Personal Rapid Transit in Stockholm: Market Demand and Economic
This research project deals with the problem of introducing a potential
demonstration-track in the form of a PRT system somewhere in the Stockholm region. The aim
of the study is to find an answer to the following main questions:
- Which is the best site for such a test-track of a PRT system for Stockholm?
- Which is the most probable demand for such a PRT system, and how much traffic would
be diverted from the private car and other modes of transport and how much would be newly
benefits, system costs and overall cost-benefit ratio?
In order to answer these three highly important and interesting questions, the study is
divided into the following major parts - a PRT Market Demand Analysis; and a PRT Economic
The research results and findings were documented in a research report (in Swedish,
130 pages) by the end of 1998. This paper describes the contents of the project, the
methods chosen for the analyses, and the results and research findings.
1. Personal Rapid Transit (PRT) - individual trips in public vehicles
Personal Rapid Transit (PRT) offers individual trips in public vehicles - a
competitive alternative to the most popular mode of urban transport - the private
automobile. PRT is developed to offer some of the advantages of the private auto:
+ It departs on demand without any timetable.
+ It runs the quickest path without any stop and without any transfer.
+ It offers a private trip alone or together with passengers of your own choice.
At the same time one would like to avoid some of the major disadvantages of the
- Noise and exhausts.
- Congestion and accidents.
- Parking demand.
PRT is a system of small, automated vehicles on their own guideway that is
demand-responsive and offers a direct trip to the destination without any stop en route.
The PRT solution with many small vehicles can be derived from many different
· The trip maker should not wait for the vehicle to come - the vehicle should wait
for the passenger· If one does not force several passengers to travel together, there
is no need for large vehicles. The load will resemble that of a taxicab.
· The track should not be larger or more expensive than what is needed. The track cost
increases with the weight of the vehicle. One has to distribute the weight. A car with 4
passengers every second gives the same capacity as a traditional train every 15-minute
with 1,800 seats.
· The stations should be short. This will be possible with a constant turnover of
vehicles and travelers, i.e. dense departures with small vehicles.
· If the vehicle is driven automatically, the only reason for large-scale vehicles
falls short. The passenger service governs the traffic performance, not the driver cost.
Small vehicles demands a very dense traffic, which means that vehicles are not allowed to
stop for boarding and alighting on the main track from capacity reasons. Also, unnecessary
stops for service reasons should be avoided. Therefore, all PRT systems are designed with
stations located on sidetracks.
· Short time slots between vehicles do not allow switches on the track; this would be
too time-consuming. Instead the vehicle chooses its route through fixed switches.
· Acceleration and deceleration does not allow standing passengers. Therefore the
system is designed to carry seated passengers only. Guaranteed seating capacity also
contributes to the attractiveness of the system. Wheel-chair passengers are foreseen to be
able to travel in all PRT-vehicles.
· A PRT ride without a stop between origin and destination station is not only
comfortable and convenient. The energy consumption is less than one fourth of that of an
2. The long-term evolution of auto and transit traffic
Transek Consultants was commissioned by the Office of Regional Planning and Urban
Transportation at the Stockholm County Council to investigate the long-term evolution of
auto and transit traffic (see Ref. 1, 2, 3 and 4). The estimated auto traffic production
in terms of vehicle-kilometers has increased by 88 % between 1970 and 1995 or by 2,5 %
annually. The corresponding transit ridership (estimated through ticket sales records) is
estimated to have increased by 18 % or by 0,7 % annually between 1973 and 1997. The
imbalance of modal development - both in the retrospective and in the forthcoming period
is shown in Figure 1.
Our conclusion from this observation is that the present type of transit systems (bus,
metro and commuter rail) is insufficient in its performance to attract new travelers to
cope with the self-service system of the automobile. There is a strong need for a
high-quality performance transit system - such as PRT - if the urban transportation
problems of too low efficiency, too high accident rates and environmental air pollution
should be curbed.
3. PRT in Stockholm - an efficient and sustainable transport system
The purpose of this chapter is to illustrate the area-wide potential of a
high-level-of-service transit system in terms of generalized travel times and market
shares - in comparison to the more traditional transit modes, such as bus, commuter rail
and subway. A second purpose of this exercise is to form a basis for the selection of the
best site for a PRT demonstration track in the Stockholm Region. Therefore, a PRT trip
demand analysis has been carried out for the entire Stockholm County Area (population;
1,775,000 inhabitants in 1998), with the simplified assumption that a PRT-station would be
(theoretically) available in every traffic zone (1,043 zones) and running on the present
major road links in the network. The existing transit modes are assumed to prevail. The
demand procedure is summarised in Figure 2.
