On Jan. 14, 2014, twenty sevenof the industry’s thought leadersfrom academia, state governmentand the private sector participatedin a roundtable discussion of someof the major issues affecting thefuture of the P3 project deliverymodel in the U.S. What follows isa summary of those talks plussome commentary and additionalinput solicited by PWF from acad-emics and other experts.

After introductory remarks,seven industry experts each wereasked to address one topic relatedto the P3 project delivery modelthat is ripe for research. Thesedirected discussions started with astatement from PWF framing theissue. That was followed by aproposition stating what onemight conclude from the state-ment. Then the experts said in 10minutes why they agreed or dis-agreed with the proposition beforeopening up the discussion to thetable for a futher 20 minutes.

PWF organized the event,recruited experts, and facilitatedthe discussions. Jack Basso ofParsons Brinckerhoff moderated.The American Road andTransportation BuildersAssociation hosted the three-hourmeeting at its Georgetown head-quarters during the TRB annualmeeting in Washington, D.C.

Opening statement fromPWF:

“For the P3 alternative deliverymodel to reach its full potential inthe U.S., academic researchers,government agencies and privatepractitioners should collaboratemore closely to define practical P3research, enhance data availabilityand identify funding resources toaddress these needs. Aligning thevarious interests to achieve this goalis complicated. The purpose of thisdiscussion is to explore P3 researchneeds and how to meet them, collec-tively.”

INTRODUCTION

Research Challenges

The discussion started with anoverview of research challenges byJennifer Mayer of Ernst & Young,who also chairs TRB’s JointSubcommittee on P3s.

Public sector challenges – Thebiggest challenge to public sectorresearch is the lack of time – practi-tioners have the most knowledge, butthe least time/incentive to write it up.Some of the most interesting aspectsmay be proprietary/sensitive political-ly. There’s also the worry that collect-ing data on public vs. private will beused to hammer the public sector.Finally, there’s a lack of “control”projects to show what would havehappened without P3/innovativeoptions.

Academic challenges – The

ROUNDTABLE ON P3 RESEARCH NEEDSSPONSORED BY PUBLIC WORKS FINANCING

JAN. 14, 2014WASHINGTON D.C.

IntroductionJennifer Mayer, Ernst &YoungMichaelo Garvin,Virginia Tech

pp.. 99The view from Ireland

Rethinking Value for Money Studies——MMiicchhaaeell PPaarrkkeerr,, EErrnnsstt && YYoouunngg——

pp.. 1100

Level the playing fieldOpinion: VfM miscounts private risk (p. 17)

The Availability Payment P3 Model——JJoonnaatthhaann SSttaarrttiinn,, HHNNTTBB——

pp.. 1133

The market for revenue risk P3s

Are Alternative Technical ConceptsWorth It?

——NNaannccyy SSmmiitthh,, NNoossssaammaann——pp.. 1155

List of roundtable participants

We Need A Center of P3 Research——JJoonnaatthhaann GGiillmmoorree,, JJaammeess MMaaddiissoonn UUnniivveerrssiittyy——

pp.. 1166

Fix the Research-to-MarketProblem

——RRiicchhaarrdd GGeeddddeess,, CCoorrnneellll UUnniivveerrssiittyy——pp.. 1166

Make PWF the clearinghouse for P3 researchworking papers

The Canadian P3 Approach——MMaatttt GGiirraarrdd,, PPlleennaarryy GGrroouupp——

pp.. 1188

Where Are We Going? ——TTyylleerr DDuuvvaallll,, MMccKKiinnsseeyy && CCoo..——

pp.. 1199

The role of academia

PWFinancing /February 2014 9

research that benefits government/industry doesn’t neces-sarily benefit grad students’ and assistant professors’ aca-demic career/tenure. TRB publications don’t count asheavily towards this; journals tend to focus on narrow,academic areas (e.g. highly technical issues in economics)with less real world application. Academics don’t often goto work at senior levels in DOTs, and vice versa.Academics may lack access to industry and governmentorganizations for hands-on knowledge/data.

Private sector challenges – perception of bias (whobelieves private-sector funded research talking about bene-fits of the private sector?), proprietary concerns (why collectdata when it might be used to help your competitors do theirjobs better and/or help the public sector make your role lessneeded?).

