House of Commons - Environment, Transport and Regional Affairs

Select Committee on Environment, Transport and Regional Affairs Appendices to the Minutes of Evidence

Memorandum by Declan Gaffney and Dr Jean Shaoul (FLU 06)

Questions raised by our analysis for the Minister, London Underground and their financial advisors (the numbers in brackets refer to the paragraph(s) in the submission):

1. Will the Minister/London Underground put the output specification and the contract in the public domain so that it will be clear exactly what work is to be carried out under the PPP? (paragraph 52)

2. Will the Minister/London Underground provide a detailed explanation of the work to be carried out with the financial costs so that a comparison can be made with past expenditure and investment schedules?

3. Our analysis, using estimates provided by PwC, revealed an affordability gap of £175 million in year two that was set to rise (paragraph 37). How will the affordability gap be bridged? Does the government intend to provide a subsidy of about £2.5 billion for the PPP as suggested by recent press reports? If so, why, since giving the same amount of money to LU without the PPP would ensure that it was spent on enhancing the network not the financial institutions?

4. If the government does not intend to bridge the affordability gap between the PPP and LU's cash flows, will the Mayor have to increase fares above the rate of inflation, something that LU has said will not happen, or divert funds from other sources? (paragraphs 46-47)

5. Will the Minister/London Underground put the full business case, including details of the risk assessment in the public domain?

6. Which risks are included and excluded? (paragraphs 49, 53 and 61)

7. What guarantees or assurances will be given to the PPP's providers of finance? What are the implications for the Mayor and TfL? (paragraphs 55-56)

8. There have been a number of failures of high profile IT PFI schemes that failed to deliver, where the public sector agencies bore significant financial costs and the risk was in effect transferred back to the public sector. Given this, how will the contracts protect London Underground in the event of a failure on the part of an Infraco(s) to deliver to specification when those for the IT projects did not protect the public sector? (paragraphs 49-54)

9. How will performance standards be defined for monitoring purposes? (paragraph 52) How will London Underground be able to monitor the performance of the Infracos, a time consuming task, if 6,000 staff, more than one third of the total workforce, are to be transferred to the PPP?

10. In the light of the deteriorating safety records of the privatised railway industry and London Underground, what steps will be taken to ensure that the Underground's safety is not jeopardised by the separation of functions and responsibilities for the infrastructure and passenger services? How will responsibilities be defined where the infrastructure overlaps between the three Infracos? (paragraphs 40 and 57)

11. Given that the investment proposals in effect do little more than prevent the tube collapsing, while leaving several threats untouched, how does the government propose to provide the funds for a modern, efficient Underground system? (paragraph 32)

12. How are additional investment projects to be awarded once the PPP is operational? What further legal difficulties are posed by additional contracts to other contractors? (paragraph 41)

1. We welcome the Committee's new inquiry into the funding of London Underground (LU) in the context of the government's proposals for a public-private partnership (PPP) to finance LU's maintenance of the infrastructure. We would like to take this opportunity to provide some financial analysis and other background information to assist the Committee. The purpose of this report is to:

— show how LU has been funded over the last 15 years, its financial performance, labour productivity, maintenance and investment programme (paragraphs 5-19);

— examine the funding of other major metro systems in New York and Paris (paragraphs 20-25);

— assess the degree to which the problems confronting LU are in fact generic to infrastructure industries rather than specific to and a result of LU failings (paragraphs 26-29);

— evaluate the extent to which the PPP proposals address and resolve the problems facing LU (paragraphs 30-46);

— estimate the affordability and risks of the PPP (paragraphs 47-58);

— assess some of the alternatives to the PPP (paragraphs 59-71);

— draw some conclusions (paragraphs 72-76);

— make recommendations about the future organisation (paragraph 75) and funding of London Underground (paragraph 77).

— raise a number of specific questions about the PPP that require detailed answers.

This submission is based on our report, Funding London Underground: Financial Myths and Economic Realities. That report was prepared for Listen to London and contains further supporting detailed financial evidence and tables based on London Underground's accounts and the briefing document prepared in December 1999 by Price WaterhouseCooper (PwC)[4], London Underground's financial advisors, plus some additional research. Our report has been publicly available since the beginning of February 2000 and none of the financial assumptions, calculations or conclusions have been contradicted by either LU, its financial advisors or the government. On the contrary, all the subsequent public discussion and announcements by government have confirmed our analysis.

3. The problems confronting LU are widely acknowledged and are not repeated here. They stem from a lack of investment, and the government proposes to redress this by a public private partnership (PPP) for London Underground. Thus the government presents the PPP as a means of replacing government funding and improving passenger services. The public-private partnership proposals involve firstly, dividing LU's operations into two. Operating the infrastructure—the network of tracks, tunnels, stations, signals and rolling stock and so on—will henceforth be separate from running passenger services. Secondly, the maintenance, renewals and capital investment in the infrastructure will be out sourced under three separate 25 to 30 year concessions to private sector consortia after a competitive bidding process. 6,000 LU staff who currently perform these functions will transfer to the infrastructure companies. Thirdly, the PPP is not intended to expand the existing network. "The Infrastructure Services [to be provided] will not include provision of new Underground lines[5], but to deal with the backlog in maintenance and renewal of the infrastructure of £1.2 billion and to upgrade or replace existing assets. Government has given no indication of how any future expansion of the network, such as completion of the long-delayed East-West Crossrail, is to be either financed or funded nor how other problems, such as London's rising water table, which threatens the Underground is to be managed. Fourthly, the rump of LU will run the passenger services and pay an annual fee or Infrastructure Service Charge (ISC) for the maintenance of the network to the three infrastructure operators (Infracos). The scheme involves no transfer of assets from LU to the Infracos. Rather, it rests upon a complex set of contractual relationships not only between London Underground and the infrastructure companies but also between the three Infracos, since their use of some of the lines, signal, stations, etc, will necessarily overlap. Each Infraco is a special purpose vehicle or consortium of companies, which in turn will have further contracts with the providers of finance. These proposals will be evaluated in some detail later.

4. It is important to note that the government justifies its policy of PFI/PPP in general and the Underground PPP in two ways. The private sector will provide firstly, the investment in the infrastructure that the government cannot afford, and secondly, the private sector managerial skills that will lead to increased efficiency, lower costs and benefits to the travelling public. The clear implication is that inefficiency is one of the major causes of the problems that daily confront passengers and that as a public sector agency, LU is less efficient than the private sector. We therefore examine LU's performance next.

