Understanding the PPP Arbiter's Draft Directions
Add to My Calendar

On 17 December 2009, the PPP Arbiter published an important document, which may turn out to be a staging post in the collapse of Tube Lines and – following 2007’s similar collapse of Metronet – of the Public-Private Partnership (PPP) itself.

As his own website explains, “The role of the PPP Arbiter was established by the Greater London Authority Act 1999. The Arbiter deals principally with disputes about the financial terms of the PPP Agreements. The Secretary of State appointed Chris Bolt as the first Arbiter in December 2002, and his appointment runs until 30 June 2011. Under the terms of the Act, the Arbiter is independent of Government and of the PPP Parties.” The Arbiter is part of the PPP set-up, and his four stated objectives are to: allow London Underground to revise its requirements; promote efficiency and economy; ensure that an Infraco that acts efficiently and economically can earn its rate of return ie. make lots of profit; and enable the Infracos to plan with reasonable certainty.

But despite being a pillar of the PPP regime, and probably without meaning to, the Arbiter’s latest document exposes many of the flaws of the PPP system itself. The Arbiter is advised by various private companies, including technical advisers Halcrow, which was part of Tube Lines at the time at which it made its PPP bid!

The “draft directions” is a lengthy document, but if you do not fancy wading through its 127 pages, here is a summary, together with comments from myself. (Paragraph and table numbers in brackets tell you where in the document that particular fact or quote comes from.)

The Background

The PPP is a 30-year contract, with review points every 7½ years. Tube Lines signed its contract to take over maintenance and improvements to the Jubilee, Northern and Piccadilly lines (JNP) on 31 December 2002, so its Second Review Period (RP2) is due to start on 1 July 2010.

The Arbiter assesses how much money Tube Lines should need to spend to meet its obligations (including its ‘obligation’ to make profit!) in RP2 and therefore how much it is allowed to charge London Underground. In December 2008, London Underground issued Restated Terms: changes that set out what work it wants Tube Lines to do in RP2. It also issued Affordability Constraints: an estimate of how much London Underground would be able to afford to pay Tube Lines for its services. On 30 June 2009, Tube Lines responded to the Restated Terms, disputing them. On 23 September, London Underground asked the Arbiter to give directions concerning the second 7½-year review period (para 1.1), particularly: the costs and revenues that an efficient Infraco could expect; the level of closures that would be needed; Tube Lines’ arguments about risk and financing impossibility; and the level of Infrastructure Service Charge (ISC) in RP2 (para 1.2). In parallel, Tube Lines referred to the Arbiter its contention that London Underground’s Restated Terms were technically unachievable because London Underground would not agree to close its services as often as Tube Lines wanted it to (para 1.12). Tube Lines also argued that the Restated Terms were financially impossible, did not comply with the PPP Agreement, and involved an increase in risk (para 1.15). Tube Lines submitted that its costs for RP2 would be £5.75 billion; London Underground disagreed, suggesting that Tube Lines costs for RP2 should be £4 billion (para 1.14).

’Notional Infraco’

The Arbiter uses the concept of a ‘Notional Infraco’, which he defines as “an Infraco with the same contractual obligations, third party contracts and financing arrangements as the actual Infraco, but which also carries out its activities in an overall efficient and economic manner and in accordance with Good Industry Practice.” (para 1.19) The Arbiter then compares the real Tube Lines with this ‘Fantasy Tube Lines’, in order to decide how much Tube Lines should be spending and earning compared with how much it actually is doing. Given that the Notional Infraco uses the concept of Good, rather than Best, Industry Practice (para 5.38), it should not be too difficult for a Real Infraco to match.

Jubilee Line upgrade

Tube Lines is upgrading the Jubilee Line’s signalling system, using a system called Transmission Based Traction Control (TBTC). In 2003, Tube Lines entered into a contract with the company Thales to deliver TBTC, based on Thales’ own SelTrac signalling system.

