Road versus rail
- Road versus rail: Facts sheet No 6: Examples of Cost escalation for railway projects in the UK.
Wp ref. Foebaab/email/06costs1. At privatisation in 1994 British Rail's forecast for track maintenance beyond 2001 was £774 million (source, page 11 of the prospectus issued in advance of sale). At 1999 prices that amounts to £0.888 billion. In contrast the network management statement for the years 1996/7 and 2000 provide an annual average for the 10 years 1995/6 to 2005/6 at 1999 prices of £2 billion per year. The 2001 network statement provides an average of nearly £3 billion per year at the 2001/02 price base for the five years 2001/2 to 2005/6. Thereafter the cost is forecast to taper off to £2.2 billion in 2010/11. The increase from £0.888 bn to nearly 3 billion, a factor of over 3 is typical of the appalling record the rail industry has for estimating its costs.2. The West Coast Main Line Modernisation Programme was to cost £2.35 billion in 1997, £2.95 billion in March 1999, £4.75 billion in October 1999, £5.56 billion in January 2000 and £5.8 billion at the start of the Public Inquiry in February 2001. The price rose to £6.3 billion during the inquiry when there were press reports that it would cost £9 billion. Now the cost is £13 billion (Source for the earlier numbers is the Overview Paper produced in May 2000 and a report to the Rail Regulator by the aptly named Booz-Allen and Hamilton dated June 2000).3. The recent reports of the cost of the proposed Train Protection System (I do not have its exact name to hand). There I believe the original cost was set at £1 billion but it is now £6 billion - over 40 times the value of the life and limb likely to be saved - see future fact sheet.My view is that these massive failures to make reasonable estimates of the actual costs may be deliberate. They are mirrored by equally massive misrepresentations to do with capacity, safety and other issues, see other facts sheets.Certainly the original cost estimates (let alone all the other misrepresentations) will have misled the Government and shareholders alike into commitments which would never have been entertained had the facts been available.Meanwhile the overall cost of Railtrack's nation wide Modernisation programme has risen from £50 billion through £60 billion to a projected £73 billion - Sufficient to build the residential accommodation for a city of 1.5 million people.Paul Withrington August 2002
- Road versus rail: Facts sheet No 7: Track costs Wp ref Foebaab/email/07WCMLM1This E-mail compares the costs of upgrading the West Coast Main Line with the cost of rebuilding the M1 from scratch.1 During the Public Inquriy intothe WCMLM it was clear thet he priject then had the raqnge 6.3 billion to £9 billion for modernisation plus essential renewal. At the Simon Maple for Railtrack said that the expenditure was concentrated on the core 1000 km of track. Hence the cost per track-km is between £6.3 million and £9 million. Now the cost is said to be £13 billion so the cost per track-km is £13 million.2 In contrast the Independent of 17th February 1999 reported a Treasury study which estimated the replacement cost of the M1 as £2.1 billion for all works and land. The lane length, assuming 3 lanes all the way from the M25 to Leeds, is 1800 km. Hence the cost per lane-km is £1.12 million - 12 times less than the present projected cost of the WCMLM. Of course the rail cost is inflated by the difficulty of working on a live railway but then the cost excludes the land and all the earthworks and much else.3 However the product of building a road or a railway is not the track but the passenger and freight flow. Hence I converted the daily train paths provided by Railtrack for the WCML to equivalent coach and lorry flows by dividing (a) 8 or (b) 12 to obtain a range of equivalent bus flows and the freight trains by 30 to generate equivalent lorries. The sum divided by 4 (because its a four track railway) provided the equivalent daily coach plus lorry flow per track. The answers were (1,600-2,100) north of Lichfield and (2,340-3,478) south of Milton Keynes.4 In contrast the all vehicle flow averaged over 50 census points spread along the length of the M1 was 15,000 per lane per day. However most of those are cars and many of the lorries are relatively small. So I converted the number to equivalent big lorries and buses by (a) multiplying the car flows by an occupancy of 1.65 and dividing by 20 (b) halving the small lorry flows (c) adding the results to the big lorries and buses. That produced an equivalent big vehicle daily flow of 3,000 per lane. The road had the advantage of offering car passengers door to door journeys whereas rail presupposes a change of mode and other journeys to and from rail stations. Additionaly the road had the potential to carry perhaps 10,000 buses plus lorries per lane per day, far in excess of rail and of any demand.5 Against that background it is fair to say the WCMLM is at least 12 times as expensive as the M1 in terms of the product.