This formed a basis for considerations of the best suitable location for a PRT
The major changes in the generalized travel times that could be achieved by the PRT
system, are mostly a dramatically reduction in the waiting and transfer times, compared to
the present day modes of mass transit.
As the PRT system operates as an automated and a demand responsive system, the time
spent waiting for the vehicles, does not differ at all between peak and off-peak time
periods; this being the opposite for today's' manually driven fixed line service. Thus,
the major travel time gains with PRT will occur during the off-peak period. The weighted
generalized time (see Footnote 1) is calculated to be reduced from almost one hour (55
minutes) in the base scenario to a little more than a half-hour in the PRT scenario (see Figure 3).
If an area-wide PRT system would be introduced in all Stockholm region, a
substantial modal shift from the auto mode (-4 % units) would occur; also a slight shift
from the walk and bike modes towards the transit modes, including the new areawide PRT
system. The transit modal split is estimated to augment from 46 to 52 % by the new PRT
system, i.e. a 13 % growth in market share (see Figure 4).
The number of auto trips is calculated to be reduced by 9 % in the peak period, with
its dramatic and positive impacts in terms of reduced congestion, air pollution and road
traffic accidents. Transit trips - including the new PRT mode - is forecast to expand by
almost one third (31%) during all day, and by 41 % in the off-peak period (see Figure 5).
4. The demand for PRT-trips in the Akalla - Kista area
The choice of the most suitable location for a potential PRT demonstration track is
based on at least six various criteria:
- Areas (in fact origin-destination pairs) with a generalized time elasticity with
respect to the demand for transit trips(numerically) above -2.0 and a minimum number of
transit trips· Areas with a travel time relationship between the transit and auto mode
of three or more and a minimum number of transit trips
- Areas with an even distribution of peak and off-peak trips and a minimum number of
- Areas with a high traffic load and a minimum number of transit trips
- Areas with a high load of estimated PRT trips per track-kilometer
- Robust areas with a combination of high densities in the number of (1) occupied
residents per square kilometer, (2) work-places per square kilometer, (3) household income
potential per square kilometer and (4) privately owned autos per square kilometer.
Maybe, the most important criteria above all, are the support from local authorities.
By coincidence, most of the areas selected according to the above mentioned six criteria,
are also preferred locations by the local municipalities:
· Handen Center
· Karolinska Institute & Hospital-Solna-Sundbyberg
· Sigtuna - Arlanda - Märsta
· Skärholmen-Kungens kurva-Huddinge C-Huddinge Hospital
· Södertälje Centre
A corridor from the cities of Sundbyberg - Solna - Karolinska and the northwestern part of
the inner city have been excluded due to political and visual intrusion points of view.
The major results for the studied PRT network alternatives are shown below (see map of
|Track length, km
|Number of stations
|Vehicle fleet size
|PRT trips per day
|Daily trips per track km
|Average trip length (km) in peak
|Average trip time (min.) in peak
The results indicate that the number of daily trips per track-kilometer increases as
the network size augment from 9 to 28 kilometers. This is an indicator of the cost-benefit
ratio, as the number of trips is associated with user benefits, and track size with its
5. A Stated Preference Study on PRT comfort and convenience
A Stated Preference survey was carried out with the aim to investigate the willingness
to pay for PRT comfort and convenience factors, such as:
- In vehicle travel time with PRT
- PRT headway
- In-vehicle travel time with bus
In all 162 persons were interviewed in the Barkarby - Kista area in the northwestern
suburbs of Stockholm, of which 50 % were auto drivers and 50 % transit users.
· The result for the onboard travel time as well as for the trip frequency (or
headway) showed no significant deviation in the travel time component value for a PRT trip
compared to a bus trip.
· To have manned stations - instead of unmanned stations - has a very high value, 0.50
US$ per trip, reflecting the insecurity of today's mostly unmanned metro and rail stations
· Travelling 5 meter above the surface with a PRT vehicle, is shown to have a slight
negative value of -7 cents per trip.
Besides, the following types of attitudinal questions also revealed some
- On the question: "I am uninterested in PRT, as it has a negative visual
intrusion (makes the city look more ugly)", only 25 % agreed. Therefore, visual
intrusion does not seem to be a major drawback for PRT.
