Saving the Rail Investment Programme

When not just one project, but a whole programme runs late and way beyond budget, there is only one option: blow the whistle and have a stand-still period. The Secretary of State was right to be shrill in the case of rail electrification.

But how should confidence – vanished from paymasters in the Treasury, and in short supply at DfT – be rebuilt? We need to find the right rail investment projects again for a sector that is still growing strongly (in the intercity sector and on peak services into the major Northern cities especially).

There are the almost inevitable re-organisation plans. Most people concur (following Nicola Shaw in her review) that devolving powers to Network Rail's Routes is at least the right direction of travel. And we have the possibility of greater contestability and new forms of funding to try to implement too (the Peter Hansford review).

But does anyone expect these changes to reduce costs dramatically (for that's what's needed to restore broken business cases)? I've written before (Transport Times, March 2017) on three key assumptions that hinder cost effective project delivery.

But here are two more revolutionary changes I think we need to try.

The first requires a look again at the starting point of project development. Engineering specifications need to follow a remit. In rail, this takes the form of a conditional output statement: "this is what we want, subject to it being affordable (that is: with an acceptable benefit/cost ratio)".
Once set, the project development process is peculiarly linear and unchangeable. Indeed, to maintain focus, a full future year train service specification (TSS) gets drafted. The remit becomes: "design for this service pattern (timetable), at least cost". This is a soluble challenge, but a pointless one. Because of one thing we can be sure: the train service specification will change. Several times, and in ways that are often hard to foresee. For economists, TSSs are like land use assumptions in demand models: something they would like to say is unchangeable to keep the transport investment appraisal 'pure'.

To avoid future budget tragedies, we need a smarter approach to planning. The first stages must be exploratory – and, please stay calm, circular. This is the problem 'shaping' stage. It allows questions like: "what can we afford?

Best of all it allows a re-framing of the problem as evidence emerges. If it looks like a major project, such as electrification is, after all, what's needed, then it allows what is (astonishingly) a question missing at present: "what can we do to make the very best use of electrification, if we had it?" To which the answer would almost certainly be to scrap the studiedly-neutral TSS and work out new fleet deployment and train and crew rosters – plans that would be adopted if and when the railway is electrified. The effect will be more benefits and better business cases.

To get reduced rail project costs, perhaps we need the second revolutionary change – and again major route electrification can serve as an example. Such investments drive a push to modernise: rationalising old track layouts (fewer track miles to electrify); immunising existing signalling systems; replacing high-risk level crossings; improving fencing. With plenty of new standards to consider, costs mount up. And there might be a need for additional capacity to accommodate expected growth (as reflected in the TSS) too.

Beneficial changes to train control systems typically have to wait years until the new infrastructure and train services are bedded in. But as David Waboso – head of Digital Railway at Network Rail – pointed out to me the other day, this could be a very expensive way of proceeding. In the case of the sub-surface lines of London Underground (Circle/District/Metropolitan) a couple of years ago, the prospect was for a huge renewal junction upgrade programme: high costs and highly disruptive. Why not, David reasoned, put in new train control systems first, and see what performance benefits flow from that (in terms of reliability, frequency, journey times)? Then what is needed in terms of civil engineering-based upgrades can be fine-tuned – and scaled back. And that's what London Underground did.

Metros are different from the national rail network which is much more of an open system. Cab-fitment of communications-based train control equipment is not so easy with right-to-roam freight locomotive fleets and the like. But DfT can discuss the challenge ahead with the freight companies. Having halved railfreight grant support, it will need to come up with some funding if the overall efficiencies from digital technologies are to be captured.

So, two revolutionary thoughts. Stop planning for fixed future year train service specifications and avoid the inevitable disappointment: major infrastructure schemes should surely allow the service offering to change! And implement digital train control systems ahead of deciding on the civil engineering needed.