Improving the PFI tendering process

Improvig the PFI tederig process REPORT BY THE COMPTROLLER AND AUDITOR GENERAL HC 149 Sessio 2006-2007 8 March 2007

The Natioal Audit Office scrutiises public spedig o behalf of Parliamet. The Comptroller ad Auditor Geeral, Sir Joh Bour, is a Officer of the House of Commos. He is the head of the Natioal Audit Office, which employs some 850 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. Our work saves the taxpayer millios of pouds every year. At least 8 for every 1 spet ruig the Office.

Improvig the PFI tederig process LONDON: The Statioery Office 13.50 Ordered by the House of Commos to be prited o 5 March 2007 REPORT BY THE COMPTROLLER AND AUDITOR GENERAL HC 149 Sessio 2006-2007 8 March 2007

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 Comptroller ad Auditor Geeral Natioal Audit Office 1 March 2007 The Natioal Audit Office study team cosisted of: Philip Airey, Be Morris ad James Robersto with assistace from Ceria Davies, Christia Petzold ad Liam Worboys. This report ca be foud o the Natioal Audit Office web site at www.ao.org.uk For further iformatio about the Natioal Audit Office please cotact: Natioal Audit Office Press Office 157-197 Buckigham Palace Road Victoria Lodo SW1W 9SP Tel: 020 7798 7400 Email: equiries@ao.gsi.gov.uk Natioal Audit Office 2007

CONTENTS SUMMARY 4 PART ONE Itroductio 8 The Public Accouts Committee has expressed 8 cocers about the tederig of PFI deals Guidace for Departmets to help resolve these 8 cocers has bee issued New public procuremet regulatios require a 10 radical chage i the way PFI deals are tedered PART TWO PFI projects are attractig fewer bidders 12 who are prepared to submit detailed bids There is evidece that PFI projects are receivig 12 fewer developed bids tha previously Legthy tederig periods ad iadequate 13 preparatio of projects have put off some bidders The applicatio of ew public procuremet 15 regulatios may make bidders more selective PART THREE Tederig for PFI deals ca be improved 16 It takes o average just uder three years to 16 teder ad close a PFI deal Almost half of the time take to teder a PFI 16 deal is take up by egotiatios with a sigle preferred bidder Tederig ad egotiatio periods could ofte 18 have bee reduced The direct cost of tederig delays was at 22 least 67 millio Sigificat chages were made to oe-third of 24 projects at the preferred bidder stage A variety of methods were used by Authorities 26 to check the cost of chages but they could ot be wholly satisfactory APPENDICES 1 Scope ad methodology of the 27 Natioal Audit Office s examiatio 2 Projects icluded i the survey work 30 3 List of bodies iterviewed for the study 32 4 Statistical methodology 33 Photographs courtesy of Digital Visio

SummARy 1 Before ew PFI assets are costructed ad services delivered, there is a process by which procurig authorities ivite teders ad select a wiig bidder for the cotract (Figure 1). I order for value for moey to be achieved i PFI deals, all stages of the deal have to be maaged effectively, icludig this tederig stage. 2 This study has arise out of cocers expressed by the Committee of Public Accouts i 2003 that the tederig of PFI projects did ot follow good practice ad was ot hadled with sufficiet skill o the part of the public sector, icurrig high costs ad riskig value for moey. New Europea Uio procuremet rules reiforce the eed for best practice i the future. 3 We examied the tederig process for all cetral Govermet Departmet PFI projects i Eglad that closed betwee April 2004 ad Jue 2006, icludig PFI schools ad hospital projects. 1 The effectiveess of this process impacts directly o the value for moey of PFI deals, though it is outside the scope of this report to provide a evaluatio of the value for moey of each sigle deal that closed over the period. 2 1 We surveyed 49 projects with a combied capital value of just uder 8 billio. We excluded a small umber of local authority projects (combied capital value of 750 millio) ad projects with idividual capital values of uder 20 millio that would ow ot be cosidered for PFI uder Treasury guidelies. 2 See Appedices 1 ad 2 for details of the methodology ad projects covered i this study. 4 ImPROvING THE PFI TENDERING PROCESS

summary 4 We foud that key elemets of the tederig process had ot improved ad i some respects had worseed: a b c There are sigs that the private sector is becomig more selective i developig detailed bids 3 for PFI projects, i part due to the cumulative impact of legthy tederig periods ad high bid costs. Oe i three projects that closed betwee 2004 ad 2006 had two detailed bids competig for the busiess, compared with oe i six authorities prior to 2004. Though with differeces betwee sectors, tederig periods overall lasted a average of 34 moths (25 moths average for PFI schools, 38 moths for PFI hospitals ad 47 moths for other PFI projects) - o better tha the average for projects that closed betwee 2000 ad 2003. 4 Shorter tederig periods are ot desirable if they are achieved at the expese of the overall value for moey of the projects. However, we foud that may of the reasos for log tederig periods (some of which may relate equally to covetioally procured projects) could have bee avoided or mitigated by the public sector, without riskig overall value for moey. Withi the overall tederig period, egotiatios to fialise deals with a sigle preferred bidder have icreased, lastig o average over a year ad i some cases as log as five years. Material chages (both upwards ad dowwards) were frequetly made by public sector project teams ad cotractors to the prices of deals durig preferred bidder egotiatios, whe the disciplie d e of competitive tesio had bee removed. I oethird of projects we examied, there were major scope ad specificatio chages (both upwards ad dowwards) durig the preferred bidder period worth just over 17 per cet of the projects preset values or a average of 4 millio a year for each project. Project teams have cotiued to pla less well tha they should for the amout of professioal advice eeded for a PFI deal. Where budgets were set, spedig was o average 75 per cet more tha aticipated, or 0.9 millio extra per project. The average cost of exteral advice for all projects was just over 3 millio per project or approximately 2.6 per cet of the capital value of the projects. I the health sector, for which data are available over time, the amouts spet o advisers as a proportio of the value of deals has goe dow margially sice 1997-2000. Although may project teams told us that they passed o lessos leared to others, systematic ways to esure that useful lessos are shared were ot always exploited. Some issues commo to PFI deals were resolved by project teams actig i isolatio. However, the issuace of guidace o stadard PFI cotract terms was oe way i which lessos were captured ad the Treasury believes that the further developmet ad applicatio of stadard terms will yield further beefits i relatio to both cotractual terms ad improved procuremet times. 1 The positio of the tederig process withi a PFI project Optio appraisal ad project preparatio Tederig process Decidig o the scope of the project Securig busiess case approval Ivitig bidders to submit ad develop teders Choosig a preferred bidder Costructio Service delivery Preparig to go to the market Negotiatig the fial agreemet Source: Natioal Audit Office 3 Detailed bids for this purpose are defied as bids submitted at the Ivitatio to Negotiate Stage. 4 The overall average for deals that closed 2000-03 was 33 moths. I o sector was there sigificat improvemet compared with the period 2000 to 2003. Our figures for tederig times differ slightly from the figures quoted i Stregtheig Log-Term Parterships (HM Treasury, 2006). This is due to a differet project selectio: further details ca be foud i Appedix 1. Improvig the PFI tederig process 5

