The financial analysis for the London Underground Public Private Partnerships

Transcription

1 The fiacial aalysis for the Lodo Udergroud Public Private Parterships REPORT BY THE COMPTROLLER AND AUDITOR GENERAL HC 54 Sessio : 15 December 2000

2 This report has bee prepared uder Sectio 6 of the Natioal Audit Act 1983 for presetatio to the House of Commos i accordace with Sectio 9 of the Act. Joh Bour Natioal Audit Office Comptroller ad Auditor Geeral 14 December 2000 The Comptroller ad Auditor Geeral is the head of the Natioal Audit Office employig some 750 staff. He, ad the Natioal Audit Office, are totally idepedet of Govermet. He certifies the accouts of all Govermet departmets ad a wide rage of other public sector bodies; ad he has statutory authority to report to Parliamet o the ecoomy, efficiecy ad effectiveess with which departmets ad other bodies have used their resources. For further iformatio about the Natioal Audit Office please cotact: Natioal Audit Office Press Office Buckigham Palace Road Victoria Lodo SW1W 9SP Tel: equiries@ao.gsi.gov.uk Website address:

3 Cotets Summary 1 Lessos for future Public Private Parterships 2 Lodo Udergroud is 3 followig a thorough process for a fiacial evaluatio The Public Sector Comparators have 3 bee carefully prepared The evaluatio of bids has bee thorough 6 ad comprehesive The Public Sector Comparators 8 eed to be iterpreted carefully There is iheret ucertaity i modellig the costs of 8 the Lodo Udergroud ifrastructure over 30 years The fiacial models aloe would provide oly 8 limited guidace to the most likely cost of a public ifrastructure operatio There is ucertaity over the most appropriate 9 fiacig approach for the cost of public ifrastructure operatio Fiacial aalysis does ot 11 cover all the factors affectig value for moey Usig the fiacial aalysis 13 i decisio-makig Glossary 15 Appedices 1. The Public Private Partership approach Methodology Natioal Audit Office ad Public Accouts 20 Committee Recommedatios o Public Sector Comparators

4 summary 1 Lodo Trasport is curretly i the process of evaluatig bids for the Public Private Parterships (PPP) for the fudig ad maagemet of the ifrastructure of the Tube. Lodo Udergroud Limited (Lodo Udergroud) is the subsidiary of Lodo Trasport resposible for implemetig the PPP approach. This approach ivolves the trasfer of trais, statios, track ad some 6000 staff to three separate ifrastructure compaies uder 30 year cotracts for differet parts of the Tube. I March 2000, the cotracts were expected to deliver billio of capital ad maiteace work i the first 15 years of the cotract 1. Uder the PPP approach, Lodo Udergroud will retai resposibility for the operatio of trai ad statio services. Appedix 1 describes the PPP cotractual framework i more detail. 2 Lodo Udergroud has ow received Best ad Fial Offers from bidders for two of the three PPP cotracts 2. The biddig process for the third cotract is at a less advaced stage. After cosiderig bids, Lodo Trasport will make a recommedatio to the Secretary of State for the Eviromet, Trasport ad the Regios, who will take the fial decisio o whether to approve the PPP cotracts. O preset plas, Lodo Trasport will trasfer its resposibilities to Trasport for Lodo i the summer of The Departmet of Eviromet, Trasport ad the Regios (the Departmet) ad Lodo Trasport ited to award the PPP cotracts to private sector bidders oly if they pass two separate tests: the safety test: whether chages to the structure of the Tube will maitai or improve curret safety stadards; the value for moey test: whether the bidders offer superior value for moey whe compared to a alterative, publicly-fuded ifrastructure operatio. 4 To iform the judgemet about value for moey, Lodo Udergroud is udertakig extesive fiacial aalysis. It has prepared estimates of the likely cost of publicly fudig the ifrastructure uder alterative fiacig scearios, icludig the issue of bods. These cost estimates, which are made for the Tube ifrastructure as a whole ad for the three ifrastructure compaies, are called Public Sector Comparators (Comparators). Lodo Udergroud has also evaluated the Best ad Fial Offers from bidders for the PPP cotracts ad estimated the likely costs of each. At the core of the fiacial aalysis is a compariso betwee these two sets of estimated costs. 5 Followig a recommedatio from the Eviromet, Trasport ad Regioal Affairs Committee 3, we decided to report o the extet to which Lodo Udergroud's fiacial aalysis resolves the value for moey test. Our methodology is explaied i Appedix 2. The safety test is beig cosidered by Lodo Trasport ad the Health ad Safety Executive ad is ot part of this report. 1 Eviromet, Trasport ad Regioal Affairs Committee, 14th report : Fudig of Lodo Udergroud, HC Bids have bee received for the BCV ifrastructure cotract, coverig the Bakerloo, Cetral ad Victoria Lies, ad the JNP ifrastructure cotract, coverig Jubilee, Norther ad Piccadilly lies. 3 Fudig of Lodo Udergroud, 14th report, HC 411 1summary