5. LU's funding and financial performance Since 1990 London Underground has been funded from two sources: passenger fares and central government grant. Fares revenue increased in absolute terms and as a proportion of income since 1984-5, partly due to increased demand and partly to above-inflation fare increases (Table 1). Travel patterns and trends are closely related to the level of economic activity with more than half of all journeys each day directly connected with work.

6. Rail revenues rose from £333 million in 1983 to £565 million in 1993, an increase of 70 per cent, less than the rate of inflation. As Table 1 shows, from 1984 to 1989, the number of rail passenger journeys rose by 45 per cent as Britain came out of recession. The number of journeys fell by 1 per cent during the recession of the early 1990s. After initially falling, rail fares per passenger km rose towards the end of the period, from 7p per passenger km in 1989 to 9.8p in 1993, thus compensating for the decline in passenger volume. This increase of 40 per cent was well ahead of the 23 per cent rise in the Retail Price Index (RPI). Thus until 1990, the increase in rail revenues was largely the result of rising demand. Between 1990 and 1993, the revenue increase was the result of fare increases. Rail revenues rose from £642 million in 1993-94 to £1,009 million in 1998-99, an increase of nearly 60 per cent, considerably more than the rate of inflation of 19 per cent for the same period. By 1999, the number of passenger journeys had increased by 20 per cent and reached the highest level ever, as Britain climbed out of recession. It was, however, also due to the increase in fares: fares per passenger km rose from 11p per passenger km in 1994 to 15p in 1999, a 27 per cent increase.

7. Thus, the Underground is a cyclical industry in which patterns of demand are closely tied to movements in the economy. Demand decreases substantially during recession. The continuous rises in LU's rail revenues both during and after the 1990s recession depended on above-inflation increases in fares.

8. Grants and subsidies have traditionally provided a significant proportion of funding for LU and other urban rail networks. The Greater London Council (GLC) funded 66 per cent of LU's grant through a precept on the rates. Since the abolition of the GLC, grant has been allocated by central government. Central government grants have been made to cover three expenditure items: (i) revenue subsidy for passenger services; (ii) renewal of the infrastructure, ie maintaining and replacing existing infrastructure assets; and (iii) enhancement and expansion of the network.

9. Revenue subsidy: This ended in1994. Since then, grants have only been made for maintaining and enhancing the infrastructure. To put it another way, grants would only be available for the network not for passenger services.

10. Grant for renewals: Since 1995-96 the grant for infrastructure renewals has declined from £398 million in 1994-95 to £160 million in 1998-99 (Table 2, column 1). For the last two years, the grant has not covered the cost of maintaining the network. Thus, fares have had to rise to absorb the full cost of running both passenger services and a significant and increasing proportion of the cost of infrastructure maintenance. According to the most recent estimates of government department spending, the grant for renewals (approximately £156 million in 1998-99) is set to disappear altogether, thereby constraining the amount available for capital maintenance, without a further round of cost cutting.

11. Grant for capital expenditure: In the allocation of government grant, funding for enhancement of the existing network is in general kept separate from funding for major developments such as the Jubilee Line Extension, which is ring-fenced by government. The Underground was starved of investment until the 1990s when there was a very modest increase. However, much of this capital expenditure went on the new Jubilee Line Extension and other new, much needed lines rather than on the existing lines. Table 2 (columns one and five) shows that according to London Transport's accounts, of the £6.32 billion government grants for capital expenditure since 1990-91, 40 per cent has gone on the Jubilee and other new developments. While the grant for capital expenditure on existing lines, including maintenance, has declined from £533 million in 1990-91 to £143 million in 1998-99 (Table 2, column 3).

12. LU has made significant reductions in operating costs. Like any railway industry, while external purchases of goods and services are low, a major internal cost is labour. Labour costs rose from £246 million in 1985-86 to £581 million in 1992-93, (Table 3) and accounted for 60 to 73 per cent of total internal costs (value added), more or less the norm for British manufacturing industry. Employment fell from 21,598 in 1985 to 19,000, a fall of 14 per cent in 1993. Between 1993-99, employment fell by a further 3,000 (16 per cent) to 16,000. Most of the jobs lost were in engineering services as the number employed fell by half from more than 6,000 in 1993 to 3,000 in 1999 (Table 3). Labour productivity, as measured by the number of passenger km per employee, rose from 0.278 million km in 1992 to 0.423 million km in 1999. As a result, between 1993 and 1999, labour's share of internal resources (value added) fell to an average of 55 per cent.

13. It is worth noting that labour's share of value added is significantly lower than that of the former British Rail (84 per cent in 1994). Yet British Rail had a labour productivity that was much higher than the average labour productivity in Europe railways and its subsidy was much lower.[6] It is therefore likely that LU's labour productivity will also compare well with its European counterparts.

14. Historically, the renewal of the infrastructure was funded by government grant, which has declined in the last period. For the last six years, renewals have been running at about £200 million per year. For the last two years, grants for renewing the network has been less than actual expenditure on renewals. In 1998-99, the cost of renewals was £244 million, considerably more than LU's grant of £108 million (Tables 4 and 2). In other words, the decline in infrastructure renewals is directly the result of successive government policy aimed at curbing public expenditure rather than LU's cutbacks on infrastructure spending.

15. Investment in new or enhanced assets during the 1980s was only slightly greater than the annual depreciation charge, ie LU was barely replacing its worn out assets, let alone enhancing the network. Since 1993, new asset formation has averaged £680 million per year. However, much of this has been for new lines, such as the Jubilee Extension rather than existing lines. In all, for the last two years, the maintenance and investment grants for existing lines have been less than actual expenditure, as evidenced by the negative sign in column 6 of Table 5. In other words, LU was funding some of the expenditure on maintenance and investment from its own operational savings (Table 6).

16. Nevertheless, expenditure falls far short of LU's requirements. There is a backlog of work arising from past under-investment of about £1.2 billion. Therefore, assets are being used beyond their design life with corresponding increases in maintenance and inspection costs and increasing unreliability. Expenditure of about £150 million a year for about seven to eight years are required to remove the investment backlog with a further £400 million a year required to maintain the system in a "steady state" and prevent the backlog reappearing.[7]

17. The split of operational expenditure on the infrastructure and train operations is crucial, since this issue lies at the heart of the PPP proposals to outsource the management of the infrastructure and rolling stock. The money spent on the infrastructure will be used to pay the ISC. However, unlike the former British Rail, LU provides no details of the cost split. Prior to the privatisation of the national railways, the infrastructure accounted for approximately 50 per cent of revenue costs. There is no reason to suppose that LU would be substantially different. The significance of this will be developed in paragraph 45.

