Why is CrossRail so late? And the multiple Londons problem.

Europe’s biggest infrastructure project. 

There’s few people in Britain and in London more specifically that don’t know just how delayed CrossRail is. To many, it’s become a poster child (like HS2) for bad project management and government bungling yet again.

After all, businesses near the stations going east to west and in central London made plans, took loans for expansion and moved premises to take advantage of the traffic. To date, the only taste of the Elizabeth line in its glorious soon-to-be purpleness is to the west of Paddington and to the east of Liverpool Street the forking rails going south of the river and out to Shenfield

It was originally planned for opening in 2018. The expected delivery date for the whole project is now March 2021, although there are some rays of hope peeking through that it’ll be done a bit earlier.

The key question, ‘Why is it so behind schedule?’ , comes with multiple answers. The official answer teeters between budget and testing, the in-depth one from on civil engineering in-the-know sources was that the original plan was pure fantasy and now the answer pieced together from all the evidence (courtesy of hindsight) seems to mount up to ‘we did bad, we know better now’.

I started with this question in mind as there was so little information in mainstream sources. Digging deeper, I found there were two main causes, one technical and one organisational. And as we’ll see, for once it was too much autonomy rather than too much government control that led the project where it is now.

Inspiration and early beginnings

In terms of inspiration and purpose, the YouTube channel RMTransit and Londonist made an interesting point that Paris’ long-distance suburban RER (Reseau Express Regional) trains acted as the model Crossrail drew from. However, the point can be made that this is merely because of the high usage on existing lines and extension of the outer suburbs further and further away from the CBD. Paris, London, Berlin and many other cities that have undergone electrification and then grew underground/train networks experience the same growing pains and consider similar paths to resolve those pains. 

Source: Pexels

The Crossrail Act received royal assent and became law in 2008. The project started in 2009 and the first tunnel broke ground in May 2012, finishing at Farringdon in 2015. But there was still hard work to be done : 50km (31mi) of track to be laid, many more miles of wire and cable, stations outfitted (6 new stations in central London + Canary Wharf, Abbey Wood, Woolwich and Custom House) and signalling software set up. On top of it all, the 26 miles of new underground tunnels in central London, while large enough to use large Bombardier trains, were still small confined spaces and this limited the speed at which parallel construction could happen. 

In 2015, the construction companies began laying the track. It was at this point that a critical decision was made. The first signs that the December 2018 deadline could not be met appeared in 2015. Given the progress made by contractors, the deadline was implausible, yet the management team chose to keep it, leading to some… strange decisions. One example was early train testing – done before the tunnels were fully operational and broadly similar to standing in the tunnels and imagining where the trains go. If you think that’s harsh, a lot of the testing was done in 2016 and 2017, before they’d come across the final boss of this project (more on this later). 

Early testing might have seemed like an effective project management decision, but it also meant construction had to be done in parallel in these confined spaces while it took place, placing yet more strains on an already hard to reach deadline.

As of March 2018, the Annual Report from the Infrastructure Projects Authority gave Crossrail the first amber after 6 continuous years of green lights. This (and a handful of warning signs in the months preceding this) was the first indication to the wider public that the project’s delivery was flawed. Although, if you didn’t read that report in 2018 and thus weren’t aware of this back then, I don’t blame you. It was later in 2018 that funding came into question and under heavy scrutiny. While the original cost was £14.6bn, by 2019 the first increases in that figure popped up. At first it was by a couple hundred million, then again and as of early 2020, the total cost estimate ballooned to over £20bn.

Source Pexels.

A marathon obstacle course

Building the first new Tube line in over 30 years going right through central London was never going to be easy. But the critical period after the tunnel construction was finished in 2015 seems to have been bungled from a project management perspective. Unfortunately it is this 2015 to 2018 window that we have very little information on to draw conclusions and have to rely on the National Audit Office’s report titled ‘Completing CrossRail’. 

One tidbit is that in 2018, the head of the NAO (National Audit Office), Amyas Morse, said: “Throughout delivery, and even as pressures mounted, Crossrail Ltd clung to the unrealistic view that it could complete the programme to the original timetable, which has had damaging consequences.”

The report adds:  “… a number of stakeholders we spoke to expressed the view that the Crossrail Ltd executive team recognised the challenges but believed this was an exceptional team capable of delivering exceptional results and overcoming these challenges…”

It was a whole year after, in April 2019, that the Board of Crossrail convened to review the progress. The meeting on the 25th of April (check pages 2-3/7) featured a request to review the EOP (Earliest Opening Programme) and DCS (Delivery Control Schedule) and have them presented at the following meeting in May 2019. 