summary 5 Public authorities are ow expected to use a ew procuremet procedure kow as Competitive Dialogue. 5 Uder this procedure, more of a PFI deal has to be agreed with all bidders before a preferred bidder is selected tha has bee the case i the past, so maitaiig competitive tesio for loger ad reducig the scope to make sigificat chages to the deal oce the competitio has bee closed, as happeed i the past. Some sector-specific guidace has ow bee issued to procurig authorities, however, the practical effects of Competitive Dialogue are ucertai at this early stage. The ehaced competitive elemet withi the ew procedure will i priciple brig beefits, but ay risk of icreased tederig costs for the private sector will eed to be maaged so that bidder iterest does ot weake. Recommedatios 6 I its March 2006 publicatio, Stregtheig Log-Term Parterships, the Treasury put forward several proposals to improve public sector skills ad procuremet support ad to reduce procuremet timescales ad costs for both the public ad private sectors. These icluded a greater emphasis o preparatio of projects before they go to the market, icreased moitorig ad scrutiy of projects, icludig a ew stage of scrutiy before selectio of preferred bidder, ad measures to address shortages of skills i the public sector. Other measures which had already bee take to address issues of lesso learig, co-ordiatio ad skill shortages icluded the creatio of sector-specific programmes such as Buildig Schools for the Future. This programme brigs together all future PFI school projects with the aim of itroducig cetralised programme maagemet. Our recommedatios are iteded to supplemet ad to complemet these proposals ad measures, ad will remai relevat i the light of the Competitive Dialogue process. Addressig the risk of PFI deals ot receivig eough developed bids for a viable competitio a The level of public sector experiece is a factor that ca ifluece bidders iterest i projects. Authorities should always esure that they ca draw o staff with experiece of complex capital procuremet ad Departmets, icludig where relevat their Private Fiace Uits (PFUs), should also have sufficiet specialist kowledge o which to draw. As part of the reform of the Govermet Procuremet Service 6, the Office of Govermet Commerce (OGC) is cosiderig how to facilitate b c d the effective recyclig of existig skills i complex procuremet across the public sector ad to promote a attractive career path i complex procuremet, such as PFI, backed by a structured traiig ad developmet programme. Public sector procuremet teams have i the past aimed to receive detailed bids from at least three bidders. Uder Competitive Dialogue, there may be circumstaces i which, after elimiatig weaker bidders, it makes sese to udertake the later stages of the dialogue with the two strogest bidders. However, where oly two viable bids for a project are received early o, or if bidders pull out of the competitio, leavig the procurig authority with oly two bids to choose from 7, there should be a review by the relevat sposor Departmet. The review should cosider whether: there are ay defects i the scopig or maagemet of the project that may explai the low level of market iterest ad could be remedied i time for a re-ru of the competitio; ad whether the bids o the table offer a good competitio ad are likely to lead to a value for moey solutio. The OGC should esure that there are cosistet priciples ad comprehesive guidace for practitioers across the public sector i applyig the Competitive Dialogue procedure to PFI procuremet. The OGC should also carry out a review after the first 18 moths of the operatio of Competitive Dialogue to idetify ay lessos, i particular to esure that private sector bidder iterest is maitaied. For the first time there is ow a database maitaied by the OGC cotaiig iformatio about plaed costructio projects across the public sector. Departmets should use this database to assess the impact of their plaed projects o the market. I additio, the OGC should use the database to provide the market with iformatio o whe projects are expected to proceed to tederig ad to moitor whether projected demad from the public sector is likely to exceed market capacity. Departmets ad, where relevat, the Treasury should also cotiue to assess whether project teams have tested likely market iterest as part of the Outlie Busiess Case approvals process. 5 This stemmed from a EU Directive, implemeted ito UK law from Jauary 2006, which added the optio of Competitive Dialogue to the existig rage of public procuremet procedures. 6 Trasformig govermet procuremet (HM Treasury, Jauary 2007). 7 I cases where oly oe bid has bee received, the Treasury has published specific value for moey guidace. 6 Improvig the PFI tederig process