5 6 The report focuses o Lodo Udergroud's methodology for the fiacial aalysis of bids ad the Comparators. Negotiatios with bidders are still uderway, so fial figures are ot available. The govermet's policy is that iformatio o the value of the overall or idividual Public Sector Comparators would risk givig bidders a advatage i the cotract egotiatios ad therefore could have a adverse impact o value for moey. I the light of this policy, our published aalysis is at the level of the ifrastructure as a whole, rather tha the idividual parts of the ifrastructure which are the subject of bids, ad o absolute values for the Comparators or bids are give. We are however keepig the whole biddig ad decisio process uder review, ad will report further as ecessary. Meawhile, we have published this first report to give aalysis ad commet at this stage for the beefit of Parliamet. 7 We cosider that Lodo Udergroud has udertake a thorough process i estimatig the costs of public sector ad PPP optios. The rages of values produced from the Comparators are of some use i guidig judgemet. But, as Lodo Udergroud recogises, a decisio take oly o the basis of where a bid lies compared to the rages for the Comparators would be usoud. The fiacial aalysis provides useful but icomplete isight ito the value for moey of alterative approaches to maagig ad fudig the Udergroud s ifrastructure. The fiacial aalysis does ot, ad could ot, take ito accout all the factors which are relevat to assessig the relative value for moey of the alterative approaches. It is essetial that decisio makers uderstad what lies behid the figures before reachig a coclusio. The mai sectios explai these coclusios further. Lessos for future Public Private Parterships 8 I some past PPP trasactios, the process of preparig a Comparator has bee separated from the rest of the PPP procuremet. I this case, Lodo Udergroud has esured that there has bee close co-ordiatio betwee its ow egieers, techical advisers, ad fiacial ad modellig experts. This co-ordiatio should serve as a example of best practice for future PPP trasactios elsewhere i the public sector. 9 As i this case, departmets should i future esure that value for moey decisios are ot based o oe-dimesioal comparisos of sigle figures. 10 Uderstadig the differeces betwee the private sector ad public sector approaches to the same output specificatio lies at the heart of assessig value for moey. I choosig the PPP optio, departmets must always esure that they uderstad how the private sector is deliverig value for moey, ad why, i their judgemet, it would ot be possible for the public sector to achieve the same value for moey. We uderstad that Lodo Udergroud iteds to udertake such a aalysis i this case. 11 Fiacial modellig is a iheretly ucertai techique, as Lodo Udergroud has clearly recogised i this case. We cosider that there may be beefits for future PPP trasactios, i terms of time, cost ad clarity, i modellig the Public Sector Comparator o a limited basis, with fewer variables ad less complexity. 2summary 12 The success of a Public Private Partership requires a geuie aligmet of iterests betwee cotractig parties to esure that partership is more tha just a statemet of itet. Achievig this aligmet meas that all parties to ad stakeholders i a deal eed to be egaged throughout the process.

6 13 Value for moey depeds o a wide rage of factors. Aalysis of the fiacial implicatios of differet optios is oe factor. This fiacial aalysis typically ivolves three elemets: the preparatio of a Public Sector Comparator. A Public Sector Comparator is a bechmark agaist which value for moey is assessed. It is typically a cost estimate based o the assumptio that assets are acquired through covetioal fudig ad that the procurer retais sigificat maagerial resposibility ad exposure to risk; Lodo Udergroud is followig a thorough process for a fiacial evaluatio a evaluatio of bids, which produces estimated costs of the private sector optio; ad a compariso betwee the two sets of costs. 14 With three separate ifrastructure cotracts to procure, ad the largest Comparator by value yet prepared by ay public etity, the fiacial aalysis i this case ivolves a large amout of complex iformatio. The effectiveess of the process adopted to prepare ad maage this iformatio is therefore a key part of the overall reliability of the fiacial aalysis. 15 We have foud that Lodo Udergroud has udertake a exceptioally thorough process i preparig the capital ad operatig costs of the Comparators ad the estimated costs of the bids. Figure 1 presets the process i graphic terms. The Public Sector Comparators have bee carefully prepared 16 The process of costructig the Lodo Udergroud Public Sector Comparators ivolved three stages: establishig a methodology; producig a ubiased estimate of the costs to meet the required level of performace over 30 years; ad estimatig the rage of risks that would face a public sector ifrastructure operatio. 17 Lodo Udergroud established a clearly documeted methodology, which they published i March This methodology stated that the Comparators would estimate the costs of the public sector meetig the PPP performace specificatio. The methodology also stated that Lodo Udergroud would estimate base costs, the risk margi to add to those base costs as a result of ucertaity ad the likely efficiecies Lodo Udergroud could achieve over time. Its itetio was to esure that risks for public sector procuremet were fully ad appropriately quatified, without either maagerial bias towards optimism or a automatic assumptio that ay past shortcomigs i maagig risk will cotiue. 18 The base costs to meet the PPP performace specificatio have bee carefully estimated, o the basis of Lodo Udergroud's existig costigs for idividual assets, after attemptig to strip out ay cotigecy or margi o 3

7 The Process leadig to the deal The comparator process The biddig process Costs adapted from Whole Life Asset Pla Risk workshops Shadow ruig of the three separate etworks (JNP, BCV ad SSL) Public Sector Comparator for ivitatio to teder stage Ivitatio to teder bids Evaluatio ad Shortlistig Bidder egagemet period Public Sector Comparator for Best ad Fial Offer stage Best ad Fial Offer bids Natioal Audit Office report Best ad Fial Offer evaluatio Recommedatio to LT Board Preferred bidder aoucemet Fial egotatios Commercial close - i priciple agreemet o cotract terms Deal close Fiacial close - cotracts siged 4

