REIGNITING THE RAILWAY CONVERSION DEBATE

Updating and expanding the paper in the Journal of the Institute of Economic Affairs

Volume 24 No 2 June 2004

Abstract

The economic functions of railways could be carried out by express coaches and lorries at one quarter the cost of the train, using 20-25% less fuel, requiring one quarter to one third of the land and imposing a casualty cost on passengers half that suffered by rail passengers. The railway conversion debate was initiated in the 1950s by the late Brigadier Lloyd and carried forward by the Railway Conversion League, subsequently renamed the Railway Conversion Campaign, until the death of its chairman, Angus Dalgleish, in 1994. The purpose of this paper is to re-ignite that debate. The government should remove all impediments to the conversion of railways to roads.  

The paper updates and expands its predecessor published in the Journal (Volume 24 June 2004).

Subsidies to rail and the value of roads

The subsidy to rail over the period 1975 to 1995 amounted to £40bn at June 2004 prices – an annual average of £1.9bn.[1] That may be an underestimate since it is extraordinarily difficult to find the actual yearly expenditure.  E.g. Gabriel Roth found that the outturn subsidy for the 16 years ending 1995 was £28bn [2].  That compares with £17bn from our source [3], which inexplicably omits Public Service Obligation Grants totalling 11bn).  Roth comments that when he was seeking such numbers the then Ministry of Transport “could not give any figure for the total of post-war rail subsidies. The issue must have been considered of so little importance that Ministry staff were not required to even quantify it”. 

Whatever the case the SRA confirmed in September 2003 that the National Rail’s wish list for capital expenditure consists of the committed Channel Tunnel Rail Link, Cross rail, Modernisation and a quarter of the East Coast High Speed line, together costing £100bn.  If that were to be repaid over 30 years at the Treasury Discount rate of 3.5% the annual bill would be 5.4bn.  If commitments were entered into and the cost escalated in line with the Treasury’s optimism bias factors the bill may rise by 40% to £7.56bn annually.  There would then be operating subsidy of £2bn per year providing a total of £9.56bn.  If, on the other hand, capital expenditure amounts to as little as £50bn for the decade then that plus operating subsidy leads to an annual bill for £4.7bn. Hence it is fair to say that the subsidy will run at between £5bn and £10bn per year for the next 30 years if not forever. 

That annual cost is equivalent to between £156,000 and £312,000 per track-km or to 12.4 and 25 pence per passenger-km or to taking taxes of between £200 and £400 per year from every household in the land forever at a time when most of them hardly ever use a train. (E.g. less than 1 in 50 motorised journeys are by rail[4] and half the population use a train less than once a year - a fact that the SRA perversely use to illustrate how importance rail is[5]).

In Contrast the annual income to the exchequer provided by taxes on road vehicles less expenditure on the roads is in excess of £35bn.  That may be given a context by expressing it as income per passenger-km (5 pence) or per vehicle-km (9 pence) or per lane-km.  The motorway and trunk road network has a lane length in the range 52,000 to 60,000 km and carries 37% of vehicle-km.  Hence the yield per lane-km for that network has the range £216,000 to £250,000 compared with a loss of between. £156,000 and £312,000) per track-km by rail.

Road and Track Costs

The Treasury has estimated the replacement cost of the M1, including land, as £2.1bn at 1999 prices.[6] The lane length, assuming 3 lanes in each direction, is 1,800 km. Hence the cost per lane km is £1.17 million. The West Coast Main Line Modernisation Programme (WCMLM) concentrates expenditure on the core 1,000 km of track (from a total of some 2,600 km).[7] The cost has risen to £10 bn.  Hence the cost per track-km is £10 million – over 8 times the £1.17 million per lane-km for rebuilding the M1 from scratch. However, the product of the expenditures should perhaps be measured in terms of equivalent vehicle flow.  If one makes reasonable assumptions about the capacity of the rail and road alternatives, in terms of flow per lane or track-km, the WCMLM is 12 times as expensive as the M1 built from scratch.