18. LU is a public sector corporation with no interest or divided obligations or statutory requirement to make a return on capital employed. Since it does not generate a surplus either to distribute to external stakeholders (interest or dividends to the providers of finance) or to pay tax, it spends the entirety of what remains after paying for purchases and labour, on maintenance (Table 5). More recently, any surplus after covering the cost of renewing the infrastructure renewals has been ploughed back into the business in the form of investment. LU would therefore be incapable of servicing debt if grants were to be replaced by interest bearing debt without subsidies, fare increases or increased labour productivity. Neither would it be able to afford any extra charges to cover the cost of the PPP.

19. This may not be readily apparent as LU's accounts can easily be misunderstood. LU classes the revenue remaining after paying for purchases and labour—but before capital maintenance—as "gross operating margin". Numerous commentators have taken this figure, positive for several years to mean that LU makes a profit. They have further concluded that LU would be able to meet the cost of debt service under the PPP or a bond issue with no need for external subsidy. This is potentially very misleading. Profit is what remains after providing for depreciation and capital maintenance. After taking these charges for capital maintenance into account, LU does not make, and indeed does not report a profit. Furthermore, it is only in the last two financial years that LU has reported positive cash flows from operating activities.

20. We examine next the funding of metro systems in other major cities. Contrary perhaps to expectations, governments everywhere have provided funds for their transport systems. This is because the capital-intensive nature of the industry makes it impossible to recover the full cost of running a metro system from user charges. Consequently, the funding of Paris and New York, cities with very different political structures and traditions, provide illustrative examples. The two organisations concerned, corresponding to London Transport or Transport for London (TfL), are transport authorities that cover both bus and suburban rail.

21. Paris: The Parisian transport authority, the Syndicat des Transports Parisiens, is funded through a levy on employers, the "Versement de travail". It also receives half the revenue raised from traffic and parking fines. The proceeds of the Versement de travail are used to subsidise fares and fund capital maintenance. Central government provides the main source of funds for line extensions and new lines with contributions from local and regional government. In 1998, passenger charges contribute only 34 per cent of the costs of public transport. The various transport companies generated 42 per cent of total revenues. Local and regional tiers of government were the main source of the remaining 58 per cent.

22. New York: The New York Metropolitan Transportation Authority (MTA) is funded through a combination of user payments and subsidies for both operating and capital costs. In 1997, fares revenue accounted for 52 per cent of funding. The State and the City of New York provide grants and subsidies, which account for 46 per cent. A large proportion of the subsidy comes from hypothecated taxation, which is in principle at least targeted at the non-fare paying beneficiaries of the system. These include "new charges on business and property owners, taxes on commercial rents, business telephone charges [and] property transfers."[8] Because of these assured future revenues, the MTA is able to raise finance through the revenue bonds that have attracted such attention in the media here (paragraphs 64-65). The New York MTA receives considerable subsidy for operating costs as well as infrastructure costs. "In calendar year 1998, State and City operating assistance, special tax-supported subsidies and reimbursements for the Transit System constituted, on a cash basis, approximately 38 per cent of the total revenues . . . not including reimbursements for capital expenses or amounts used for debt service."[9]

23. Sophisticated arrangements exist for spreading the burden of costs between passengers and beneficiaries of the system and between different levels of government. In both cities, transport companies are allowed to borrow and receive subsidies to meet the costs of debt service. Thus in neither case is national taxation the sole source; rather, forms of local taxation are used to raise revenue from people who benefit from the existence of an urban transport system. The costs of public transport are thus spread between users, beneficiaries, regional and national taxation in both the Paris and New York systems.

24. An important consideration here is the highly centralised public expenditure system in the UK where most of the taxation is raised by or under central government control. Central government also exercises final control over the application of funds, leaving considerably less room for the kind of distribution of costs that takes place in Paris and New York. It should be noted that until relatively recently the funding of the Underground was much closer to the systems adopted in Paris and New York. However, with the abolition of the GLC in 1984 the system of domestic rates in 1990, control of funding for London Transport was taken over by central government, which then reduced both the types and level of funding. Prior to the abolition of the GLC, the proportion of public funding was almost identical to the present arrangements for the Paris metro.

25. The devolution plans for London will not create the kind of structures whereby arrangements similar to those in Paris and New York can be developed. Central government, when setting up the GLA, as in its fiscal settlements for the devolved assemblies in Scotland, Wales and Northern Ireland, retained control over both the level of resources available and the way resources can be used. In particular, the Treasury restricts capital expenditure by controlling the amount of borrowing by public authorities. This restriction has led to PPP/PFI becoming "the only game in town" for public bodies seek to renew their crumbling infrastructure. Although the Mayor will have the power to levy new taxes on road traffic, borrowing against these revenues will remain under central government control. Consequently, the methods used to fund the metro systems in Paris and New York are not available. The funding of the Underground is therefore as much a political issue concerning the relationship between central and devolved government as it is about cost sharing and equity.

26. The essential problem for railways, including the Underground, is that they are unable to generate the revenues from fares that cover the full costs of and investment in the infrastructure and train operations, ie distribution network and passenger services. This is despite the fact that fares are among the highest in Europe. A single adult fare of £1.50 for zone 1, the inner zone, is higher than the fare for one journey of any length in both Paris (50p) and New York (£1). Full cost recovery would entail fares that would choke off demand, bring economic and social life in the capital to a standstill and be politically unacceptable.

27. While LU has significantly increased labour productivity, and it is difficult to increase it further without additional investment, no amount of increased productivity can release the resources to fund the necessary levels of investment as the nationalised railways showed, LU has had to cover an increasing proportion of its costs from user charges as successive governments have scaled back the subsidies for passenger services, and cut back on grants for maintenance and investment in the core business. However, despite greatly increasing the contribution it makes to infrastructure costs from fares income, LU is currently faced with a backlog of £1.2 billion, representing infrastructure assets (track, signalling, rolling stock etc) which are being used beyond their design life.

