Upload
wilbur
View
35
Download
0
Tags:
Embed Size (px)
DESCRIPTION
THE ARLANDA AIRPORT RAIL LINK – LESSONS LEARNED FROM A SWEDISH PPP CONSTRUCTION PROJECT. Lars Hultkrantz, Urban Karlström, Jan-Eric Nilsson. www.vti.se. VTI presentation sv 0503 / sid2. Uppsala. A – the four-track section B – the Arlanda extension, including under-ground stations - PowerPoint PPT Presentation
Citation preview
VTI presentation sv 0503 / sid2 www.vti.se
Lars Hultkrantz, Urban Karlström, Jan-Eric Nilsson
THE ARLANDA AIRPORT RAIL LINK
– LESSONS LEARNED FROM A SWEDISH PPP CONSTRUCTION PROJECT
The Project
C
B
A
Stockholm C
Uppsala
Arlanda Airport
A – the four-track sectionB – the Arlanda extension, including under-ground stationsC – the north-bend extension
The process towards project start
• First proposal in 1987; initiated political process leading to…
• …competitive procurement in 1993-94 with– prequalification (more than 30 companies)– final bidding (four consortia)
• The incumbent’s consortium came second• The contract was awarded to Arlanda Link Consortium
– NCC, Siab och Vattenfall– Mowlem and GEC Alsthom
Contractual structure
• Life-cycle contract, i.e. the entrepreneur builds and maintains infrastructure and operates services.
• 45 + 10 year contract.• (Much of) investment costs are supposed to be
paid for by ticket revenue.• The government could cancel the contract after
15 years (i.e. 2010).
Was this a good project design
• … financially?
• … economically?
And is there something to learn from the perspective of industry organisation?
Construction costs, million SEK
Section Ex ante(1992)
Ex post(1999)
A 1 900 2 400
B 2 600 2 700
C 850 850
Rolling stock 600 850
Comment on costs
• No price level adjustment due to low inflation/ downturn in business cycle.
• Information about costs for building section C is not available– the amount paid to the consortium.
• The ex ante and ex post projects were different, in particular with respect to station and tunnel design.
Costs and funding
• Total costs for section B SEK 2700 million– ”conditional loan” from government SEK 1000
million– SEK 1100 million in commercial loans– SEK 600 million risk capital
• In addition, the consortium got SEK 850 mkr to build section C.
Costs; conclusions
The private consortium accepted the full cost risk;It designed the investment in order to balance
investment costs against future maintenance costs.
Has this enhanced cost efficiency?The government
- could ”save” SEK 1,7 billion- did not have to face any cost overruns- and the project was on time!
Revenues
• Investment costs are to be paid by ticket revenue.• The Consortium carries the full revenue risk.• It faces harsh competition from busses.• Revenues decided by
– Price
– Number of passengers.
• No. of passengers strongly depends on no. of airline passengers.
Table 2: Million Arlanda airline passengers. 1990 forecast and outcome.
Forecast no of travellers Actual no. of travellerswithout train with train
1988 10,8
1998 16,1
1999 17,1
2000 20,2 21,6 18,3
2001 18,1
2002 16,4
2003 15,1
2020 31,5 33,5
Market share (%) for different modes of transport to and from Arlanda. Based on
surveys made by Luftfartsverket.1999 2001 2003
A-Train - 19 19
Other train - 4 5
Coach (Stockholm) 24 14 13
Coach (Uppsala) - 2 2
Other coach - 4 4
Taxi 23 22 21
Car 35 35 35
Other, no answer 18 3 4
No. of passengers with A-Train; actual numbers and 1993/94 forecast*.
Passengers Employees Total
2000 1 700 000 400 000 2 100 000
2001 2 500 000 400 000 2 900 000
2002 2 400 000 350 000 2 750 000
2003 2 200 000 350 000 2 550 000
2004 2 500 000 365 000 2 865 000
2005 5 100 000*
2020 7 400 000*
Poor revenues due to…
• External changes affecting the whole air transport market.– Is it efficient to transfer the full revenue risk to the
concessionaire?
• The pricing policy:– A priori expectation a price a par with busses, but– Bus SEK 90 (40 minutes);– A-Train SEK 190 (20 minutes).
A-Train’s poor financial result
• For year 2004 (2003) – revenues SEK 402 (359) million– operating costs SEK 314 (310) million– net financial costs SEK 155 (100) million.
• The balance – SEK 68 (51) – added to the company’s debt.
Economic aspects
• More congestion and higher environmental costs than if p=mc for using tracks.– Much traffic on the road that could/should be lured over to
use train.
– Other train operators are charged for using tracks and station.
– No inter ticketing, meaning that it is costly to combine standard railway tickets with the A-Train fare.
– No coordination with commuter train services, in spite of that this is a different market segment
But…
• What if lower fares would improve the financial result?
• What if A-Train could negotiate a deal with SL over commuter services?
The search for an efficient industry structure
• Europe; vertical separation, subsidies, competition on or for track access.
• America; vertical integration, competition between (some) class A freight operators; limited and subsidised passenger services.
• Is the Arlanda solution, an island of integration in a sea of vertical separation, viable?
This would require…
• A large and captive (no outside opportunities) market in order to pay for costs.
• No interest from “outside traffic” in using the facilities.
• Malmbanan fits better in on these criteria,
• … but Arlanda could possibly be fixed?
To conclude
• Transparency problems!
• Private money and improved (?) cost efficiency may be paid for by reduced allocative efficiency.
• How could PPP deals be designed in order to safeguard overall efficiency improvements?