Maintenance of rail is also more expensive. Railtrack's Network Management Statement for 2001 provides an estimate of £3 bn per year at 2000 prices for track maintenance. Even on the most generous assumptions, the cost per vehicle-km by road was one ninth of that by rail.

Signalling costs

Significant amounts of money are being invested in rail safety mechanisms. Stopping the SPAD (Signal Passed at Danger) problem was reported to have a cost as high as £6bn.  The latest figure for the European Train Management System (ERTMS) is £3.6bn.[8]  If that is to be repaid over 30 years and if the interest rate is set at the old Treasury Discount Rate of 3.5% then the annual cost (without optimisation bias is £190 million.  If the annual maintenance of the system is to cost as little as 5% of the £3.6bn that will add £180 million to the annual bill, leading to a total of £370 million per year.  Railtrack’s Network Management Statements for 2000[9] suggest that these systems will save two lives per year.  A death is valued at some £1.2 million at 2001 prices.  However, deaths account for only about 22% of casualty costs in train accidents.  Hence the actual value of life and limb saved may be close to £11 million per year, or less than 3% of the annual capital plus maintenance cost.  In comparison motor roads (a) seldom have any signals despite carrying flows many times higher than achieved by rail (b) have lower system-wide casualty costs.

(The SRA claims that the ERTMS will allow higher capacity and hence use.  It is then claimed that the additional passengers will be from those previously travelling by road so saving perhaps 10 lives per year.  However, these assumptions seem over generous. Firstly, the hoped for growth seems unlikely; secondly, there is no evidence that the additional rail travellers would otherwise be travelling by car; thirdly, it is unrealistic to assign all the growth in rail traffic to the ERTMS; and fourthly, the casualty cost associated with the Killed and Seriously Injured category by rail is double that suffered on the motorway network (see below)).

Rolling stock

The capital cost of a railway carriage is in the range £(1-1.5) million. It may have 75 seats. If capital and interest at are to be repaid over 30 years at 3.5% and if annual maintenance amounts to 7.5% of the capital then the annual cost has the range £1,720 to £2,580 per year per seat. In contrast a bus or coach costs £120,000 to £250,000. It may offer 50 seats and may last 15 years. If maintenance is again set to 7.5% of capital and if capital is to be repaid at 3.5% over the vehicle life then the annual cost has the range is £390 to  £810 per seat – a fraction of the rail costs.

Comment and belief

The above figures should start alarm bells ringing. We know that the tax revenue from road users does not necessarily reflect the economic value of roads: that can only be discovered by a proper system of road pricing. However, it would seem that, in order for a subsidy to rail of the above magnitudes to make any sense, the so-called social benefits or positive externalities from rail travel would have to be huge. As we see below, they demonstrably are not. The identification of this issue demands a serious consideration of the relative costs and benefits of rail and road travel. 

The cost, and the consequent drain on the taxpayer, is tolerated because of the strongly held belief that:

• Rail has a far higher capacity than road.
• Rail commuters to central London could not get there any other way.
• Rail uses much less fuel than road transport.
• Rail is far safer than road ever can be.
• High-speed rail is essential.
• Rail is in some way "sustainable" compared with road transport.
• Railways are far too narrow to be converted to roads.

Do those beliefs have any basis in fact? The most important of these issues are considered below.

Capacity and use

In the morning peak hour some 50,000 passengers alight at Waterloo Main Line station. There are 4 pairs of tracks on the approaches.  The 50,000 could all find seats in 1,000 50-seat coaches - sufficient for one lane of a motor road managed to avoid congestion. With a demand as high as 50,000, probably 70 or 100-seat buses could be used, cutting the vehicle flow to between 500 and 700 per hour.  That illustrates that on the line haul, the express coach would have 3 to 4 times the capacity of the train to move people. The fact is consistent with American research, which concluded long ago that there is no movement corridor in the world where demand cannot be satisfied by a single express bus lane[10] ^[11].  A practical example of that is the approach to the New York Bus terminal via the Lincoln tunnel. There, 700 45-seat coaches use one lane 3.2 metres wide in the peak hour – offering over 31,500 seats.  Clearly 90–seat coaches would double that to 63,000.