28. The problem is not so much one of efficiency but the capital intensive nature of the industry where a high proportion of the costs is associated with providing the network, relative to the ability to generate income. In this, the Underground and the railways are similar to other public utilities such as gas, electricity, telecoms and water where the main costs relate to the distribution network. These industries, are however able to recover their full costs because they have a very different demand pattern. They are essentially universal providers.

29. Thus the problem is an income one. Yet under current government expenditure plans, grants will play no role in the future funding of the Underground with the possible exception of new lines and extensions, meaning that all operating and infrastructure costs will have to be met by passengers. The Underground will then be the only metro system in Europe where the entire cost of maintaining services is borne by passengers.

30. Under the concession agreements, each of the three PPP companies will charge London Underground a basic "Infrastructure Service Charge" (ISC) which will be fixed for the first eight year period. The agreements are to be subject to periodic reviews at roughly eight year intervals, where the level of charges for the next period will be subject to negotiation not regulation. Although little detail has been made available on the intended form of the concession agreement, it seems clear that the review process is not intended to allow LU to end the concession agreements or to award them to other contractors within the 25 to 30 years of the agreement. In the event that LU and a PPP company fail to agree a level for the ISC, a Statutory Arbiter will set the ISC. The Mayor, TfL or LU will not be able to end any of the 30 year agreements, unless of course a PPP operator fails to perform adequately against the performance specifications, an issue we will return to later.

31. The ISC will depend upon the amount of work to be carried out under the concession agreement and the PPP's costs, including the cost of capital. The Infrastructure Service Charge (ISC) will be set for the first eight year period but "stepped up" when major enhancements are delivered.[10] Thus the ISC will increase several times during the first eight years. While as yet there is no information about the likely cost of the ISC, we estimate this by reference to the amount of investment, and the amount of private sector finance required after taking into account the cash available from LU after paying for passenger services. This information is presented in Table 7, provided by LU's financial advisors.

32. There has been some confusion about the amount of investment projected under the PPP. As we explained at the outset, the plans, although not insubstantial, are for little more than day to day maintenance, the £1.2 billion backlog of maintenance, and some minor works to maintain the system in steady state and prevent the backlog from reappearing. In other words, these are not proposals that will radically transform the Underground into a modern, efficient system but will simply prevent the system collapsing. Government and others have presented various figures, ranging from £7.5 billion to £15 billion as to the levels of investment projected under the PPP. Because LU's financial advisors believe that the private sector, despite never having run an Underground system before, can make efficiency savings that LU cannot, they estimate the PPP companies will spend £12.53 billion over the first 15 year period (Table 7). Of this, however, £4.15 billion consists of day to day maintenance and other operating costs. The planned investment is, on the most recent projections, £8.38 billion, an average of around £560 million a year.

33. The funding available for the PPP, the difference between LU's income (Table 7, line 1) and its costs (Table 7, line 2), is the "operating cash flow" (Table 7, line 3). Private finance (Table 7, line 5) bridges the gap between this available funding and the PPP's expenditure on the infrastructure (Table 7, line 4). Since LU's advisors project that rail receipts will rise by 40 per cent over the 15[11] year period, most of the investment will be funded directly by fares, leaving only £2.44 billion to be borrowed. However, this borrowing, which is effectively substituting for government grant, will have a significant impact on London Underground's costs.

34. The ISC will be set so as to recover the costs of the PPP companies. The main costs to be recovered are those of maintenance and investment in the network (Table 7, line 4) and the costs of dividends on shareholders' capital and debt service interest (line 6) and repayment of principal. Repayments are not shown as PwC does not include principal repayments for each year. PwC estimate an annual "cost of capital" of 10 per cent. PwC's figures therefore imply an affordability gap: the sources of funding do not match the costs of the PPP in each year up to year 14. In other words, there is a shortfall equivalent to the interest and dividend payments in each of the 15 years.

35. We have estimated the cost of repayment of debt principal taking year two of the contract as representative of the first eight years. In year two of the contract, PwC calculates interest and dividend payments of £110 million. A 10 per cent cost of capital implies that the capital raised by year two is around £1.1 billion. Assuming 20 per cent of this is in the form of equity, this leaves a gross debt of £980 million. If this debt is paid off evenly over 15 years, then £65 million is required in year two, in addition to interest and dividend payments of £110 million. Thus the PPP's total costs are £1,075 million as shown in Table 8. Of the £1,075 million, external borrowing will provide £360 million, leaving £715 million to be raised through the infrastructure charge.

36. However, the funding stream identified by PwC is not enough to cover the ISC. In year two for example, PwC predict operating cash flows of £540 million and borrowings of £360 million. This leaves an affordability gap of £175 million as Table 9 shows. Even this does not take account of the extra costs of financing and legal fees incurred by the private sector in undertaking the PPP and securing investment, already estimated at £26 million by one bidder, and which they would seek to recover through the ISC. Neither does it take account of the requirements of the PPP companies to build up reserves. In other words, it is likely to be an underestimate. The PPP might be able to defer but not resolve the funding problem by the use of sophisticated financing mechanisms.

37. In short, the PPP is not affordable to LU on the basis of available revenues (income less costs). By year two, there is a gap of nearly £175 million between available revenues and the level at which the PPP charge would have to be set to allow for debt service and returns to shareholders at the beginning of the contract period. There is no reason to expect that gap to diminish as the private sector takes out more debt and seeks to recover the cost via the ISC. This gap can only be met by increasing fares, which are already higher than in other comparable cities, or increasing the subsidies as in the case of the national railways. While there have been press reports[12] that the Treasury will provide a subsidy of £2.5 billion, there has been no explicit announcement to this effect. If this is the case, this means that our calculations are indeed correct. In that case, the government had either not worked carefully through the financial implications of the PPP for LU or had hoped that no one would notice the shortfall until after the contracts were signed when it would become the responsibility of the Mayor and TfL. But even more importantly from the perspective of financial prudence, such a sum would finance the Infracos' interest and dividend charges necessitated by the PPP, not improvements to the Underground.

38. But these projections of future revenues make assumptions about fare income, capacity and costs that are not unproblematic. If we consider fare income first, the projections assume constantly rising income from fares without any increase in the fares level (price). In LU's projections, it is assumed that this will be affordable because of increased revenues. However, up until the first review period, the ISC will rise whether fares revenue rises or not. The ISC will be the first claim on LU's revenues, so any shortfall will be borne by LU, not by the PPP companies.

