Engineering costs need to be cut

Privatisation has brought many benefits to our industry – new trains, more frequent services, brighter stations, additional routes, competitive ticketing, increased freight traffic – but the stark reality is that the railway now costs much more to the taxpayer than it did in BR days – £4 billion today against £1.4 billion in 1995. Why this should be is a question on many minds and with the country facing public expenditure cuts on a massive scale, the money to be spent on railways will be reduced as Government patience runs out.

Recent studies and commentary have suggested that European railways are more efficient in the way they do their engineering. But is this true? A recent Railway Engineers Forum seminar looked at the perceived problem of engineering costs with four speakers from different sections of the industry putting forward some interesting considerations.

engineering costs west coast
Costs escalated on the West Coast project due to the lack of a 'controlling mind'

Infrastructure elements

Setting the scene was Dr James Martin from Mott MacDonald, someone with extensive geographic and industry experience, and now responsible for the transport business within the company. He suggested several factors as to why costs are so high and how they might be reduced.

Lack of a ‘controlling mind’

The standard dimensions of time, cost and quality all remain but the latter is the big challenge. The traditional practice of having a Chief Engineer in the Civil, Mechanical and Signalling disciplines is no longer the fashion. These ‘big men’ gave direction and coherence with consequential decisions being centred on delivery to time and cost. The electricity and oil industries still have such people and benefit accordingly. It is not a perfect solution – there isn’t one – and things will go wrong on occasions, but the mechanism for recovery is much better.

The WCML example was cited. Originally estimated at £3 billion, it eventually cost £8 billion. At one point, a figure of £13 billion was mentioned. But the appointment of Stuart Baker as a de facto controlling mind saw the project through to a focussed conclusion. Crossrail threatens to have similar problems. Even now, the simple decision of platform lengths remains undecided.

Contractors and risk

With one possible exception, no contractor in the UK is big enough to carry the cost of risk. As a result, the tendered price reflects 100% chance of the risk happening. More equitable risk share has to happen and Network Rail plus London Underground need to be part of this. Both are relatively small companies – whilst the power company EON has an annual capital spend of £30 billion, Network Rail plans £22 billion of investment in five years but seeks to put all project risk on its contractor base. Hence costs are much higher than they need be.­

Barriers to entry

Why do the big European contractors not enter the UK market? Specifications place too much unnecessary prescription with far too much focus on past engineering practices to make the contract worthwhile. S&C designs were cited – these still required square-headed coach bolts to be used, a legacy from the Victorian era, when other cheaper and more effective methods were available. The Javelin train was likened to a Honda car but having to have Nissan brakes. Non-UK signal companies encouraged in the recent past to compete for UK business know all too well the problems of compliance with outdated UK practice.

Safety-critical obsession

Safety layering in the control and communication discipline is excessive. The railway is not the only industry that operates safety-critical systems. In extracting offshore oil, toxic gases are emitted that have to be piped ashore. These are routed to processing plants often through built-up housing areas. The consequences of leakage are horrendous, yet the sensors and valves are controlled by proprietary equipment. The control system is proportional to safety risk and much less expensive than anything the rail industry would consider.

Rolling stock considerations

Different factors seem to impact on rolling stock costs according to Ian Walmsley, the Engineering Development Manager at Porterbrook. Train build costs have dramatically increased – a Class 158 was £450,000 per vehicle in 1988 whereas a similar vehicle today is around £960,000. Why has this happened?

Trains have more features nowadays – air conditioning, passenger information systems, disabled facilities, greater crashworthiness – but there is more to it than just this. Five main elements seem to be the crux of the problem –


The privatised railway needs to generate about 15% plus payment of dividends and funds for reinvestment.


This has to be built-in to the price of safety, audits, potential legal costs and the TOC-ROSCO relationship. Some risk can be offloaded onto subcontractors which, if not needed, can produce more profit.


Real competition between potential franchisees has almost disappeared. However franchise bidding costs – typically £20 million a time – plus competing for skilled labour resources both represent significant on-costs.


Double and triple checking of engineering changes and alterations plus DfT micro-management all have to be paid for.

Supply chain

Train purchase procedures are much more complex – accreditation of suppliers, subcontractors and hierarchy levels within parent company groups all add cost.

The franchising system has attracted criticism as the short duration of these promotes secrecy, discourages cooperation and gives little opportunity for leveraging Network Rail. Furthermore, the UK has relatively few spare trains compared to Europe and many of the repair workshops have closed. Where have all the savings gone?

Some radical solutions were suggested to put matters right –

•        fewer TOCs to give economies of scale without destroying the competitive edge

•        long-term rolling stock build plans perhaps in conjunction with electrification

•        big continuous orders of similar trains – far too many bespoke requirements emerge from the TOC organisations that yield no obvious benefit, something train builders should be stronger in resisting

•        limit and define safety responsibilities

•        trust and partnerships instead of legal contracts which only benefit lawyers

•        vertical integration, possible in selected areas so as to give focussed accountability

•        competent leadership from on high – DfT and SRA left much to be desired.

engineering costs train building
Train builders should resist TOC requests for bespoke features

Signalling and safety approvals

Signalling systems can now deliver much more performance but the downside is increased complication – powerful computers and high speed communications are needed. Programmable electronic systems are complex and signal engineers are sometimes guilty of making them more so, according to Rod Muttram, Vice President, Quality & Safety for Bombardier Rail Control Systems. There is near universal agreement that signalling costs are too high but what are the factors that influence this. Four main reasons were suggested.