summary Reducig the legth ad cost of tederig e f We have foud examples of well maaged ad properly resourced projects that have take 18 moths to teder, icludig preferred bidder egotiatios lastig less tha six moths. This suggests that a target of betwee 18 to 24 moths would ot be ureasoable for may projects, although it may be urealistic for particularly complex, oe-off PFI deals. Departmets should agree with the Treasury what would costitute a appropriate target time for their sector, ad idividual procuremet teams should be boud to this uless they ca satisfy the Departmet that the target would be urealistic i their case, eve with more up-frot preparatio. To achieve much tighter timescales while maitaiig good value for moey, Authorities should: obtai commitmet to the project from all key stakeholders at a early stage; develop better output specificatios, icludig greater dialogue with potetial bidders about the desig of assets, before approachig the market; establish the affordability of the project before it is brought to the market ad agai before a preferred bidder is selected. I establishig affordability, authorities should calculate available resources agaist a rage of scearios; ad agree the commercial basis of a deal as well as key aspects of the detailed desig prior to selectig a preferred bidder ow a requiremet uder the Competitive Dialogue procedure. h i j Parterships UK maitais a database of iformatio o PFI projects, parts of which are available o its website. The Treasury should cosider whether iformatio o the legth of time take to procure idividual deals should be published, possibly i the form of a aual league table, to help motivate project teams to achieve shorter tederig times. The same issues arise repeatedly across projects, legtheig procuremet periods ad icreasig costs uecessarily. Departmets should idetify lessos from recetly closed PFI projects of relevace to subsequet projects, revisig sector specific guidace ad stadard specificatios where ew issues recur across projects i a particular sector. Where appropriate, a programme approach to PFI projects such as that uder Buildig Schools for the Future ca facilitate the trasfer of experiece from earlier to later deals. There should also be a more structured process of learig ad sharig lessos across sectors ad public authorities, which icludes: Departmets esurig that post-project evaluatios are completed as a matter of course ad ay lessos shared with other public sector procuremet teams; a more co-ordiated ad targeted approach to sharig good practice by cetral advisory bodies such as the OGC, PUK, 4ps ad the Project Review Group withi the Treasury; ad a forum through which procurig Authorities ca share their experieces ad raise queries, to complemet the existig PUK helplie. g As part of good project ad programme maagemet usig appropriate methodologies, the moitorig of projects as they progress through procuremet should be ogoig ad highlight where target times are likely to be missed. I these cases, the likelihood of missig the target time should serve as a alarm sigal ad trigger a actio pla from the procurig authority. Improvig the PFI tederig process 7

PART ONE Itroductio This part of the report sets out the reasos for carryig out this study o PFI tederig, ad provides the cotext of ew procuremet rules that came ito force i Jauary 2006. May procurig authorities had appoited a preferred bidder eve though importat issues remaied uresolved. Negotiatios with preferred bidders eeded to be kept to a short ad tight timetable. 1.1 The Private Fiace Iitiative (PFI) was lauched i 1992. Sice the successive govermets have cotiued to egotiate projects that draw o private sector expertise, icludig the raisig of fiace, to provide services which had bee provided directly by the public sector. I total, over 750 PFI deals have ow bee siged with a combied capital value of 55 billio. The Public Accouts Committee has expressed cocers about the tederig of PFI deals 1.2 I 2003 the Public Accouts Committee published a report 8 that highlighted commo problems i obtaiig value for moey from the PFI ad expressed cocers that: Procurig authorities did ot have staff with the right skills critical to good project maagemet. There was a eed for authorities to give much greater emphasis to developig such skills ad to adopt best practice more widely. The cost of employig advisers was very high ad i may cases exceeded budgets by a substatial margi. Procurig authorities eeded to drive dow advisers costs ad esure that sesible budgets were adhered to. They also eeded to be midful of costs to bidders. Imposig excessive costs o bidders was likely to result i higher charges i the log ru ad risked deterrig firms from biddig. The Committee cocluded that the taxpayer was ot always gettig the best deal from PFI cotracts because good procuremet practice was ot beig followed. Guidace for Departmets to help resolve these cocers has bee issued 1.3 Figure 2 shows the pricipal measures take by the Govermet sice 1997 to improve PFI procuremet. Some of the more importat measures iclude: the developmet of sector-specific stadard forms, itroduced to improve the quality ad cosistecy of PFI cotracts ad to reduce egotiatio times; the itroductio of ew sector-specific delivery models such as the Buildig Schools for the Future programme; ad the itroductio of a lower limit of 20 millio (capital value) below which tederig costs were cosidered to be disproportioately high. However, the Govermet has recetly recogised that challeges remai, acceptig that procuremet timeframes are still uecessarily log. 9 It has committed to a series of further measures desiged to stregthe frotlie procuremet skills, to icrease cetral scrutiy of projects ad to esure that projects are better prepared before they are brought to the market. 1.4 To test whether the issues highlighted by the Public Accouts Committee i 2003 are beig addressed, we collected iformatio o the tederig process for projects which closed betwee 2004 ad 2006 ad with a combied capital value of 7.78 billio. 10 Details of our methodology are set out i Appedix 1. 8 28th report from the Committee of Public Accouts, Deliverig better value for moey from the Private Fiace Iitiative (HC 764, Sessio 2002-2003). 9 HM Treasury (2006),. 10 The work covered all projects procured either directly by cetral Govermet or as part of major ad well-defied cetral Govermet programmes coverig PFI hospitals ad schools. It excluded local govermet projects such as waste, social housig ad street lightig. See Appedix 1 for more details. 8 ImPROvING THE PFI TENDERING PROCESS