8 those costigs. This approach ivolved cosiderig the scope of physical works desiged to meet the performace specificatio to produce a Log Term Ivestmet Pla. It appears to us that Lodo Udergroud's approach to estimatig the capital spedig programme has bee a reasoable oe, although lack of past kowledge about asset coditios meas that margis for ucertaity are ievitably substatial. 19 These base costs have bee adjusted to reflect possible risks ad efficiecies that could arise for a public sector operatio. The risks were estimated o the basis of a aalysis by egieerig experts of Lodo Udergroud's historic record of cost overrus. For example, o desig ad costructio risk this ivolved cosideratio of some 242 idividual projects, with a average overru of 20 per cet for various reasos icludig cliet chages of scope ad fudig costraits. Egieers judged that some of the more extreme overrus observed o projects were geuiely oe-off ad could ot be take as idicative of likely future achievemet. Other overrus were judged to be irrelevat, such as the costs of ew tuellig o the Jubilee Lie Extesio sice o ew tuellig work is evisaged i the PPP requiremet. 20 Egieers also estimated the likely efficiecies that Lodo Udergroud could reasoably expect to achieve over the ext 30 years. They cocluded that some moderate cost savigs could be aticipated, especially for ivestmet i the statio ifrastructure. However, i other areas, such as ivestmet i ew rollig stock, they assumed that little improvemet was possible because of the extesive use of PPP-style cotracts i those areas already. 21 Oe of Lodo Udergroud's key judgemets was about its ability to meet performace targets. Lodo Udergroud cocluded that i the past, it has failed to meet the performace improvemets expected from ew ivestmets, because of its failure to itegrate systems effectively. It further cocluded that these performace failures would occur i future ivestmet projects, because learig from oe project was ot easily trasferred to the ext. As a result Lodo Udergroud cocluded that the Comparator should iclude the otioal additioal costs that such failures to meet the performace specificatio would impose o Lodo's passegers, usig its stadard cost beefit techiques. This adjustmet also applies to bidders. 22 Figure 2 summarises the risk ad efficiecy categories used by Lodo Udergroud. 2 The risk ad efficiecy categories used by Lodo Udergroud Factor Descriptio Impact Relative price effect Reflects the ucertaity surroudig the impact Small icrease i of iflatio o differet categories of cost costs Uit cost ucertaity Reflects the ucertaity surroudig the estimates Small icrease i of uit costs withi the basic cost estimates costs Ivestmet cost ucertaity Reflects ucertaity over the proposed scope of Large icrease ivestmet projects, which, i Lodo i costs of more Udergroud's experiece, always icreases as tha 10 per cet the project progresses Maiteace ucertaity Reflects ucertaity over the proposed scope of Icrease i costs maiteace projects Performace ucertaity Reflects the risk that Lodo Udergroud's Large icrease i ivestmet projects fail to deliver expected costs, mostly improvemets, leadig to a failure to meet the estimated as the PPP performace specificatio ecoomic cost of delays to passegers Efficiecy ucertaity Reflects efficiecies that Lodo Udergroud Reductio i costs could reasoably expect to achieve over the ext thirty years. Source: Natioal Audit Office 5

9 23 Lodo Udergroud did ot quatify all of the key risks that it idetified. Istead, it summarised them i a commetary o the Comparators provided to decisio-makers. For example: the risk that legal disputes may arise uder the PPP approach has ot bee quatified formally because it would be very difficult to estimate reliably. the potetial positive impact of the itroductio of the PPP structure o the efficiecy of the operatig compay retaied i the public sector has also ot bee quatified. 24 Lodo Udergroud has updated the Comparators sice March Revisios have bee well documeted, ad revisig Comparators as more iformatio becomes available complies with Treasury guidace ad best practice. As a result of these revisios, the value of the Comparators has icreased sigificatly betwee the shortlistig stage ad the Best ad Fial Offer stage. The mai reasos for the icrease have bee: et icreases i the operatig base costs (aroud 20 per cet of the icrease) as a result of more up-to-date iformatio o the costs of the ifrastructure services ad revised estimates of admiistrative costs, for example relatig to existig PFI cotracts for power, ticketig ad commuicatios; chages i the scope of the PPP specificatio that bidders are beig asked to meet (aroud 20 per cet of the icrease); ad brigig the method of quatifyig performace risk i lie with that proposed for the evaluatio of bids (aroud 60 per cet of the icrease). 25 Lodo Udergroud is obtaiig idepedet assurace o the preparatio of the Comparators from several sources. Lodo Udergroud was supported by its egieerig cosultats, Ove Arup, who stated that they cosider the Comparators represet a fair estimate of the costs of meetig the output specificatio, ad by its fiacial advisers, PricewaterhouseCoopers. KPMG, Lodo Udergroud's statutory fiacial auditors 4, cocluded that the methodology for the Comparators complied with relevat Treasury guidace ad is expected to report that the fial Comparators represet a reasoable basis for estimatig public sector costs. Fially, Lodo Udergroud has also acted o our ow commets about its approach to the fiacial aalysis. The evaluatio of bids has bee thorough ad comprehesive 26 Ruig parallel to the preparatio of the Comparators, ad formig the secod part of the overall fiacial aalysis, Lodo Udergroud has bee udertakig a bid evaluatio process. 27 Lodo Udergroud prepared ad is followig a thorough ad comprehesive bid evaluatio strategy to idetify which bidder has submitted the most ecoomically advatageous bid. The competitio for the two deep tube ifrastructure cotracts is further advaced tha that for sub-surface lies. Six bidders submitted iitial bids i March for two deep tube cotracts 5 o the basis of a fixed Ifrastructure Service Charge ad target performace levels. The iitial evaluatio resulted i two bidders beig shortlisted i July for each ifrastructure cotract. The shortlisted bidders the submitted Best ad Fial Offers i November. 6 4 KPMG are also auditors of Lodo Trasport ad its successor, Trasport for Lodo. 5 Three for each of the deep Tube cotracts.

10 28 The iitial evaluatio sought specified qualities i the bidders. These were: ability to deliver the required performace; asset stewardship, icludig evidece of safety ad evirometal competece; stewardship of Lodo Udergroud's staff; acceptace of the priciples of partership; ad value for moey compared to the Comparators. The Best ad Fial Offer istructios are desiged to build upo the iitial bid submissios. 29 Durig the evaluatio of the Best ad Fial Offers, Lodo Udergroud is cosiderig bid deliverability, together with techical, fiacial, legal ad commercial ad risk maagemet factors, icludig fiacial robustess. It is also cosiderig orgaisatioal factors to judge whether the bidder would be a suitable log-term parter for Lodo Udergroud. The fiacial evaluatio calculates et preset values of paymet streams implied by each bid ad estimates the rages of highest ad lowest values for each bid. 30 Fiacial adjustmets are beig made i four categories: fiacial, techical, impact o Lodo Udergroud's costs, ad legal ad commercial. These will take accout of areas where there are doubts about or qualificatios to aspects of the bids. Bids will also be adjusted for the additioal or reduced customer beefits flowig from the differet levels of performace bid to the extet that these are judged to be credible. 31 The estimated rages of each bid are beig compared to the Comparators to iform the value for moey decisio. Lodo Udergroud is comparig these costs over the whole 30 year cotract period, ad also over the first 7½ years. The purpose of the 7½ year compariso is to cosider the reasoableess of bid profiles ad to protect agaist bidders attemptig to bid low to wi the cotract ad the adjust prices upwards followig the first periodic review after 7½ years. 7

