Sheffield to Rotherham tram-trains Contents

1Tram-train pilot project

1.On the basis of a report by the Comptroller and Auditor General, we took evidence from the Department for Transport (the Department) and Network Rail on the Sheffield to Rotherham tram-train project.1

2.The Sheffield to Rotherham tram-train project will be the first passenger service in the UK to use the existing street tramway and national rail network. The Department first expressed an interest in this technology in 2007, having identified that tram-trains offer the potential to reduce the cost of transport services and create growth by improving access to city centres. It agreed with Network Rail that a pilot project would be the best way to see if the new technology could work in the UK and deliver these benefits. The Department selected the Sheffield to Rotherham route in September 2009 as it was considered to offer lower costs and less disruption than alternative routes.2

3.The tram-train project requires modifications to both the national rail infrastructure and tram network, as well as the purchase of new tram-train vehicles. The Department is the project sponsor and is responsible for overseeing overall progress with the project. Network Rail is responsible for the project to modify the national rail infrastructure. South Yorkshire Passenger Transport Executive (SYPTE) is responsible for modifying the tramway section and buying new vehicles, while Stagecoach Supertram will test the vehicles and run the passenger service.3

4.The Department’s rationale for the pilot project was based on its potential to deliver wider strategic benefits of testing the tram-train technology and rolling out similar schemes elsewhere in the UK. It identified that tram-trains could help reduce urban congestion by bringing more people into city centres via public transport. The Department accepted that the wider financial benefits of the project were very uncertain. Its assessment of the project business case also concluded that it offered low value for money. The project had a benefit-cost ratio of 1.0, based on the benefits to local transport users in the Sheffield area.

5.HM Treasury approved the project on an exceptional basis to allow a more detailed value for money assessment of tram-train schemes. After receiving HM Treasury’s approval, the Department gave the project the formal go-ahead in May 2012.4

Cost increases and delays

6.In May 2012, Network Rail forecast that its works to modify the national rail infrastructure would cost £18.7 million and be completed by the end of 2015. The Department and Network Rail agreed a budget of £15 million on the assumption that Network Rail could make efficiency savings during the project. However, Network Rail identified significant cost increases and delays in November 2014 and July 2016. By December 2016, it forecast the works would cost of £75.1 million, an increase of 401% on the original estimate, and it would not be complete until May 2018.5

7.Network Rail accepted that its early cost estimates for its work on the project were inadequate as the design and specification of the project were not advanced enough. It told us that it had experienced similar issues on the Great Western railway upgrade project but the Sheffield-Rotherham tram-trains project was riskier as it involved trialling technology that was new and untested in the UK. Despite this, neither the Department nor Network Rail prepared any quantified risk assessment for the cost estimates. Network Rail acknowledged that it did not understand the level of technical risk before 2014, some two years after the project was approved.6

8.The Department admitted that it had accepted Network Rail’s initial cost estimates without an adequate level of understanding of the project and without the necessary level of assurance on costs. The Department recognised that the nature of the project, as a pilot, meant that it was inherently risky, and that it would expect some costs not to become apparent until later in the project. But it accepted that it had been overly optimistic about the forecast costs and Network Rail’s ability to deliver efficiency savings. The Department told us that under current arrangements it would require independent assurance of the costing and schedule, and would challenge Network Rail harder.7

9.The Department and Network Rail told us there had been no deliberate manipulation of the forecast cost of the project in order to gain approval. However, both Network Rail and the Department conceded that, at the time, the regulatory environment and lack of public scrutiny of Network Rail’s accounts created circumstances where there was insufficient rigour in cost estimation. Network Rail told us that the regulatory oversight in place in 2012 allowed it to move away from its initial cost estimates, provided that it could demonstrate to the regulator that the actual cost was efficient. Network Rail conceded that this created perverse incentives and “was not an appropriate way to run a major capital programme”. The reclassification of Network Rail as a public body in late 2014 caused it to be much more disciplined in its cost estimation.8

10.Network Rail told us that it has changed its approach to developing projects which will prevent a repeat of making cost forecasts too early in the project cycle. It will now only provide cost and scheduling forecasts after completing adequate design and specification assurance work. Network Rail has also committed to developing a deeper understanding of the technology used in each project.9

The decisions to continue with the project

11.The Department reconsidered whether the project should go-ahead on two occasions when Network Rail reported delays and cost increases. The first occasion was in November 2014 when Network Rail reported the cost of works had increased to £44.9 million and would not be completed until October 2016. Network Rail reported that the works were more complex and the condition of assets was worse than it expected. The Department’s Permanent Secretary allowed the project to continue and agreed that the Department would provide full funding to cover the increase. The Department did not ask for a revised business case to inform this decision.10

12.In June 2016, Network Rail reported a further cost increase of £25 million and a further delay of 18 months. The Department’s officials recommended that the project should be cancelled at this point because many of the lessons about using tram-train vehicles in the UK had been learned and there was no commercial justification to provide further funding to the project The Rail Minister allowed the project to proceed. The then-Permanent Secretary did not seek a Ministerial Direction as he was satisfied that the strategic case was still valid and there was a respectable benefit-cost ratio, when based on sunk costs. The Department subsequently recalculated the benefit-cost ratio, which had fallen to 0.31, but did not request a revised business case.11

13.The Department told us that it had explicitly considered the increases in costs and the falling benefit-cost ratio on both occasions when the project was allowed to proceed. However, the Department’s main reason for allowing the project to continue was the broader strategic case; i.e. that completing the pilot project would facilitate future tram-train schemes. The Department conceded that, had it known the actual cost of the project at the outset, it would have provided greater challenge on Network Rail’s cost estimates, sought greater assurance over these estimates, and required greater proof to demonstrate the pilot’s broader benefits.12

1 Report by the Comptroller and Auditor General, The Sheffield to Rotherham tram-train project: investigation into the modification of the national rail network, HC 238 (2017−19), 4 July 2017

2 Q 66, C&AG’s Report, paras 1.3−1.8

3 C&AG’s Report, paras 1.1−1.2

4 Q 24, 25, C&AG’s report paras 2, 1.10−1.15

5 C&AG’s report, paras 2.15 & 3.2

6 Qq 2, 12,13, 19, 20

7 Qq 14, 15, 20, 63

8 Qq 16, 17, 22

9 Qq 8−11

10 Q 35, C&AG’s report, paras 4, 2.5−2.9

11 Q 35, C&AG’s report, para 2.15−2.20

12 Qq 38, 63, 71

13 December 2017