part oe 2 Measures take by the Govermet to improve PFI procuremet Date Sice 1997 Govermet Measures Developmet of Private Fiace Uits (PFUs) Descriptio Dedicated uits providig advice ad co-ordiatig the use of PFI withi Departmets. Sice 1999 Itroductio of stadard PFI guidace ad associated sectorspecific cotracts Itroduced to promote a commo uderstadig of the mai risks which are ecoutered i a stadard PFI project, to reduce the period ad costs of egotiatio ad to ecourage a cosistet approach betwee projects. Compliace with Stadardisatio of PFI Cotracts (SOPC) was made madatory from May 2004. 2000 Establishmet of the Office of Govermet Commerce (OGC) Works with Govermet Departmets to improve value for moey i commercial activities. 2000 Establishmet of Parterships UK A Public Private Partership with a public duty to improve the delivery of parterships betwee the public ad private sectors. 2001 Itroductio of the Gateway process A cofidetial assessmet maaged by the OGC of the deliverability of projects, icludig PFI projects. 2003 Meetig the Ivestmet Challege A policy documet which explaied the Govermet s approach to PFI, set ew limits o its appropriate use, ad proposed improvemets both to the process of assessig value for moey ad to the systems for deliverig projects. 2004 Value for Moey Assessmet Guidace The Treasury published guidace o assessig value for moey i PFI trasactios i August 2004. This guidace was updated i October 2006. 2004 Establishmet of the Buildig Schools for the Future (BSF) programme This is iteded to improve the delivery model for PFI ad o-pfi schools ad is beig ru by a ewly-created dedicated body, Parterships for Schools. 2005 Reform of the Project Review Group (PRG) Itroduced i 1998 to act as a approval mechaism for local govermet projects, the PRG was reformed i 2005 to iclude the itroductio of a secod stage review prior to the appoitmet of a preferred bidder. 2006 Stregtheig Log Term Parterships A documet which idetified a umber of issues i the procurig of PFI projects ad which proposed measures to address them, icludig a re-affirmatio that projects eeded to be properly developed before goig to the market, the stregtheig of Departmetal PFUs, greater scrutiy of projects before preferred bidder selectio ad improvig procuremet skills through better traiig. Source: Natioal Audit Office Improvig the PFI tederig process 9

part oe New public procuremet regulatios require a radical chage i the way PFI deals are tedered 1.5 The procuremet of goods ad services by public authorities i the UK is govered by Europea Uio Directives, desiged to promote ad ecourage trasparet ad fair competitio betwee cotractors i EU member states. Chages to these Directives have bee implemeted i UK law from 31 Jauary 2006. Promiet amog the chages is the ew procuremet procedure of Competitive Dialogue for complex projects, such as PFI deals (Figure 3). 1.6 Oe of the mai features of the ew procedure is that there is less scope to make chages to a project after a preferred bidder has bee selected. Although there is flexibility withi Competitive Dialogue for bidders ad the authority to discuss how the output specificatio will be met, oce the competitive phase has closed, bidders ca oly be requested to fie tue, specify ad clarify their bids. Ay chage to the preferred bid must ot substatially modify what had bee agreed ad must ot distort the outcome of the competitio, though the defiitios of the scope or other chages that will be allowed durig the preferred bidder period have yet to be tested i practice. 1.7 From 31 Jauary 2006, the Competitive Dialogue process codifies i law what should have bee good practice uder the previous procuremet procedure, for example: up-frot plaig with a clearly defied scope ad a well-developed draft output specificatio before goig to the market; ot choosig a preferred bidder util all bidders offers have bee fully probed uder competitive coditios. 1.8 The itroductio of Competitive Dialogue should have beefits i terms of ecouragig better procuremet practice ad stregtheig the competitive elemet of the PFI procuremet process, reducig the scope to make sigificat chages to the deal oce the competitio has closed. This should reduce time take ad costs i the preferred bidder stage, but there are risks as well. I particular, there is a risk that the eed to egotiate more of the deal with a greater umber of bidders prior to preferred bidder selectio will icrease overall tederig costs for both the public ad private sectors, ad that this, i tur, will weake bidder iterest. 10 Improvig the PFI tederig process

part oe 3 The ew ad old procuremet processes (illustrative) Competitive dialogue procedure Negotiated procedure Plaig Plaig Prepare for Procuremet Icludes preparatio of Outlie Busiess Case ad the project specificatios Prepare for Procuremet Icludes preparatio of Outlie Busiess Case ad the project specificatios Go to market (issue OJEU otice) Go to market (issue OJEU otice) Competitive Phase Pre-qualify bidders Pre-Qualificatio Questioaire (PQQ) leadig to short-listig Competitive Phase Pre-qualify bidders Pre-Qualificatio Questioaire (PQQ) leadig to short-listig Ivitatio to Negotiate (ITN) Ivitatio to Participate i the Dialogue (ITPD) ad ope Competitive Dialogue Shortlist bidders usig Outlie Solutios Detailed solutios, clarificatio ad dialogue over all aspects Bidders produce detailed solutios based o full project specificatios [Sometimes split ito two tederig rouds: a prelimiary ITN ad a fial ITN] [Sometimes further short-listig of bids ad further rouds of tederig to develop bids e.g. a Best ad Fial Offer (BAFO) roud] Assess readiess to close dialogue/ cotiue refiig solutios Appoit Preferred Bidder Close Competitive Dialogue ad call for fial teders Closure Closure Preferred bidder egotiatios Appoit Preferred Bidder Parties oly allowed to clarify, specify ad fie tue existig agreemet Preferred bidder egotiatios Negotiatios to reach agreemet o may aspects of the deal Fiacial Close Fiacial Close Source: Natioal Audit Office Improvig the PFI tederig process 11