11 The Public Sector Comparators eed to be iterpreted carefully 32 I spite of Lodo Udergroud's thoroughess described above, our aalysis has show that there are limits, which Lodo Udergroud recogises, to the weight that ca be put o the figures emergig from the Comparators exercise. These derive from three areas: there is iheret ucertaity i modellig the costs of the Lodo Udergroud ifrastructure over 30 years; the fiacial models aloe would provide oly limited guidace to the most likely cost of a public ifrastructure operatio; ad the costs of public operatio are iflueced by the choice of fiacig sceario, icludig the availability ad impact of bod fiace ad the costs of covetioal public fiace. Yet there is ucertaity about what some of these costs are ad how they should be assessed. There is iheret ucertaity i modellig the costs of the Lodo Udergroud ifrastructure over 30 years 33 Estimatig Lodo Udergroud's ifrastructure costs over a 30-year period is a iheretly difficult exercise. There is ievitable ucertaity resultig, for example, from the upredictability of asset lives ad the possibility of chages i the performace required from those assets. Furthermore, the performace specificatio requires a iovative approach to estimatig costs sice Lodo Udergroud has ot traditioally estimated costs to meet a defied performace specificatio, so its historic records may ot provide a ideal basis for estimatig costs over the ext 30 years. 34 It is difficult to model strategic optios available to public sector etities i the future. For example: the Comparators assume a persistet level of overru i ifrastructure ivestmet projects over the 30 year period. It is perhaps ulikely that Lodo Udergroud's maagers ad owers would tolerate persistet failure without decisive actio. But it is difficult to defie i a model exactly what actio would be take ad whe. it ca also be difficult to capture i a model the value of public sector flexibility to reflect chagig priorities, or the opposite risk that a public sector etity repeatedly revises its targets ad icurs excess trasitioal costs as a result. The fiacial models aloe would provide oly limited guidace to the most likely cost of a public ifrastructure operatio 8 35 The fiacial model used for the Comparators produces a rage of possible values. The rage of the combied Comparators for all three ifrastructure cotracts is several billio pouds 6 betwee the lowest ad highest values. The model also produces sigle expected values, but we, Lodo Udergroud ad its advisers have cocluded that sigle values should ever be relied o i a fiacial aalysis, especially give the particular characteristics of this exercise. 6 For reasos explaied i paragraph 6, we have ot give absolute values. The rage referred to here is the rage of et preset values at a real discout rate of 6 per cet over 30 years.

12 36 As oted i paragraphs 18 to 22, the model depeds o fixed iputs of base costs, which are multiplied by iputs of plausible rages for risk ad efficiecy factors. These iput rages are used to geerate several hudred possible values for the Comparators, from which the output rages have bee geerated. The risk ad efficiecy factors are therefore the drivers of the rages of values, ad reflect Lodo Udergroud's best judgemets about the ucertaities facig a public sector ifrastructure operatio. 37 The value of each ucertaity factor depeds o the elicitatio of upper ad lower values from egieerig experts, as described i paragraphs 18 to 22 above. For example, for sigallig ivestmet, egieers aticipated a upper level of overru of 27 per cet ad a lower level of 13 per cet. I the absece of ay clearer iformatio, Lodo Udergroud assumed that the most likely value was the mid-poit of the upper ad lower rage (for example, 20 per cet i the case of sigallig ivestmet). As a result of usig that assumptio, the fiacial model might be take to suggest that the most likely value of the public sector costs lies i the middle of the rages which the model produces. That coclusio would be implausible. As is well kow, major projects sometimes are delivered below budget, but very rarely by a wide margi below. But major projects do sometimes overru their budgets very cosiderably. That should mea that the most likely outcome would be i the lower half of the rage ad ot i the middle as i this case. 38 Furthermore, the modellig process appears to have cosidered ucertaity factors o a idividual, lie-by-lie basis, rather tha a holistic basis. Because this comprehesive approach has bee used, a large umber of variables over 30 years have bee iput ito the model. I such circumstaces, it is difficult to be sure that there has bee o double-coutig of ucertaity i differet factors. A smaller model, with fewer key variables, might have made possible a more itegrated aalysis of the risks facig Lodo Udergroud. 39 As a result of these issues, we ad Lodo Udergroud have cocluded that the modellig provides some useful iformatio about the upper ad lower bouds of public sector costs over the ext thirty years but caot reliably be take to produce a sigle expected value withi those boudaries. There is ucertaity over the most appropriate fiacig approach for the cost of public ifrastructure operatio 40 The compariso of the Public Sector Comparators ad the PPP bids ivolves a series of fiacig ad ecoomic assumptios. These cover: the source of fudig for the public sector operatios - essetially covetioal fudig or bod fudig; the cost of capital relevat to each source of fudig (for example, the iterest costs o bods); ad the govermet's time discout rate, which is used to discout future costs ad preset them as et preset costs. 41 Lodo Udergroud's covetioal fudig scearios assume that the ifrastructure operatios are fuded through a mixture of fare reveues ad geeral govermet reveues. Lodo Udergroud has supplemeted this with bod fudig scearios, which evisage bods issued by Trasport for Lodo i three traches ad with a credit ratig based o a idicative view provided by a major credit ratig agecy. Icludig a bod sceario reflects Lodo Udergroud's view that covetioal fudig ca impose costraits o operatios because, with oly at best a three-year horizo of secure fudig, it ca hider log-term plaig of capital works. Raisig bods 9