Review of Current P3 Research

Michael Garvin, a PhD engineer and associate professorat Virginia Tech, provided a brief analysis of academicresearch in P3s, excluding privatization. He found no lack ofliterature on the topic—50,000 hits on transportation P3salone in Google Scholar. P3 research is highly fragmentedby discipline and into various topic areas such as risk analy-sis, economic benefits and social value, governance, andregional considerations.

Research activity was strongest in the U.K. in the late1990s. It gained momentum later in Australia, NorthAmerica and elsewhere as data became available and modelsfor evaluating performance were improved. The 2008 reces-sion slowed market activity considerably, and subsequentresearch tended to examine its impacts. Garvin commentedthat P3 research “is a kind of trendy topic. A lot of academicshave explored this topic, had their say and moved on.”

As a visitor to the United States itwas hugely instructive to observe theP3 Research Roundtable discussionorganized by Public Works Financing onJanuary 14th.

The vast majority of P3 eventsthat I have attended over the last 10-15 years have been dominated byeither academics or P3 practitioners.It was therefore encouraging toobserve a plurality of representationat the Roundtable discussion. Thisencouraged a lively debate abouthow best to share the research andanalysis of academics with the widerP3 community.

The discussion illuminated how theincentives provided for academics(tenure, promotion etc.) are not com-patible with prompt dissemination ofvaluable research findings. Academics

are driven by the need to publish inpeer-reviewed journals that can takeyears to come to print. There istherefore a strong need to find analternative means of disseminationthat is compatible with the con-straints faced by academics and theneeds of P3 practitioners. It wasstriking to observe the Roundtableconsensus against a proposed nationalcenter for P3 research. It was broadlyagreed that that such a move mightdiminish the necessary competitivetension that already exists betweenacademics and universities. Amongthe worthy suggestions was astraightforward solution to create arepository where P3 academics canpost working papers. This need notbe confined to domestic contributors.There is much scope for ongoingexchange of ideas and experiencesfrom inside and outside the US.

In this regard it is apparent thatongoing P3 challenges faced in the USare highly correlated with thoseobserved elsewhere. Among themany unresolved issues are: the use-fulness of value for money (VfM)exercises and the efficiency of the P3procurement process. The debatearound VfM is not confined to theU.S. It continues in mature P3 mar-kets such as the UK where it is wide-ly recognized that overreliance onVfM limits the consideration of a suf-ficient range of procurement options.Speeding up the P3 procurementprocess and reducing transactioncosts also continues to challenge P3stakeholders. The trade-off betweenthe desire for innovation and highertransaction costs makes policy choic-es difficult in all jurisdictions.

Notwithstanding these issues theconsensus from the Roundtable isthat the panoply of partnershiparrangements that fall within the P3rubric will continue to grow. The P3debate is set to rage on. �

10 PWFinancing /February 2014

MEETING THE CHALLENGES OF U.S. P3S

– A VIEW FROM IRELAND

Dr. Eoin Reeves, Kemmy Business School, University of Limerick, Ireland

Director of the Privatisation and PPP Research Group

visiting scholar, Cornell Program in Infrastructure Policy

PWFinancing /February 2014 11

DIRECTED DISCUSSIONS

• Rethinking Value for Money (VfM)

Public Works Financing: Value for Money (VfM) hascome to imply a very specific type of risk-adjusted present-value analysis which often implies more precision than itreally has. It also implies that there are two binary choiceswhen there are myriad choices each with their own pros andcons.

Proposition: We need a different way to analyze pro-curement options in the U.S. that does not rely on such arigid process for making important public policy deci-sions.

Discussion leader Mike Parker, Ernst & Young

Value for money (VfM) has been used to evaluate onlya handful of U.S. P3 procurements but there was a consen-sus that VfM is an imperfect tool for evaluating deliveryoptions for a number of reasons:

Governments don’t budget on a net present value (NPV)basis. An NPV is a metric but typically cannot be a sole oreven primary basis for decision-making. Regardless ofwhich option (P3 or not P3) appears to provide better value,agencies still have to be able to afford the nominal amountof payments in each year for their projects. Public agenciesare responsible for an entire work program, not just a singleproject. In addition, they typically face debt capacity andother capital investment constraints. A P3 may offer oppor-tunities to optimize a work program and/or agency capitalstructure that would make a project more affordable or pos-sible to build sooner; these are not captured in an NPVanalysis of a single project.

