, D.C., and transferring staff to three regional
operating centers.  Amtrak hopes to improve efficiency by
establishing operating centers in Philadelphia for the Northeast
Corridor, Los Angeles for West Coast service, and Chicago for the
rest of the nation.  These changes and reductions in staff, however,
involve too few dollars to substantially affect Amtrak's financial
condition.  After severance costs, Amtrak expects these cuts to save
approximately $30 million in fiscal year 1995. 

Realizing that additional actions needed to be taken, in December
1994 Amtrak announced plans to cut expenses by reducing service, as
we discussed in chapter 2.  If implemented, these actions could bring
Amtrak's operating costs in line with the revenues and grants for
1995.  Nonetheless, these planned actions will not solve the
corporation's longer-term problems.  Revenues will continue to fall
short of expenses on most routes, and Amtrak estimates that operating
expenses will exceed operating revenues and the federal operating
subsidy by $1.3 billion between 1996 and the year 2000.  Furthermore,
Amtrak will still need over $4 billion to replace worn-out equipment
and infrastructure. 

Achieving about 26 percent of the $364 million in annual net savings
that Amtrak anticipates from these actions might require collective
bargaining and/or legislative change, according to Amtrak.  Also, in
eliminating routes, Amtrak will incur labor protection expenses to
compensate workers who lose their jobs or are placed in lower-pay
positions.  Amtrak estimates that labor protection costs due to the
proposed changes could be between $80 million and $158 million. 

Amtrak has identified additional legislative changes that it believes
could improve the corporation's long-term viability.  These changes
include

including Amtrak in a federal transportation trust fund;

limiting Amtrak's provision of state-assisted rail passenger service
to situations in which the state is willing to provide the actual
cost of such service;

amending section 405 of the Rail Passenger Service Act to permit
negotiations on labor protection issues without the statutory
rigidity that currently limits those negotiations;

further amending section 405 to remove constraints on contracting for
work;

eliminating from Amtrak's budget the mandatory payments for railroad
retirement and unemployment benefits incurred by non-Amtrak
employees\1 ;

eliminating Amtrak's obligation to pay federal fuel taxes;

requiring that Amtrak's federal operating grant be provided on the
first day of the fiscal year to reduce overall cash flow costs and
requirements;

limiting punitive damages assessed against Amtrak through tort
reform;

providing Amtrak with the authority to issue tax-exempt debt;

providing tax incentives to freight railroads for the revenues they
earn from on-time performance payments from Amtrak; and

clarifying Amtrak's exemption from local regulations on permits for
work to improve the Northeast Corridor. 

At the time of our review, Amtrak officials had not determined the
extent to which these proposed changes to legislation would reduce
the corporation's expenses, but they plan to do so.  In addition,
Amtrak has not identified the costs and benefits associated with
these changes, including the impact of tax reductions or credits on
the national debt and the impacts on other affected parties.  Nor
have we assessed the impact of these actions on Amtrak or other
affected parties. 


--------------------
\1 Amtrak also believes that part-time station custodians should not
be considered railroad employees for purposes of this payment. 


   IMPROVING EXISTING RAIL SERVICE
   WOULD REQUIRE INCREASED FEDERAL
   FUNDING
---------------------------------------------------------- Chapter 5:2

Increased funding, especially capital investment, is essential if
Amtrak is to continue its nationwide service at the present level. 
Over the longer term, increased funding offers Amtrak the potential
to increase revenues and improve service quality on its existing
routes, increase the efficiency and productivity of its operations,
and, possibly, introduce high-speed service on some routes. 

To substantially improve its deteriorating financial condition,
Amtrak must increase its passenger revenues, which have been
declining in real terms since 1990, as noted in chapter 4.  By
investing in new passenger cars and locomotives, Amtrak could
increase its seating capacity on the routes where the demand is
greatest.  As we discussed in chapter 3, Amtrak recently purchased
245 passenger cars and 72 locomotives for nearly $1 billion. 

If the Congress more than doubled Amtrak's current capital subsidy to
$500 million annually from fiscal years 1995 to 2000--which may be
difficult within the current federal budget--Amtrak could improve its
maintenance facilities, stations, and information systems, among
other things.  However, even with the gains in efficiency and
ridership expected from these improvements, Amtrak would still need
more than $400 million in annual operating subsidy from state or
federal governments through the year 2000.\2 (See table 5.1.)



                                    Table 5.1
                     
                        Estimated Revenues, Expenses, and
                      Operating Grant If Amtrak Receives an
                     Annual Capital Subsidy of $500 Million,
                              Fiscal Years 1995-2000

                              (Dollars in millions)

--------  --------  --------  --------  --------  --------  --------  ----------
Estimate   $1,427    $1,646    $1,778    $1,944    $2,099    $2,239   $11,133
 d
 operati
 ng
 revenue
 s
Estimate   $1,960    $2,107    $2,247    $2,382    $2,523    $2,649   $13,868
 d
 expense
 s
Operatin    $533      $461      $469      $438      $424      $410    $2,735
 g grant
 require
 ment\a
--------------------------------------------------------------------------------
\a The operating grant requirement is the difference between the
estimated operating revenues and estimated expenses. 

Source:  Analysis of Amtrak's data for GAO by Snavely, King &
Associates, Inc. 


--------------------
\2 App.  I describes how we made these estimates.  These estimates
were made using data available from Amtrak during the summer of 1994. 
Since that time, Amtrak has announced its strategic business plan,
which fundamentally changes some of the forecasts used in our
analysis.  According to Amtrak, its revenue outlook is now much less
optimistic than it was in the summer of 1994, and its cost outlook is
more conservative.  However, Amtrak believes that updating our
estimates would not produce results that differ substantially from
the estimates we report. 


   REDEFINING SYSTEM AND
   REALIGNING SERVICE COULD
   IMPROVE AMTRAK'S FINANCIAL
   CONDITION AND SERVICE QUALITY
---------------------------------------------------------- Chapter 5:3

The Congress could redefine Amtrak's role and mission by
restructuring and realigning Amtrak's route system to a smaller basic
network that would continue to be eligible for federal funding.  This
basic network could be augmented by regional routes fully funded by
states.  This approach has the potential to improve Amtrak's
financial condition and service quality.  Additionally, it may lead
to reductions in federal funding over the longer term, although
capital subsidies will continue to be needed.  An analysis of
Amtrak's current market--the routes with the largest ridership and
revenues--could be used as a starting point in determining a basic
route network that would be eligible for federal funding.  We believe
that the specific routes included in this restructured system could
be determined by Amtrak or by a temporary commission such as the one
formed to evaluate military base closures. 


      CRITERIA FOR DEFINING A
      REDUCED ROUTE NETWORK
-------------------------------------------------------- Chapter 5:3.1

A basic route network could be defined by determining where Amtrak's
market is currently strongest.  As a starting point, we identified
the (1) routes with the largest ridership and (2) routes that
generated the greatest revenues.  We found that 12 of Amtrak's 44
routes accounted for 70 percent of Amtrak's 22 million riders in
fiscal year 1993.  The top five routes--all in the Northeast and
southern California--each had over 1 million riders and accounted for
56 percent of all riders.\3 (See fig.  5.1.)



(See figure in printed edition.)Figure 5.1:  Amtrak's Major Ridership
Markets, Fiscal Year 1993

The routes with the most riders roughly coincided with the routes
generating most of the passenger revenues.  Eleven routes (each with
revenues greater than $30 million in fiscal year 1993) earned 68
percent of Amtrak's passenger revenues in fiscal year 1993 and
accounted for 61 percent of the fully allocated costs.\4

The top five routes generated 47 percent of the total train revenues
and accounted for 38 percent of the costs.\5 The other six routes
extend service on the East Coast, provide a third route from Chicago
to the West Coast, and serve the West Coast.\6 (See fig.  5.2.)



(See figure in printed edition.)Figure 5.2:  Amtrak's Major Revenue
Markets, Fiscal Year 1993

Coast-to-coast service could be maintained if the Chicago-New York
City/Boston route--the next largest in revenues--is added to the
network.  Generally, however, the data on ridership and revenues
indicate that there is not a strong demand for passenger rail service
between the East Coast and Chicago, largely because rail service
cannot compete in time with air travel.  The daily train leaving
Washington, D.C., at 4:40 p.m.  arrives in Chicago at 9:10 a.m.  the
next day--a trip of over 16 hours.  By comparison, a flight between
Washington, D.C., and Chicago takes about 2 hours. 