Terminal capacity is a separate issue and more difficult to demonstrate simply. However, commuters with prepaid tickets may board at the rate of one per 1.5 per second and alight at the rate of one per second, providing a total of two minutes for a 50-seat bus. If 5 minutes are allowed 1,000 buses per hour require 85 bus bays. If these were spread over 3 levels there would be some 30 bays on each. At Waterloo there are 21 platforms plus a very large concourse area. Probably the bus would use terminal space 3 to 4 times as efficiently as the train, in line with the track capacity.

Fuel consumption

Rail is often held out as environmentally kind compared with road transport, but does that have any basis? In 2003 rail returned the equivalent of 108 passenger miles per gallon for Network South East, 123 for Regional services, 123 for Intercity, an average of 115 for all services.  In comparison a small diesel powered car containing two people may return 120 passenger miles per gallon and an intercity coach containing 20 people may return 200 passengers-miles per gallon. Figures for freight show rail more fuel efficient than rail, but if the national railway function, were carried out by express coaches and lorries on uncongested rights of way then 20-25% less energy would be required.[12]

Of course, from the environmental standpoint energy consumption is only part of the story. Other factors include the location of the emissions and their relative toxicities. However, it is clear that, at least on the energy measure, efficient mass transit road transport outperforms rail by a significant margin.

Safety 

Sir Robert Horton said in Railtrack's annual report of 1998/9 that "rail is 27 times as safe as road (measured by fatalities and serious injures) and is becoming steadily safer". Since 1996 injuries to rail passengers have been defined as either fatal, or “passenger” – meaning taken to hospital from the scene of the accident.  That is similar to a serious injury by road, defined as spent the night in hospital or at the discretion of the police. Casualty costs per passenger-km may be estimated by adding the deaths and passenger injuries for the period 1996-2002, multiplying the totals by the government’s values for the Killed and Serious Injured (the KSI) and dividing by the number of passenger-km. The value per million passenger-km for rail in the envelope bounded by the ticket barriers came to £2,000 (excluding those hurt falling over packages, hit by barrows, falling down steps or escalators or in the category “other”).  The corresponding value for passengers in buses on non-urban roads, including an allowance for those hurt a short time before boarding or after alighting, came to £550. A caveat in the calculation is that the passenger-km by bus used in calculations depended on vehicle occupancy, set to 16.  That compares with 9 for the national average and with 25 for the occupancy claimed for coaches leaving Victoria coach station.

If staff, post office workers, those hurt at level crossings and as trespassers (but not suicides) are added then the system-wide casualty cost for rail rises to £6,550 per million passenger-km (excluding passengers hit by barrows, falling over or down steps etc and excluding injuries to staff at trackside or in shunting accidents). The corresponding value for motorways came to £3,360. For all roads the value was £13,500, including £6,160 attributable to pedestrians, cyclists and people on motorbikes, classes seldom met with on railway alignments.  If those classes are excluded the value for all roads came to £7,300.[13]

Hence it is clear that Sir Robert's "rail is 27 times safer...than road…" has no basis except in so far as the total number of selected casualty types may provide a spurious coat hanger.  Like has to be compared with like thus, when coming to a view about safety. One should compare the casualty costs of rail with those of mass transit by coach.

Alternatively one may limit calculations to death rates. The headlines there are:

For passengers there are 0.4 deaths per bn passenger-km in the envelope bounded by the ticket barriers.  Our estimate of the corresponding value by bus and coach on non-urban roads is 0.2.  If the cordon is drawn more tightly so as to capture only those within the envelope bounded by boarding and alighting and including injury by the doors the death rates are, 0.26 by rail and 0.16 by bus and coach.  [14]

System-wide there were 4 deaths (including trespassers but not suicides) per bn passenger-km by rail compared with 1.58 on the motorway network.

