What Are The Key Risks Associated With Private Investment In Start Up Toll Road Projects In Developing East Asian Economies

Henley Management College What are the key risks associatedwith private investment in start-uptoll road projects in Developing East Asian Economies? Richard F. Di Bona ID No.: 1005661 Dissertation submitted in partial fulfilment of therequirements of Master of Business Administration 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) ACKNOWLEDGEMENTSI am indebted to many for assistance and advice given during the preparation of thisDissertation. Firstly, to my supervisor, David Parker; also to all the staff of the HenleyHong Kong office, and to Ken Bull in Henley.Within transport planning and associated professions, there are simply too many peopleto thank individually. I believe I have learnt something from almost everyone I haveworked with over the last 14 years, who afforded me the opportunity to work across afascinating mix of countries. Over the last couple of years I have picked the brains ofmany colleagues and clients, past and present; and due to frequent commercialsensitivity, many comments and discussions have been on an anonymous basis. Manyalso acted as disseminators of my questionnaire and as “sounding boards” to discussideas and informally corroborate “ball park” figures used in the Monte Carlo risksimulations.I should also like to thank Consolidated Consultants in Amman, for their assistance withprinting the Dissertation.Finally and most importantly, I must thank my wife Mariles for her moral supportthroughout the course of my MBA studies and our daughter Vanessa (for helping metake my mind off of my studies for essential relaxation).DissFinal i December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) DECLARATIONI confirm that this Dissertation is my own original work. It is submitted in partialfulfilment of the requirements of Master of Business Administration in the Faculty ofBusiness Administration of Henley Management College. The work has not beensubmitted before for any other degree or examination in any other university.DissFinal ii December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) ABSTRACTSince the 1980’s there has been a resurgence in private sector involvement ininfrastructure, especially in tolled highways, including in developing economies(Malaysia, Mexico and Thailand were early adopters). Activity expanded during the1990’s across much of Latin America and East Asia, the latter region being where theauthor has worked extensively. Following a slowdown in the aftermath of the 1997Asian Financial Crisis, activity has recently picked-up again.The 1980’s and 1990’s were characterised by generally declining price inflation andinterest rates; whereas now there is evidence of them increasing. Based on theKondratieff Wave (long-term business cycle; a.k.a. “K-Wave”), price inflation andinterest rates could be expected to trend upwards significantly over the coming 10-15years. This Dissertation seeks to determine whether this will significantly change thenature of project risk. Thus the specific hypothesis is: “There is a significant change in the nature and extent of project finance risks for private stakeholders in East Asian toll roads during a period of increasing price inflation and interest rates”The focus is on inter-urban toll roads in Cambodia, Mainland China, Indonesia, Laos,Malaysia, Myanmar, the Philippines, Thailand and Vietnam.The Literature Review begins with basic taxonomy and a review of infrastructureprivatisation trends (globally and in East Asia), illustrating likely future demand.Financial valuation methods are reviewed, suggesting that whilst FIRR and NPV can beused, the upfront capital-intensity of toll roads makes annual ratios such as Return onCapital Employed less relevant to ex ante project evaluation. Generic project risks areDissFinal iii December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)then investigated, showing that most project-risks are “front-loaded” on toll roads. TheKondratieff Wave is then introduced and its potential applicability discussed, followedby Kuznets’ work on both infrastructure development cycles and developmenteconomics. The implications of cycles on over-investment are then discussed, withspecific emphasis on the genesis and aftermath of the 1997 Asian Financial Crisis.Transport modelling theory is presented, followed by discussion of traffic risks andforecasting issues, resulting variously from uncertainty, institutional risks andmethodological weaknesses, but also demonstrating the primacy of economic growth onoutturn performance. Construction risks are also considered, followed by a briefdiscussion of other issues (primarily related to governance and business norms).Forecasts of toll road demand and construction cost have often been unreliable, withserial underestimation of cost and overestimation of demand.Environmental analyses of the East Asian countries studied are then presented, usingPESTLE and stakeholder analysis. Focussing on Thailand (for consistency with theLiterature Review’s analysis of the Asian Financial Crisis), recent economicperformance is assessed, suggesting that recovery is underway. Potential growth invehicle ownership and the demand for roadspace is then considered, benchmarking thestudied countries against more developed economies; this shows substantial up-sidepotential. The performance of a number of Chinese expressways