--------------------
\3 The top five routes for passengers were the Metroliner (which
operates high-speed trains between Washington, D.C., and New York
City), conventional service in the Northeast Corridor (between
Washington, D.C., and Boston), New York City-Philadelphia, New York
City-Albany-Niagara Falls, and Los Angeles-San Diego.  The other
seven routes were New York City-Florida through coastal North and
South Carolina, New York City-Florida through inland North and South
Carolina, Chicago-Milwaukee, Chicago-Seattle, Chicago-Oakland, Los
Angeles-Seattle, and Oakland-Bakersfield. 

\4 In fiscal year 1993, Amtrak's train revenues were about $1
billion. 

\5 In fiscal year 1993, Amtrak's fully allocated costs were nearly
$1.9 billion. 

\6 The top five routes for revenues were Metroliners (high-speed
service between New York City and Washington, D.C.), conventional
service in the Northeast Corridor (between Boston and Washington,
D.C.), Chicago-Seattle/Portland, Chicago-St.  Louis-Oakland, and the
Autotrain.  The other six routes were New York City-Albany-Niagara
Falls, New York City-Florida through coastal North and South
Carolina, New York City-Florida through inland North and South
Carolina, Chicago-Los Angeles, New York City-New Orleans/Mobile, and
Los Angeles-Seattle. 


      FURTHER ANALYSIS COULD MORE
      PRECISELY DEFINE BASIC
      SYSTEM
-------------------------------------------------------- Chapter 5:3.2

The criteria for identifying the routes that could constitute a new
basic system needs to balance the requirement that the routes be
well-patronized with the costs associated with providing the service. 
Our analysis provides a preliminary look at a basic route network. 
Some of the routes with the most riders and revenues, however, also
have the largest losses.  Further analysis of the revenues and costs
associated with routes is needed to determine (1) what revenues could
potentially be lost as a result of losing passengers who are now
connecting from routes that would be eliminated and (2) whether
entire routes or only segments generate high revenues and ridership. 
In addition, the direct costs for each route, which are currently
allocated using a formula, need to be measured more precisely and
directly allocated. 

Amtrak compiles boarding and destination information on its
passengers from tickets collected on the trains.  These data could be
analyzed to determine the potential effect on revenues and ridership
of the loss of passengers now connecting from other routes.  Such an
analysis would reveal, for example, the number of passengers on the
Chicago-San Francisco route who begin their trips at stations that
could potentially be eliminated.  Additional information would need
to be gathered on transportation alternatives for interconnecting
passengers and the number who may still ride Amtrak after taking a
car or bus to the nearest alternative station. 

The data from tickets could also be analyzed to determine revenues
and ridership on segments of routes.  For example, most of the
revenues and ridership on the Chicago-Seattle route might be commuter
traffic between Chicago and Milwaukee and/or intrastate travel
between Spokane and Seattle. 

Until mid-December 1994, Amtrak did not know precisely the direct
costs associated with its individual routes.  Even costs that could
be directly measured, such as fuel and labor, are computed and
allocated to routes on the basis of formulas developed by Amtrak. 
Such formulas are a reasonable way to account for expenses on a
systemwide basis.  However, this method may provide insufficient
information to allow management to make judgments on the relative
performance of individual routes.  For example, this method does not
consider operating conditions that may be specific to routes, such as
the costs of turning a train around at the terminal stations.  These
costs are allocated to all routes, although turning trains around is
not necessary on some routes.  In addition, consideration must be
given to the sensitivity of financial performance on a given route to
the age and type of equipment assigned to it. 

Finally, Amtrak currently allocates costs to routes differently
depending on whether the route is considered "basic" or "incremental"
service.  For example, Amtrak designates the route between
Washington, D.C., and New York City as basic service and allocates a
portion of all operating costs to the route.  In contrast, the
Philadelphia-New York City route is designated as incremental
service, and only the costs of adding cars to trains on the basic
Washington-New York City route are allocated to it.  In addition,
Amtrak does not allocate the cost of locomotives to incremental
routes. 

Amtrak officials have recognized these problems with their cost
information and acknowledge that this method of allocating costs may
be making some routes appear much more unprofitable than they really
are.  During the fall of 1994, Amtrak undertook an in-depth analysis
of the costs and revenues of individual routes in developing its
December 1994 strategic business plan.  As part of its new business
plan, Amtrak intends to measure more direct costs and allocate them
to routes rather than allocating all costs according to formulas. 


      DETERMINING A NEW ROUTE
      STRUCTURE
-------------------------------------------------------- Chapter 5:3.3

Specific, alternative route structures and expected levels of federal
support could be developed and recommended to the Congress by Amtrak
or a temporary commission established by the Congress for that
purpose.  The commission or Amtrak could identify several alternative
networks, depending on the level of funding available.  A commission
to restructure passenger rail could operate in a manner similar to
the Department of Defense's (DOD) Base Realignment and Closure
Commission.\7 After the Congress identifies specific future goals and
objectives, the commission could obtain route-specific data from
Amtrak on ridership, revenues, and costs and define basic route
networks commensurate with the different funding levels. 

After identifying a potential basic network, specific route
structures could be assessed by considering (1) all the fixed costs
that could be reduced or increased if a large number of routes were
eliminated\8 and (2) the increased revenues that could be gained by
redeploying equipment to routes where the demand would support
increased service. 

Additional factors that the commission or Amtrak could assess include
the availability of alternative intercity transportation and the
impact of the proposed network on energy consumption, pollution, and
traffic congestion.  However, if these nonfinancial factors are used
to determine the route structure, it will be difficult to make
improvements in service quality and Amtrak's financial condition
without higher federal subsidies. 


--------------------
\7 DOD's current Base Realignment and Closure Commission, established
by law in 1990, is responsible for reviewing the recommendations on
base closures made by the Secretary of Defense.  On the basis of
those recommendations, the commission proposes the closure and/or
realignment of military bases.  In establishing this commission, the
Congress outlined specific procedures to be followed, including
provisions that all bases be compared equally against selection
criteria developed by DOD and a plan for the structure of the armed
forces for the following 6 fiscal years.  In addition, DOD is
required to certify the accuracy of the data it presents to the
commission.  Finally, all recommendations for base closures and/or
realignments are subject to public comment and are then forwarded to
the President.  If the President approves, the recommendations are
forwarded to the Congress.  Either the Congress must reject all
recommendations within a specified period of time or the Secretary of
Defense is directed to implement the recommendations. 

\8 Fixed costs could increase as a result of the labor protection
costs--up to 6 years' wages--that Amtrak would incur for employees
who lose their jobs when routes are eliminated. 


      ADDITIONAL SERVICE COULD BE
      PROVIDED WHERE STATES ARE
      WILLING TO SUBSIDIZE IT
-------------------------------------------------------- Chapter 5:3.4

After the Congress agrees on a basic rail service network that would
be eligible for federal funding, individual states could be given the
option of adding specific service to Amtrak's basic network if they
are willing to fully subsidize the added service.  Some states have
already entered into such agreements with Amtrak.  For example,
Washington State contracted with Amtrak to operate upgraded service
between Seattle and Portland, Oregon.  All seats on the train were
reserved, the train included a dining car that served local
specialties such as salmon, and the train featured on-board telephone
service and video entertainment.  By comparison, Amtrak's regular
service between Seattle and Portland offered only snack and beverage
service.  The Washington State Department of Transportation collected
all revenues for this service and paid all costs. 

Nonetheless, a state's flexibility in using federal funds for
intercity passenger rail service is constrained under current law. 
Passenger rail service competes for limited transportation funds and,
unlike aviation, highways, and mass transit, it does not have a
federal trust fund.  State and local governments have some
flexibility to allocate their federal transportation funds among
different modes, but not to intercity passenger rail operations. 


      EVEN IF IT RESTRUCTURES,
      AMTRAK WILL CONTINUE TO NEED
      CAPITAL FUNDS
-------------------------------------------------------- Chapter 5:3.5

Even if it completes a major restructuring of its service, Amtrak
will continue to need capital funds for equipment and infrastructure. 
In Europe and Japan, where competitive conditions are more conducive
to rail travel, intercity passenger service requires substantial
public support, including significant investments in the
infrastructure.  For example, France plans to invest nearly $25
billion in its railroad during the 1990s.  This amount includes $6.8
billion for rolling stock, $5.3 billion for investments in
infrastructure on high-speed lines, and $1.1 billion for other
investments in infrastructure.  Germany plans to invest over $70
billion in its main railway lines in the 1990s, including $28.8
billion for improvements in infrastructure, $18.5 billion in other
upgrades, and $8.2 billion in equipment. 