One question that I only found an answer from the NAO report is where the cost overruns came from – was it extra salaries for workers and equipment hire because of the deadline extension? The report points at the contractual structure of the Crossrail organisation and the high degree of autonomy for contractors, meaning lower downward cost pressure. In effect, even as the contractors asked for more money, Crossrail management were powerless to say no. One of the main issues with private sector involvement in public works is the incentive to bid low and then ask for more money close to the deadline, after the point of no return. 

“Crossrail Ltd had to compensate individual contractors for delays that occurred on other contracts, on which their work depended, and had to engage in costly change control negotiations.” the report points out.

It was around the original deadline of December 2018 that the public relations and media presence of Crossrail kicked into high gear. Between public inquiries and possibly demands from the government (local and national) for more transparency, you could find Crossrail filming every detail of construction. And yet, at the time, I found this did little to explain the delays.

Photo by Dom J from Pexels

Signalling and the many Londons problem

London has a long history and if you look at East, West and Central London, you come across many differences. Cultural, economic, demographic, you name it. A less well- known difference is signalling technologies. Signalling technology has advanced throughout the 20th century and depending on when track and equipment was set up, different sections of the Underground use completely different systems. As a result, an east-to-west Elizabeth line must make the two main pieces of signalling software play nice with the new Siemens one in central London in order to keep trains in one piece. 

The section between Paddington and Heathrow uses ECMS (European Control Management System), while west of Heathrow and east of Stratford it’s TPWS territory (Train Protection and Warning System). The central underground bit in between Paddington and Stratford is where the Siemens solution came in. It’s called CBTS (Communications-Based Train Control) and is the modern version compared to the legacy ECMS and TPWS.

The hurdle is making sure a train going across London communicates with all three systems without any glitches. A number of bugs caused software development to return to the train testing phase, according to Crossrail CEO Mark Wild in 2019. This is what’s referred to as Dynamic Testing and one of the main drivers of delays and cost overruns.

The economics of large infrastructure projects and a joke about pens and pencils.

Many large scale projects financed by the government ripple far beyond the project itself. In many cases, they are exercises in Keynesian economics, aimed at stimulating local and national economies via construction, supplier networks and the fabled multiplier effect. In that respect, I have no doubt Crossrail rippled throughout the construction and transport industries (in ways that HS2 or any construction project will too, if it keeps going long enough).

There’s a space-age joke that the Americans got to orbit and their pens wouldn’t work without gravity. So they spent millions, researched advanced materials and created a hundred versions of a zero-G pen with ink that dries instantly. The Soviets used a pencil. 

As funny as the jab at bureaucracies goes, the Americans would have developed entire industries and created jobs out of this in-joke endeavour (see Skunkworks at Lockheed Martin for a similar effort). I say ‘would have’ because the joke is untrue. NASA used mechanical pencils and besides, the Soviets are famous for many things but not sure the efficiency of their state apparatus is one. Oh, and broken pencil tips can be a hazard in space while ink isn’t.

So there are benefits to Crossrail, that’s undisputed. More so as it is an important lesson for anyone planning construction in London, with its myriad authorities, tunnels and history encumbering anything from a house extension to a decade-long east to west line to supplement that overworked horse, the Central line. If you read my post on nuclear energy, you’ll remember that part of the increasing costs was the lack of local expertise given fewer and fewer reactors were built. With the amount of transport projects going on, I don’t doubt construction firms in the UK and project management the world over have learned heaps. 

But, as far as £20bn (£5bn more than the initial projections) goes, it was an expensive lesson, with the final bill added to the tax-payers for years to come and delayed revenues (to the tune of £190m a year in 2019) for TfL stretching an already thin budget post-lockdown. London’s traffic will justify a lot of the final cost and bankers trekking Canary Wharf to Heathrow will surely be legion. But if management had taken measure of their failures back in 2015 instead of trying to stick to an impossible deadline, the cost would’ve been lower and outrage at its budget and the deadline overrun easier to manage in due time. 

Photo by Luca Nardone from Pexels

The sum-up

I said in the beginning that I found two main causes and all the literature seems to point at software bugs in the signalling software and a lack of accountability (don’t worry, I shudder at the sound of buzzwords too) between 2015 and 2018. 

In the HS2 case study I pointed out that the longer a project, the more need for a robust plan in the beginning. The budget is almost always geared towards optimism. Expected cost savings and synergies hardly ever materialise and there are hordes of unexpected cost increases. When was the last time you heard of an unexpected efficiency, though? The latter takes much longer to become visible, while cost increases are immediate and often urgent.

Now, what was that old adage? Those who do not learn from history…

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s