Volume and unrealistic expectations

Signalling is misaligned with the commercial market. Mass-market devices such as laptops and mobiles have no recurring costs in their life expectancy. Signalling equipment, now using similar technology, is expected to have a 25-year life which is not a realistic design criterion. Signalling software should be non-recurring and hardware should be treated as a disposable commodity to be seamlessly replaced with something every so often.

Poor specifications

Customer dictation of signal system architecture is full of past practices and takes no cognisance of what modern equipment can do. The optimum location of interlockings was given as an example.

Lack of real standardisation

Whilst progress has been made to rationalise standards, there are still far too many. Preference engineering still abounded and could be seen in the number of ERTMS options having to be catered for. The lack of standardised operating rules across Europe did not help.


Project timescales in the UK and Europe take too long with too much time being taken up in agreeing finance. As a consequence, implementation phases are compressed, driving expensive mistakes. Compare this to China where 1,000km of ERTMS L2 has been achieved in less than two years.

All of this has been made worse by an emerging and onerous ‘Safety Assessment and Approvals’ process. An industry has grown up around these tasks which has generated its own inertia. Not knowing when to stop and the tendency to syndicate safety have led to techniques designed for managing large risks being used on even the smallest of schemes. The situation has been made worse by European legislation, increased media and public expectations, a nervousness amongst individuals of being made scapegoats, a lack of common understanding and the emergence of multiple ISAs, NoBos and other bodies.

An IRSE working party, chaired by Rod Muttram, has set down 14 principles for making safety approval more straightforward. These are now published and should help signal engineers to plan, design and implement new projects with the minimum of external accreditation, thus reversing the increase in assurance costs.

A European comparison

The building of the Brussels-Amsterdam high speed line has encountered many difficulties but the lessons learned should be of use to the people planning HS2 in the UK. Jo Urlings from BAM Rail in Holland explained the principal steps to getting the line built.

•        A PPP contract was awarded by the Dutch government on the basis of a 25-year concession of which five years was the build time. The expected cost of €1.2 billion is yielding a €3.0 billion revenue stream.

•        Finance came from industrial and institutional investors, mainly banks, but this produced lots of ‘advisors’ who have become the controlling party.

•        A consortium was formed with ten main companies.

•        A whole-life cycle approach was taken, leading to the adoption of a slab track system, Rheda 2000, as this gave a better payback over 30 years compared to ballasted track.

•        The project costs are split as follows – track 42%, electrification & control 44%, project management 11%, bid costs 3%.

•        Construction maximised the use of pre-built factory units with a mix of traditional/manual and automated/mechanised methods on-site. This gave an optimum track build rate of 498 metres per day.

•        Payment to contractors was only made when trains could run – i.e. 99% complete.

•        The maintenance philosophy is based on continual measurement and work is only carried out when needed.

engineering costs dutch high speed
A Dutch high speed line benefited from the use of pre-built factory units

In the end, the final build cost was €4 billion, part of this being to accommodate political ‘sillies’. A tunnel for cows cost €400 million. The project also ran three years late mainly due to the different interpretation of ERTMS requirements across the Dutch-Belgian border. Other lessons learned were –

•        verification and validation must be programmed in from the start

•        health & safety and RAMS knowledge/skills need to be built-in at the tender stage

•        a systems engineering approach will deliver the best quality.

When comparing the Dutch experience (100km) with the planning of HS2 (206km), the current estimates for the latter are realistic.

Conclusions and the way forward

The day gave a fascinating insight into the state of rail engineering in the UK today. Whilst the reasons for the high costs were well explained, not much was put forward as to how the situation could be reversed. Encouraging engineers to be more innovative and less blinded by inappropriate standards and over-zealous safety procedures seemed to find favour. How to get this in place is less clear.

Is going back to vertical integration likely to cost less than the present vertical separation? A DfT speaker commented that reducing costs was the number one goal, probably by adopting more standardised products and by eliminating in-fighting between the various rail factions.

To this must be added the need to understand costs – first costs, whole-life costs and system costs. Specifications must be appropriate and affordable. Changes, once underway, lead to a disproportional cost increase. The control and understanding of risk is key and with this comes the need to put in place controlling minds within the engineering community to bring focus and accountability.

It was perhaps unrealistic to expect the perfect answer to emerge from the day – indeed some mixed messages came across. Those advocating a changed structure for the industry should remember the old rule – every time you reorganise, you bleed. Building efficiency into what we currently have would seem the best way forward by cutting out duplication, waste and unnecessary layers. Nonetheless, time is running out and it would be better if the engineering community steered its own solutions rather than having something inappropriate forced upon it.


Please enter your comment!
Please enter your name here

More like this...