Paul Withrington August 2002
- Road versus rail: Facts sheet No 7: Track costs We ref Forebear/email/07WCMLM1This I-mail provides compares the costs of upgrading the West Coast Main Line with the cost of rebuilding the M1 from scratch.1 During the Public Inquiry into the WCMLM the project then had the cost range 6.3 billion to £9 billion for modernisation plus essential renewal. Simon Maple for Railtrack said that the expenditure was concentrated on the core 1000 km of track. Hence the cost per track-km is between £6.3 million and £9 million. Now the cost is said to be £13 billion so the cost per track-km is £13 million.2 In contrast the Independent of 17th February 1999 reported a Treasury study which estimated the replacement cost of the M1 as £2.1 billion for all works and land. The lane length, assuming 3 lanes all the way from the M25 to Leeds, is 1800 km. Hence the cost per lane-km is £1.12 million - 12 times less than the present projected cost of the WCMLM. Of course the rail cost is inflated by the difficulty of working on a live railway but then the cost excludes the land and all the earthworks and much else.3 However the product of building a road or a railway is not the track but the passenger and freight flow. Hence I converted the daily train paths provided by Railtrack for the WCML to equivalent coach and lorry flows by dividing (a) 8 or (b) 12 to obtain a range of equivalent bus flows and the freight trains by 30 to generate equivalent lorries. The sum divided by 4 (because its a four track railway) provided the equivalent daily coach plus lorry flow per track. The answers were (1,600-2,100) north of Lichfield and (2,340-3,478) south of Milton Keynes.4 In contrast the all vehicle flow averaged over 50 census points spread along the length of the M1 was 15,000 per lane per day. However most of those are cars and many of the lorries are relatively small. So I converted the number to equivalent big lorries and buses by (a) multiplying the car flows by an occupancy of 1.65 and dividing by 20 (b) halving the small lorry flows (c) adding the results to the big lorries and buses. That produced an equivalent big vehicle daily flow of 3,000 per lane. The road had the advantage of offering car passengers door to door journeys whereas rail presupposes a change of mode and other journeys to and from rail stations. Additionally the road had the potential to carry perhaps 10,000 buses plus lorries per lane per day, far in excess of rail and of any demand.5 Against that background it is fair to say the WCMLM is at least 12 times as expensive as the M1 in terms of the product.
Paul Withrington August 2002
- Road versus rail: Facts sheet No 8: Track maintenance costs
Wp ref. Foebaab/email/08maint1 This note shows that track maintenance for rail transport costs between 5 and 10 times as much as for road transport.2. Tables 5.11 and 5.12 of Transport statistics Great Britain 2001 editions provides 39.2 billion passenger-km and 18.1 billion tonne-km. Dividing the passengers by 20 to convert to buses and the tonnes by 15 to convert to lorries and adding the two yields 3.17 billion (bus plus lorry) vehicle-km. The Network Management statement now puts the maitenance bill at £3 billion per year. Hence the cost per bus/lorry vehicle-km is 95 pence. Alternatively divide the £3 billion by the 32,000 km of track length and get £94,000 per track-km per year.3 For roads all maintenance costs are here assigned to the heavier class of lorries, namely 4 axle rigids, and all artics. That is because nearly all the damage to roads is by those classes of vehicle. The corresponding flow across the Motorway and Trunk Road network in the whole of Great Britain amounted to 11.5 billion in 1998. The corresponding road maintenance (current and capital accounts) was 1.14 billion. Hence the cost per heavy lorry vehicle-km was 10 pence, slightly more than one tenth the value of 95 pence by rail. Alternatively divide the £1.14 billion by the lane length estimated at 55,000 km and get £21,000 compared with £94,000 for rail.4 The rail maintenance cost of £3 billion is the average for the next 5 years and may be high, due to neglect. That cost is forecast to fall to £2 billion by 2010. Similarly the road maintenance cost may be about to rise if the newspapers are to be believed. If the latter is increased by 30% and if the rail cost is reduced by one third then the ratio of rail to road costs falls to 5.5. However, the calculation is biased favour of rail since (a) buses are included in the equivalent rail flows but not on the roads so depressing the unit cost for rail (b) substantial elements of road maintenance such as signs, lighting, verges, hard shoulders, winter maintenance, and some wear and tear should be attributed to other vehicles on the roads. That has inflated the costs attributed to lorries .6 Detail, slightly superseded, is in Appendix 5 to main evidence presented to the Public Inquiry into the West Coast Main Line Modernisation Programme. None of it was challenged.Paul Withrington May 4th 2002
- Road versus rail: Facts sheet No 9: Passenger Rolling Stock costs
Wp ref. Foebaab/email/09rstock1 This note shows that the annual cost of passenger rail rolling stock is 3 to 4 times that of the equivalent motor coach.2. The annual costs of capital is here defined as that sum which if paid annually will repay capital plus interest, set at the Treasury Discount rate of 6%, over the lifetime of the vehicle.3. A railway carriage costs between £1 million and £2 million. It lasts perhaps 30 years. The capital recovery factor is 0.0726. Hence the annual cost of the capital has the range £72,600 to £145,200.4. A bus or coach has the price range £150,000 to £250,000. The vehicle may last 15 years. The capital recovery factor is then 0.1030. Hence the annual cost of the capital for the bus/coach has the range £15,450 to £25,750. If the bus is 70% the size of a railway carriage then, for comparative purposes, these numbers become £22,100 to £36,800.5. It follows that the annual capital cost of a bus/coach is 3 to 4 times less than equivalent floor space by rail.6. Probably I have underestimated the cost of rail rolling stock, e.g. by omitting the cost of the traction unit on intercity, but some may say I have overestimated the size of a bus. Motorway coaches have 50 somewhat cramped seats. If one row were removed leaving 46 seats the cramp would be eliminated except for the tallest or fattest since knee room would be increased by 3 inches. Some double deck buses such as the Olympian three axle bus may have 90 seats or perhaps 80 comfortable ones. A railway carriage may have 80 seats. However. I would welcome any more exact seat counts or dimensions or estimates of usable floor area and costs.Paul Withrington May 8th 2002 postsed to World 17h Ocotber 2002
- Road versus rail: Facts sheet No 10: Rail safety costs, value for money?