13 directly from capital markets could provide Lodo Udergroud with more secure log-term fudig ad ehace log-term plaig. Lodo Udergroud did ot prepare a sceario with a govermet-backed bod. 42 I the bod sceario, Lodo Udergroud assumes additioal efficiecy savigs arisig from secure fiace ad the ed of aual fudig egotiatios, amoutig to 7 per cet per aum off ivestmet costs ad 1.5 per cet off maiteace costs. The bod scearios also assume some improved performace by the public sector. There is ievitably little by way of quatified evidece available o the level of possible efficiecies with bod fiace, ad Lodo Udergroud recogises that the 7 per cet figure is a matter of subjective judgemet based o its cosideratio of the limited evidece. 43 Some of Lodo Udergroud's scearios assume a govermet cost of capital ad discout rate both at 6 per cet real. This is i lie with stadard ivestmet appraisal practice i govermet, although a o-stadard approach has also bee examied i this case. The o-stadard approach arises from cocers about the appropriateess of the 6 per cet real rate. I 1997, the Committee of Public Accouts oted 7 that the Treasury's assumed discout rate of 6 per cet was at the top ed of the rage of values that could be justified. The Committee was cocered that the use of too high a discout rate would itroduce a bias i favour of PPP-style deals over covetioal fudig. The Committee observed that usig more recet figures for iterest rates o govermet debt would lead to a sigificatly lower figure for the discout rate. The Committee recommeded that it would be good practice for authorities to examie the sesitivity of fiacial aalysis of this sort to the choice of discout rate. 44 I accordace with this recommedatio, ad give that the stadard 6 per cet rate is sigificatly higher tha the prevailig rate o govermet debt (betwee 2 ad 3 per cet) ad professioal ecoomists also questio it as a time discout rate, Lodo Udergroud is examiig the impact o the fiacial aalysis of usig a discout rate of 3.5 per cet ad has developed a matrix of assumptios usig various discout rate ad iterest rate assumptios. At the time of writig, Lodo Udergroud had oly just completed iitial figures for some of these assumptios, so we are uable to reach ay coclusio o whether the figures require further refiemet, their impact o the fiacial aalysis ad whether this impact has bee fairly take ito accout by Lodo Udergroud. 45 Where appropriate, some of these scearios iclude a adjustmet to capture the impact o the govermet's reputatio for prudece of extra borrowig, kow as reputatioal exterality 8. There is some theoretical basis for makig such a adjustmet, but also, i our view, cosiderable scope for argumet over its size. We ote that the issue largely disappears if the discout rate is set at a more realistic level of 3.5 per cet. 46 The various fiacig ad discoutig scearios are techically complex, ad would ot be ecessary for most PPP cotracts. They have sesibly bee chose as scearios for this trasactio because it ivolves uusually large sums of moey, ad because alterative fudig sources are coceivable for Lodo Udergroud, with secure fare reveues that could be used to fiace direct borrowig th report, The stadard Treasury approach uses a cost of capital ad discout rate at 6 per cet. Scearios which use a lower cost of capital tha the discout rate create a bias i favour of public borrowig by lowerig the implied iterest costs of such borrowig. But Treasury cosiders that icreasig public borrowig has a exteral cost o the govermet's reputatio for prudece, ad the reputatioal exterality is a calculatio to reflect this assumed exteral cost.

14 47 The fiacial aalysis described i the precedig sectios of this report is ot sufficiet o its ow to ascertai which optio offers best value for moey. There is a wide rage of other factors, which are either difficult or impossible to quatify i fiacial terms, ad which could impact directly o the value of the differet optios. Such factors cover both the structure of the PPP ad the implemetatio of that structure. They iclude: Fiacial aalysis does ot cover all the factors affectig value for moey strategic issues: the strategic decisio to award log-term cotracts to the private sector ivolves risks ad beefits. Lodo Udergroud cosiders that the expected beefits iclude potetial access to private sector efficiecies ad skills i asset maagemet ad i maagig ifrastructure; ad clearer icetives through a cotract tha exist with a wholly public sector operatio. Possible risks arise from the divisio of resposibility for the Tube etwork ad a potetial loss of flexibility as it may become harder to re-set objectives for the ifrastructure i the face of chagig circumstaces, such as a icrease or decrease i passeger demad or shifts i the patter of commutig from oe part of Lodo to aother. the PPP cotractual framework: the value for moey of the PPP approach depeds o the flexibility ad eforceability of the cotracts to be siged with wiig bidders. Give that the cotracts ru over 30 years, some flexibility is required, but too much flexibility could allow bidders to raise prices durig the course of the cotract as uexpected cotigecies arise. To the extet that the cotract is ueforceable, bidders may i future be able to depart from prices or commercial terms agreed ow i egotiatios with Lodo Udergroud. Lodo Udergroud therefore eeds to assure itself that the proposed cotracts are robust, that bidders will have adequate fiacial capacity ad that they will ot be able to demad excessive icreases i paymet as ew requiremets or uexpected cotigecies emerge. the icetives i the performace specificatio: the PPP cotractual framework icludes fiacial icetives to improve performace i three areas beyod bechmark levels: the capability of the ifrastructure to support trai services, the actual availability of trai services ad the ambiece of the eviromet provided to passegers. The bechmark levels are set above curret performace, (except for availability, which is set at 95 per cet of curret performace). Bidders are icetivised by a paymet mechaism to meet these targets ad they icur pealties if performace declies. The value for moey of the PPP approach depeds o the extet to which this performace specificatio avoids perverse icetives ad successfully icetivises private sector etities to deliver log-ru improvemets i the service provided to Lodo Udergroud's passegers. 9 9 The specificatio balaces capability, availability ad ambiece. There are trade-offs betwee the three areas. For example, passegers may tolerate poorer ambiece, such as less pleasat eviromets from which to board trais, i retur for a higher trai frequecy. 11