39. The revenue generated by fares depends on four factors: the demand for the service, the capacity of the network, the level at which fares are set and the structure of the fares system. The two last factors will, once the GLA takes over TfL, no longer be within LU's control. Fare receipts are expected to rise from £997 million in 1998-99 to £1,170 million in two years time and to keep rising every year thereafter. LU's revenue projections for the next 15 years assume no increases in fares more than 1 per cent above inflation after 2001-02. However, they do assume a real increase in fares revenue of 40 per cent due to increases in passenger traffic. The case for the financial viability of the PPP proposals turns on this increase in revenue from passengers.

40. Unless all this extra demand is confined to off-peak times, these revenue projections clearly depend on increases in network capacity. At current capacity levels, LU has difficulty meeting existing demand at peak hours. The increase in the number of passenger places has fallen far short of the increase in the number of passenger journeys (Table 1), leading to severe overcrowding. Overcrowding in turn reduces capacity through the delays it introduces at stations (jammed doors, struggles to get on, off). It can also contribute to injury rates. In public transport (bus or rail), injury-producing accidents can occur without vehicle or train collisions. Sudden or sharp breaking and acceleration can cause accidents as passengers fall, particularly in overcrowded trains with standing passengers. In this context it is worth noting that serious injury rates have doubled in the last six years (Table 10).

41. The Central London Study commissioned in 1988 concluded that to increase capacity to meet forecast demand in the central area, two new lines would be required.[13] Yet the PPP proposals do not involve significant expansion of the network. The new lines which, at least until recently, have been regarded as essential to reducing unacceptable levels of congestion in the central area would have to be funded separately, and no plans to undertake this work have been announced. Furthermore, the complexity of the legal arrangements if the contracts were not awarded to the existing concessionaires would be enormous. But that in turn means that the Infracos would be the preferred candidates for any new lines of substantial investment programmes on existing lines, with all the consequent implications for monopsonistic bargaining power vis a vis LU/TfL, which does not appear to have been considered.

43. How then do the PPP plans envisage dealing not only with existing congestion but the greatly increased demand which is so central to the scheme's viability? LU's advisors, PwC have explained that a number of initiatives already in place, will increase both capacity and generate the increase in passenger receipts between 1998-99 and the start of the PPP.[14] These include the opening of the Jubilee Line Extension, the new rolling stock on the Northern Line and a number of measures being taken to alleviate bottlenecks. They also believe that the planned capital expenditure, and improved asset management, will increase capacity by reducing journey times, time spent in stations etc. The performance regime under which the PPP companies will operate, it is argued, will tie their payments directly to increases in capacity resulting from improvements of this kind. The assumption on which the plans turn is thus that extra demand can be accommodated by a more intensive use of the existing network.

44. These projections assume that, as with new roads, increased capacity on the Underground will itself generate increased ridership, thus securing the extra fares revenue needed for the commercial viability of the scheme.[15] There are two problems with this. The analysis of LU's rail receipts over the last 15 years shows that the pattern of demand is closely tied to movements in the economy (Table 1). If there is a downturn in the economy, passenger numbers may decrease, as they did during the last two recessions. Secondly, if passenger numbers do rise to meet all available capacity, how will the existing overcrowding on the network be relieved? Thus even in the most optimistic scenario, it is unclear that the PPP can in fact ameliorate the problems passengers presently face. Yet that was one of the justifications for the PPP.

45. But the financial projections rest not only upon demand and capacity but also upon costs. They assume that costs will continue to fall relative to income. In this context it is worth noting that costs are expected to be considerably less at the beginning of the PPP than they were in 1998-99. This is despite the increases in labour productivity and outsourcing that have already occurred. LU's financial advisors explain this in several ways.[16] For the PPP, LU's costs refer to passenger services only, which are necessarily lower than LU's present costs which include the cost of the infrastructure as well as passenger services. A number of investment initiatives are under way that will reduce costs and wok done on the backlog of maintenance work will remove the additional costs imposed by the poor condition of the infrastructure. But it is unclear how removing the maintenance backlog will reduce the cost of passenger services.

46. But if demand fails to rise sufficiently, or the network is unable to absorb sufficient demand, or costs cannot be further reduced, the Mayor will be faced with the choice of increasing fares or providing subsidy. Any significant increase in passenger fares—coming on top of fare rises of more than twice the rate of inflation in the last 10 years—would be deeply unpopular and politically difficult—quite apart from the risk of choking off demand. If LU is entirely dependent on its passengers for revenues, either scenario could be disastrous. This has implications for the financial viability of any long-term recurrent expenditure commitments.

47. Several further questions arise. Firstly, in the context of the transfer of responsibility for LU/LT to the Mayor, the question arises of who will bear the cost of any affordability gap? Will the DETR commit itself to meeting the gap through grant over the concession period of 30 years? Or—as is surely more likely—will the Mayor and the GLA have to choose between fare increases (for which they will be responsible under the GLA Act) or diverting resources from other users?

48. Secondly, it is also worth considering the position of LU, and by implication the Mayor and the GLA, if the Infracos's performance is unsatisfactory. One of the main arguments in favour of the PPP policy is that it allows the risk of things going wrong to be privatised. In other words, all the claims on LU's revenues arising from the PPP are held to be conditional on the performance of the PPP companies. There are several inter-related issues here: the financial implications of poor performance and the termination of the contract.

49. We consider first poor performance. The failed PFI Information Technology projects for the Home Office and the Department of Social Security provide a precedent. All PFI projects are predicated upon risk transfer to the private sector corporations. In the event of failure, the private sector is assumed to carry the cost of failure. Both the Passport Agency and Social Security computer projects have gone massively over budget and over time, yet the private sector will pay fines that are nothing more than loose change in the context of the additional costs to the public sector. In the case of the Passport Agency, Siemens Business Systems will be fined £69,000 (as opposed to the maximum £400,000 possible), whereas the costs to the public sector are estimated at about £15 million.[17] Additional staff will have to be taken on and the public will have to bear the cost of higher fees for a new passport. In other words, the risk has been transferred back to the public sector, as the Public Accounts Committee noted. Thus while it is possible to specify the risks to be transferred on an a priori basis, it is an entirely different matter to transfer the risks in a legally enforceable contract.