Key benefits of a P3 can be greater long-term cost andperformance certainty. The responsibility for the perfor-mance and lifecycle cost of a project is borne by theinvestors who fund its construction, which also indirectlyprovides a very long-term warranty. However, VfM doesn’tdirectly value cost certainty nor the implied long-term war-ranty.

Complex VfM methodologies are based on “seemingrigor” but are highly subjective, and findings are vulnerableto bias. Undertaken initially 1-2 years before a financialclose, VfM analyses require estimates of future interestrates, construction and lifecycle costs, assumptions about

risk allocation before a concession agreement has been writ-ten, and often adjustment factors about cost savings from aP3 in states and localities that have never undertaken a P3.A straw-man VfM analysis can anticipate a design-bid-buildapproach in the public sector comparator, when a design-build may well be used. One of the most significant inputsto a VfM analysis is whether or not a TIFIA loan should beassumed and, if so, on what terms. Changes in USDOTTIFIA policy can make a project’s VfM change significant-ly.

The choice of discount rate(s) for NPV calculations canimpact a VfM outcome, and is not always rigorously under-stood. Some practitioners account for risk in both the cashflows and the discount rate, a form of double counting.Interest rates are a curve and yet practitioners and govern-ments routinely use static discount rates. In addition, pub-lic and private financing options can be of different dura-tions. In addition, the pledges offered (e.g. general obliga-tion versus subordinated, appropriation risk) in public sec-tor and P3 alternatives can also be quite different.

Challenges to VfM results – particularly for reportsundertaken prior to a procurement – will increase as com-plexity grows. The more complicated the mousetrap, the

Groundbreaking thinking

Infrastructure: one of the biggest and most complex challenges of

the 21st century. An estimated US$40 trillion of investment will

be needed by 2030 to sustain global growth. Our Global

Infrastructure practitioners, on-site in the United States,

Canada, and around the world, advise governments,

developers, and investors across the life cycle of

projects—from strategy and financing to delivery

and hand-back.

Dig deeper at kpmg.com/

infrastructure

© 2011KPMG LLP, a Delaware limited liability partnership and the U.S. member firm of the KPMG network of independent member firms affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved. Printed in the U.S.A. The KPMG name, logo and “cutting through complexity” are aregistered trademarks or trademarks of KPMG International. 63481NYO

12 PWFinancing /February 2014

more opportunities there will be for “academics to evisceratethese reports.”

Value for money is a methodology that was purposely builtto look at the merits of paying for a project’s upfront cost using

public debt or availability payment obligations.Toll concessions financed with revenue risk debtshould not be evaluated using the same methodsas availability-financed projects.

The Federal Highway Administration isdeveloping guidance on VfM methodologies,but states are developing their own approachesthat in many cases take into account both thequantitative factors above as well as qualitativeones. At this time single, national approachwould not be practical nor would it reflect thespecific circumstances of the procuring agen-cies.

Alternative approaches for evaluating projectdelivery options are being developed. (1) Somestates want only a break-even analysis based onthe availability of PABs and TIFIA. (2) Absent

hard data, adopting a “reasonableness factor” for key variablesand presenting findings in a format that’s useful to publicclients, was suggested. (3) An empirical approach was brieflydiscussed that would compare delivery approaches based onmanagement capacity, timely completion, long-term war-

IRIDIUM CONCESIONESDE INFRAESTRUCTURAS, S.A.(Formerly Dragados Concesiones de Infraestructuras, S.A)

.

WORLDTRANSPORT

LEADER

1994-20052007-2011

56 HIGHWAYS & ROADS 3,861 miles16 AIRPORTS18 PORTS & TERMINALS5 RAILROADS & SUBWAYS 987 miles

26 BUILDINGS & FACILITIES3 PUBLIC TRANSPORT HUBS

21 COUNTRIES

VfM is commonly used in Canada,the UK and other countries wherethere is no government tax subsidyfor buyers of public debt. As aresult, the analysis of procurementoptions is done on a relatively levelplaying field in terms of cost offunds. In some cases in Canada, thegovernment issues high-quality debtto subsidize P3 project financings inorder to obtain a low cost of fundsoverall while also obtaining the risktransfer benefits of the P3 contract-ing approach. On a similar note,some of the AP projects in the U.S.,including the financing of theGoethals Bridge, have adopted ahybrid approach, funding milestoneor completion payments using pub-licly sourced funds.