Amtrak will continue to need federal grants to meet its capital
needs.  A 1991 study\9 conducted for FRA evaluated four scenarios for
Amtrak's future--a system in the Northeast Corridor only, the
Northeast Corridor plus a few routes connected to the corridor where
losses are relatively small, Amtrak's current route network, and the
current network minus several routes where losses are the largest. 
All scenarios assumed that Amtrak would require a capital subsidy. 
This study further assumed that Amtrak would improve its efficiency
in certain areas, such as its maintenance facilities.  Even though
some assumptions were overly optimistic in predicting improvements in
Amtrak's efficiency, this study estimated that by the year 2000,
Amtrak would still require a federal operating subsidy under all
scenarios.  The smallest subsidy would be required for the current
network minus the routes with the largest losses.  The requirement to
pay up to 6 years' wages to employees who lose their position if a
route is eliminated was a major factor in determining the cost under
the alternative scenarios. 


--------------------
\9 Amtrak Review, prepared for the U.S.  Department of Transportation
by Battelle et al., draft (Oct.  17, 1991). 


   PRIVATIZATION OF AMTRAK IS AN
   UNLIKELY OPTION UNDER PRESENT
   CIRCUMSTANCES
---------------------------------------------------------- Chapter 5:4

If the Congress chooses to eliminate or greatly reduce federal
subsidies for Amtrak, it could privatize the operation and make it
subject to market forces.  Amtrak's route network would be
reconfigured so that only those parts of the system that had the
potential to cover their costs would continue to operate.  While some
rail passenger services conceivably could be taken over by the
private sector, significant federal investment would be needed before
any part of the current system could be privatized, and nationwide
service as it exists today could not be offered. 

Privatizing Amtrak might be complicated by a number of factors. 
First, it is not clear what would be privatized.  Amtrak owns very
little track outside the Northeast Corridor.  Also, most of the
stations Amtrak serves are owned either by other railroads or by
local governments.  Even many of Amtrak's passenger cars and
locomotives are leased.  Second, the term "privatize" is sometimes
used to mean "defederalize"; that is, to shift the responsibility for
subsidizing intercity train operations to state and local
governments.  Passenger train services might be inherently
unprofitable.  Therefore, private, for-profit firms are unlikely to
be interested in such business without some government assistance. 
The private railroads in this country were unable to operate
passenger service profitably, and throughout the world, intercity
passenger trains are heavily subsidized.  If the public benefits of
intercity passenger train services are largely local or regional,
these services might more appropriately be offered or supported by
state or local governments.  Third, even in the Northeast Corridor,
where Amtrak controls significant assets, different degrees of
privatization are possible.  For example, programs to privatize
freight rail service in the Netherlands envision privatizing the
operations but not the infrastructure.  The government will continue
to provide capital to maintain and develop the track and other
facilities. 

In addition, privatizing those parts of Amtrak that could potentially
be profitable might still require substantial initial investment to
create a saleable asset, similar to what occurred with Conrail.  In
that instance, many unprofitable routes were abandoned and, after
substantial federal investment, a profitable core business was
established.  However, the analogy with Conrail may be limited
because, on the basis of experience in the United States and
throughout the world, we have found no evidence that intercity
passenger rail operations can cover all costs and generate a return
on investment.  Therefore, even if privatized, Amtrak will continue
to need federal or state funds to meet its capital needs.  Even the
Northeast Corridor would be difficult to sell to a private firm until
the upgrade between New York and Boston is complete and significant
repairs have been made to the segment between New York and
Washington, D.C. 

Finally, privatizing Amtrak is not likely to result in successfully
preserving a nationwide passenger rail system.  Under this option,
passenger rail service could be reduced to a few regional corridors
because only a few well-traveled routes could ever generate
sufficient revenues to cover the substantial operating costs. 


   CONCLUSIONS
---------------------------------------------------------- Chapter 5:5

Amtrak and the federal government face a difficult set of choices. 
We believe that continuing the present course--the same funding level
and basic route system, even with the proposed service cuts--is
neither feasible nor realistic because Amtrak will continue to
deteriorate.  The Congress is confronted with numerous budget
decisions that make substantial increases in Amtrak's subsidy
unlikely.  If increases are not forthcoming from federal and state
sources of funds, Amtrak's viability may depend on restructuring
operations to reduce the route network.  Even so, Amtrak will
continue to require government subsidies, especially to meet its
capital needs.  Amtrak and the federal, state, and local governments
need to make important decisions about the quality and extent of
intercity passenger rail service to be provided and the long-term
funding of such an operation.  First, the Congress needs to decide on
the nation's expectations for intercity passenger rail service and
the scope of Amtrak's mission to provide that service.  These
decisions require defining a national route network, along with
determining the extent to which the federal government would fund
operating losses and capital investments and the way any remaining
deficits will be covered.  We believe that Amtrak or a temporary
commission could provide the Congress with specific options that
would define a national route network consistent with the available
funding.  Finally, once the Congress decides on a national route
network, Amtrak could develop and provide to the Congress a long-term
financing and operating plan (5 to 10 years).  This plan should
provide realistic expectations for repairing and maintaining Amtrak's
fleet, replacing aging infrastructure, and meeting increases in
expenses that can be reasonably anticipated. 


   RECOMMENDATION TO THE CONGRESS
---------------------------------------------------------- Chapter 5:6

In light of Amtrak's financial and operating problems, we recommend
that the Congress consider whether Amtrak's original mission of
providing nationwide intercity passenger rail service at the present
level is still appropriate.  If the Congress decides to reassess the
scope of Amtrak's mission, it could direct Amtrak or a temporary
commission, similar to the one established to close military bases,
to make recommendations and offer options defining and realigning
Amtrak's basic route network so that efficient and quality service
could be provided within the funding available from all sources.  The
Congress could then make the final decision on Amtrak's future route
network. 


   RECOMMENDATIONS TO THE
   PRESIDENT OF AMTRAK
---------------------------------------------------------- Chapter 5:7

To ensure that Amtrak accurately communicates its operating and
financial conditions and its need for federal funds to the Congress,
we recommend that the President of Amtrak

provide detailed information in federal grant requests on how
revenues from intercity passenger service have been estimated;

incorporate into federal grant requests dollar estimates of the costs
of future accident and weather-related contingencies;

develop and present to the Congress a plan outlining the costs and
benefits of participating in high-speed rail service outside the
Northeast Corridor, including the impact on Amtrak's annual grant
request;

undertake a comprehensive review and/or audit of the costs associated
with its commuter rail and other ancillary activities, identifying
the costs associated with these activities, the way these costs are
allocated to individual commuter rail contracts, and the overall
profit or loss of each activity as well as assessing the
appropriateness of any formulas used to allocate costs; and

provide the Congress with proposed legislative changes that could
improve Amtrak's long-term viability, along with estimates of the
expected effect of each proposal on Amtrak's finances and a
discussion of the other parties that will likely be affected and to
what extent. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 5:8

Amtrak said that our draft report accurately presented the railroad's
financial and operating status and correctly portrayed the
corporation's capital investment problems.  However, Amtrak had four
principal points. 

First, Amtrak stated that the draft report understated the impact of
actions adopted by its Board of Directors in December 1994 and said
the board had, in effect, implemented our recommendation to redefine
and reduce the route system consistent with the funding available. 
We revised our report to specifically highlight Amtrak's proposed
operating changes, including plans to reduce staffing, reduce the
frequency of service on some routes, and eliminate service on others. 
In addition, Amtrak acknowledges that even with the planned service
reductions, its operating deficits will exceed operating subsidies by
$1.3 billion through the end of the century.  While we believe that
the new business plan is an important first step, it does not
implement our recommendation, nor does it resolve Amtrak's financial,
capital, and service quality problems. 

Unless sufficient funds are available to support Amtrak's current
operations and provide the necessary capital, we recommend that the
Congress reassess Amtrak's scope of operation and mission and direct
either Amtrak or a temporary commission to provide the Congress with
options for a route network that is consistent with the level of
funding available.  Not only do the service reductions announced in
the board's plan still leave a large gap between the deficit and the
subsidies from the federal and state governments after 1996, but
about a quarter of the planned savings will require negotiations with
organized labor and/or legislative changes.  Also, Amtrak's plan does
not resolve how the corporation will meet its capital needs, now
totaling about $2.5 billion in the Northeast Corridor alone. 
Although the board set a goal that could ultimately eliminate the
federal operating grant by 2002, this was made contingent on Amtrak's
receiving (1) "sufficient capital funding to achieve a
good-state-of-repair," (2) the current level of operating grants
until 2002, and (3) increased funding from the states to cover
operating deficits for service they receive.  These are significant
assumptions about funding that will require congressional endorsement
as part of the process of defining options for Amtrak's future route
network. 