15 log term partership: Lodo Udergroud will eed to cosider how well bidders would take to a log-term parterig arragemet with Lodo Udergroud ad Trasport for Lodo. I a ew 30 year arragemet, co-operatio o all sides is crucial ad may alleviate some of the strategic ad cotractual risks. Buildig partership i this case is complicated because of the hadover of overall resposibility for the Tube to Trasport for Lodo, whose staff have ot yet bee directly ivolved i the procuremet process, although they have bee cosulted at certai poits. risk maagemet: the successful implemetatio of improvemets to the Tube ifrastructure requires effective risk aalysis ad maagemet. Ay value for moey decisio should be supported by a comprehesive risk aalysis coverig both operatioal risks (for example, the risk of deterioratig performace uder the PPP framework) ad cotractual risks (for example, the risk of cotract termiatio ad chage of cotrol of wiig bidders). This aalysis should be clear about which party bears each risk ad who is best able to maage it. 48 Lodo Udergroud has cofirmed that it will cotiue to take all these factors ito accout i its fial value for moey appraisal. We ited to prepare a further report after the fial decisio has bee take. 12

16 49 The precedig sectios have cosidered the process udertake by Lodo Udergroud ad the extet to which the figures from the Comparators ca be relied o for decisio-makig. This fial sectio provides our view of the steps that Lodo Udergroud should take before it ca make a fial recommedatio to Lodo Trasport ad the Departmet. Usig the fiacial aalysis i decisiomakig 50 Lodo Udergroud recogises the limitatios of the Comparators exercise. It has explicitly addressed how the fiacial iformatio ca best be iterpreted by decisio makers, i the followig ways: it has recogised the eed to focus o the risks ad beefits of the various optios, ad ot just o the headlie et preset values. it has recogised that the fiacial model relies o subjective judgemets, ad it has ow decided ot to preset sigle expected values. It has used rages istead of sigle values to provide a guide to the boudaries of the Comparators. it proposes to explai the cotet ad sigificace of the PPP bids ad the Comparators, rather tha just the fiacial differeces. it also proposes to preset to decisio-makers a rage of plausible fiacial figures, uder differet operatig ad fiacig scearios. As part of this work, it has idetified the largest factors drivig public sector costs - ivestmet i statios, sigallig, track ad rollig stock, ad maiteace of rollig stock - ad assessed the impact of chages i these factors o overall et preset value. Lodo Udergroud also iteds to aalyse the fiacial impact of assumig that the public sector ad the private sector achieve idetical levels of performace to illustrate relative cost differeces. 51 Despite the fact that the fiacial aalysis has limited use, it icludes valuable iformatio. Lodo Udergroud should, as iteded, esure that it fully uderstads the source of ay fiacial differeces betwee the public sector ad PPP optios. This level of aalysis will help to explai what, if ay, extra efficiecies private sector etities propose to brig to the Tube, the extet to which these claims are credible ad why they may ot be achievable by a public sector operatio. 52 I comig to a decisio about the likely value for moey offered by each of the optios, due weight should be attached to all relevat factors. The fiacial aalysis examied i this report is oly oe of these factors. It is ot a pass/fail test, ad the ucertaity surroudig the fiacial figures must be made clear. Lodo Udergroud should cotiue to address the wider issues fallig outside the fiacial aalysis ad discussed i paragraphs 47 ad 48 of this report. 53 There are other factors o which Lodo Udergroud places sigificat weight, icludig safety, bidders' ability to strike a timely ad acceptable deal ad the importace of developig a geuie partership. It should also esure that it makes clear the assumptios lyig behid ay affordability aalysis that feeds ito the decisio. Ad if ay material chages to the fiacial or 13

17 commercial terms arise durig egotiatios with a preferred bidder, Lodo Udergroud should coduct a further aalysis of the value for moey offered compared to public sector provisio. 54 Lodo Trasport's decisio to award each cotract will be made idepedetly of the other cotracts. This meas that there could be a mix of public ad private provisio, with oe or two of the three ifrastructure compaies maaged by successful bidders ad the other(s) retaied i the public sector. If this optio is to be give serious cosideratio, Lodo Udergroud will eed to esure that sufficiet aalysis of the impact of this mix has bee udertake ad that plas for copig with these outcomes exist. Such a solutio would etail additioal complex cotractual ad operatioal iterfaces betwee public ad private sectors, but may also offer strategic beefits by allowig more detailed bechmarkig of ifrastructure costs. 55 Regardless of which mix of public ad private sector provisio is selected, the successful implemetatio of improvemets to the Tube requires effective risk maagemet. As part of this process, Lodo Udergroud should cotiue to esure that it assesses the capacity of the bidders ad of the public sector to maage risks; that it has udertake a comprehesive risk aalysis coverig the whole 30 year period; ad that effective maagemet ad moitorig arragemets for the 30 year period are i place. 14

18 Glossary Ambiece Availability Base costs BCV Best ad fial offers Capability Cost of capital Deep tube Discout rate Efficiecies/efficiecy factors Health ad Safety Executive A output measure coverig the quality of the eviromet for passegers, icludig the clealiess ad geeral coditio of trais ad statios, ad the provisio of passeger iformatio. A measure of passegers' total additioal jourey time resultig from disruptio caused by icidets attributable to the Ifraco. Lodo Udergroud's estimate of what it would sped to ehace, maitai ad maage the ifrastructure over 30 years i accordace with the PPP performace specificatio. Bakerloo, Cetral, Waterloo ad City, ad Victoria lies. The fial bids made i the competitio betwee private sector bidders. A performace measure of the ifrastructure's ability to support trai services. It is based o average jourey time per passeger, for a give time of day, ad for a give lie or part of a lie. The rate of retur implicit i the sums repayable by a orgaisatio to those bodies that provide it with fudig. BCV ad JNP lies. The percetage rate applied to cash flows to eable comparisos to be made betwee paymets made at differet times. The rate quatifies the extet to which a sum of moey is worth more today tha the same amout i a year's time. Moetary adjustmets to the public sector comparator base costs ad the risk adjustmets to reflect the extet to which maagemet ca reasoably foresee beig able to deliver the ifrastructure services i the public sector at a lower cost i the future. The statutory body which icludes HM Railways Ispectorate ad is resposible for acceptig ad eforcig Lodo Udergroud's railway safety case (the statemet which sets out how it will hadle safety). Ifrastructure compaies (Ifracos) The three orgaisatios (BCV, JNP ad SSL) resposible for deliverig ifrastructure services to Lodo Udergroud uder the PPP cotracts. Ifrastructure Ifrastructure service charge Ifrastructure operatio/service JNP Railway, trais, statios ad depots o Lodo Udergroud's etwork (icludig track, sigals, tuels, bridges, embakmets, platforms, escalators, lifts). The amout payable uder the PPP cotract by Lodo Udergroud to the Ifraco, as adjusted from time to time, to cover the Ifraco's costs of maitaiig, reewig ad upgradig the ifrastructure, icludig overheads, profit ad fiacig costs. Work to maitai, reew ad upgrade the ifrastructure so as to deliver the outputs required uder the PPP cotract, to progressively reduce ad elimiate the existig backlog i asset maiteace, ad to deliver a cotiuous overall improvemet i asset health, capability ad reliability i service. Jubilee, Norther ad Piccadilly lies. 15