PART TWO PFI projects are attractig fewer bidders who are prepared to submit detailed bids This part of the report examies the stregth of the market for PFI projects as it affects tederig ad cosiders the potetial impact of ew EU public procuremet regulatios. There is evidece that PFI projects are receivig fewer developed bids tha previously 2.1 Strog competitio is essetial if PFI deals are to achieve the optimal mix of price, quality ad risk trasfer. We therefore looked at how effective the biddig process durig tederig has bee i geeratig ad maitaiig competitive tesio, focusig especially o the poit i the process where bidders are ivited to substatially develop their bids (the ivitatio to egotiate or ITN stage). It is importat that the public sector has strog competig bids to cosider at this poit. 2.2 We compared projects that closed at differet poits i time to explore treds i the umber of developed bids submitted. Figure 4 shows that 85 per cet of PFI projects icluded i our sample that closed prior to 2004 attracted three or more developed bids. However, the figure for a comparable umber of more recetly tedered projects has reduced to 67 per cet. Oe third of the projects icluded i our cesus, closig betwee April 2004 ad May 2006 (with o differeces betwee sectors), attracted oly two bidders at the poit they were requested to submit detailed bids. 11 2.3 I the schools sector, there is a ew delivery programme kow as Buildig Schools for the Future. Early idicatios from Parterships for Schools, the dedicated body set up to oversee the programme, suggest that competitio has bee relatively strog so far. Fiftee out of the 17 schemes that have reached ITN (or its equivalet uder the Competitive Dialogue rules) attracted three or more developed bids at that stage. 4 Percetage of projects 100 90 80 70 60 50 40 30 20 10 0 The proportio of projects attractig oly two viable bids has icreased i recet years 2003 or earlier Four or more bids Three bids Two bids Year of fiacial close 2004 to 2006 Source: Natioal Audit Office surveys 2006 11 i.e. at the Ivitatio to Negotiate (ITN) stage. 12 ImPROvING THE PFI TENDERING PROCESS

part two 2.4 I the majority of the projects receivig two developed bids that closed betwee April 2004 ad May 2006, this was because of isufficiet bidder iterest i the project. Either there were oly two bidders from much earlier i the tederig process, or existig bidders withdrew from the competitio. It was comparatively rare for these procurig authorities to choose to elimiate weaker bids: the choice was effectively out of their hads (Figure 5). Procurig authorities which received oly two viable bids at ITN argued that the impact o value for moey was limited. They cosidered that the two bids left o the table were of high quality, ad that two bids still allowed for competitive tesio (i cotrast with sigle bidder situatios). 2.5 It may become more commo uder competitive dialogue for procurig authorities to choose to go dow to two bidders at a earlier stage i the process tha has bee usual up to ow because of the eed to reach a greater level of agreemet prior to the selectio of the preferred bidder. Nevertheless, the absece of a third bid removes the possibility of havig a secod opiio bechmark o value for moey. It also leaves the procurig authority vulerable if oe bidder subsequetly pulls out. Legthy tederig periods ad iadequate preparatio of projects have put off some bidders 2.6 As Figure 6 overleaf idicates, the umber of developed bids received by procurig authorities has bee urelated to the size of the projects beig procured. Larger deals have bee as likely to receive relatively few developed bids as smaller deals. 12 2.7 Istead, our discussios with private sector cotractors, advisers ad others idicated that there were two pricipal reasos why bidders may have bee more selective i developig detailed bids for PFI projects: High bid costs ad legthy tederig periods, which reduced the umber of projects for which cotractors were prepared to bid i ay particular year (Box A overleaf). It was commo for cotractors to set a aual budget for how much they were prepared to sped o biddig for PFI ad they would ot exceed this. Greater iteratioal opportuities, which led to some compaies reducig their exposure to the UK market. 5 Most projects received two bids at ITN because of isufficiet bidder iterest (projects closig April 2004-May 2006) Oly ever two bidders Three bidders, oe withdrew at ITN leavig two Four or more bidders, but withdrawals at ITN leavig two Three bidders: Authority elimiated weakest bid to leave two bids Four or more bidders: Authority elimiated weaker bids to leave two bids Oly 4 projects received 2 bids at the ITN stage as a result of a deliberate choice by the procurig authority 0 1 2 3 4 Number of projects Source: Natioal Audit Office cesus 2006 12 See Appedix 4 for the statistical basis of this statemet. Improvig the PFI tederig process 13

part two 6 Number of bidders at ivitatio to egotiate stage 10 8 6 4 2 0 0 For projects closig betwee 1995 ad 2006 there has bee o relatioship betwee the size of a project ad the umber of viable bids received Source: Natioal Audit Office surveys 2006 BOX A 200 400 600 800 Capital value ( millio) 1,000 The views of the Busiess Services Associatio, a represetative body of PFI cotractors 2.8 I additio, there were four mai reasos which cotractors reported that led them to withdraw from competitios or to avoid particular projects altogether: i) Perceptios of the level of skills ad experiece cotaied withi public sector procuremet teams 1,200 I some ways capacity is limited because biddig compaies ow work to the assumptio that procuremet periods will typically be loger tha plaed for by Authorities Delays are limitig the desire of compaies to bid. 13 2.9 Cotractors regularly assessed the level of skills ad experiece withi public sector procuremet teams to help reach a view o the viability of a project ad the likelihood of delays. Just uder oe-third of procurig authorities reported that they had had isufficiet resources or i-house expertise for part or all of the tederig process. This will have had a adverse impact o the competece of the public sector team as see by potetial bidders. ii) Perceptios of the quality of preparatio for particular projects 2.10 I the view of cotractors, the adequacy of project preparatio was very importat i determiig whether to bid, as evideced by the quality of the project specificatios ad the overall teder documetatio. Cotractors thought there was scope to improve project preparatio ad also stressed to us the eed for certaity about the affordability of projects. iii) Iadequate plaig of deal flow 2.11 I some cases, projects attracted fewer bids because of iadequate pre-ojeu up-frot warig about the timig of projects beig brought to the market or because of iadequate co-ordiatio of deal flow (Box B). I either case, the result was that cotractors were ot i a positio to respod whe deals came to the market. Some cotractors also commeted that eve where the deal flow was well-maaged withi sectors, there was little co ordiatio betwee sectors. BOX B Kirklees City Coucil the importace of maagig deal flow Kirklees City Coucil first brought its special schools PFI project to the market i March 2002, just before the ed of that fiacial year. However, may other projects wet to the market at the same time, ad the Coucil oly received two bids. Followig cosultatio with the Departmet for Educatio ad Skills, the Coucil decided to re-submit the project to the market i May 2002. This time it received seve bids, of which six were viable bids. 13 Iterview with Busiess Services Associatio, 28 February 2006. 14 Improvig the PFI tederig process

