by James Meek
But in January 1995, most of Europe’s state-owned railways – nineteen of them – came to the joint conclusion that moving block was not ready to be used in the real world, and a simpler, transitional form of technology should be the next step. Again, Railtrack ignored the warning. In fact, Railtrack may never have heard it: at this time the firm had barely any contact with Europe. It was an extraordinary situation, of which the public was ignorant: a group of Anglo-American consultants and executives took for granted European support to develop a technology which those same Europeans openly declared to be premature.
In March 1995, Railtrack and the government went public with their plans for the west coast main line. Despite the warnings, they endorsed, with few reservations, the consultants’ recommendations to make moving block central to the modernisation. John Watts, the Tory rail minister, spoke of an ‘innovative signalling and control system’ which would be ‘at the heart of the proposals’ for the WCML. (Interviewed by phone, Watts said he couldn’t remember details of this period.)
The consultants’ report remained secret. On 22 March Railtrack released what it said was a summary of the report. By and large, it was. But there were important changes and omissions in what the public was told compared to what the consultants had said. The consultants warned, for instance, that if moving block turned out not to work, and conventional signals had to be used, any attempt to try to get trains to run faster than 125 mph – always Railtrack’s intention – would incur ‘exceptionally high costs’. This fateful warning, one of the keys to Railtrack’s eventual demise, was omitted from the public summary.
At the beginning of March, signalling-company bosses had told MPs that they might – might – be able to have moving block working and installed within ten years. In their secret report, the consultants talked, with optimism, of ‘a five-year programme’ for the development and fitting of moving block. Yet this is what Railtrack told the public: ‘The development programme is anticipated to take between three and four years.’
Most surprisingly, Railtrack inserted a line into the public summary that had never appeared in the consultants’ report. ‘Most of the hardware for this train control system already exists,’ it read, ‘the technology required being relatively mature.’
I read that line out to Arthur when I met him. He said: ‘Mm. That’s interesting.’
Did he find it misleading? Arthur paused for a long time. Eventually he said: ‘I don’t think there’s any doubt, sitting here now, that it was not as far developed as we thought it was … I’m surprised by those words, I really am … I am amazed at that statement. Because I don’t know where they had any proof of that.’
Neither Horton nor Edmonds would comment on the discrepancies. I asked another senior industry figure, with close knowledge of the subsequent attempt to make moving block work, what he thought. ‘To say the technology is mature – yes, I think that was a bit adventurous, certainly in 1995.’
What happened? The consultants’ report never mentioned it, but there was one other factor in the back of the minds of Railtrack and the Tories. In 1994 there had been a painful strike by railway signalling workers. Bringing in the new technology would be another step towards ending the unions’ leverage, by getting rid of thousands of signalling staff.
Some in the rail industry are inclined to give Horton, Edmonds and their colleagues the benefit of the doubt. ‘I think dishonesty does matter, but my suspicion is that it wasn’t a question of dishonesty, it was more a question of misjudgment,’ Michael Beswick, the regulator’s director of rail policy until he retired in 2013, told me. ‘It was assumed that technology would move very quickly, but it doesn’t. That was the problem. That is a bit of an indictment of the calibre of the people running the show at that time in Railtrack.’
Another senior rail industry figure said: ‘John Edmonds was keen to get the company privatised and wanted to say things that would encourage people to believe it was bold and dynamic. It was all about creating confidence, which requires bold statements, sometimes.’
Before I knew I would get to see the consultants’ report, I spoke to a former senior Railtrack executive. He didn’t have a copy of the report, and was trying to remember its conclusions. As I found out later, his memory didn’t quite reflect the report; rather, perhaps, it reflected the real back-room conversations going on in Railtrack at the time. ‘The basic conclusion was that it was impossible to upgrade the west coast at any sensible cost if you went for conventional signalling,’ he said. ‘And the only way forward – whether it was feasible or not – was to bring in twenty-first-century signalling technology.’ Feasible or not: the decision was made, and Railtrack began unconsciously to weave its downfall.