Second, Amtrak expressed concern about the timing of our
recommendation for a commission because the board's recent analysis
showed that reducing service frequency was more economical than
closing routes and because Amtrak is a commercially driven
corporation, which should not act like the Department of Defense in
making decisions about closures.  Amtrak also states that little more
can be done beyond the board's recommendation, short of closing down
the national system.  Amtrak's point on timing is well taken if the
railroad's requirements for capital and operating funds for the
national system, as presently constituted, are met.  If they are not,
cutbacks in service frequency and routes beyond those announced by
the board will be required.  In addition, much of the expected
savings comes from reductions in frequency on long-distance routes. 
Amtrak believes that revenue losses will be minimal because most
long-distance train riders are discretionary travelers who are not
time-sensitive and, therefore, will adjust their travel plans to
accommodate Amtrak's reduced service level.  Unfortunately, Amtrak
lacks experience on the effect on ridership of reducing the frequency
of long-distance service.  Amtrak's estimates of significant net
savings assume that ridership and revenue losses will be minimal. 
This is an important assumption.  With respect to Amtrak's view that
the economics of intercity rail passenger service preclude cutbacks
beyond those already planned unless the goal of providing a national
system is abandoned, we believe that Amtrak must first define what it
means by a national system.  If it means a cross-country network of
interconnected routes, then it may not be possible to support such a
system with the available funds.  However, there may be numerous
routes, either densely traveled corridors or segments of existing
long-distance routes, on which service could be continued with these
funds.  In addition, with appropriate congressional approval, Amtrak
might enter into partnerships with individual states or groups of
states to "reconnect" the federally supported parts of the system. 

As for Amtrak's reservations about having a commission offer options
to the Congress, we recognize that Amtrak could provide the Congress
with options for realigning and closing routes, and we explicitly
provide for that option in our recommendation.  We recognize that
Amtrak has superior knowledge of the economics of its operation, and
Amtrak's recent analytical efforts could provide the starting point
for considerations about restructuring the system.  However, we also
realize the difficulties inherent in deciding which states and
locations should receive service.  Because the commission would be
independent, it could help eliminate some of the problems normally
associated with reducing service in different areas of the nation. 
In addition, if the Congress chooses, such a commission could take
into account factors beyond revenues and costs, such as highway and
airport congestion relief, in deciding how to realign the route
network. 

Third, Amtrak believed that the role of the states in intercity rail
passenger service and the need for access to federal transportation
funds deserved emphasis in our report.  We revised the report to
reflect the fact that unlike aviation, highways, and mass transit,
intercity passenger rail has no trust fund and that the ability of
the states to use the trust funds that exist for other modes for rail
is extremely limited.  In 1991, the Intermodal Surface Transportation
Efficiency Act provided the states with some flexibility to use
highway dollars for mass transit, but it did not authorize the direct
use of such funds for intercity passenger rail service.  Although the
states' access to federal transportation trust funds is a key element
of Amtrak's plan to have the states cover significantly more of the
railroad's costs, it is a policy judgment for the Congress to make
whether and to what extent such flexibility should be extended to
intercity passenger rail.  Amtrak also wanted assurance that we
consider the federal funds for Amtrak's capital needs as an
investment and not a subsidy.  Our report clearly notes that capital
expenditures are an investment. 

Finally, Amtrak believes that if it is freed from certain legislative
restraints, it could operate more as a competitive commercial entity. 
Amtrak envisions changes to labor laws to give it greater latitude to
negotiate on such matters as severance pay, contracting, and work
processes.  Amtrak also wants (1) an exemption from requirements to
pay federal fuel taxes, (2) authority to issue tax-exempt debt, (3)
inclusion in a federal transportation trust fund, (4) relief from
requirements under the Railroad Retirement Act, and (5) limits on its
liability for punitive damages after accidents.  We added a
recommendation to our report that Amtrak provide its proposals to the
Congress, along with the estimated effect of each proposal on
Amtrak's finances and on other affected parties.  This information
will provide a vehicle for congressional deliberation on the merits
of each proposal and allow for consideration of opposing views. 

Amtrak also provided comments that clarified certain technical
information or statements made in a draft of this report.  We found
Amtrak's suggestions useful and incorporated these changes in the
report where appropriate.  Amtrak's written comments and legislative
proposals are presented in full in appendix IV. 


OBJECTIVES, SCOPE, AND METHODOLOGY
=========================================================== Appendix I

In March and April 1994, during House and Senate hearings to
authorize and appropriate Amtrak's federal grant for fiscal year
1995, we testified that Amtrak's financial and operating condition
had deteriorated and that the railroad's ability to provide an
acceptable level of service was seriously threatened.\1 Five
committees expressed concern about Amtrak's financial viability and
future need for federal funding:  (1) the Senate Committee on
Commerce, Science, and Transportation; (2) the Subcommittee on
Transportation and Related Agencies, Senate Committee on
Appropriations; (3) the Subcommittee on Transportation and Related
Agencies, House Committee on Appropriations; (4) the Subcommittee on
Transportation and Hazardous Materials, House Committee on Energy and
Commerce; and (5) the Subcommittee on Information, Justice,
Transportation, and Agriculture, House Committee on Government
Operations.  After the hearings, we were directed to continue to
assess (1) Amtrak's financial and operating conditions, (2) the
likelihood that Amtrak can overcome its financial and operational
problems, and (3) alternative actions that could be taken in deciding
on Amtrak's future mission and on commitments to fund the railroad. 

We conducted this review under basic legislative authority provided
by the Rail Passenger Service Act, which authorizes the Comptroller
General to conduct performance audits of Amtrak's activities and
financial transactions.  The review also responded to some of the
requirements in Senate Conference Report 103-150, supporting the
Department of Transportation's appropriations legislation for fiscal
year 1994, that we review (1) Amtrak's staffing, including a review
of salary, promotion, and job classification policies and procedures;
(2) the adequacy of Amtrak's budget resources, capital expenditures,
and operating expenses, including equipment overhauls and maintenance
expenses; (3) revenues from Amtrak's management of real estate
properties; and (4) ticketing policies and revenues from Amtrak's
yield management system. 


--------------------
\1 See GAO/T-RCED-94-155, Mar.  17, 1994, GAO/T-RCED-94-145, Mar. 
23, 1994, and GAO/T-RCED-94-186, Apr.  13, 1994. 


      AMTRAK'S DECLINING FINANCIAL
      AND OPERATING CONDITION
------------------------------------------------------- Appendix I:0.1

To assess Amtrak's declining financial and operating condition, we
interviewed officials and reviewed documents throughout Amtrak.  We
spoke to numerous officials in Amtrak's Finance and Administration
Department, including the Chief Financial Officer, Treasurer,
Director of Corporate Accounting, and Director of Resource
Management, as well as auditors from the two independent auditing
firms (Arthur Andersen & Co.  and Price, Waterhouse & Co.) that
Amtrak has employed over the past
5 years.  We analyzed Amtrak's annual financial reports, federal
grant submission to the Congress, capital acquisition plans, and
other financial documents such as annual budgets, revenue forecasts,
and minutes from meetings of the Board of Directors.  In addition, we
analyzed supporting data on Amtrak's projections of expenses and
revenues, including current debt levels and projections of the cost
of servicing debt, and procedures for managing cash.  We discussed
Amtrak's grant disbursement process with officials of the Federal
Railroad Administration (FRA) who are responsible for overseeing
Amtrak's drawdowns of federal funds.  To evaluate Amtrak's cost
accounting methodology, we also reviewed Amtrak's cost accounting
reports on route profitability and the internal guidance on preparing
those reports.  Finally, we reviewed the documents generated by the
independent auditors in rendering their opinion on the fairness of
Amtrak's financial statements. 

As part of our review of Amtrak's finances and to evaluate the
railroad's procedures for developing the annual request for federal
funds, we examined the models Amtrak uses to estimate its internal
budget and federal grant request.  In particular, we analyzed
Amtrak's aggregate demand model to determine its (1) structure, (2)
input variables and data, and (3) accuracy in estimating passenger
revenues.  However, we did not independently verify the accuracy of
Amtrak's data, nor did we test the equations used by Amtrak to
determine statistical or predictive accuracy.  We also obtained memos
on the revenue projections used in developing Amtrak's federal grant
request for fiscal years 1989-94.  Using these data, we analyzed the
differences between each year's forecast and the actual revenues
realized to determine the error due to yield alone and the error in
forecasts of passenger miles alone. 