19 Lodo Trasport Lodo Udergroud Log Term Ivestmet Pla Output(s) Periodic review PPP PPP performace specificatio Public sector comparator Reputatioal exterality Risk adjustmets/ risk factors Sub-Surface Lies (SSL) Trasport for Lodo Lodo Regioal Trasport - a atioalised idustry curretly resposible for Lodo Udergroud ad aswerable to the Departmet of the Eviromet, Trasport ad the Regios. Lodo Udergroud Limited - a subsidiary of Lodo Trasport - resposible ow ad i the future for operatig passeger trais ad statios ad beig resposible for safety. It iteds to award PPP cotracts to private sector bidders ad will be their public sector parter uder those cotracts. After PPP cotracts are siged, it will remai i the public sector but will be trasferred to Trasport for Lodo. Lodo Udergroud's 30 year pla for the major projects ad other capital expediture which they estimate will be required to meet the PPP performace specificatio; private sector Ifracos are ot boud to adhere to this pla. Capability of the trai service; availability ad ambiece of trai services ad statio services; improvemets i asset health. A provisio i the PPP cotract which, every 7½ years, allows Lodo Udergroud's paymets to the Ifracos to be reset to take accout of chaged circumstaces ad cost icreases which a "ecoomic ad efficiet" Ifraco would icur. Public Private Partership. Specifically, the partership betwee the public sector Lodo Udergroud ad the private sector ifrastructure compaies uder the 30-year PPP cotracts. The levels of output which the Ifraco is required to deliver uder the PPP cotract, coverig "ambiece, availability ad capability" for trai ad statio services. A bechmark agaist which value for moey is assessed. It is typically a cost estimate based o the assumptio that assets are acquired through covetioal fudig ad that the procurer retais sigificat maagerial resposibility ad exposure to risk. I this case, it is Lodo Udergroud's estimate of the cost to the public sector of procurig the maiteace, reewal ad upgrade of the ifrastructure so as to deliver the PPP outputs. Separate comparators exist for raisig fiace through (a) traditioal fudig such as public sector grat ad (b) log term bods. The estimated cost impact that additioal borrowig would have o the Govermet's reputatio for fiacial prudece. Moetary adjustmets to the public sector comparator's base costs to reflect the probability that, due to uforesee evets, services will ot be delivered by the public sector Ifraco at the cost show i the base estimates. These adjustmets are oly icluded for risks that Lodo Udergroud believes it will trasfer to the private sector uder the PPP cotracts. Circle, District, Metropolita, East Lodo ad Hammersmith ad City lies. These lies are oly just below groud level, havig bee built i 'cut ad cover' tuels. The body appoited uder the Greater Lodo Authority Act 1999 which has take over Lodo Trasport's resposibilities, apart from Lodo Udergroud. After the PPP cotracts are siged, Lodo Udergroud Limited will be trasferred to Trasport for Lodo ad will be aswerable to the Mayor of Lodo for the service it delivers. 16

20 Appedix 1 The Public Private Partership approach 1 O 20 March 1998 the Deputy Prime Miister aouced that the Govermet would be itroducig Public Private Parterships (PPP) to brig stable, icreased ivestmet ito Lodo Udergroud. The plas ivolved dividig Lodo Udergroud ito a publicly-owed operatig compay resposible for deliverig services to customers, ad three privately owed ifrastructure compaies. The objectives were to deliver 8 billio worth of ivestmet i Lodo Udergroud over the ext 15 years, to maitai ad upgrade the ifrastructure, improvig services for passegers ad with the itetio of providig value for moey for the taxpayer. 2 Uder the PPP, trai ad statio services would cotiue to be plaed ad operated by a publicly owed, publicly accoutable Lodo Udergroud. Lodo Udergroud would be resposible for safety o the whole of the Udergroud. Private compaies have bee ivited to bid for the resposibility to maitai ad upgrade the track, tuels, sigals, statios, lifts, escalators ad trais uder 30 year cotracts to Lodo Udergroud. At the ed of the cotracts, the assets would retur to the public sector. 3 The lie groupigs that have bee offered to the private sector are: 3 Bidders for the deep tube cotracts BCV JNP LINC: Tube Lies Group: Bombardier Bechtel/Halcrow Mowlem Amey Fluor Daiel Hyder Alcatel Jarvis Aglia Water Metroet: TubeRail: Adtraz Brow & Root WS Atkis Alstom Balfour Beatty Amec Seeboard Carillio Thames Water BCV - Bakerloo, Cetral, Waterloo ad City ad Victoria lies. JNP - Jubilee, Norther ad Piccadilly lies. Sub-surface - Circle, District, Metropolita, East Lodo ad Hammersmith ad City lies. 6 The private sector would take over three ifrastructure compaies (Ifracos), ad the resposibility for deliverig a large ifrastructure ivestmet programme. These compaies are beig asked to provide log term ivestmet plaig, professioal project maagemet ad effective delivery of day to day maiteace. 4 Two bidders were shortlisted for the BCV ad JNP cotracts i July 2000, LINC ad Metroet for BCV; ad Tubelies Group ad Tuberail for JNP. Members of each cosortium are show opposite. 5 This report is oly cocered with the BCV ad JNP groups, as the SSL bid process is o a later timescale. 4 How the Tube would be ru i 2001 Passegers New Lodo Udergroud 7 Uder the PPP, the private sector would ot have ay operatioal resposibilities; o the etwork they would maily be resposible for the work that goes o overight, while the system is shut dow. Operatios - icludig drivers, statios staff ad resposibility for safety - would remai i the public sector. 8 Trasport for Lodo (TfL) is a ew body, reportig to Lodo's Mayor. It is resposible for a wide rage of trasport i Lodo, primarily implemetig the Mayor's itegrated trasport strategy. O 3 July 2000 TfL took o all of Lodo Trasport's previous resposibilities apart from Lodo Udergroud. Oce the competitios are completed, Lodo Trasport will be woud up ad Lodo Udergroud will be trasferred to TfL ad the Mayor. Ifraco Sub surface Ifraco BCV Ifraco JNP Source: Departmet of the Eviromet, Trasport ad the Regios. 17