- On the question: "I am uninterested in travelling by PRT if I have to share my trip
with other passengers in peak hours", only 13 % seem to think this might be any
problem. More than two thirds of the respondents denied this would be a problem.
- Of all respondents, about half of them felt insecure travelling in a driver-less
vehicle, of which 15 % had a very strong expression against it; while 30 % declared this
was no problem. This shows there is a need for more information to the customers of this
new kind of driver-less transit service (which is not in operation anywhere in Sweden so
far) Professor Elsa Rosenblad's focus group interviews in Gothenburg show that this fears
for automation disappears after a proper information about it (Ref. 5).
- On the question: " I feel unsafe travelling 5 meters above the ground", only
20 % confirmed this negative statement. As many as 60 % expressed their view, this was no
problem to them. As the average monetary value was slightly negative, we conclude that
there is a minority with a very strong negative feeling for going elevated (there is no
such transit system in Sweden except for ski lifts).
- The last question was "If a PRT system would be built between Barkarby and Kista,
how often could you imagine to go with it"? Almost 65 % or two-thirds could imagine
going by PRT regularly or sometimes and only 16 % answered 'seldom' and just 3 % said
'never'. These positive results are well in accordance with the research findings from
Professor Elsa Rosenblad's study in Gothenburg (Ref. 5).
6. A Cost-benefit Analysis of a PRT network in the Akalla-Kista Area
Several cost-benefit analyses have been carried out for the five various PRT-networks
(described in section 4 above). Our findings reveal that the best cost recovery is
obtained for the largest PRT network, i.e. the Akalla-Husby-Kista-Helenlund-Sollentuna
Investment cost data were obtained from Raytheon's PRT2000, and from two conceptual
Swedish systems - Swedetrack's FlyWay (a suspended PRT system) and SkyCab (a supported
system). A high (0,24 US$) and a low (0,17 US$) operating cost per passenger-kilometer is
also associated with the US PRT 2000 and the two Swedish conceptual systems, respectively.
The analysis is carried out over the calculated economic lifetime of the PRT project,
60 years. In our recommended cost-benefit analysis procedure, we consider higher values of
time, comfort, safety and environmental impacts over the total time span for the project.
This is related to the assumed average long-term economic growth rate of 1-2 % annually
(GNP or household disposable income per capita). A present value and related annuity
benefits and costs are then calculated.
As a consequence of these assumptions, the first year's benefits from the PRT project
will increase over time due to the fact that the travelers will evaluate the benefits at a
higher value each year, as prosperity grows in the future years to come. As a sensitivity
analysis we have also calculated the benefits without an adjustment of the behavioral
values over time (not presented in this paper).
The table below shows that a PRT demonstration network in the presented Akalla - Husby
- Kista -Helenelund - Sollentuna area of Stockholm would be economically viable and well
justified in the low cost alternative. The cost-benefit ratio is calculated to be 1,5,
which means that one dollar spent on PRT in this area yields one dollar and 50 cents in
total benefits. Even the more expensive Raytheon PRT 2000 system would yield 70 cents per
spent US dollar at its full-calculated price.
With a 25 % reduction (covering engineering, construction, management, administration,
start-up and testing , see Footnote 2), also the PRT 2000 system would balance benefits
and costs (benefit-cost ratio equals 1,0).
Summary result: Benefit - Cost Analysis of PRT in
Annual Costs, MSEK (see Footnote 3)
||PRT 2000 (less 25% overhead)
|Capitalized Investment costs
|Annual Operating costs
|Cost of public capital; shadow price
|VAT tax burden
|TOTAL ANNUAL COSTS
|Transit travel time gains, incl. PRT
|Ticket revenues, incl. less public capital
|Traffic safety gains from less auto trips
|PRT comfort & convenience gains
|Less congestion due to less auto traffic
|Health and Environmental gains
|TOTAL ANNUAL BENEFITS
|NET BENEFITS (Benefits - Costs)
Another Swedish conceptual system - the SkyCab PRT system - has also been examined.
The SkyCab system has an investment cost of 894 MSEK and a total annual cost of 175 MSEK.
With the same assumed performance and benefit level of 339 MSEK as the other systems in
the table above, the SkyCab system would reach a benefit-cost ratio of 1.9, which means a
yield of 1.90 US$ per 1 US$ spend on the system - a highly viable system.