is then examined. Theopportunities and threats facing the studied countries are discussed, grouping thecountries into three categories corresponding to risk-versus-potential characteristics.Finally, analysis of gold price and treasury bill rates are used to postulate the currentglobal economy’s position on the K-Wave, showing that it is likely in the early stages ofan upswing.DissFinal iv December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Next, practitioner perceptions, expectations and experience were tested using aquestionnaire survey (which generated over 160 responses; respondents having a meanof 20.6 years’ working experience). These showed that legal and political factors weredeemed most significant; but once detailed evaluation (i.e. modelling) commences,economic factors predominate. As expected, data were perceived as less available andreliable in developing economies. However, no strong preferences regarding the choiceof modelling method were shown; rather that the approach should be tailored to eachproject in turn. Under-forecasting demand seemed rare and over-forecasting it relativelycommon, in line with Literature Review findings. There was evidence of transportmodellers being pressured by clients to adjust forecasts. There was also evidence thatmany forecasters do not appreciate differences between equity- and debt-side evaluationrequirements. NPV and FIRR are both widely used in evaluation. Based on perceptionsof individual countries’ prospective toll road markets, the country categorisationsproposed in the environmental analysis were broadly supported (with the exception ofIndonesia being seen more bearishly by respondents). Interestingly, respondents seemedto generally expect many symptoms of the K-Wave upswing, in terms of rising interestrates and price inflation. However, they were not that convinced of the impacts of theseparameters on forecast performance.Consequently, Monte Carlo risk simulation modelling was employed to quantitativelytest likely impacts of different risk elements. The model comprised traffic/ revenueforecasts and financial analysis for a notional inter-urban start-up toll road facility.10,000 model runs were undertaken, with each run tested over three economicscenarios, namely:DissFinal v December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) “Conventional Case” based on recent previous forecast modelling assumptions (e.g. interest rates, price inflation and economic growth at levels similar to recent years); “Respondents’ Case” based on expectations gauged from the questionnaire survey (with slightly higher economic growth, interest rates and price inflation, but markedly higher fuel cost inflation); and, “Kondratieff Case” based on K-Wave upswing conditions (higher economic growth, interest rates and price inflation; though fuel price inflation at the same level as the Respondents’ Case).The Respondents’ Case tended to give the most optimistic results, but results were morevariable than in the Conventional Case. Meanwhile, results from the Kondratieff Caseappeared quite volatile, tending to support theory. Furthermore, interest rates wereshown to become substantially more important to overall risk as they rise; and priceinflation may also increase in importance. Under Kondratieff Case conditions, ifeconomic growth outstrips the impacts of rising price inflation and interest rates, thenprojected returns can be quite significant.What the above implies is that the nature and extent of project finance risks for privatestakeholders are indeed likely to change as price inflation and interest rates increase.However, if investors can lock-in fixed-rate debt (e.g. issuing bonds) before interestrates increase significantly, these risks can be mitigated. Price inflation subsequent tothe issuing of bonds would also serve to decrease the real value of debt outstanding. Butdownstream refinancing is likely to prove increasingly costly (versus experience duringthe 1980s and 1990s when cheaper refinancing was often available as a consequence ofDissFinal vi December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)declining interest rates). In summary, therefore the hypothesis is broadly supported byevidence. Approximate word count of main text is 16,900 words. KEYWORDS Infrastructure project finance Demand forecasting Developing countries Risk analysis Long wave business cycle (Kondratieff wave) Economic growth Price inflation Interest rates Transport planning Start-up toll road facilitiesDissFinal vii December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) TABLE OF CONTENTS1. Introduction ............................................................................................................. 11.1 Terms of Reference/ Personal Development............................................................ 11.2 Applicability and Hypothesis ................................................................................... 21.3 Geographic Scope .................................................................................................... 31.4 Research Approach and Dissertation Structure ...................................................... 52. Literature Review.................................................................................................... 62.1 Historical Perspective and Basic Taxonomy ........................................................... 62.2 Economic Benefits of Transport Infrastructure Development ................................. 