Wp ref. Foebaab/email/10sigThis note provides data which suggests that for every 3 lives (plus the associated lesser casualties) saved on the railways about 150 fatalities (plus their associated lesser casualties) may have been sacrificed on the roads or elsewhere.1 The latest reported cost of the proposed safety systems for national rail is £6 billion. If the system lasts 30 years then, with the interest rate set to 6%, the annual cost of the £6 billion is £435 million per year. That annual costs is net of maintenance and will save train accidents only.2. Over the 18 years ending 1999 an average of 6 passengers per year were killed in train accidents. My data suggests that that should perhaps be increased by 15% to allow for staff. If half would be saved by the signalling system we are left with 3.5 fatalities per year saved.3. Facts sheet 04 suggests that fatalities in train accidents account for only 22% of casualty costs in train accidents, the rest being attributable to serious and slight injuries. The value of a life is set at £922,874, at the 2000 price base, by the government. Hence the value of the casualties saved per year will be approximately: 3.5 x 922,874 / 0.22 = £14.7 million.4. It follows that the annual cost of the system, £435 million, will exceed the annual value of the casualties saved, £14.7 million, by a factor of at least 30; an amazing misuse of public or private money if ever there was one.5. If the recent newspaper reports, suggesting a saving of some 80 fatalities over 40 years, are to be believed then, instead of the 6 fatalities per year saved suggest above, the system may save only about 2. In that case the annual cash value of the saving is £8.4 million, 41 times less than the annual cost of the capital.6. That implies a great loss of life and limb elsewhere, such as in our hospitals or on the roads. E.g. the casualty cost per fatality on the roads is £2.8 million at 2000 prices (note this is not the cost of a fatality; instead it is the cost per fatality of the fatality plus all the associated lesser casualty types). Hence to justify an annual expenditure of £435 million on the roads there would have to be 155 lives saved (plus the associated lesser injuries).7 Meanwhile, if the railways were converted to roads, limited to buses, coaches and lorries, the flows would seldom justify more than advance direction signs - go see any motorway or rural trunk road. Additionally the casualty cost of previous rail passengers per passenger-km would be reduced by a factor of 3.
Paul Withrington As at August 2002
- This is the last of it
Facts sheet 13 - Summary
The previous 13 Facts sheets show that, very much against political
sentiment and public expectation, there is an overwhelming case for
converting the UK's national rail network to a network of motor roads
managed to avoid congestion. That is to say the rubber tyred option would
offer 3 to 4 times the capacity at one quarter the cost of rail while using
30 to 40% less energy and reducing casualty costs suffered by rail
passengers by a factor of 3.
The problem with the case is that (a) it is indeed very much against
expectation (b) the numbers are so overwhelming as to inspire disbelief
rather than belief (c) few people have ever seen a motor road managed to
avoid congestion - The UK road network is a collection of access roads never
designed for motor traffic. (d) rail is so romantic.
Certainly it took me months to come to terms with the data back in the late
1960s, so I sympathise with other's disbelief.
In support of the propositions, all the statements were tested at the Public
Inquiry into the West Coast Main Line Modernisation Programme. There, all
Railtrack's immensely expensive Inquiry Team could do in the face of the
facts was to quack and Moo. Any person who doubts that may have copies of
my and Railtracks's Closing Inquiry Statements.