The consultants weren’t reckless. They never imagined that Railtrack would manage and finance the WCML modernisation itself. In some detail, they outlined a scheme whereby Railtrack would get a big, experienced civil-engineering consortium to raise money for and manage the project, thereby assuming most of the risk. But when Norman Broadhurst, then Railtrack’s finance director, studied the numbers, he thought the returns looked too juicy to be given away. He thought Railtrack should do it, and talked his colleagues round. One member of the board described how he’d argued against the proposal. ‘I said at the time Railtrack did not have the management capability to bring that in house,’ he said. He was proved right. No sooner had Railtrack committed itself to moving block than it began to waste time bringing it about. In mid-1995, its rump signalling team had dwindled to the extent that it was possible for it to move office in a single taxi. The following year, a move to Birmingham caused further losses of personnel. One senior Railtrack figure at the time said: ‘In resources terms, two years were lost.’ It wasn’t until March 1996, just a few months before Railtrack was privatised, that the firm picked two consortia of engineering multinationals to develop alternate prototypes of the moving-block system. Soon the consequences of Edmonds’ determination to gut Railtrack of its in-house engineering and project-management expertise became apparent. ‘What Railtrack did in 1996 was quite exceptional, which was to take a really high-calibre engineering team on the BR system and destroy it,’ said Chris Green.
Railtrack had assumed that the two signalling consortia would develop similar types of moving-block technology. It assumed their work could then be pooled to provide the foundation for a system that actually worked. But it didn’t happen that way. The consortia saw each other as rivals.
‘Not unexpectedly, their work tended to diverge rather than converge,’ said a senior figure in the signalling industry at the time. This wouldn’t have mattered so much, except that while the moving-block research was meandering, Railtrack made a catastrophic decision. It invited Richard Branson to hold a gun to the company’s head.
In February 1997, Richard Branson’s Virgin Trains had won the franchise to run fast inter-city services on the WCML. In October, after the newly elected Labour government backed away from its commitment to renationalise the railways, Branson and Railtrack announced how the WCML project was going to be financed. They painted a wonderful picture for inter-city travellers. Railtrack would spend £1.5 billion to restore the worn-out railway to basic reliability, and, in exchange for a slice of Virgin’s profits, would lay out another £600 million to install moving block and other improvements to create a high-speed line. By 2002, new Virgin tilting trains would travel the line at 125 mph; in 2005, they would accelerate to 140 mph. London and Manchester would be only an hour and forty-five minutes apart, London and Glasgow less than four hours. Suddenly, Railtrack found itself locked in a contract with Virgin to deliver a non-existent signalling system on the entire west coast main line by a firm date – 2005 – with crippling financial penalties if it failed to do so.
‘They thought, “Oh crumbs, we’ve just signed [the Virgin deal], we’d better get on with this.” There was a realisation the talking had to stop,’ said the signalling source.
Yet Railtrack was locked in
a catch-22 situation. It couldn’t put a contract out to tender to develop a thing if it couldn’t specify what the thing was. And it couldn’t specify what the thing was until it had signed a contract for somebody to develop it. Meanwhile the two signalling consortia had made the technology look less, rather than more, certain. ‘Because we went on divergent routes, because we saw ourselves as rivals, it didn’t enable Railtrack to say, “It’s obvious, here’s the specification,” ’ said the signalling source. ‘What’s at the core of this, I think, is that Railtrack did not have its own sufficiently strong in-house knowledge and expertise to be able to use industry for what it was good at, to gather their views in and make a judgment.’
Railtrack did have one ace in the hole. It had, in fact, long since recruited an expert in moving-block signalling. Way back in 1995, the then Conservative government had worried that Rod Muttram’s lack of railway experience would put off City investors. At the government’s insistence, Railtrack appointed a champion of moving-block technology to be its engineering director. Brian Mellitt was a clever, experienced specialist who’d already supervised the early stage of introducing moving block on London Underground’s Jubilee line extension, then under construction.