50. The first point to consider is that the courts will not enforce the penalty clauses included in contracts for failing to deliver on time, to budget or to specification. Most commercial contracts rest instead upon the concept of "liquidated damages". Under "liquidated damages", the contractor is only liable for the financial losses incurred by the contracting organisation where the organisation in turn has a contract with the public and produces evidence of financial loss. But most public agencies do not have such contracts with the public that enable, at least in principle, the public to make a claim against it. In the case of the Passport Agency, the public does not have a statutory right to a passport and could not therefore make a claim against the government for failure to provide a passport. In the event, the government made ex-gratia payments to the public for the inconvenience and these took in any event a minor second place to the additional costs incurred by the late delivery and operation of the computer system. Thus, neither LU nor the travelling public will be able to seek redress from the contractors or LU respectively.

51. It is noteworthy that in none of the failed IT PFI projects has the government attempted to seek redress via the courts. While this may be because the government is anxious to get PFI contracts off the ground, it may also be because they stand little chance of winning. While the PPP is not a PFI contract, many of these contractual issues are the same. There is no reason to assume that London Underground's lawyers will be any more successful than those of government departments. In any event, recourse to the courts consumes a huge amount of financial and human resources, which are in short supply. Having consulted lawyers, we have found a widespread belief that public agencies would be unable to enforce liquidated damages clauses in practice. This is why the Treasury has issued guidance for contracts that do not rely on "liquidated damages".[18]

52. Instead, the government has decided that contracts will rely on a reduction in the annual payment for failing to deliver according to the contract. But output specifications that depend upon quality as opposed to quantity are notoriously difficult to determine, monitor and enforce. This was one of the issues in relation to the failure of IT PFI projects. London Underground has refused to make the output specification for the invitation to tender publicly available on the grounds of "commercial confidentiality". The failure to put that and any (future) final contract in the public domain means that it is impossible for the public or Parliament to scrutinise, comment or hold LU and the PPP to account.

53. But reliance on withholding payment by a fixed amount according to the output specification and contract led the Passport Agency to withhold about £5 million from Seimens. This was exceeded by the costs to the Passport Agency of about £15 million.[19] These costs of course leave out the cost to the public at large then or later, any other public agencies that might suffer the knock on effects of failure to deliver, as well as the wider political damage, which in the case of the Passport Agency, were not inconsiderable. It is worth noting that since the Passport Agency is required to cover the full cost of operating its services, and the passport departments of consular services overseas, from the travelling public, the cost of passports has had to rise. Thus it is the public who has borne the risk resulting from failure. All these risks are excluded from any risk assessment, which is only required to consider the risks to the two parties to the agreement.

54. We consider next the issue of termination. Although logically distinct from the preceding discussion, the same issues will apply if the amount withheld from the annual payment were to threaten the financial viability of the Infraco. Most of the finance to be raised will be in the form of debt, with a minor contribution from shareholders in the PPP companies. Debt service and returns to shareholders constitute completely new claims on LU's revenues, and the nature of these claims needs to be understood. PFI/PPP deals involve a three-way relationship between the public sector, the PPP contractor and the lenders providing debt to the contractor. The interest of the lenders is to ensure that there is an assured stream of future revenues to meet the cost of debt service. It is not, therefore, in their interest to make it easy for the PPP contract to be terminated due to poor performance by an Infraco. Under Treasury guidance, public agencies are required to take this into account when drawing up contracts. Thus, for example, contracts should make provision for sufficiently generous "rectification periods" where the Infracos have a chance to put things right rather than having the contract terminated immediately.

55. Furthermore, if previous PFI arrangements are anything to go on, LU/TfL will, implicitly or explicitly, guarantee the debt even in the event that the contract is terminated due to default on the part of the Infraco. An implicit acceptance of liability by the public sector allows PPP companies to force down interest rates and thus reduces the overall cost of the arrangement. This is particularly crucial, as an Infraco is Special Purpose Vehicle whose only activity is the LU concession, not a corporation with multiple revenue streams. Some earlier PFI contracts have guaranteed debt by simply including compensation clauses in the contract specifying that debt will be repaid. The agreement for London Underground will take a different form but will have to achieve a similar effect: if the PPP company loses the contract, and if the lenders fail to find another contractor to take it on, London Underground will have to buy the contract back. While the lenders may lose out to a certain extent under this arrangement, they can count on LU meeting most of the cost of the outstanding debt. There is also a commercial logic to these provisions. If the public sector client could simply walk away from PPP debts, there would be the possibility of "windfall" gains where the public sector would benefit from any investment undertaken by the PPP companies without, in effect, having to pay for them.

56. However, while this type of arrangement may be commercially fair, it is not unreasonable to question the fairness of the distribution of political costs it implies. Even if the PPP failed to meet any of the performance targets set out in the contract LU would still have to maintain debt service on loans it has not itself taken out. In this event, the decision on how the cost was to be handled would rest with the Mayor and the GLA who will not, of course, be party to the negotiations between LU and the PPP companies. Since the concession agreement is expected to run for 25 to 30 years, the PPP will impose long term liabilities on the Mayor and GLA.

57. Our analysis has concentrated on the financial implications of the PPP. But little thought appears to have been given to the consequences of fragmenting the Underground in this way. All the evidence from the privatised rail industry shows that the problems of planning and co-ordination have increased while safety has decreased. The financial constraints under which LU operates have already led to a doubling of the injury rate. This can only increase further as the infrastructure and operations are parcelled out to four separate entities and financial pressure from the market.

58. In short, our analysis suggests that the PPP is likely to be very much more costly than both PwC's and our initial estimates. But either way, the scheme is simply unaffordable. That in turn means that either the travelling public and/or the taxpayers will bear the cost. All the evidence suggests that the legal contracts will fail to protect the public sector agencies. Several lawyers have said that the real winners in all this will be the lawyers.

59. All the alternatives to PPP have thus far focused on replacing government funding. The one that has attracted public attention is the proposal that investment could be financed through revenue bonds, which have been used successfully in New York. Its proponents argue that this would be cheaper than the PPP proposals, as bonds issued by a public body would incur lower interest charges than borrowing by a private company, and there would be no need to fund returns to shareholders.