We also reviewed the statistical and forecasting results of
alternative specifications of the demand forecasting model that
Amtrak developed as part of efforts to improve the model's
forecasting accuracy.  These alternative specifications (1) deleted
some variables from the original model and/or (2) added variables
representing consumer confidence, national unemployment levels, and
ticket class (first class vs.  coach). 

We obtained information on the cost of Amtrak's operations and the
condition of the railroad's assets from officials in Amtrak's
Engineering, Mechanical, Corporate Planning, and Passenger Marketing
departments.  We analyzed data on the current condition of Amtrak's
equipment, facilities, and rights-of-way, including the historical
and current costs of operating and maintaining these assets,
deferrals of maintenance and renovation projects, and requirements
for capital investment.  We observed the condition of Amtrak's
overhaul facilities at Beech Grove, Indiana, and Wilmington and Bear,
Delaware, and of repair and maintenance facilities in Chicago,
Illinois, New York, New York, and Washington, D.C., during tours
conducted by the officials responsible for the facilities' operation
and management.  We also obtained information on environmental
pollution at Amtrak-owned sites and the estimated costs of cleanup. 
We evaluated Amtrak's labor costs by interviewing officials in
Amtrak's Labor Relations and Personnel offices and in the 14 labor
unions that represent Amtrak's employees.  In addition, we
interviewed officials at seven major freight railroads to obtain
information on potential cost increases for Amtrak's use of these
railroads' track and for liability agreements (see
app.  II). 


      ABILITY TO OVERCOME
      FINANCIAL AND OPERATIONAL
      PROBLEMS
------------------------------------------------------- Appendix I:0.2

To assess the likelihood that Amtrak can overcome its financial and
operating problems, we evaluated the functions within Amtrak, in
addition to its intercity rail service, that appear to have the
potential for generating revenues.  This evaluation included
interviewing officials in Amtrak's Passenger Marketing and Sales
Department, Corporate Planning Department, and Northeast Corridor
High-Speed Rail Improvement Project.  We reviewed all of Amtrak's
contracts to provide commuter rail service and interviewed officials
in several of the commuter rail authorities to which Amtrak is
contracted.  We discussed the potential for real estate development
with officials of the Real Estate and Operations Department and
obtained information on the three Amtrak stations that have been or
are being developed (Union Station, Washington, D.C.; 30th Street
Station, Philadelphia, Pennsylvania; and Chicago Union Station,
Chicago, Illinois).  We also reviewed financial reports and business
plans concerning Amtrak's efforts to generate revenues through
right-of-way leases for telecommunications lines and mail and baggage
service. 

We reviewed the potential for increased revenues from high-speed rail
operations by evaluating information provided by Amtrak's Director of
Capital and Business Planning and Director of Route Planning and
Strategy and by talking with members of FRA's Railroad Development
staff.  We also reviewed data supplied by the Michigan Department of
Transportation on estimates of costs and revenues for a planned
high-speed rail link between Detroit, Michigan, and Chicago,
Illinois. 


      ALTERNATIVE ACTIONS FOR
      DECIDING ON FUTURE MISSION
------------------------------------------------------- Appendix I:0.3

To assess the alternative actions that could be considered in
deciding on Amtrak's future mission and on commitments for funding
the railroad, we analyzed data on passenger and route profitability
provided by Amtrak's Chief Financial Officer.  We also contracted
with Snavely, King & Associates, Inc., to estimate the amount of
federal operating grant that Amtrak might need in the year 2000 under
two scenarios:  (1) if it receives an annual capital subsidy of $500
million for fiscal years 1995 through 2000 and (2) if it receives an
annual capital subsidy of $195 million for those years.  For its
analysis, Snavely, King & Associates used Amtrak's actual expenses,
revenues, and federal subsidies for fiscal years 1991 through 1993;
Amtrak's estimate of how the capital subsidy would be used from 1995
through 2000; and Amtrak's estimate of the expected increases in
revenues, reductions in costs, and/or increases in expenses to repay
debt.  Snavely, King & Associates estimated Amtrak's operating grant
requirement (the difference between estimated operating revenues and
estimated expenses) for fiscal years 1995 through 2000 on the basis
of trends before 1994 and the following assumptions: 

average annual economic activity includes a 2-percent growth in real
gross domestic product,

unemployment is at least 5-6 percent nationwide,

disposable income rises by 2-3 percent per year,

inflation increases by 4 percent per year, and

increases in the consumer price index remain below 5 percent per
year. 

For the scenario in which Amtrak receives an annual capital subsidy
of $195 million, costs and revenues were estimated to increase by 4
percent per year.  At an annual capital subsidy of $500 million,
costs were estimated to increase by 4 percent per year and revenues
were made one percentage point higher than the compounded cost
inflation rate during each year to estimate the benefits achieved
from the higher level of service quality resulting from the higher
level of capital investment. 

As shown in table I.1, Amtrak's operating subsidy remains over $500
million through the year 2000 under the scenario of $195 million in
capital funding.  As table I.2 shows, the operating subsidy declines
from fiscal years 1995 through 2000 under the scenario of $500
million in capital funding. 



                                    Table I.1
                     
                         Detailed Estimates of Revenues,
                       Expenses, and Operating Subsidies If
                        Amtrak Receives an Annual Capital
                      Subsidy of $195 Million, Fiscal Years
                                    1995-2000

                              (Dollars in millions)

--------  --------  --------  --------  --------  --------  --------  ----------
Base      1,400.0   1,400.0   1,400.0   1,400.0   1,400.0   1,400.0   8,400.0
 revenues
Revenue     27.0     126.2     174.5     231.7     274.8     303.3    1,137.5
 increas
 es from
 project
 s
Revenue      0        61.0     128.5     203.7     284.5     369.0    1,046.7
 increas
 es from
 inflati
 on
Total     1,427.0   1,587.2   1,703.0   1,835.4   1,959.3   2,072.3   10,584.2
 operati
 ng
 revenue
 s
Base      1,885.7   1,885.7   1,885.7   1,885.7   1,885.7   1,885.7   11,314.2
 expense
Expense     13.3      54.3      80.2     117.1     142.2     165.6    572.7
 increas
 es from
 project
 s
Expense      0        77.6     160.4     250.1     344.5     444.4    1,277.0
 increas
 es from
 inflati
 on
Expense     60.9      72.2      95.5     106.0     102.9      99.9    537.4
 increas
 es from
 interes
 t
 payment
 s
================================================================================
Total     1,959.9   2,089.8   2,221.8   2,358.9   2,475.3   2,595.6   13,701.3
 expenses
Operatin  (532.9)   (502.6)   (518.8)   (523.5)   (516.0)   (523.3)   (3,117.1)
 g
 subsidy
 require
 ment
--------------------------------------------------------------------------------


                                    Table I.2
                     
                         Detailed Estimates of Revenues,
                       Expenses, and Operating Subsidies If
                        Amtrak Receives an Annual Capital
                      Subsidy of $500 Million, Fiscal Years
                                    1995-2000

                              (Dollars in millions)

--------  --------  --------  --------  --------  --------  --------  ----------
Base      1,400.0   1,400.0   1,400.0   1,400.0   1,400.0   1,400.0   8,400.0
 revenues
Revenue     27.0     167.2     229.0     312.7     379.0     425.1    1,540.0
 increas
 es from
 project
 s
Revenue      0        78.4     149.2     231.0     320.0     413.7    1,192.3
 increas
 es from
 inflati
 on
Total     1,427.0   1,645.6   1,778.2   1,943.7   2,099.0   2,238.8   11,132.3
 operati
 ng
 revenue
 s
Base      1,885.7   1,885.7   1,885.7   1,885.7   1,885.7   1,885.7   11,314.2
 expense
Expense     13.3      70.8     103.2     149.1     139.9     220.1    750.4
 increas
 es from
 project
 s
Expense      0        78.3     162.3     254.1     353.2     456.2    1,304.1
 increas
 es from
 inflati
 on
Expense     60.9      72.2      95.5      92.6      89.8      87.1    498.1
 increas
 es from
 interes
 t
 payment
 s
================================================================================
Total     1,959.9   2,107.0   2,246.7   2,381.5   2,522.6   2,649.1   13,866.8
 expenses
Operatin  (532.9)   (461.4)   (468.5)   (437.8)   (423.6)   (410.3)   (2,734.5)
 g
 subsidy
 require
 ment
--------------------------------------------------------------------------------