21 Appedix 2 Methodology 1 The Natioal Audit Office examied the Comparators ad their applicatio to iform the decisio to select a preferred bidder for the proposed Lodo Udergroud Public Private Partership. 2 We used a issue aalysis approach to desigig the scope ad ature of evidece required to complete the examiatio. That is, we set a series of highlevel audit questios that we cosidered it would be ecessary to aswer i order to assess the Comparators ad their use, ad collected evidece accordigly. For each of the top level questios, we set a subsidiary group of questios, liked logically to the mai questio, i order to direct our detailed work ad aalysis. 3 The top level questios we set were: whether the relevat guidace had bee followed; whether the Comparators were soud; whether the evaluatio of the bidders' prices was beig coducted soudly; ad whether the compariso betwee the bid prices ad the Comparators were likely to be soud. Figure 5 summarises these primary issues ad the more detailed areas cosidered uder each. 4 Our mai evidece came from detailed examiatio of the Comparator models, ad iformatio provided by the Departmet ad Lodo Udergroud's files. We also udertook iterviews with key staff, ad discussios with Lodo Udergroud's fiacial advisers, PricewaterhouseCoopers. Fially we discussed issues of cocer with major stakeholders i the project, such as the Mayor of Lodo, the Trasport Commissioer for Lodo ad bidders. 5 I additio, we appoited a Review Pael, which cotaied experts from outside the Natioal Audit Office. This pael provided advice o issues ad aspects of the report. 6 Exteral members of the Review Pael were: Professor Vic Barett, Nottigham Tret Uiversity Kigsley Maig, Maagig Director, Newchurch Limited Professor Toy M. Ridley CBE, FREg, Emeritus Professor of Trasport Egieerig, Imperial College Michael C. Spackma, Natioal Ecoomic Research Associates George Steel, Maagig Director INDECO (Iteratioal Maagemet Cosultats) Limited Bruce O. B. Williams 18

22 5 The Natioal Audit Office s Issue Aalysis Has cosideratio bee give to areas of guidace with which the Natioal Audit Office disagrees? Have the requiremets of the Gree Book ad Treasury Task Force bee followed? Icremetal adjustmets reflected i base costs? Reasoable expectatio that capital programme will meet PPP requiremet ad alterative capital programmes evaluated? Factors sesibly adjusted whe costs chage? Iput base capital ad operatig costs reasoable ad revisios robust ad defesible? Risk ad efficiecy factors affect ability to meet specificatio? Has relevat guidace bee followed Adjustmets avoid double coutig? Risks ad efficiecies fairly idetified ad evaluated ad assumptios valid? Passeger demad, iflatio ad taxatio treatmet reasoable? Bod fiacig optios rigorously costructed ad assumptios valid? Required outputs effectively produced ad uderstadable? Ecoomic ad fiacig assumptios reasoable? Are the Public Sector Comparators soud? Risk ad ucertaity treatmet reasoable? Hardwired iputs evideced ad Model logical ad free from codig error? Modellig soud ad reasoable? Mai elemets of bids uderstood by Lodo Udergroud? Robust evaluatio methodology developed ad followed? Assessmet of each bid price fair ad reasoable thus far? Will the compariso betwee the Best ad Fial Offer bids ad the Public Sector Comparators provide a soud guide to the fiacial aspects of the decisio for the Lodo Udergroud Public Private Adjustmets to bids evideced ad defesible? Curret rakig of the bids o cost soudly based o et preset cost ad beefits? Is the evaluatio of the bidders prices beig coducted soudly? Comparable level of detail to allow uderstadig of cost drivers? Sigificat differeces i iput profiles? Performace ad over-performace outcomes comparable ad evaluated reasoably? Bid prices stated i comparable fashio ad bids comparable with the Public Sector Comparators? Use of expected values ad discoutig rate ad method valid? Methods beig used to compare the bid prices ad the Public Sector Comparators valid? Is the compariso betwee the bid prices ad the Public Sector Comparators likely to be soud? Treatmet ad calculatio of ucertaity valid ad relevat sesitivity aalysis udertake? Sigificat differeces betwee bid prices ad the Public Sector Comparators curretly uderstood ad explicable to Lodo Udergroud? Public Sector Comparator Performace variatios compared? Causes of differece cofirmed as valid ad egotiated with bidders? All quatifiable exteral differeces idetified ad adequately priced? Coclusios beig draw from the compariso reasoable? Decisio makers provided with adequate iformatio for decisio? Clear basis for decisio o fiacial grouds? Impact of o-quatifiable factors beig cosidered? Coclusios appear cosistet with available iformatio? 19