part two iv) Geographical locatio 2.12 Figure 7 idicates that, with very few exceptios, projects that reached fiacial close i the two years up to May 2006 ad which were tedered i Lodo, the North West ad North East had three or more bidders at the ivitatio to egotiate (ITN) stage. More tha half the projects tedered elsewhere i Eglad durig that period were dow to two bidders at ITN. 14 7 Regio The umber of developed bids received for projects which closed betwee April 2004 ad May 2006 has bee partly iflueced by geographical locatio Number of developed bids received Midlads South/South East South West North West Natiowide Lodo North East 0 2 4 6 8 10 Two or fewer at ITN Number of projects Three or more at ITN 2.13 Some PFI advisers ad cotractor cosortia poited to impracticalities i biddig i certai parts of the coutry, due to the small umber of regioally based sub-cotractors who were acceptable to bidders ad capable of takig o the work. I additio, some cotractors felt that they would be at a competitive disadvatage i a regio i which they had o existig liks with sub-cotractors. The applicatio of ew public procuremet regulatios may make bidders more selective 2.14 As Part 1 of this report sets out, the procuremet of PFI projects is ow subject to a ew procuremet procedure kow as Competitive Dialogue. This procedure requires that a greater part of the deal is agreed durig the competitive phase of tederig with all bidders, before a wiig bid is selected. There are potetial beefits from maitaiig competitive tesio for loger. However, there is also a icreased risk that the private sector will become more selective i the face of potetially higher bid costs ad loger periods without certaity of wiig the competitio. This makes it particularly importat that bidders cocers are met by procuremet teams: esurig that projects are properly prepared before they are brought to the market, that teams are suitably resourced ad that there is access to existig experiece of complex procuremet, so as to brig dow tederig times ad costs. Source: Natioal Audit Office cesus 2006 14 However statistical aalysis has ot cofirmed a correlatio betwee geographical locatio ad the umber of developed bids received, possibly because of the small umber of cases ivolved. See Appedix 4 for further details. Improvig the PFI tederig process 15

PART THREE Tederig for PFI deals ca be improved This part of the report shows that tederig for PFI deals takes loger ad costs more tha expected. May of the delays ad excessive costs would be avoided by a more professioal, coordiated approach to public sector tederig. It takes o average just uder three years to teder ad close a PFI deal 3.1 Tederig periods for all the PFI deals with a capital value of over 20 millio that closed betwee 2004 ad 2006 lasted a average of 34 moths. 15 Betwee 2000 ad 2003, the comparable figure was 33 moths. There were variatios betwee sectors, with the average times for schools, hospitals ad other deals beig 25, 38 ad 47 moths respectively. The shortest overall tederig period was 16 moths ad the logest 73 moths (Figure 8). 3.2 It is importat for the ultimate success of a PFI deal that sufficiet time is take to esure that tederig is doe well, but despite Treasury ad Departmetal measures to improve ad make the PFI procuremet process more efficiet, the average legth of time take to teder the projects which closed betwee 2004 ad 2006 was o shorter tha that for earlier projects, closig betwee 2000 ad 2003. The lack of ay official bechmark for tederig times meas there is o guidace o this issue, ad comparisos with covetioally procured projects are misleadig. PFI deals require agreemet ot just o the costructio of buildigs or other assets but also o the services to be provided throughout a 25 to 30 year cotract, ad are thus iheretly more complex to procure. 16 3.3 I 2004, the Departmet for Educatio ad Skills set up a joit veture with PUK, called Parterships for Schools, to ru the tederig of schools projects (icludig PFI projects) ad to address some of the issues highlighted i this report. The joit veture has developed ew procuremet documetatio, provided a cetral source of procuremet expertise ad cosulted extesively with the market. It is too early, however, to determie its impact o programme delivery as most projects tedered uder the programme have yet to reach fiacial close. Almost half of the time take to teder a PFI deal is take up by egotiatios with a sigle preferred bidder 3.4 Because of the complexity of PFI deals, ad because the backig of leders has to be fully secured, bidders are usually uable to make fial bids which are ucoditioal. As a result, it has bee commo practice for there to be a period of exclusive egotiatios followig selectio of a sigle, preferred bidder before the deal is siged. Uless the procurig authority ca keep careful cotrol of this part of the process there is a high risk that the terms of the deal will become less favourable. 3.5 Prior to 2004, aroud half of PFI projects ivolved preferred bidder egotiatios of a year or more. Betwee 2004 ad 2006, two-thirds of PFI projects had preferred bidder egotiatios lastig a year or more, ad the average legth of preferred bidder egotiatios for all projects that closed durig that period was 15 moths (see Figure 9). This is twice as log as Authorities expected whe they selected a preferred bidder. O average, egotiatios with a preferred bidder took almost as log as the competitive phase that preceded it. 15 This is based o our cesus data. The 11 local authority projects closig betwee April 2004 ad May 2006 that we did ot survey had a very similar average tederig time of 35 moths. 16 OGC data for covetioally procured projects with a capital value of 20 millio or greater idicated that average tederig times were approximately half those of PFI projects. 16 ImPROvING THE PFI TENDERING PROCESS