Speed and service frequency

Before the introduction of speed limits on motorways express buses never hesitated to cruise at 150kph (90mph). Today they are limited to 100kph (60mph) on dual carriageways and motorways and to 80kph (50mph) on single carriageways, 16kph (10mph) slower than cars. Lorries have the same limits as buses except that articulated lorries are limited to 64kpg (40mph) on single carriageways. In comparison intercity trains achieve 175kph (110mph) and aspire to 225kph (140mph). Hence, it would appear that express coaches would be slower than rail. However:

• The average passenger rail journey is only 41km (25 miles) long and 90% are less than 130km (80 miles) long. Over these distances the headline speeds of the intercity express are rarely relevant.
• Provincial services seldom average more than 64kph.
• The express coach would offer a service frequency up to 10 times as great as rail.

Hence, coaches would offer comparable journey times except for the longest journeys.

Weather, pranks and accidents

The susceptibility to rail to ‘leaves on the line’ or the ‘wrong sort of snow’ is legendary. In comparison, road transport carries on in most conditions. For example, on the Monday after a weekend of flooding in November 2000 the entire rail network came to a virtual standstill. On the following Tuesday the Today programme interviewed a major road haulier. He said that his organisation had reached virtually all its customers. No doubt there was disruption but probably 95% of road journeys were unaffected compared with a completely paralysed rail system. Indeed rail is so fragile that plastic hawks were installed to prevent birds nesting in the gantries - a falling nest could short the system, causing hours of delay on an entire route.[15] Even schoolboy pranks, such as a pound coin placed upon a rail track in the hope of obtaining a flattened one, can cause disasters. Meanwhile, an accident such as Hatfield or Ladbrook Grove disrupts rail for weeks, if not months.

In comparison the disruption which arises when there is a major motorway accident seldom lasts more than a few hours.

The Fundamental Problem

There are technical reasons for the extra costs of railways compare with road transport. The steel-tyred wheel and steel rail leads to a stopping distance up to 4 times that required for road vehicles;[16] rail provides an inflexible mode of operation offering no possibility of serving local destinations beyond terminals; very high point loads between wheel and rail lead to high maintenance costs; and there is a need for a meticulously laid and maintained track.

Widths and headroom

The track width for a train is “4 feet 8.5 inches”.  It is derived from the carts being dragged by pit ponies 150 years ago.  However, bridge abutments, tunnels and viaducts on double track railways offer a clear width seldom less than 7.3m (24 feet) - the same as the carriageway width required for a two-way trunk road. Elsewhere double-track railways offer a level width of 8.5m (28 feet) on tangents and more on bends.  Single-track railways offer 4m (13 feet) between bridge abutments but many were built on double track formations. Hence the widths of most railways would accommodate carriageways the same width as those for new (single carriageway) trunk roads but not the 3m verges that form part of the design standard for green field construction. However, effective verges are generally absent on most ordinary roads and would serve little or no purpose on railway alignments.

The European carriageway width for a two-way road is 7m and at one time the standard for Scottish Trunk Roads was 5.5m (18 feet). It may also be noted that there are many “A” roads 6 metres wide where lorries and coaches operate without difficulty.

As to headroom, clearances above rail top are normally 4.16m (13 feet 8 inches) increased to 4.77m (15ft 8 inches) where there is overhead electrification.  Road level would be 300mm (1ft) below rail top. Hence, without altering tunnels and bridges, the clearances available are 4.46m up to 5.07m. In many parts of the world the required headroom is 4.5m

Although car transporters in the UK are often 4.9m high nearly all container lorries are less than 4.3m high.  Further, the standard height for international transport is 4m.  Double deck buses range in height from 3.9m to 4.44m. A headroom clearance above the vehicles of 200mm is adequate.

Against that background the notion that railways are too narrow or lack adequate headroom should be dismissed.