72.3 East Asian Transport Infrastructure Privatisation Trends ...................................... 82.4 Financial Valuation ................................................................................................. 92.5 Project Risk Analysis ............................................................................................. 142.6 The Kondratieff Wave ............................................................................................ 162.7 Kuznets Cycle, Kuznets Curve and S-Curves ........................................................ 182.8 Infrastructure Development, Cycles and Crises .................................................... 192.9 Transport Modelling .............................................................................................. 232.10 Traffic Risks and Forecasting Issues ..................................................................... 252.11 Construction, Operations and Maintenance.......................................................... 332.12 Other Considerations ............................................................................................ 352.13 Summary of Key Issues .......................................................................................... 373. Environmental Analysis ....................................................................................... 393.1 Introduction and PESTLE Analysis ....................................................................... 393.2 Political, Legal and Stakeholder Issues................................................................. 403.3 Economic Recovery ............................................................................................... 423.4 Vehicle Ownership ................................................................................................. 463.5 Traffic Performance of Existing Toll Roads .......................................................... 483.6 Opportunities and Threats ..................................................................................... 513.7 Postulated Position on K-Wave ............................................................................. 534. Questionnaire Survey ........................................................................................... 554.1 Purpose .................................................................................................................. 554.2 Design Concept and Sample Selection .................................................................. 564.3 Questionnaire Design and Survey Execution ........................................................ 574.4 The Survey Sample ................................................................................................. 584.5 Tollway Appraisal.................................................................................................. 624.6 Transport Modelling Issues ................................................................................... 654.7 Forecast Performance and Evaluation Criteria .................................................... 674.8 Countries’ Outlooks ............................................................................................... 704.9 Economic Outlook ................................................................................................. 734.10 Other Comments .................................................................................................... 754.11 Key Conclusions from the Questionnaire Survey .................................................. 755. Risk Simulation Modelling ................................................................................... 775.1 Introduction ........................................................................................................... 775.2 The Case Study and Its Parameterisation ............................................................. 785.3 Methodology .......................................................................................................... 825.4 Comparison of Cases under “Base Run” .............................................................. 845.5 Comparison of Simulation Results between Cases ................................................ 85DissFinal viii December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)5.6 Analysis of Individual Risks ................................................................................... 885.7 Discussion of Results ............................................................................................. 916. Discussion and Conclusions.................................................................................. 926.1 Introduction ........................................................................................................... 926.2 Evaluation Criteria and Implications of the Time-Nature of Risk ........................ 936.3 Macro-Level Risks and Opportunities ................................................................... 946.4 Market Risks .......................................................................................................... 966.5 Forecasting Risks................................................................................................... 986.6 Is the Market Anticipating a Change in the Rules-of-the-Game? ....................... 1006.7 What Lessons for Practitioners? ......................................................................... 