part three 8 The legth of tederig varied widely betwee projects ad betwee sectors Tederig times (moths) 80 70 60 50 40 30 20 10 0 Schools Mea tederig time: 25 moths Hospitals Mea tederig time: 38 moths Other Mea: 47 PFI projects that closed betwee 2004 ad 2006 Source: Natioal Audit Office cesus 2006 NOTE The black lies represet the upper ad lower quartiles. 9 Preferred bidder egotiatios for projects were legthy, lastig a average of 15 moths Legth of preferred bidder egotiatios (moths) 35 30 25 20 15 10 5 0 PFI projects that closed betwee 2004 ad 2006 Source: Natioal Audit Office cesus 2006 Improvig the PFI tederig process 17

part three It is hard to fid systematic explaatios for variatios i the legths of tederig periods, although there are differeces betwee sectors 3.6 As illustrated by Figure 10, there was o direct correlatio betwee the time it takes to teder a deal ad the capital value of a deal. The type of deal, for example whether a deal is for a school, hospital or other serviced buildigs did have a sigificat ifluece o time take to teder. Overall, however, oly about a third of the variatio i tederig times was explaied by factors such as deal size, type of deal ad the experiece of procuremet teams i maagig PFI projects. 17 Tederig ad egotiatio periods could ofte have bee reduced 3.7 We asked procurig authorities, Departmetal Private Fiace Uits, advisers ad cotractors what issues they thought had bee importat i legtheig tederig times. Figure 11 summarises these explaatios, dividig them ito causes of delays that were uavoidable, causes that were capable of beig mitigated (whether by the procurig authorities or by the public sector more widely) ad causes that were avoidable. Although this 10 Legth of tederig period (moths) 80 70 The legth of tederig for PFI projects was urelated to the size of those projects report focuses o PFI tederig, some of the causes of delay (such as policy chages ad uforeseeable evets) would also have caused delays i projects procured covetioally i the same period. 3.8 Factors cited as importat icluded uforeseeable evets which were etirely out of the cotrol of public sector procuremet teams. For istace, the evets of 11 September 2001 had wide ramificatios for the isurace market. Equally, some projects were affected by the fiacial collapse of Jarvis plc, a major cotractor for PFI schools i particular (see Case Example 1). 11 Commoly cited causes of log tederig periods Out of the cotrol of the public sector Uforeseeable evets (such as the impact of the evets of 11 September 2001 o the isurace markets ad the fiacial collapse of Jarvis plc) Private sector admiistrative ad approval processes Could be partly mitigated by the public sector Negotiatios surroudig chages to the bid proposed by the preferred bidder Delays caused by public sector admiistrative ad approval processes Delays caused by plaig procedures Delays caused by havig to adapt to policy chages (such as Ageda for Chage i the health sector) Reiteratios of desig, where these are caused by a poor respose from bidders Lack of expertise, experiece or resources withi the public sector procuremet teams 60 50 40 30 20 10 0 0 500 1,000 1,500 Capital value ( millio) Source: Natioal Audit Office cesus 2006 2,000 Could be avoided by the public sector Chages made by the public sector procuremet team to the scope or specificatios of the project Reiteratios of desig where these are ecessary because of authority-led scope ad specificatio chages or affordability issues Isufficiet developmet of specificatios prior to the project goig to the market Poor process maagemet Revisitig of affordability issues Source: Views expressed by project maagers of public sector teams, Departmetal Private Fiace Uits, advisers ad cotractors 17 See Appedix 4 for the statistics to support this paragraph. 18 Improvig the PFI tederig process

part three case example 1 Kirklees City Coucil the impact of Jarvis fiacial difficulties o the PFI tederig process Kirklees Metropolita Coucil selected Jarvis plc to be the preferred bidder for its special schools project i November 2003, followig a competitive process. It reached commercial close i March 2004, ad was ready to sig the cotract i July 2004. However, Jarvis fiacial difficulties were made public a few days before fiacial close was due ad its fuders pulled out of the deal. Throughout the summer, the Coucil worked with all iterested parties to develop a way forward. As Jarvis was ot able to provide the Coucil with the ecessary reassuraces about its fiacial viability, the Coucil de-selected Jarvis as its preferred bidder towards the ed of September 2004. The bidder which had come secod i the competitio agreed to take Jarvis place, usig some sub-cotractors which had bee part of the Jarvis cosortium. However, the Coucil had to accept the additio of costructio iflatio to the cost of the project ad some re-pricig of risk. This led to affordability issues ad, as there was o further moey available ad a judgemet was made that there should be o further compromises o desig quality followig lessos from a previous PFI project, a decisio was take to remove oe school from the project to restore affordability. 3.9 Other factors might have bee oly partially withi the cotrol of public sector procuremet teams, or outside of the direct cotrol of procuremet teams but withi the cotrol of the public sector as a whole. I particular, policy chages ca affect PFI projects quite sigificatly, i terms of both cost ad time, by requirig alteratios to the services required, which will exted tederig periods if they occur durig procuremet. Oe third of NHS Trusts which closed PFI projects i the past two years cosidered that policy iitiatives, such as Ageda for Chage ad the itroductio of the Retetio of Employmet model, had prologed their procuremet timetables. May of the remaiig causes of log tederig ad egotiatio times, however, fell more directly uder the cotrol of public sector procuremet teams ad could possibly have bee avoided by the public sector with better plaig, maagemet ad exploitatio of existig experiece. Projects are ot always prepared sufficietly for the tederig stage 3.10 The Treasury has commeted i its most recet assessmet of PFI that procurig authorities are ot allocatig sufficiet time ad resources to adequately prepare ad develop their projects before formal egagemet with the market. 18 May professioal advisers we spoke to believed that more preparatio would help to prevet large-scale chages to projects or uplaed extra rouds of tederig, both of which were see as major reasos for tederig delays. Two-thirds of procurig authorities that closed projects betwee April 2004 ad May 2006 made scope or specificatio chages to their projects after goig to the market. 3.11 The amout of pre-tederig preparatio by authorities varied betwee aroud five moths ad early five years. 19 There was o correlatio betwee preparatio time ad size of deal; or was there a sigificat correlatio betwee preparatio time ad the subsequet legth of the tederig process. I the latter case, this may partly reflect the fact that legthy preparatio periods were ot always focused or backed with sufficiet resources ad levels of experiece. Experiece was ot shared as widely as would be useful 3.12 May Authorities aticipated issues i advace of preferred bidder egotiatios, but ofte foud that they proved more difficult to resolve tha expected. These issues iclude matters such as fialisig the detailed desig, obtaiig plaig approvals, lad ad property issues (such as coveats ad deeds), employmet, pesios ad eergy issues. Although these issues were commo to may PFI deals, they teded to be resolved by idividual Authorities actig i isolatio. There may have bee scope for improvig the tederig process through further developmet of stadard specificatios ad guidace frameworks. For istace, some NHS Trusts told us that egotiatio times would have bee reduced had there bee a stadard market approach across schemes where retaied estate is icluded, a stadard defiitio of latet defects ad a commo approach towards clearig backlog maiteace. 3.13 May Authorities reported that they passed o ay lessos leart followig the completio of their PFI tederig process to Departmetal Private Fiace Uits (PFU), ad sometimes to other Authorities. Such itelligece ca be extremely effective, as the example i Box C overleaf suggests, but there was o systematic way of esurig that useful lessos were shared withi or across sectors. Although may Departmets require projects to complete self-evaluatio templates kow as post-project evaluatios, these were ofte ot completed ad were ot pursued by Departmets. 18 Stregtheig log-term parterships, p103. 19 We calculated this from the date of project iitiatio, defied as the poit at which the decisio was take to begi the substative work towards Outlie Busiess Case. Improvig the PFI tederig process 19