The Arbiter has found that “At the point of contract signature, however, significant elements of the approach, including the application of London Underground’s standards, were not settled and responsibility for many risk items not apportioned.” (para 2.11) The Arbiter looked at line upgrades carried out on underground railways in other countries, and concluded that a Notional Infraco would have adopted a different strategy from the one that Tube Lines did (para 2.12), “to commence procurement only when the functional specification for the project had been developed and agreed with London Underground” (para 6.23). In other words, Tube Lines should not have offered the job out until it actually knew what it involved! Rather, Tube Lines should have spent two years developing the conceptual design and procuring the contract before signing it, then a further two years on the detailed design for the upgrade (para 6.24). If it had done so, Tube Lines would have completed the Jubilee line upgrade by the end of RP1, at a lower cost, and would have made more progress with the Northern line upgrade during the first 7½ year period (RP1), so leaving less to do during RP2 (para 2.13) and finishing it by 7 January 2012, on time (para 6.25). In other words, Tube Lines did the wrong thing, and ended up spending more money than it needed to and unnecessarily delaying the Northern line upgrade.

How does Tube Lines explain the delays and costs of the Jubilee line upgrade? It argues that because the signalling contract was competitively tendered, the costs are efficient – as if by definition, organising a bid-out for a job guarantees it will be done at the best price. It argues that London Underground caused the delays by insisting on significant modifications to Thales’ original SelTrac design, by omitting items from its estimates (para 6.16), and by frustrating Tube Lines’ access to work on the Jubilee line. London Underground denied causing the delays, citing a report commissioned by Tube Lines itself (which you can read here) which gave six reasons for the delays, only one relating to London Underground (para 6.13). London Underground sees the real causes of the delay as “failure of planning and risk management by Thales and an equivalent failure by Tube Lines to manage its signalling contractor (para 6.14).

London Underground’s behaviour

The PPP expects that London Underground acts as a good client and “facilitates the delivery of the efficient work programme” (para 6.7). Tube Lines argued that London Underground has not done this; London Underground said that it had (para 6.8). Tube Lines argued that London Underground’s behaviour as a client increased its costs, with an impact greater than it could recover through claims (paras 2.16, 6.44).

The PPP Agreement has a Dispute Resolution Agreement, and Tube Lines made claims against London Underground totalling £727 million during RP1; further claims totalling more than £500 million are currently pending resolution (para 2.18). The Arbiter disagreed with Tube Lines: “That is not to say that its [LU’s] behaviour has not in some cases added to costs. Rather, the Arbiter considers that these are matters which are properly the subject of claims” (paras 2.17, 6.46).

Work in the Second Review Period

Line upgrades

Tube Lines has to upgrade the Northern line by 7 January 2012. It also has to upgrade the Piccadilly line, phase one (Cockfosters to South Ealing and South Harrow) by 11 October 2014 and phase 2 (Heathrow loop and Rayners Lane) by 10 October 2015. There is an option to extend the new signalling system beyond Rayners Lane to Uxbridge, but as this is an option, the Arbiter does not consider it (para 7.7). TfL, rather than Tube Lines, will finance the new Piccadilly line trains (para 2.36). The Arbiter has allowed Tube Lines some substantial “on-costs”, including 10% for project set-up costs; 15% for project management costs; and 30% for “factors specific to the Piccadilly line” (para 7.11). Stations Tube Lines now has to carry out barely a third of the station refurbishments in RP2 that it originally had to. The Restated Terms halved the number, from 100 to 50, and now the figure stands at only 38 (para 7.32). Just four stations will have the staff accommodation refurbished (papa 7.31).

Why? The Arbiter revealed that “stations work was underpriced by all bidders” (para 2.10). Because the private companies under-stated the cost of their work in order to get their contract, Underground passengers and staff must continue to put up with 62 shabby stations which should have been refurbished.


Tube Lines’ approach is based on replacing key components but leaving the existing escalator truss in place (para 7.40), whereas London Underground advocates replacing escalators (para 7.42). London Underground refers to plans for 14 “machine-room-less escalators” (para 7.41), lacking the room which all escalators currently have, from which staff access the physical workings of the escalator. Union representatives for both station and engineering staff are concerned about how workers can safely operate and maintain escalators without a machine room. Tube Lines and London Underground have jointly submitted that the notional asset life of the Jubilee Line Extension’s escalators should be extended from 25 to 40 years, and the “major intervention cycle” to 14 years (para 7.42). This could see escalators kept in service rather than replaced when they are aged and unreliable.