But in 1997, just when his expertise was most needed, Mellitt’s pet project was suffering a horrible public failure. Quite simply, moving block on the Jubilee line didn’t work. Computers that were supposed to talk to each other couldn’t. Costs soared. Desperately trying to complete the project in time for the opening of the Millennium Dome, engineers had to come up with a crash programme for an old-style conventional signalling system. The number of trains an hour was slashed in half. Westinghouse, the key signalling company responsible, took much of the blame. John Mills, who in 1995 had told parliament that moving block on the west coast main line might be possible in ten years, was removed as chief executive.
Throughout 1998, as news of the problems on the Jubilee line began to emerge, the newly arrived Gerald Corbett – who took over from Edmonds in late 1997 – became increasingly uneasy about the moving-block plan he’d inherited for the west coast. He didn’t get on with Mellitt. Corbett was, if anything, even more hostile towards the engineering profession than his predecessor. Railtrack didn’t sign a contract to develop the system with the British-French company GEC-Alsthom (later Alstom) until July 1998, more than a year behind schedule. But even this wasn’t a proper contract to deliver something; it was a nine-month contract to define the thing that should be delivered.
And the signalling was only part of the WCML modernisation. Much more needed to be done. Tracks needed to be relaid, tunnels modified, bridges altered. This work, too, was behind schedule and over budget. Plans changed constantly. At one point, Railtrack paid £10 million to a contractor to develop a new kind of transformer, only to abandon it later and go back to the original type. With every passing day, contractors were finding out how much more seriously the line had deteriorated than Railtrack and the consultants had understood.
At the same time, because Railtrack had shed so much of its engineering and railway operations expertise, it had little ability to judge whether the prices its myriad contractors were charging were fair. ‘It is easy to understand why certain elements within the industry took advantage of that situation,’ said one rail industry figure. ‘It would have been a great temptation.’
In one iteration of a recurring pattern, Railtrack turned to a US company for project management expertise, but managed to botch that, too. Railtrack hired Brown & Root, a subsidiary of the US engineering group Halliburton, then run by Dick Cheney. Whether Brown & Root, which had grown fat on Pentagon contracts from the Vietnam war and beyond, could have managed the job is unknown, but Railtrack never allowed it to try.
‘Railtrack was half in bed with them and half not,’ said an executive who saw the process from the inside. ‘Brown & Root had a lot of experience in oil and gas contracts. Railtrack said, “Fine, show us how,” but they got cold feet and never signed up.’
Meanwhile, Corbett and his team were becoming horribly aware of the other set of baroque errors that had been made by their predecessors: the Virgin contract, presented to the new Railtrack boss on his accession as a fait accompli.
Despite Virgin Trains’ initial reputation for lateness, there was a general fondness for the Railtrack-Virgin plan, not just in Railtrack, but among politicians, the media and the public. Tilting trains going at 140 mph; London to Glasgow in time for lunch; it sounded good. It sounded like progress. But there was a severe problem.
Again, the lay observer would think it was obvious: other trains needed to use the same line. And these other trains did not travel at 140 mph. Some of them, freight trains, for example, struggled to reach half that speed. In all, the WCML was used by 120 trains a day. Two thousand different trains travelled up and down its various lines, and another 4,000 crossed it at some point. They were operated by fourteen private train operators, using at least nineteen kinds of locomotive, a mix of regional passenger services, local trains and freight – a contractual nightmare. But contracts had been signed. And, in the course of 1998, Railtrack began to come to the terrible realisation that it couldn’t keep to its Virgin contract without breaching some of the others.
The introduction of faster, more frequent Virgin trains on the route meant that Railtrack was obliged to carry out extra work to guarantee the other, slower trains would still be able to do their job and earn their money. When Railtrack began to look more closely at the Virgin contract, and compared it with the work it had promised to do to the railway, it saw that it was unable to draw up future timetables without breaking at least one of its contractual commitments. Or rather, Railtrack looked, but it refused to see. It was the beginning of the end.