60. We consider first the criticisms of bond finance made by the government, LU and PwC. Firstly, they claim that efficiencies under the PPP would easily outweigh the higher cost of finance. This rests upon the briefing issued by London Underground's financial advisers, PwC in December 1999, which modelled the relative costs of the PPP and a bond financed alternative. This document does not make the case to support the government's argument, It assumes from the start greater inefficiency on the part of the public sector, leading to higher costs over the contract period and thus the need to borrow more than would be necessary under the PPP. As the briefing assumes what it is held to demonstrate—the lower cost of the PPP option—it provides no evidence to support the argument. Others point to the "efficiency savings" made by other privatised industries, particularly rail and water. But these are "own goals". Services, however defined, have declined in the national railways, as infrastructure investment by Railtrack has not risen commensurate with need, expectation or commitments.[20] The much vaunted "efficiency savings" that were assumed to follow water privatisation came not so much from lower operating costs but from a lower level of capital expenditure on both renewals and investment than that predicted at privatisation.[21] These in part at least were the result of inflated cost estimates that provided the basis for price increases. More importantly from the perspective of this study, while some performance targets were set for the water companies, many of the targets were not achieved,[22] as the failure of the public water supply to West Yorkshire in 1995 and leakages running at a higher level than pre-privatisation testify. Thus the "efficiency savings" were made at the expense of consumers and the public at large. Yet proponents of the PPP believe that private sector operators can make similar savings in investment and maintenance expenditure.

61. Secondly the government has argued that bond finance would risk leaving the public sector (and thus passengers and/or the Mayor) with the bill for any cost overruns. The example of the Jubilee Line extension, a major piece of civil engineering whose costs are notoriously difficult to predict accurately is frequently cited to this effect. But this is scare mongering rather than argument, since the government is not comparing like with like. The PPP involves numerous relatively small-scale projects not new tunnelling and construction work. Risk can be shared between the public sector and contractors by means of appropriate contracts, without any need for private finance. In the PPP, the presence of private finance militates against efficient transfer of risk because in practice, LU cannot simply walk away from the debt taken out by the PPP companies, whatever the cost overruns due to incompetence or negligence, as we have seen. It also needs to be remembered that cost overruns can occur for many different reasons, many of which are unpredictable, such as new safety requirements, environmental factors, including, just by way of example, London's rising water table, and the unknown condition of many of the Underground's assets. It is very unlikely that the PPP companies will be taking on all such risks (and it would be prohibitively expensive if they did). In other words, there are still plenty of reasons why costs may still overrun under PPP, leaving the Mayor and TfL to pick up the tab.

62. The third argument against bonds, that bond issues would leave LU burdened with £8 to £10 billion of debt applies equally to the PPP. To say or imply that PPP borrowing is "investment", whereas borrowing through revenue bonds is "debt", is pure sophistry. Under the PPP, LU will have to underwrite and pay the PPP companies' debt service costs as well as funding returns to shareholders through the ISC. The difference between bond finance and the PPP is in essence little more than whose balance sheet will show the liability for the debt—LU or the Infracos.

63. We made a direct comparison of the cost of financing the investment through bonds on a like for like basis with the PPP. We used the same assumptions about efficiency savings, that LU would spend the same amount as the private sector, but have replaced their 10 per cent private sector cost of capital with a conservative public sector borrowing rate of 5 per cent. The figures should not be taken as predictions of the cost of the investment programme, any more than the models produced by PwC should be taken as predictions of the cost of the PPP. They are crude estimates based on the interest payments on the accumulated debt in each year. The results of our calculations show the difference in total costs over the 15 year period is around £1 billion, on capital raised of £2.4 billion (Table 1). In other words, as one might expect given the lower cost of public sector capital, public sector borrowing costs half that of the PPP. In reality, this underestimates the difference, as the interest rates at which the Mayor could borrow are likely to be lower than 5 per cent. We have not included the debt repayments to estimate total costs but since the Underground does not make a profit on its existing operations, bonds, like the PPP, are simply unaffordable. While revenue bonds involve lower financing costs, in the absence of government grants, fares would almost certainly have to increase to meet the cost of debt service, albeit to a lesser extent than under the PPP.

64. But bond issues have to be secured against future revenue streams, and no such revenue stream is currently available. Securing debt against fares alone is not a realistic option. Debt financing does not replace public subsidy, it requires it. The frequently cited example of New York's Metropolitan Transportation authority provides a useful example. The revenues against which the bonds issued by the MTA bonds are secured include both fares and subsidies, in the form of hypothecated local taxes. Because some of the MTA are secured against fares, there has been a misconception that it meets the costs of its investment programme, at least in part through fare revenue, and this has encouraged comparison with London. However, while half of the bonds issued by the MTA are secured against and payable out of fare revenues, fares alone are insufficient to meet even the running costs of the Subway. In other words, the fares against which bonds are issued are subsidised, and without subsidy, fares would have to rise to meet the costs of debt service.

65. Without an assured stream of public funding to supplement fares revenue, the Underground is in no position to take on the financing costs of an investment programme. This is all the more true for the more expensive PPP option that is equally dependent on finance. The New York example demonstrates, and herein lies its significance, the importance of public funding, rather than the advantages of bond issues.

66. Under the GLA Act, revenues from proposed congestion and workplace parking taxes would be earmarked for public transport. To the extent that these provide a predictable stream of future revenues their benefit could be maximised by borrowing against them. Bond issues on a similar model to New York might then be possible.

67. However three aspects of such taxes need to be borne in mind: firstly, their successful implementation requires increased capacity on public transport, which means that capacity would have to increase before the revenues from the taxes began to come in. In other words, there would still be a problem in funding investment in the early years of such a scheme. Secondly, such charges would be politically unpopular particularly when it becomes clear that the revenues will be used to fund returns to the financial institutions not improvements to the Underground. Thirdly, it should not be assumed that the aim of reducing motor traffic dovetails neatly with that of bringing in revenue for public transport. The more successful road taxes are in reducing traffic, the less they may bring in if people walk, cycle, work from home via the Internet, or relocate. Moreover, if the revenues from road taxes become the main source of funding for transport other than fares, other measures for reducing traffic, which do not yield revenue such as pedestrianisation, reduced speed limits, and bans on traffic during certain hours, may come to be unfairly neglected.

68. Furthermore, while revenue raising via bonds seems to offer of allowing the Mayor a certain independence from the Treasury in making investment decisions, no function is more jealously guarded by the Treasury than the power to borrow. Neither the GLA nor TfL will be able to borrow to finance investment without approval by central government. GLA and its functional bodies (which include TfL) will be subject to the local government financial regime and will therefore require credit approvals from the Secretary of State for the Environment, Transport and the Regions before undertaking any borrowing.