To get its work done, an Infraco needs London Underground to sometimes close lines, sections of lines, stations, parts of stations, lifts and escalators. An Infraco would want the maximum access to work sites and therefore a high allocation of closures, whereas London Underground should want to minimise disruption to passengers and therefore to limit closures. The extent of closures is measured in Lost Customer Hours (LCH). London Underground suggested 15.5 million LCH for Minor Closures and 6.1 million LCH for Lifts and Escalator Closures, but Tube Lines wanted more – 35.6 million LCH for Minor Closures and 9.4 million LCH for Lifts and Escalator Closures (table 11.2).

The Arbiter decided that a Notional Infraco would learn from international practice and therefore need fewer closures for the Piccadilly line upgrade than for the Jubilee and Northern upgrades (para 2.53). Metro de Madrid and Metro de Paris both introduced major signalling upgrades with significantly lower amounts of access than similar projects in London (para 11.19), and although London’s requirements are more onerous and access more restrictive (para 11.22), these examples are still relevant (para 11.21). The Arbiter concluded that Tube Lines should be allocated 21.5 million LCH for Minor Closures and 7.8 million LCH for Lifts and Escalator Closures (table 11.2), significantly less than Tube Lines wanted. Tube Lines has referred its dispute over Base Allocation of Minor Closures to external adjudication (para 5.17).

Costs for Second Review Period (RP2) 

Costs are split into Base Costs and Overlays. Tube Lines submitted that its Base Costs would be £5,189 milliion; London Underground suggested £3,621 million; the Arbiter ruled £3,943 million (table 2.2). RP2 costs allow for the London Living Wage (para 7.38), so there is no excuse for cleaners, or anyone else, to continue being paid less than this. Tube Lines claimed that it needs £21 million during RP2 to prepare for RP3 (the third review period). The Arbiter responded that it needs just £10 million (para 7.65). Tube Lines claimed that it needs £217 million for overheads during RP2; London Underground reckoned £113 million; the Arbiter settled on £151 million (para 7.66).


In its Restated Terms, London Underground introduced an indemnity for Infracos’ pension costs. Tube Lines’ contribution rate is capped at 21.5% of pensionable salaries, with any excess repaid by London Underground (para 9.3). But with Tube Lines’ section of the Transport for London Pension Fund in deficit (table 9.1), the Arbiter has allowed costs of £73 million in RP2 for a recovery plan. Tube Lines had asked for £60 million; London Underground suggested nil (para 9.18).

Overlays are added in respect of efficiency, differential inflation and risk (para 2.26).

Efficiency and Secondment fees

The Arbiter has caught Tube Lines paying over-the-top “secondment fees” to its own shareholders, Bechtel and Amey, describing these as “high and potentially outside of market norms for this sort of contractual arrangement” (para 6.31). The Arbiter dealt with this by assuming an Infraco would only pay such high fees if doing so led to efficiency savings elsewhere (para 6.33). The Notional Infraco’s central costs are £243 million lower than Tube Lines’ actual costs.

‘Differential Inflation’

The PPP uses RPIX (Retail Price Index, excluding mortgage payments) as its inflation figure. But it also protects Infracos financially if the real prices of things that Tube Lines spends its money on, such as materials, rise by more than official inflation. This is known as ‘differential inflation’, and would be a very useful concept for workers’ pay negotiations. However, when RMT representatives have argued for real price rises above the official inflation rates to be considered in pay talks, management have generally dismissed the idea! However, it seems that even this protection is not enough, as Tube Lines and London Underground are now discussing protecting Tube Lines from differential inflation differing from predictions (para 1.18).

The Arbiter assessed that Tube Lines needs just over £170 million for differential inflation during RP2 (para 2.30), plus a “significant risk premium” of £160 million (para 2.33). I can find no evidence that if real prices go up by less than official inflation, then Tube Lines has to give money back!


London Underground argued that the PPP set-up protects Tube Lines from risk through the Extraordinary Review process, and that “the process is asymmetric and there is therefore a potential windfall gain for the Notional Infraco if outturn costs are lower than forecast” (para 7.109). So PPP protects the Infraco against losses if its gambles fail, but allows it to pocket the proceeds if things go well. Either the Infraco wins or the public sector loses. Tube Lines proposed a risk allowance of £302 million, “the minimum that the lenders’ technical adviser would accept” (para 7.115). But the Arbiter believed that Tube Lines “overstated” the increase in risk in some areas, and London Underground asserted that Tube Lines failed to consider contract changes that reduce risk (para 12.10). The Arbiter assessed £200 million for asset risk (para 2.32) and £59 million for corporate risk (para 2.34).