Here are a selection of the main points and a list of the facts sheets.
Capacity and use
1. Rail has one third to one quarter the capacity to move people
compared with motor roads managed to avoid congestion - go look at Waterloo
2. National Rail carries an average flow per track equivalent to only
300 buses plus lorries per day. It is difficult to find a minor road
anywhere in the country so lightly loaded in terms of vehicles.
3. The density of use achieved by the National Rail system is one third
to one fifth that obtained from the Motorway and Trunk road Network.
4. The energy consumption of passenger rail per passenger-km in 1990
was 2 to 3 times as high as required by express buses capable of doing the
5. In contrast rail uses half the fuel per Tonne-km required by road
but only if the drag in and out is ignored. When the latter are considered
freight by road and rail appear to have similar energy consumptions.
6. If the National Rail function were carried out by buses and lorries
on motor roads managed to avoid congestion there would be an energy saving
Journey lengths, speed and fares
7. 50% of passenger rail journeys are less than 25 miles long, the same
as by express bus. 90% are less that 80 miles long. For all those journeys
the express coach would match the train for journey time particularly after
taking account of a service frequency up to 12 times that of the train.
8. Fares by express coach are generally a fraction of those by train
despite the coach paying taxes and making a profit. If rail were to operate
without subsidy fares would have to double at least without loss of
9. The death rate to passengers per passenger-km by rail is
historically higher than the corresponding rate by buses and coaches on
10. The casualty costs attributable to deaths in Train accidents amounts
to about 4% of all casualty costs to passengers on rail.
11. The casualty cost suffered by passengers per passenger-km on rail is
3 times as high as suffered by passengers in buses and coaches on non-urban
12. System wide, including trespassers on rail and pedestrians etc. on
roads, the death rate by rail is 0.36 per billion passenger-km compared with
0.46 for all roads. Of the deaths on the roads nearly half were
pedestrians' cyclists and motor bikers. If those are removed from the sum
on the basis that rail enjoys a largely segregated system then the road
environment would offer the lower death rate.
13. Sir Robert Horton said in Railtrack's 1998/9 annual report that rail
was "27 times as safe as roads in terms of fatalities and serious
injuries.........". Sir Robert forgot to tell us that the roads carry 17
times as many passenger-km as do the railways. Hence, form the start, Sir
Robert was exaggerating the relative safety of rail by a factor of 17. Sir
Robert also forgot to say that over 40% of road deaths are to pedestrians,
cyclists and motor bikers, classes of people singularly absent from railway
rights of way.
14. The £6 billion to be spent on rail safety systems is equivalent to
£435 million per year for 30 years. The corresponding value of the life and
limb likely to be saved is less than 15 billion per year, or some 30 times
less than the annual capital cost - illustrating what must surely be the
grossest misapplication of resources on offer.
Widths and headroom
15. A two track railway typically offers room for a UK standard 7.3
metre carriageway with one metre marginal strips but no other verges.
16. On the approaches to towns and cities there is often room for a dual
two or three lane highway.
17. Where there is overhead electrification headroom would often be
adequate for a triple decker
18. The annual capital cost of rail passenger rolling stock is 3 times
as high as is equivalent floor space for express bus
19. Track maintenance costs per tonne/passenger-km for rail are at least
10 times higher than for equivalent road transport.
20. The cost per track-km of the West Coast Main Line Modernisation
programme is 10 times higher than the cost per lane-km of building the M1
from scratch including the cost of land.
21. The rail Modernisation Programme is to cost £70 billion. Its target
is to increase passengers by 50%, e.g. from 6% to 9% of current
passenger-km, and to increase rail freight from 11% to 17% of tonne-km.
That will have a negligible effect on car traffic and a small effect on road
freight. Hence the £70 billion will have been almost entirely wasted -
equivalent to burning the residential accommodation for a city of 1.5
22. In contrast, replacing the railway lines by a road surface managed
to avoid congestion would cost at most £12 billion. The effect would be to
offer faster journey times for all but the longest journeys at fares a
fraction of those charged by rail.
1. Flow, density of use and capacity.
2. Safety, definitions and unit costs.
3. Death rates comparisons
Extra 03 - proportion of road casualty costs attributable to pedestrians,
cyclists and motor cyclists.
4. Casualty cost comparisons
5. Fuel consumptions
6. Cost escalation for railway projects in the UK.
7. Track costs
8. Track maintenance costs
9. Passenger Rolling Stock costs
10. Rail safety costs, value for money?
11. Journey Length distributions/Speed and fares
12. Costs of conversion and rates of return
Railways Myth and Maths: A paper read to the British Association for the
Advancement of Science and to a fringe meeting of Friends of the Earth and:
Nearly endless Inquiry evidenced.