ORGANIZATIONS CONTACTED BY GAO
========================================================== Appendix II


      AMTRAK
------------------------------------------------------ Appendix II:0.1

Engineering Department
Executive Offices
Finance and Administration Department
Government and Public Affairs Department
Labor Relations Department
Law Department
Marketing and Business Development Department
Mechanical Department
Passenger Services Department
Transportation Department


      FREIGHT RAILROADS
------------------------------------------------------ Appendix II:0.2

Burlington Northern
Consolidated Rail Corporation (Conrail)
CSX Transportation
Illinois Central Railroad
Norfolk Southern Corporation
Southern Pacific Railroad
Union Pacific Railroad


      UNIONS
------------------------------------------------------ Appendix II:0.3

AFL-CIO Transportation Trades Department
American Federation of Railroad Police
American Train Dispatchers Association
Amtrak Service Workers Council
Brotherhood of Locomotive Engineers
Brotherhood of Maintenance of Way Employees
Brotherhood of Railroad Signalmen
International Brotherhood of Firemen and Oilers
Joint Council of Carmen, Coach Cleaners, and Helpers
Railway Labor Executives Association
Sheet Metal Workers International Association
Transport Workers Union
Transportation Communications International Union
United Transportation Union


      FEDERAL AGENCIES
------------------------------------------------------ Appendix II:0.4

Federal Railroad Administration
Federal Transit Administration
Interstate Commerce Commission
U.S.  Redevelopment Corporation


      ASSOCIATIONS
------------------------------------------------------ Appendix II:0.5

Association of American Railroads


      COMMUTER RAIL AGENCIES
------------------------------------------------------ Appendix II:0.6

Maryland Department of Transportation
Massachusetts Bay Area Transit Authority
Metrolink
Virginia Railway Express


      AUDITING FIRMS
------------------------------------------------------ Appendix II:0.7

Arthur Andersen & Co.
Price, Waterhouse & Co. 


HOW AMTRAK DEVELOPS ITS REQUEST
FOR ITS FEDERAL OPERATING GRANT
========================================================= Appendix III

To develop its request for the federal operating grant, Amtrak
estimates both revenues and operating costs for the upcoming year. 
The estimate of revenues includes revenues from both passenger
service and other activities, such as commuter rail and real estate. 
Revenues from passenger trains constitute about two-thirds of
Amtrak's revenues.  Amtrak estimates its other revenues and operating
costs through its annual budget preparation process.  The amount
Amtrak requests for its operating grant is determined by subtracting
projections of cash operating expenses from estimates of total
revenues.\1 Figure III.1 compares Amtrak's initial requests with the
actual funds received in fiscal years 1990-95. 

   Figure III.1:  Comparison of
   Amtrak's Operating Grant
   Request With Actual
   Appropriations, Fiscal Years
   1990-95

   (See figure in printed
   edition.)

Note:  The figure for fiscal year 1993 includes a supplemental
request by Amtrak of $27.5 million and a supplemental appropriation
of $20 million. 

Source:  Amtrak. 


--------------------
\1 Capital depreciation, which is included in Amtrak's operating
expenses and operating deficit, is a noncash expense and therefore is
not included in calculations to determine Amtrak's operating grant. 


   ESTIMATES OF REVENUES FROM
   PASSENGER RAIL OPERATIONS
------------------------------------------------------- Appendix III:1

To forecast passenger revenues, Amtrak uses an econometric model to
estimate the number of passenger miles throughout the system
(excluding commuter rail operations).\2 The estimated passenger miles
are multiplied by Amtrak's projected yield (ticket revenues per
passenger mile) to forecast revenues for the passenger trains.  The
results, after adjustment according to the informed judgments of
Amtrak's analysts (and henceforth referred to here as initial model
results), help the Congress determine the amount of federal subsidy
that Amtrak needs.  Table III.1 outlines the model's principal
elements.\3



                                   Table III.1
                     
                          Amtrak's Model for Forecasting
                               Systemwide Ridership

Purpose                    Structure                  Variables
-------------------------  -------------------------  --------------------------
To prepare short-and       Multivariate, linear       Dependent variable is
long-range forecasts of    regression model using     passenger miles.
systemwide passenger       quarterly, seasonally
miles (and revenues) for   adjusted time series       Independent variables are
                           data                       disposable income (in
 federal grant request,                              constant dollars); retail
                           Ordinary least squares     gasoline price (in
 internal budgets,        estimation using a         constant dollars);
                           procedure to correct for   Amtrak's yield (in
 long-range plans for     serial correlation         constant dollars); ratio
equipment, and                                        of Amtrak's yield to
                           Estimated regression       airlines' yield; dummy
 5-year business plan.    elasticities calculated    variables to reflect
                           from estimated regression  events such as holidays,
                           coefficients               weather, strikes, and
                                                      derailments; dummy
                                                      variables to reflect
                                                      quarterly seasons; no
                                                      constant term.
--------------------------------------------------------------------------------
Note:  Data for the model's independent variables are obtained from
Data Resources, Inc., Amtrak's own internal analyses, and other
sources. 

Source:  Amtrak. 

The most important determinant of rail passenger miles is the
strength of the national economy, which is represented in the model
by disposable income (in constant dollars).  A 1-percent increase in
disposable income is expected to yield a 1.8-percent increase in
intercity rail passenger miles. 

Each time during the year that a new forecast is generated, the model
is recalibrated (that is, the regression coefficients are
statistically reestimated).  The model's specification--the variables
and functional form of the equation--has not changed for the past
several years except through deletion or addition of the variable for
the retail price of gasoline. 


--------------------
\2 Amtrak did not use the model to estimate its fiscal year 1996
grant request. 

\3 Amtrak operates two other models to forecast ridership.  One model
forecasts the annual number of passengers and passenger miles between
pairs of cities outside the Northeast Corridor that are served by
Amtrak.  This model estimates changes in ridership and revenues on
the basis of changes in service, such as schedule, frequency, arrival
and departure times, and total travel time.  A second model is used
mainly for forecasting ridership demand in the Northeast Corridor. 


      AMTRAK'S PROCESS FOR
      DEVELOPING THE ESTIMATE OF
      PASSENGER REVENUES
----------------------------------------------------- Appendix III:1.1

Amtrak's forecast of passenger revenues for the federal grant request
starts in the Market Planning and Forecasting Group (Market Planning)
in October as part of Amtrak's overall budget preparation for the
next fiscal year.  In October and November, Market Planning produces
an initial model forecast.  Market Planning analysts then use their
judgment to adjust the forecast up or down for such factors as
expected new services, expected delays in putting new passenger cars
in service, and anticipated work stoppages.  Market Planning forwards
these adjusted forecasts to the Resource Management Department, which
makes additional refinements.  Senior Amtrak officials then either
approve these estimates for the budget or make further changes. 
Amtrak's Board of Directors then approves the overall budget, which
becomes the basis for the grant request. 

Since 1991, Amtrak's initial model computations and subsequent
revisions by management have increasingly overestimated actual
revenues.  (See table III.2.) From fiscal year 1990 to fiscal year
1994, the model overestimated actual revenues by a total of $467
million, and the grant request--which includes subsequent adjustments
by management--overestimated revenues by $561 million.  These
overestimates were significantly larger in fiscal year 1994 than they
were in the previous years. 



                                   Table III.2
                     
                       Amtrak's Actual Passenger Revenues,
                        Initial Model Estimate, and Later
                                 Revenue Estimate

                              (Dollars in millions)


                      Actual    Estimated   Difference    Estimated   Difference
Fiscal year         revenues     revenues  from actual     revenues  from actual
--------------  ------------  -----------  -----------  -----------  -----------
1990                    $930         $889         -$41         $892         -$38
1991                    $964       $1,005         +$41       $1,064        +$100
1992                    $929       $1,035        +$106       $1,035        +$106
1993                    $943       $1,080        +$137       $1,092        +$149
1994                    $880       $1,104        +$224       $1,124        +$244
================================================================================
Total                 $4,646       $5,113        +$467       $5,207        +$561
--------------------------------------------------------------------------------
\a Model estimate for federal grant request, as adjusted by Amtrak's
Market Planning and Forecasting Group. 

Source:  GAO's analysis of Amtrak's data. 