Also Professor J. Edward Anderson's TAXI 2000
concept has been studied on a coarse level. His estimates are based on our largest system
with 28 km of guideway, 25 stations and 275 vehicles, and it shows a cost level (USm$
118.2 or 887 MSEK) i.e. very similar to SkyCab system costs. The TAXI 2000 vehicle has a
seating capacity of three, which might lead to a need for a somewhat larger vehicle fleet.
Anyhow, the TAXI 2000 system seems most likely to be a highly viable PRT system for the
Stockholm PRT demonstration-track.
Altogether, we have studied four different systems (TAXI 2000, SkyCab, Flyway and PRT
2000 (including two cost variants), and all four systems have proved to be
well-justified and economically viable from a social benefit-cost perspective.
A PRT System in the Akalla - Kista area of Stockholm would yield a wide range of positive
and desired impacts:
· Travel time and comfort and convenience gains for PRT users
· A modal shift from auto to transit (including PRT) modes of transport
· Traffic safety gains· Eased congestion from less auto traffic
· Health and environmental gains
From the analysis, one could estimate the maximum investment cost per system-kilometer
for a PRT network of the relevant size to be about 115 MSEK/km (corresponding to 15
million US$ per track-kilometer). The desired minimum peak load should amount to at least
500 passengers per peak hour and track-kilometer.
From our areawide PRT demand study (section 3 above), we have indicators of the
cost-benefit ratio for 14 potential areas within the Stockholm region. As a rough estimate
we have used the number of daily trips per track-kilometer. Bearing in mind, that this is
just a crude indicator of economic viability, one could however conclude that there might
be at least six potential areas with an even higher possible return in terms of social net
benefits over costs. These areas are, in order of their cost-benefit ratios:
- Odenplan - Karolinska Institue & Hospital - Solna
- Bergshamra - University of Stockholm - Odenplan
- Solna Center - Sundbyberg
- Solna Center - Bergshamra
- Barkarby - Akalla· Södertälje C
1 Tegnér, G. ; Loncar-Lucassi, V., (1997) "Demand for Road Use, Accidents and
their Gravity in Stockholm: Measurement and analysis of the Dennis Package". Transek
AB, Stockholm, 18 p.
2 Tegnér, G.; Loncar-Lucassi, V. (1997) "Time-series Models for Urban Road
Traffic and Accidents in Stockholm". Paper presented at the Aalborg University,
3 Tegnér, Göran B., Loncar-Lucassi, Vesna M. (1997): "An Analysis of Urban Road
Traffic Safety in Stockholm - The use of aggregate time-series models with the TRIO
programme". Proceedings, the European Transport Forum Annual Meeting,
1-5 Sept, 1997, London.
4 Tegnér, G; Loncar-Lucassi, v; Nilsson, C., TRANSEK Consultants; and Holmberg, I.,
Gothenburg University, Department of Statistics, School of Economics and Commercial law,
(1998) "The Demand for Public Transport Trips in Stockholm County - a two-stage
aggregate, non-linear time-series model. Paper presented at for Economics and Institutions
of Transport, May 25th - 27th, 1998, Borlänge, Sweden.
5 Rosenblad, E., (1997) "Spårtaxiresenärer - brukares möte med ny
teknik", (in Swedish), CTH Rapport 1997:10, Chalmers Tekniska Högskola, Göteborg,
6 Tegnér, G., (1997) "Market Demand and Social Benefits of a PRT System: A Model
Evaluation for the City of Umeå, Sweden". Infrastructure, Vol. 2, No. 3,
pp. 27-32, 1997, John Wiley & Sons, Inc.
The weights are 2 for the walk, wait and transfer travel time and 1 for the in-vehicle
travel time (see Reference 6).
These figures are based on the SeaTac study: Personal Rapid Transit (PRT) Feasibility
Project-Executive Summary and Technical Appendices, City of SeaTac, August 1997)
One million Swedish Crowns roughly corresponds to 125,000 US$ (exchange rate 1 SEK =
0,13 US $)
One million Swedish Crowns roughly corresponds to 125,000 US$ (exchange rate 1 SEK =
0,13 US $)
Contact Information: Göran Tegnér , Business
Manager Transportation Studies, TRANSEK Consultant Company, Solna, Sweden. Phone:
+46-8-735 20 10; fax: +46-8-735 20 30
Last modified: August 17, 2002