A sensible standard for conversion would offer 3.5m lanes and 4.5m of headroom. Marginal strips could flank the carriageways where exiting widths enabled that. Departures from standards (e.g. a reduction in lane width to 3m) should be allowed where the full standard may impose unacceptable costs.

Rates of return and costs

The Hall/Smith Report "Better use of Railways" (submitted in accordance with Contract 466/3 made between the Department of the Environment and the Department of Geography of the University of Reading on 29th August 1974) provides first year rates of return ranging from 18% to 500% for 6 conversion schemes in and around London, except for one scheme where the value of surplus land exceeded conversion costs.[17] The study attracted vitriolic criticism from the railway lobby, criticism which is exposed as risible in the companion volume "Comments and Rejoinders".[18]

Conversion of the costs quoted to 2002 values using the Road Construction Price Index yields the range £52 per sq m to £124 per sq m.  The mid range value is £88. Applying that to the 32,000 km of track and a lane width of 5 metres provides an all-in conversion cost for the entire rail system of £14bn. That is perhaps one-fifth the cost of the modernisation programme, quoted as rising to £73bn. In fact, the actual cost of conversion could be cheaper than these estimates. A 7.3 metre carriageway built on a railway alignment at Southport through open country cost £140,000 per km at 1991 prices.[19]  That yields £31 per sq m at 2002 prices.  If that should be the target for costs and if an asphalted lane width of 3.65 m is selected, instead of 5 m, then the estimate for conversion falls £3.6bn. Optimism bias at 40% could be added to those numbers but the same could be said for the originally proposed railway modernisation programme.  It follows that the railway modernisation programme is several times as expensive as railway conversion.

As to disruption – we should not forget that many parts of the network have been subjected to blockades lasting months and to years of significant disturbance with substitute buses in use both at weekends and during weekdays. Additionally preserving railways as railways is a very effective disruption in its own right.  It prevents many thousands of lorries and other vehicles from diverting from the unsuitable rural roads and city streets that they currently use to railway rights of way.  Those rights of way are substantially empty, have vastly superior alignments and would be free of pedestrian, cyclists and people on motorbikes.  They are also relatively remote from housing.

Politics

When something which obviously should be happening does not happen there must be powerful forces preventing it.  These are not hard to find.

• Perhaps the most potent is the railway myth driven by the romance of steam and the glory of a pair of railway lines curving away into the distance in the morning sunshine - the impression of order, the peace of the scene (meaning nothing is happening) let alone all those children’s railway stories.
• Perhaps it is that when expenditures become as vast as in the past, and as now, it is impossible to admit of a mistake.

Certainly the Department for Transport or its predecessors are the institutions that have done more than any other to perpetrate the railway myth and to prevent change, particularly the conversion of disused or lightly used railways to roads.

In any event Stewart Joy, chief economist to British Railways in the 1960s wrote in his book the Train that Ran Away[20], that “there were those in the British Transport Commission and the Railways who were, cynically, prepared to accept the rewards of high office in return for the unpalatable task of tricking the Government on a mammoth scale.  Those men were either fools or knaves”.  There were no libel actions but Joy had been forced out – too honest to work with railway men.

To illustrate - It would appear that those men would prefer to see a railway abandoned to hedgehogs and rabbits than converted to a road. Examples of that include:

1. The Inverness to Wick line which in 1955 Sir David Robertson, the MP for Caithness, tried to have converted to a single carriageway road.  Detailed costs were prepared by Balfour Beaty.  The Department of Transport killed the idea by producing an estimate for an elaborate dual carriageway that raised the cost from £4.5 million to £12 million.  40 years latter the old road was rebuilt  - as a single carriageway.
2. The Great Central line which has that lain fallow these last 40 years and had its route integrity destroyed.
3. The 13 year delay forced on the construction of Yeadon Way Blackpool by the unwillingness of the Railway owner of the derelict right of way to sell at a reasonable price.
4. The 14,400km (9,000 miles) of high-grade right of way lost to the nation following the Beeching cuts.