part three BOX C The beefits of sharig lessos betwee projects Oce they had closed their deals i 2004, both St George s Hospital Morpeth ad Addebrooke s Elective Care, Diabetics ad Geetics Cetre shared some lessos they had leared with two Trusts which were i the process of tederig for PFI hospitals. Both the Trusts which had received this advice had tederig periods sigificatly shorter tha average, ad both succeeded i keepig tederig costs to budget. BOX D The difficulty of cofirmig the affordability of a project whe the tederig period has bee legthy Affordability curves ca be very dyamic. This ca lead to a vicious circle developig: the loger the procuremet period, the more likely that the affordability equatio will chage, forcig scope chages ad further procuremet delays. (NAO iterview with a PFI adviser) 3.14 The Treasury s cotract stadardisatio guidace ad additioal sector-specific guidace ad stadard cotracts are a mechaism through which lessos have bee leart ad passed o to future projects. A example of this is the ew isurace provisios which have resolved difficulties experieced by a umber of projects followig the evets of 11 September. Although there are clear beefits from the itroductio of a stadard cotract, icludig greater cosistecy, we did ot fid that it led to a improvemet i overall tederig times for the projects we examied. Oe reaso for this was that some of these projects were delayed by the trasitio betwee versios two ad three of the stadard cotract durig 2004. Delays are ievitable if a authority decides that it caot afford the deal o offer 3.15 Authorities have to make a assessmet of the likely cost of a project ad cofirm that it is affordable before takig the project to the market. This is a requiremet for the approval of Outlie Busiess Cases by all Departmets. However, affordability issues ofte subsequetly arise durig the procuremet, either as a result of chages i a Authority s requiremets or because the bids received are more expesive tha the authority expected. We foud that, eve at the poit of preferred bidder selectio, affordability had ofte ot bee completely resolved. Ay chages made to the project or chages to the fiacial situatio of Authorities ecessitated a recosideratio of affordability, resultig i delays to the fialisatio of a deal ad sometimes requirig further chages to the project. A umber of advisers put this poit to us, icludig a adviser with extesive experiece of procurig deals (see Box D). I additio, affordability problems ofte arose as a result of a loger tha aticipated procuremet process. A major example was the St Bartholomew s ad Lodo Hospitals re-developmet PFI scheme, where delay ad substatial public ad private sector costs were icurred while decisios were made (Case Example 2). case example 2 Barts ad the Lodo NHS Trust affordability ad plaig cocers led to delays durig the preferred bidder egotiatios At a capital value of 1,072 millio, this 42 year project is the largest PFI hospital deal to have bee siged i the UK. It will provide a ew teachig hospital ad a ew cacer cetre of excellece through the substatial redevelopmet of the Royal Hospital of St Bartholomew ad the Royal Lodo Hospital. Fiacial close was reached i April 2006, followig a tederig process that took just over 4 years, icludig a 28 moth preferred bidder period. The preferred bidder period was expected to last just over a year. It took twice as log for two mai reasos: The redevelopmet ivolved listed buildigs i a coservatio area ad required approval from a umber of bodies, all of whom idividually eeded to provide statutory cosets. The preferred bidder s desig was accepted without major chages at St Bartholomew s but the desig of the Royal Lodo Hospital required modificatio. Subsequet egotiatios with plaig authorities over chages to the project desig caused a delay of te moths. Although the Trust cosidered it had reached a affordable positio both prior to ad after preferred bidder selectio, the size of the scheme meat that the deal was subject to review by the Departmet of Health ad the Treasury. The full busiess case was submitted to the Departmet ad the Treasury i November 2005. This coicided with a review of NHS fiaces that icluded all prospective PFI schemes. As a result of this review, which ivolved a delay of three moths, the scope of the deal was scaled back. I total, the costs of the delays durig the preferred bidder period added aroud 5 millio to the aual uitary charge. However, this was offset by adjustmets to the scope of the deal which resulted i a et reductio to the aual uitary charge of aroud 12 per cet (equivalet to 9 millio a year). 20 Improvig the PFI tederig process