69. Borrowing will therefore be limited to the credit approvals set by the Secretary of State. The possibility of extending those approvals by finding a formula under which the borrowing would not count towards the Public Sector Borrowing Requirement or the General Government Financial Deficit may appear attractive. However, the frequently cited example of the Channel Tunnel Rail Link, where the government argued that its guarantee of a loan should not count towards the PSBR because of the limited risk of the guarantee being called in, is not applicable because the government was guaranteeing a loan on the part of the private sector. No such arrangement is possible with TfL because TfL will itself be a local authority and therefore part of the general government sector. As such its borrowing will count towards any measure of government debt (PSBR and GGFD) whether guaranteed by central government or not.

70. It is of course possible for government to make an exception to its own accounting rules in order to allow LU to undertake borrowing without affecting the PSBR/GGFD. This has already been done for investment in local authority owned airports. However, this option is only available to central government, not the GLA. Thus while revenue bonds might allow the future Mayor of London to divert new forms of taxation to the financial institutions via subsidies to London Underground, this will only be possible if central government agrees to relax its control over the finances of devolved administrations. As in the financial settlement for the devolved assemblies in Scotland, Wales and Northern Ireland, the nominally extensive powers of the mayor are restricted by the existing public expenditure control system.

71. While our analysis therefore refutes three of the criticisms raised against bonds, it does show that a bond issue does not provide an affordable, sound or political alternative. Indeed, the apparent popularity of the bond proposal owes more to the unpopularity of the PPP and the refusal of the Treasury to fund the much needed investment in the Underground despite running a fiscal surplus, than any merits of bonds per se.

72. To conclude, London Underground is a high fixed cost industry where a high proportion of the Underground's costs are associated with providing the rail network as distinct from running the trains. Like the former public utilities (gas, electricity, telecoms and water) its main costs and much of the investment relate to the distribution network. In effect, all these industries operate two businesses, the distribution network and the supply of services. Because of the high cost of maintaining the network, the supply side of the utilities is more profitable than the distribution as the gas, electricity and telecoms industries demonstrate.

73. However the utilities differ on the demand side. They supply an (almost) universal, essential and unavoidable service in ways that are reasonably predictable. It is therefore possible to devise methods of payment (monthly direct debit, pre-payment cards, social security allowances, etc) that spread the burden over time and do not require consumers to pay up at the point of use. Thus, both their pattern of demand and method of payments allow full cost recovery from consumers.

74. In contrast, while many people do use rail services, not all do, or do so very infrequently. Furthermore, pre-payments, season tickets and other payment devices that spread the cost of user charges through time are not as well developed in the rail industry. As a result, many people pay at the point of use. LU also faces competition from other forms of transport. In the absence of adequate demand, full cost pricing may be self-defeating in that it chokes off demand, quite apart from the implications for equity. Thus the essential problem for the Underground is that it is unable to generate the revenues that cover the cost of the network and the train operations, even without the requirement to provide a return on capital employed. Operating costs and the (inadequate) annual capital spending on the Underground network exceed total revenue, and the Underground currently depends on annual grants to make up the difference. The use of public grants to support rail networks and operations in this way is, as we have shown, the norm not the exception in industrialised countries.

75. Government-instituted funding, commercial and charging policies have prevented the former state-owned railways and London Underground from developing an appropriate cost structure that could be used to apportion costs and set prices. Given that rail is essentially two businesses, the network and passenger services, which combined are unable to make a profit or break-even, the solution is to set prices to cover only passenger services. Public funding, paid for from taxation, would cover the cost of maintaining and enhancing the network which would be free at the point of use to the travelling public in the same way as roads. The fact that road infrastructure is free is, of course, why road haulage and commercial bus and coach services are able to run at a profit. While this would not be any cheaper, it would mean that the Underground would be the beneficiary not the financial institutions as in the case of the PPP or bond finance. It would lead to lower fares, passenger traffic and receipts, a more rational service, greater accountability, and train operations that at least broke even, with all the associated environmental and social benefits. This pattern is adopted in five Western European countries, where the state owned railways are funded as two distinct entities.

76. While the PPP proposals in effect split the Underground into its two constituent parts, the network and passenger services, it does so in completely the opposite way. Instead of providing a network at a lower (or no) cost that will enable the Underground to operate efficient passenger services, the PPP will raise the infrastructure costs and hence the cost of passenger services, and exacerbate the existing problems.

77. Our recommendation for the development of a modern, safe and efficient Underground is that expenditure on the enhancement of existing lines and the construction of new ones is significantly increased beyond that proposed for the PPP. As the Acheson Report on health inequalities showed, public transport is an economic, social, environmental and public health issue that affects the lives not simply of those who use the services but of those who do not[23]. The most economic and socially efficient way of running the Underground is to retain its integrated structure while at the same time providing the required level of funds. These are conditions that the proposals for the PPP and bond finance will not and cannot satisfy. The basic economic realities of this industry, both in Britain and overseas, show that these conditions can only be met by funding the full cost of the infrastructure through taxation, which may be local, national, or a mixture of both, without charge to London Underground.

Declan Gaffney,
School of Public Policy, University College London.

Dr Jean Shaoul,
School of Accounting and Finance, Manchester University

4 PwC, "London Underground", a briefing report, 6 December 1999. Back

5 Public Private Partnership for London Underground, briefing for bidders. Back

6 Shaoul J, "Railpolitik: A Stakeholder Analysis of the Railways in Britain", Manchester University, paper presented at the CIMA Public Sector Accounting Workshop, University of Edinburgh, September 1999. Back

7 House of Commons, Transport Committee Seventh Report: London Underground, Vol 1, p viii. Back

8 Robert A Gerard "New York can teach us a lot about our Tube", London Evening Standard, 25 November 1999. Back

9 Metropolitan Transportation Authority, Combines Disclosure Fillings 1999. Back

10 PwC, Briefing Note, December 1999. Back

11 PwC Briefing Note, December 1999. Back

12 See for example the Guardian 13 March 2000 "Taxpayers to bail out tube". Back

13 As cited in MMC, London Underground Ltd, a report on passenger and other services supplied by the company, HMSO, 1991, p 120. Back

14 PwC, personal communication. Back