1016.8 Conclusions: Evaluation of Hypothesis ............................................................... 103References: Literature ................................................................................................ 105References: Internet Resources ................................................................................. 117Appendices ................................................................................................................... 119 LIST OF TABLESTable 2.1: Investment and Maintenance Needs in East Asia, 2006-2010 ......................... 8Table 2.2: Bain and Polakovic Forecast Performance Statistics ..................................... 26Table 2.3: Bain and Wilkins Ramp-Up Revenue-Adjustment Profiles .......................... 30Table 2.4: Estimated Expressway Construction Costs .................................................... 34Table 2.5: Operations and Maintenance Costs ................................................................ 34Table 2.6: Summary of Key Risks and Issues ................................................................ 38Table 3.1: Highlights of PESTLE Analysis .................................................................... 39Table 3.2: Vehicle, Trip and Expressway Patronage Income Elasticities....................... 48Table 4.1: Aggregated Respondent Experience Categories ............................................ 58Table 4.2: Respondents’ Mean Years’ Experience in Various Fields ............................ 60Table 4.3: Respondents with Experience in Study Area ................................................. 61Table 4.4: Rankings of Macro-Level Risks by Respondent Category ............................ 63Table 4.5: Rankings of Project-Level Risks by Respondent Category ........................... 64Table 5.1: Basic Link Characteristics of Case Study Network ....................................... 79Table 5.2: Assumed Trip Distribution (% by O-D Pair) ................................................. 79Table 5.3: Comparison of “Base” Runs between Cases ................................................. 85Table 5.4: Summary Results from Simulation Runs ....................................................... 86Table 5.5: Rankings of Risk Categories’ Importance by Case ....................................... 89 LIST OF FIGURESFigure 1.A: Map of East Asia ........................................................................................... 4Figure 1.B: Research Approach ........................................................................................ 5Figure 2.A: Standard & Poor’s Risk Pyramid ................................................................. 14Figure 2.B: Transport Concession Risks ......................................................................... 15DissFinal ix December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Figure 2.C: Kuznets Curve and S-Curve......................................................................... 18Figure 2.D: Indexed Thai Real GDP and M2, 1991-1999 .............................................. 19Figure 2.E: Baht-US$ Exchange Rate 1994-2001 .......................................................... 20Figure 2.F: Dollarised Thai GFCF 1994-2001 ................................................................ 21Figure 2.G: Demand, Revenue and Price Elasticity of Demand ..................................... 27Figure 3.A: Typical Concession Stakeholder Map ......................................................... 40Figure 3.B: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) .................... 43Figure 3.C: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) in US$ ....... 43Figure 3.D: Thai GFCF, GDP and M2 in Baht, Indexed to 1995 ................................... 44Figure 3.E: Thai GFCF, GDP and M2 in US$, Indexed to 1995 .................................... 44Figure 3.F: Thai GFCF, GDP and M2 in US$, Indexed to 2000 .................................... 44Figure 3.G: Currency Performance since 1994 ............................................................... 45Figure 3.H: Currency Performance since 2001 ............................................................... 45Figure 3.I: Relationship between Wealth and Roads Per Capita .................................... 47Figure 3.J: Relationship between Wealth and Road Density .......................................... 47Figure 3.K: Traffic Growth on Shanghai-Nanjing Expressway...................................... 50Figure 3.L: Traffic Growth on Shanghai-Hangzhou-Ningbo Expressway ..................... 50Figure 3.M: Interest Rates, Nominal Gold Price and Kondratieff Wave ........................ 54Figure 4.A: Respondents by Experience Type ................................................................ 59Figure 4.B: Respondents by Years of Experience .......................................................... 59Figure 4.C: Respondents’ Global Experience ................................................................. 60Figure 4.D: Respondents with Experience in East Asia ................................................. 61Figure 4.E: Attitudes to Macro-Level Risks ................................................................... 63Figure 4.F: Attitudes to Project-Level Risks .................................................................. 64Figure 4.G: Data Availability and Reliability ................................................................. 