Had the Infraco delivered the Jubilee line upgrade on time, it would have received £15 million bonuses (para 8.17). Revenues for the Northern line upgrade will be £15 million; for the Piccadilly, £16 million (para 8.18). Simply doing these jobs on time would earn Tube Lines bonuses of £286 million for the Northern line (para 8.19) and £171 million for the Piccadilly (para 8.20). London Underground suggested that the Infraco could deliver the Northern line upgrade early and receive a £325 million bonus (para 8.19).


Adding Base Costs and Overlays, Tube Lines reckoned it needs £5,766 million for RP2; London Underground suggested £3,991 million; the Arbiter ruled that the appropriate sum is £4,394 million (table 2.3) – much closer to London Underground’s view and nearly £1.4 billion short of what Tube Lines says that it needs. This is the key calculation that now pushes Tube Lines into financial crisis and the wounded, staggering PPP to the edge of the cliff. This may well hasten the return of Tube maintenance to the public sector.

Who pays?

London Underground pays Tube Lines for its work via the Infrastructure Service Charge (ISC). The Arbiter calculates the total that London Underground needs to pay by adding to his assessment of RP2 costs (£4,394 million) amounts for Treasury fees, inflation and ‘fixed amounts’, giving a total of £7,104 million (table 3.1). However, London Underground also gives its Affordability Constraints, its “best estimate” of what it will be able to afford to pay in ISC (para 3.4), as £6.412 million. This gives an ‘affordability gap’ of £463 million (table 3.1).

So who pays? Both London Underground and Tube Lines “consider it would be better value for money for TfL to raise additional finance than for Tube Lines to do so” (para 3.10). The government argued the case for PPP on the basis that the private sector could raise the money needed to maintain and improve the Tube more easily than the public sector could. Intentionally or not, London Underground’s and Tube Lines’ conclusion here undermines the whole case for PPP. Even if Tube Lines raised the money, London Underground would need to repay it through a higher ISC (para 3.10). Moreover, because Tube Lines is classified as a public corporation, any new debt it raised would count on the public sector’s balance sheet, and would count against TfL’s funding from the government (para 14.7).

So whether TfL or Tube Lines raises the finance, the public sector still pays. Again, this undermines the case for PPP, which was largely motivated by a desire to get London Underground funding “off the public sector’s books”.

Damning Tube Lines?

The Arbiter’s document reflects badly on Tube Lines in several respects. He rejected Tube Lines’ claim that the Restated Terms are technically unachievable (para 15.12). He judged that Tube Lines had made major mistakes in the Jubilee line upgrade project, had overpaid its own shareholders, and had not acted as an efficient Infraco should. The Arbiter refused to accept either Tube Lines’ assessment of its costs for the next 7½ year period, or the extent of its demands for closures to enable access.

However, the Arbiter also makes several judgements that do not bode well for services or for workers. He accepts Tube Lines’ reduction in inspection frequencies, as “consistent with the Notional Infraco continuing to reduce costs toward the levels achieved by an Infraco applying Good Industry Practices.” (para 7.15) Tube Lines wants to employ a larger response team than the Arbiter thinks is appropriate, and to run ten maintenance depots (para 7.17). The Arbiter also advocates improving the “flexibility” of the workforce by having them work on all lines rather than one (para 7.19).

Not surprisingly, Tube Lines disagrees with the Arbiter’s conclusions (statement, Tube Lines media centre, 17 December 2009). Chief Executive Dean Finch complained that “The figure of £4.394 billion is very demanding for the work that has to be undertaken by us. A settlement at this level is not conducive to private sector involvement in the Underground, nor does it reflect the reality of the Underground working environment.” Perhaps Mr Finch believes that working in hard conditions with limited resources is the public sector’s job!

What happens next?

The deadline for responses to the draft directions is 1 February 2010, and Tube Lines has already pledged to make representations “robustly” (statement, Tube Lines media centre, 17 December 2009). On 4 March, the Arbiter will publish his final directions and guidance on risk, costs, performance revenue and access allowances; plus draft directions on financial issues (ISC, base/eligible finance, ability to finance, fixed amounts and RPIX).

The deadline for responding to these draft directions is 6 April, and on 29 April, the Arbiter will publish final directions. By then, Tube Lines, and the PPP itself, may be history – or they may still stagger on.