      AMTRAK'S YIELD FORECAST
----------------------------------------------------- Appendix III:1.2

While Amtrak forecasts passenger miles through statistical modeling,
it estimates average yield judgmentally, on the basis of economic and
competitive factors.  Amtrak considers factors such as airline fares
and increased airline competition in determining expected yield. 
Over the 1990-94 period, Amtrak estimated that its average yield
would increase at the rate of inflation as measured by the Consumer
Price Index.\4

In recent years, Amtrak has substantially overestimated yield.  Real
yield fell each year over the 1990-94 period; from 1992-1994, even
nominal yield fell.  Amtrak's real yield fell at an annual average
rate of 3.2 percent over the 5-year span from 1990 to 1994. 


--------------------
\4 Amtrak forecasts the increase in yield, not from the yield
forecast of the previous year's initial model, but rather from the
yield estimated in a later internal budget forecast.  In recent
years, these later yield estimates have typically been adjusted
downward. 


      ANALYSIS OF FORECAST ERRORS
----------------------------------------------------- Appendix III:1.3

Amtrak's initial model forecast of passenger revenues totaled $467
million more than actual revenues from fiscal years 1990 to 1994.  We
found that this difference resulted predominantly from Amtrak's
overestimation of yield and, to a lesser extent, from the model's
overestimation of passenger miles.  To analyze errors in the model's
estimates of passenger revenues, we used the initial forecasts--the
model results as initially adjusted judgmentally by Amtrak's Market
Planning analysts.\5

Table III.3 compares Amtrak's forecasts of passenger miles and
average yields with corresponding actual values over the 1990-94
period.  Amtrak underestimated both passenger miles and average yield
in fiscal year 1990.  Since then Amtrak has usually overestimated
both passenger miles and average yield.  From 1992 through 1994, both
passenger miles and average yield were substantially overestimated. 
Over the entire 5-year period, the average forecast errors (using
absolute values) were 4.9 percent for passenger miles and 6.7 percent
for average yield.\6



                                   Table III.3
                     
                     Comparisons of Amtrak's Actual Passenger
                     Miles and Average Yield With Forecasts,
                               Fiscal Years 1990-94

                      (Passenger miles in millions and yield
                          in dollars per passenger mile)


                               Percent
                      minu    forecast                                   Percent
                         s     differs                    Forecast      forecast
Fiscal                actu        from                       minus  differs from
year                    al      actual                      actual        actual
------  ------  ----  ----  ----------  --------  ----  ----------  ------------
1990    6,057   5,98   -77      -1.3 %  $0.1535   $0.1    -$0.0049        -3.2 %
                   0                               486

1991    6,273   6,26    -9      -0.1 %  $0.1537   $0.1    +$0.0067        +4.4 %
                   4                               604

1992    6,091   6,31  +221      +3.6 %  $0.1525   $0.1    +$0.0115        +7.5 %
                   2                               640

1993    6,199   6,59  +392      +6.3 %  $0.1521   $0.1    +$0.0118        +7.8 %
                   1                               639

1994    5,921   6,70  +781     +13.2 %  $0.1486   $0.1    +$0.0161       +10.8 %
                   2                               647

================================================================================
Total   30,541  31,8  +1,3    +4.9\b %     \c       \c          \c       +6.7\b%
                  49    08
--------------------------------------------------------------------------------
\a Yield is ticket revenues divided by passenger miles. 

\b Average percentage error:  the sum of the absolute values of the
percentage errors divided by 5 years (1990-94). 

\c Not applicable. 

Source:  GAO's analysis of Amtrak's data. 

Table III.4 compares actual and forecast passenger revenues over the
same period.  The combined effect of the forecast errors for
passenger miles and yield results in even larger errors in
revenues.\7 In fiscal year 1994, Amtrak overestimated passenger
revenues by 25.5 percent, for a 5-year average (using absolute
values) of 12 percent. 



                                   Table III.4
                     
                     Errors in Forecast Passenger Revenues as
                       a Result of Forecast Average Yield,
                      Forecast Passenger Miles, and Residual
                           Error, Fiscal Years 1990-94

                         (Dollars and passenger miles in
                                    millions)


                                                        Revenue
                                       Revenue         estimate    Error
                     Total            estimate   due      using   due to  Residu
Fisc               forecas  Percen       using    to   forecast  passeng      al
al                       t       t    forecast  yiel  passenger       er  Error\
year               error\a   error     yield\b   d\a    miles\c  miles\a       d
----  -----  ----  -------  ------  ----------  ----  ---------  -------  ------
1990  $930   $889     -$41  -4.4 %        $900     -       $918     -$12     +$1
                                                 $30

1991  $964   $1,0     +$41  +4.3 %      $1,006  +$42       $963      -$1     -$0
               05

1992  $929   $1,0    +$106   +11.4        $999  +$70       $963     +$34     +$2
               35                %

1993  $943   $1,0    +$137   +14.5      $1,016  +$73     $1,002     +$59     +$5
               80                %

1994  $880   $1,1    +$224   +25.5        $975  +$95       $996    +$116    +$13
               04                %

================================================================================
Tota  $4,64  $5,1    +$467  +12.0\      $4,896  +$25     $4,842    +$196    +$21
l       6      13              e %                 0
--------------------------------------------------------------------------------
\a Forecast passenger revenues minus actual passenger revenues. 

\b Forecast yield multiplied by actual passenger miles. 

\c Forecast passenger miles multiplied by actual yield. 

\d The "residual" errors in forecast revenues after accounting for
errors resulting from forecast yield and forecast passenger miles. 
These "residuals"--the product of the errors in average yield and
passenger miles--were small and of the expected signs. 

\e Average percentage error using absolute values. 

Source:  GAO's analysis of Amtrak's data. 

Table III.4 shows that, from 1990, the largest source of error in the
forecast of passenger revenues each year was generally the error in
estimating average yield.\8 For example, in fiscal year 1990, the
error in the average yield forecast accounted for $30 million of the
total underestimate of $41 million in passenger revenues.  From 1991
to 1993, the error in average yield accounted for $185 million of the
total 3-year overestimate of $284 million. 

The forecast errors in passenger revenues accounted for by the
underestimates in passenger miles were relatively small for 1990 and
1991.  Since then, however, passenger miles were increasingly
overestimated and accounted for a substantial portion of the
overestimates in passenger revenues.  In 1994, both the total
forecast error and that portion accounted for by the error in
forecasting passenger miles were very large.  We believe, however,
that this result was likely an aberration since a sizable portion of
the reductions in revenues and passenger miles was due to an
uncharacteristically large number of accidents and physical
disasters.\9

Amtrak's tendency to be overly optimistic about factors affecting
passenger rail demand may explain, at least partially, why it has
overestimated both average yields and passenger miles in recent
years.  For example, Amtrak does not estimate the effects of events
that can potentially depress revenues, such as accidents and bad
weather, in its initial model forecast.  While such events cannot be
specifically forecast, the expected total effect of such events each
year likely can be.  In particular, Amtrak has consistently assumed
that nothing with a negative effect on revenues would happen each
year; but negative events, in fact, occurred regularly. 

Similarly, over the past 5 years, Amtrak has consistently been too
optimistic in its forecasts of the revenue-enhancing effects of its
new services.  The assumptions about revenues used in formulating the
federal operating grant request over fiscal years 1991-94 assumed
that several new trains would begin operations each year.  In fact,
only a few new trains were initiated over the entire period. 
Therefore, revenues were correspondingly lower than forecast.  For
example, in 1992 only two of five expected additional trains were
added, leading to a shortfall in revenues of $6.9 million.  Amtrak
staff have stated that the unexpectedly slow recovery from the last
recession was an additional cause of their recent overestimates. 


--------------------
\5 Teams using econometric models routinely employ varying degrees of
judgment to produce their forecasts.  Therefore, we evaluated
Amtrak's model forecast after it had been adjusted by technical
analysts. 

\6 A consultant to Amtrak reviewed this model's results for the
period 1983-91.  The consultant stated that the model's estimated
forecast error of 4.3 percent for passenger miles seemed reasonable
for simple forecasting models of this type. 

\7 The yield estimate also indirectly affects the revenue forecast
because it is included in the model in two explanatory variables--
Amtrak's average yield and the average yield ratio for Amtrak and the
airlines.  The revenue errors resulting from an inaccurate yield
forecast would be even larger except for a partially offsetting
indirect effect.  For example, because Amtrak's yield has been
overestimated in recent years, the forecast of passenger miles is
correspondingly lower, leading to a lower forecast of passenger
revenues than would result otherwise.  However, the direction of the
indirect effect of the Amtrak/airline yield ratio is unclear, except
that its effect is likely small. 