In the U.S. P3 market, the widefederal tax exemption for municipaldebt (the public sector comparator)has been offset to a large degree bytwo narrow federal programs, PABsand TIFIA, that subsidize the debtissued by P3 developers (which sav-ings is passed on to the procuringagencies). Nearly all U.S. P3s haveaccessed these programs.

PWF data show that between2008 and November of last year, $20billion in project financing has beenarranged in the U.S. by private devel-opers for 13 transportation P3s inVirginia, Texas, Florida, Colorado,Indiana and New York. These financ-ings include public funds, privateequity and both revenue risk andavailability payment debt.

Of the 13, nine projects werefinanced using tax-exempt privateactivity bonds ($4.3 billion) issuedfrom a limited pool of federal debtcapacity reserved for transportationP3s. Twelve of the 13 obtained low-interest federal loans ($6 billion)from the US DOT’s TIFIA program.Together with state subsidies ($4.5billion), the two federal financingsubsidy programs lowered the costof funds for these 13 projects suffi-ciently so that procurement deci-sions could be made on the meritsof the P3 delivery approach.

Many VfM analyses presume thefuture availability of TIFIA on histori-cal terms, but evolving USDOT poli-cies on both credit standards andrepayment periods can negativelyimpact the affordability of projectdelivery approach that depend onFederal credit. �

PABS AND TIFIA LEVEL THE P3 PLAYING FIELD

FOR U.S. TRANSPORTATION P3S

PWFinancing /February 2014 13

ranties, schedule acceleration and other pragmatic issues

Any federal guidance will have to address whether or notto assume the future availability of federal financing subsidiesfor P3s. These come in the form of private activity bonds(PAB) and TIFIA loans – let alone evolving terms the TIFIAprogram elects to require. Both programs will be reconsideredthis year as part of the federal MAP-21 transportation reautho-rization bill.

Absent these subsidies in the future, VfM comparisons willbe based on efficiencies and life-cycle cost advantages forwhich comparative data are not well developed. An appealwas made to FHWA to make available its database of majorU.S. transportation projects estimates and results in order toallow “honest discussions about performance” in comparingvarious models of project delivery.

The point was made that ”one of the great failures academ-ic economists” in the U.S. has been to ignore the social wel-fare component of the federal tax subsidy for state and localgovernment debt. Because the subsidy must be paid withtaxes, the result is “a wash for society.” Yet, the subsidy is adisincentive to P3s, which are a benefit to society. Economistsshould be questioning the social value of this distortion.

Federal income tax payments paid by private developers ofP3s are included in VfM comparisons as a negative (becausethey have a cost but would not inure to the state/local agenciesthat are making the decisions) but the inclusion of these taxpayments do make P3s a better deal from a Federal perspec-tive. For example, shareholder income tax payments back tothe public sector for three toll concessions in the Dallas-FortWorth region were estimated (NPV @ 5%) to add up to $3.5billion by 2062.

• The Fast Growth of AvailabilityPayment P3s

Public Works Financing: The P3 development of trans-portation projects in the U.S. is trending sharply away fromdemand risk and toward availability payment modelswhere developer financing is repaid with government fundsonce performance requirements are met.

The availability payment-for-performance modeloffers timely completion, long-term budget certainty,contractually defined performance standards, taxpayerprotection from cost overruns, built-in warranty for defec-

tive design and construction (since government pays onlywhen lanes are available for use), and lower life-cyclecosts derived from design choices and regular mainte-nance.

Discussion leader Jonathan Startin, HNTB

(At PWF’s request, the summary comments below wereprovided by Startin after the roundtable due to sound-

problems with the webcast.)

Proposition: Much more risk is held by the public sec-tor in an AP project than in a revenue risk concession.