‘They had committed to the rail regulator and said that it would work, but everybody knew it was impossible,’ said Stuart Baker of the SRA. ‘A 140 mph train still catches up with a 75 mph freight train or a 50 mph commuter train rather quickly. So it was a bit of an illusion. The capacity wasn’t there for the contracted service without a new railway.’
The speed issue was not all. Railtrack was gambling its future on the highly risky maintenance philosophy that said components should be replaced only when they looked about to break, rather than at fixed intervals. When specialists looked into Railtrack’s web of contracts with Virgin and other train operators on the WCML after the company collapsed, they were astonished to see that Railtrack had promised to build a railway that would not need any maintenance from 2005 to 2012.
‘I don’t think anyone denies that the commitment made to [Virgin] was a terrible mistake,’ said Michael Beswick. ‘It was one of the factors that brought the company down, ultimately.’
In 1999, the bubble burst. When Chris Green, who took over as head of Virgin Trains in February, went out and about to see how the epic west coast reconstruction project was going, he received a shock. ‘I was surprised at how little physical activity there was. There was nothing happening on the track. I was looking for the big yellow machines ripping up track and signals being replaced and wires being renewed, but everybody seemed to be in endless debate about the scope of the work. There’s no doubt that two years were lost. I think as I arrived it was dawning on Railtrack that this was going to be an incredibly expensive and complicated project.’
He quickly found out about the timetabling crisis. ‘The basic railway skill ought to be train timetabling, oughtn’t it? Railtrack ought to have been able to timetable it themselves and I’m sure they did. When they found it didn’t work, they also found they were in a spectacularly tight contract, with massive penalty clauses, up to £250 million, with Virgin … every time they went round that situation they found they couldn’t afford to break the contract. Which is why they had to implode.’
By the spring of 1999, even as Railtrack shares hit a high of more than £17, the project was in turmoil. Brown & Root was dropped and Railtrack drafted in a fresh team of consultants, the
Nichols Group, to take stock. The rail regulator was hammering on Railtrack’s door, demanding to know how it was going to make the west coast timetables work. Alstom’s quest to try to make moving block work was going badly. Nine months of studies hadn’t produced a clear design for the system, and amid the uncertainty even Railtrack was not about to give them the £750 million they sought to finish the job. In another blow to the technology, Mellitt quit.
‘With him gone, there was nobody championing moving block,’ said an insider on the signalling contract, ‘and at this time, the Jubilee line was unravelling.’
On 9 December 1999, at a meeting code-named Black Diamond Day, Railtrack finally took the decision it had avoided for so long. It accepted that moving block, the technology on which the company had staked so much, was a mirage. The Nichols Group report was painful in its clarity. There was no more than one chance in twenty that the system would be ready in time. The Jubilee line fiasco had shown the risks. Alstom had never been able to explain what would happen to a moving-block system in the event of an accident, or if there was a disruption of radio signals. One possibility was that the entire railway network between London and Glasgow would simply grind to a halt.
Nor had Railtrack ever fully grasped the enormous sums it would have to pay train operators for permission to take 2,000 locomotives out of service to fit the new equipment and retrain 4,000 drivers. The implications for Railtrack were catastrophic. They would now have to introduce a different, conventional signalling system, at vastly greater cost. Installing conventional signalling would mean stretches of railway being closed off, which meant huge compensation payments to the train companies. All Railtrack’s contracts with train operators were predicated on moving block. The contracts would either have to be torn up – more huge compensation payments – or a costly extra programme of works set in motion to reconfigure the entire railway. Instead of the expected £2 billion, the new headline figure was a staggering £5.8 billion. It quickly turned out that this, too, was a wild underestimate. Without a massive taxpayer bail-out, Railtrack was doomed.