We comment, then as now, for the Transport Committee has recently completed its inquiry into the “The Future the Railways”.  The brief that the Committee set itself contained the following four questions:

• Is the Regulator right, or is rail an outmoded form of Transport?
• Is the present Network the right one: if not how should it be changed?
• What sort of traffic is the network best suited for?
• How does our network compare with other railways, and what are the lessons we can learn from other countries?

Evidence given by the SRA and others led the Committee to believe that rail is overwhelmingly safe compared with road and has by far the higher capacity.  For example:

1. Bombardier said “to carry 50,000 people per hour in one direction we would require a 35metre road used by buses or a 9-metre rail track bed for a metro or commuter railway”. In contrast to that we have before us the New York express bus lane delivering 31,500 seats per hour in a width of 3.2 metres at a time when the peak hour passenger flow on the main-line railways serving central London amounts to only 10,000 crushed passengers per pair of tracks.  Those passengers would fill 200 50-seat coaches – sufficient to occupy one quarter of the capacity of one lane of a motor road.
2. The Committee’s report, paragraph 186, says “The figures comparing road and rail fatalities are telling. In 2002 3421 people were killed on the roads while, according to the HSE, ‘a total of 50 passengers, rail staff and other members of the public were fatally injured and 256 people died as the result of trespass and suicide on the railway’. (Further) “The SRA points out that on average more road users die in accidents each day than rail passengers in a year”.  We comment, these statements exaggerate the relative safety of rail by a factor of 18 by ignoring usage and by a similar factor by comparing passengers killed in so called Train Accidents with all those killed, system-wide, on the roads.

The Chair of the Committee was the formidable Gwyneth Dunwoody.  She was in the Cabinet in the early 1980s when Lord Marsh, then the Minister for Transport, put forward an unanswerable case for closing the 150 km of the Central Wales Line[21].  He explained that it would be cheaper to provide each passenger with a chauffeur-driven Bentley than to keep the line open.  The then Welsh Secretary piped up: “but Minister, this line runs through 6 marginal constituencies”.  Needless to say the line is still open.  It was closed for a year when a bridge was swept away.  The inhabitants managed well enough but no one had the courage to say “well that’s it, lets forget the bridge and save the money”.

Hence it is not surprising that, instead of dealing with the brief that the Committee addressed to itself, the Committee’s report majors on re-organisation [22].  The grasp demonstrated there was very significant.  In contrast it is clear that the Committee lacked any grasp of the underlying numbers. That is proving immensely costly to the nation.  The consequence is that, over the decade, or perhaps over 20 years, £100bn of capital may be sunk in the railways – a system which is, and always will be, quite worthless in the market place.

Conclusion

The case for removing obstacles to the conversion of the national rail network to a motor road system managed to avoid congestion appears overwhelming. This article has concentrated on the accounting costs of different modes of transport, to conclude the argument we need to tie it into political economy. One conclusion of this work would be that all government subsidy to rail is wasteful and should be phased out. If that were to happen, because the road network is not properly priced, efficient mass transit alternatives may not develop. Thus the phasing out of government subsidy to the rail network should happen alongside the development of a national system of road pricing. If the road network were properly priced and rail subsidies halted, road users in congested areas, where the rail network is most dominant, would be charged economic prices for road use.  Users of all form of transport would have an incentive to find the most efficient method. There seems to be little doubt from this analysis that this would not involve rail travel. The government should ensure that there are no obstacles to the sale of land used for railway lines for conversion to roads. Such roads could be privately owned and priced and compete with the already existing network.  Removing such obstacles would require the re-establishment of a system of private property rights in railway infrastructure and the removal of any statutory duties to provide rail services. Current policy and recent policy developments are inimical to the provision of an efficient and innovative mass transit service.

Meanwhile if the Government wishes to subsides some well deserving group of would be travellers it should pay the money directly to them leaving them to make to the choice as to how or if to travel.  At least that would be preferable to subsiding the product directly – a product that is, for the most part, used by the better off.

PDF: 

reigniteupdate.pdf