Response: In an availability payment (AP) deal theprivate sector must absorb design errors, shoddy con-struction and high maintenance costs. These risks haveled to large overruns on conventionally financed pro-jects. The AP model is the right choice for agencies thatwant to have the private sector price and manage perfor-mance of the project over the long term. Transferring

“While rightly applicable for certainprojects, Availability Payment (AP) pro-jects don't yield the full value propositionof true, market risk PPPs. This could leadto PPPs becoming commoditized and, fur-ther, could leave more long-term fundingand other project risks with public clientsand ultimately, on taxpayers. That defeatsthe entire concept of PPP. Eventually, thiscould also impact the attractiveness of pri-vate investment in AP structured PPPs."

Karl Reichelt, Skanska

“I see availability payment P3s as a mas-sively positive development.”

Richard Fierce, Fluor

revenue risk on toll-eligible projects is a separate deci-sion.

Although revenue risk transfer encourages an enter-prising approach, taps private sector insights into cus-tomer preferences and priorities, and spurs radical newideas for scope and layout of the most attractive pro-jects, public agencies face the current reality ofreduced interest by bidders in the toll concessionmodel. Using the AP model to achieve life-cycle trans-fer risk offers huge benefits in itself and doesn’talways have to be accompanied by revenue risk trans-fer. Where revenues are retained, the public sector hasan opportunity to manage its toll receipts and availabil-ity payment liabilities sustainably.

Proposition: The wholesale move to the AP modelthreatens to undermine the P3 value proposition somuch that political support could be affected.

Response: In the bidding process for an AP deal thepublic sector has confidential competitive dialog with

three or four teams. Project constraints are explored andassumptions are tested as each proposer seeks the win-ning combination of lowest initial price coupled withaffordable operations and maintenance.

This competitive value engineering process is not gen-erally available to the same extent under the toll conces-sion model. The AP model encourages public agencies tobe disciplined in defining requirements without ambigui-ty, focusing on essential elements of the scope and usingperformance specifications that allow equally valid butoften very different interpretations of the core require-ments.

Far from undermining the P3 value proposition, theAP model is a key to widening the use of long-term P3contracts in a highly competitive environment, allowingthe benefits to be more widely recognized.

Proposition: APs leave more long-term funding andother project risks with public clients and, ultimately, ontaxpayers, defeating the entire concept of P3s.

14 PWFinancing /February 2014

Few developers in the U.S. P3market are able to take partial or fullrevenue risk themselves, lead theproject, and arrange nonrecoursedebt financing.

Cintra, Skanska, ACS, Vinci, andpossibly OHL, SNC-Lavalin, andSamsung have the corporate backingto bet long term on corridor-specifictraffic growth or on regional GNP.

But there aren’t many revenue-risk procurements underway, partlydue to concern that there won’t beenough bidders to get the competi-tive tension owners want.

The acid test is North CarolinaDOT’s I-77 HOT lanes project inCharlotte. Four strong teams wereshortlisted a year ago and financialoffers are due March 13. But there isconcern that the project needs pub-

lic subsidy to pencil out andNCDOT is betting all-in on a fullrevenue-risk toll financing with nobackup plan.

According to Fitch Ratings, rev-enue-risk managed lanes projects areonly ratable in the most heavily con-gested corridors in Texas, Virginia,Florida and California.

The difficulty Transurban is havingwith its 495 Express managed lanesproject in Virginia and Cintra withSH 130 in Texas underscore theequity risks inherent in greenfield tollroads that were project financed onassumptions made before the greatrecession. Future projects willrequire less leverage, more publicsubsidy, and maybe co-financing bygovernments, moving revenue riskprojects closer to the risk profile ofthe availability payment model.

There is ample evidence that therevenue-risk model in the U.S. worksfor government, that when develop-ers say they take traffic risk, theymean it, and that when scholars sayrevenue risk concessions take risksout of the public sector, they areright.

Also, large revenue-risk develop-ers have staying power because theytypically work with a portfolio ofassets. 407 ETR in Toronto distrib-uted a dividend of C$680 million toall shareholders in 2013. Cintra’s 41%share of that is about six times thetotal operation funding gap from itstroubled SH 130 project in Texas.

And financial exposure to under-performing projects is limited by thereliance on project specific non-recourse financing. Concession pro-jects may fail financially, but arestructured to remain strong opera-tionally and from a service point ofview. And at no financial risk to tax-payers or to the road’s users. �

THE MARKET FOR REVENUE RISK P3S