\8 We "decomposed" Amtrak's revenue forecast errors into three
components:  (1) the error in forecast yield, (2) the error in
forecast passenger miles, and (3) the "residual" error resulting from
errors in average yield and passenger miles combined.  In this simple
decomposition procedure, we account for the direct but not the
indirect effect of yield, which may be partially offsetting.

By definition, passenger revenues equals yield times passenger miles. 
Therefore, the error in Amtrak's passenger revenues forecast (i.e.,
forecast passenger revenues minus actual passenger revenues) can be
divided algebraically into these three components as follows:

PRA = YA * PMAwhere,PR = passenger revenues
PRF = YF * PMFY = yield
PM = passenger miles
A = actual
F = forecast

let,

diff(Y) = YF - YA
diff(PM) = PMF - PMA

by definition,

PRF = [YA + diff(Y)] * [PMA + diff(PM)]

 = [YA * PMA] + [PMA * diff(Y)] + [YA * diff(PM)]
  + [diff(Y) * diff(PM)]

Therefore, forecast passenger revenues differ from actual passenger
revenues by the second, third, and fourth terms of the last equation. 
The first term in this equation equals actual passenger revenues. 
The second term is the difference in the revenue estimate from actual
when forecast yield (from table II.4) is used, because when actual
passenger miles are multiplied by forecast yield, the remaining terms
all equal zero.  Similarly, the third term is the difference in the
revenue estimate from actual when forecast passenger miles (from
table II.4) is used, because when actual yield is multiplied by
forecast passenger miles, the remaining terms all equal zero. 
Accordingly, the "residual" error in table II.4--the fourth term of
the equation--equals the total error in the passenger revenues
forecast minus the second and third terms of the equation. 

\9 Amtrak has estimated that in 1994, the negative revenue impacts of
derailments resulted in about $70 million in lost revenues; the
impacts of ice storms and the earthquake, about $11 million in lost
revenues.  These occurrences directly reduced passenger miles but
affected average yield much less since both revenues and passenger
miles were reduced when trains could not run. 


      AMTRAK'S PREVIOUS MODEL
      RESEARCH AND SUGGESTED
      FUTURE IMPROVEMENTS
----------------------------------------------------- Appendix III:1.4

Market Planning officials stated that they have tried to modify and
improve the model on a regular basis.  They have used different or
additional explanatory variables, including (1) a lagged dependent
variable (passenger miles), (2) time, (3) the unemployment rate, (4)
the employment rate, (5) on-time performance, (6) the Consumer
Confidence Index, and (7) seat miles and train miles.  They have also
deleted from the model the average yield ratio for Amtrak and the
airlines.  Other experimentation has included exponential smoothing
and separating the model into Northeast Corridor and non-Northeast
Corridor, and into first class and coach. 

Amtrak found none of these alternative specifications to be
consistently superior to the model's current specification either on
statistical grounds or in forecasting accuracy.  For example,
Amtrak's results using the Consumer Confidence Index and unemployment
variables yielded slightly better statistical results but worse
forecast results (greater overforecasting in fiscal year 1992).  When
the model was decomposed into first class and coach, the accuracy of
the forecast was better in fiscal year 1992 but worse in fiscal years
1990 and 1991. 

Amtrak analysts believe that a key variable missing from the model is
their customers' perceived quality of service.  These analysts
believe that a decline in perceived service quality and safety has
increasingly discouraged potential ridership.  Amtrak is conducting
customer surveys to establish baseline information and plans to track
customer satisfaction monthly.  In addition, Amtrak analysts believe
that the ridership model might be better specified in a multiequation
form to better reflect the interdependencies between independent
variables.  According to Amtrak officials, they have not performed
such research because of budget limitations. 


   ESTIMATES OF REVENUES FOR OTHER
   LINES OF BUSINESS AND COST
   ESTIMATES
------------------------------------------------------- Appendix III:2

Amtrak obtains other estimates for its federal grant
request--revenues from other lines of business and expenses--through
its internal budgeting process.  In October, Amtrak uses its
current-year budget as a basis for projecting the next fiscal year's
grant request.  Next, Amtrak adjusts the budgeted expenses and
additional revenues on the basis of the projected rate of inflation
and any planned changes in operations.  To make operating adjustments
to its base budget, Amtrak requests departmental heads to provide
revised estimates on the basis of anticipated changes, such as a
salary freeze, an increase in employee benefits, a reduction in train
service, or anticipated new business. 

After these adjustments, the estimate of expenses for the grant
request is compared with the projected revenues (passenger and other
revenues) and Amtrak's anticipated federal operating subsidy.  If the
projected operating deficit exceeds the anticipated federal operating
subsidy, Amtrak again revises its expense estimates.  After Amtrak
balances its projected operating deficit with its anticipated federal
operating subsidy, it submits its grant request.  Amtrak will submit
its fiscal year 1996 grant request by February 1, 1995. 

The estimate of expenses submitted in the grant request is based on
the estimated operating budget and not on actual prior-year results. 
As Amtrak receives its revenues and incurs expenses throughout the
fiscal year, its current fiscal year budget is constantly updated to
reflect actual results.  Furthermore, budgeted and estimated expenses
are driven by the revenues because Amtrak constantly adjusts its
expenses to meet shortfalls in revenues so that its operating deficit
equals its federal operating grant.  These continual adjustments
during years of declining revenues have resulted in Amtrak's
estimated expenses being higher than actual expenses.  (See table
III.5.)



                         Table III.5
           
              Comparison of Estimated and Actual
                Expenses, Fiscal Years 1990-94

                    (Dollars in millions)

                        Estimated expenses         Estimated
                              submitted in          expenses
Fiscal                       federal grant     compared with
year         expenses              request   actual expenses
----------  ----------  ------------------  ----------------
1990          $2,012                $1,934              -$78
1991          $2,081                $2,116               $35
1992          $2,036                $2,122               $86
1993          $2,134                $2,215               $81
1994          $2,400                $2,243             -$157
------------------------------------------------------------
Source:  GAO's analysis of Amtrak's data. 


   SUMMARY
------------------------------------------------------- Appendix III:3

There are three main reasons why Amtrak has inaccurately estimated
its passenger revenues over the past 5 years.  The primary cause of
the inaccurate estimates is the assumption that yield would track the
Consumer Price Index.  Yield affects the passenger revenues forecast
both directly and indirectly.  The direct, and most important, effect
occurs because yield is multiplied by passenger miles to obtain the
forecast of passenger revenues.  However, the indirect effect, in
which yield affects the passenger miles forecast, may at least
partially offset revenues errors as a result of the direct effect. 

Second, the overestimates of passenger revenues submitted to the
Congress in recent years have been made larger by the upward
adjustments to the initial revenues estimates by Amtrak's senior
management. 

Third, Market Planning makes no allowance in its revenue forecasts
for events that may depress revenues; similarly, the effects of
revenue-enhancing events have been overestimated.  This
overoptimistic approach has contributed to the revenue overestimates
in the last several years. 

Finally, the overestimates of passenger revenues affect Amtrak's
estimate of expenses for the federal operating grant request.  The
estimated expenses for the upcoming year are adjusted to equal the
projected revenues plus the anticipated federal operating subsidy. 
Since actual revenues have been lower than forecast, estimated
expenses have been higher than actual expenses. 




(See figure in printed edition.)Appendix IV
AGENCY COMMENTS
========================================================= Appendix III

Note:  GAO comments on Amtrak's letter are found in chapter
5. 



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix V


   RESOURCES, COMMUNITY, AND
   ECONOMIC DEVELOPMENT DIVISION,
   WASHINGTON, D.C. 
--------------------------------------------------------- Appendix V:1

Ronnie E.  Wood, Assistant Director
Francis P.  Mulvey, Assistant Director
Teresa F.  Spisak, Project Manager
Sharon E.  Dyer
Lynne Goldfarb
Richard A.  Jorgenson
Deborah L.  Justice
Edmond E.  Menoche
Joseph J.  Warren
Daniel G.  Williams


   ACCOUNTING AND INFORMATION
   MANAGEMENT DIVISION,
   WASHINGTON, D.C. 
--------------------------------------------------------- Appendix V:2

Gregory D.  Kutz
Donald R.  Neff
Glenn A.  Thomas


   OFFICE OF THE GENERAL COUNSEL
--------------------------------------------------------- Appendix V:3

Michael R.  Volpe