65Figure 4.H: Attitudes to Transport Model Types ............................................................ 66Figure 4.I: Perceptions of Forecast Performance ............................................................ 68Figure 4.J: Which Forecast Outputs are Considered? ..................................................... 69Figure 4.K: How Often Are Which Criteria Considered?............................................... 69Figure 4.L: Perceived Tollway Market Opportunities by Country ................................. 71Figure 4.M: Impact of Experience on Country Perceptions ........................................... 71Figure 4.N: Country Perceptions by Respondent Category ............................................ 72Figure 4.O: Economic Expectations ............................................................................... 73Figure 4.P: Economic Expectations by Respondent Group ............................................ 74Figure 5.A: Case Study Notional Network ..................................................................... 79Figure 5.B: Volume/Capacity-to-Speed Relationships ................................................... 83Figure 5.C: Cumulative Probability Distribution of FIRR (excluding FIRR<0%) ......... 87Figure 5.D: Cumulative Probability Distribution of Payback Period (years) ................. 87Figure 5.E: Cumulative Probability Distribution of NPV at 16% ($m) .......................... 87DissFinal x December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) GLOSSARY OF TERMS AND ABBREVIATIONS ADB Asian Development Bank, Manila ASEAN Association of South East Asian Nations BOO Build-Own-Operate (concession form) BOOT, BOT Build-Own &/or Operate-Transfer (concession form) Billion One thousand million, being the international financial standard (as opposed to the strict/ traditional British definition of a million million) China For the purposes of this Dissertation, China is analogous to Mainland China, being the People’s Republic of China, excluding the Special Administrative Regions of Hong Kong and Macau and also excluding Taiwan. CIA Central Intelligence Agency, United States of America DBFO Design-Build-Finance-Operate (concession form) EIRR Economic Internal Rate of Return comprising FIRR plus social impacts Factory Gate Referring to prices of goods once manufactured but not transported, either to port or end user. FCO Foreign and Commonwealth Office, United Kingdom FDI Foreign Direct Investment FIRR Financial Internal Rate of Return FOB Free On Board: being the price of cargo loaded onto a maritime vessel GMS Greater Mekong Subregion, comprising Cambodia, Laos, Myanmar, Thailand, Vietnam plus Guangxi and Yunnan Provinces of China Guanxi meaning connections, a term covering business networks, political connections and a broad sense of developing and maintaining goodwill; see Appendix 6 for full definition HHI Hopewell Highway Infrastructure Limited IBRD International Bank for Reconstruction and Development, analogous with WB IPFA The International Project Finance Association IRR Internal Rate of Return, taken to be analogous to FIRR JBIC Japan Bank for International Cooperation and Development, Tokyo JICA Japan International Cooperation Agency K-Wave Kondratieff Wave or Cycle KOICA Korea International Cooperation Agency Kondratieff Spelling adopted for Kondratieff; alternative Latin spellings include Kondratyev, Kondratiev (original Russian: Кондратьев) NESDB National Economic and Social Development Board, Thailand NPV Net Present Value PBA Parsons Brinckerhoff (Asia) Ltd. PPP Public Private Partnership (when discussing project financing models) PPP Purchasing Power Parity (when discussing national income accounting concepts, such as GDP and GDP per capita), this in contrast to figures derived based on official exchange rates ROT Rehabilitate-Own/Operate-Transfer (concession form) SWHK Scott Wilson (Hong Kong) Ltd/ Scott Wilson Kirkpatrick (Hong Kong) Ltd (including joint-consultant reports with Scott Wilson as one of the authors) UNESCAP United Nations Economic and Social Commission for Asia and the Pacific, Bangkok, Thailand US$ United States Dollars VOT Value of Time: equivalencing time and money in behavioural models. WACC Weighted Average Cost of Capital WB The World Bank, Washington, D.C.DissFinal xi December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)1. Introduction1.1 Terms of Reference/ Personal DevelopmentFor 14 years, I have worked in transport planning, economics and demand forecastingacross 20 countries/territories, mostly on transport infrastructure scheme appraisal, oftenfor privatisation, and usually in East Asia (covering rich, “tiger” and poor economies).One reason for pursuing the MBA, the Business Finance Elective and this Dissertationtopic was to gain a more comprehensive understanding of projects’ financial risks.Hopefully to make me a “better” demand forecaster and broader project appraiser.During the course of my MBA I rekindled interest in aspects of economics, mostnotably business cycles, leading me to the Kondratieff Wave. This postulates a cycle of48-60 years duration; comprising inter alia phases of increasing interest rates andcommodity prices followed by decreases in same. Given recent increases in FederalReserve interest rates and commodity prices, Kondratieff theorists posit acommencement of an “upswing” phase, qualitatively different from the “downswing” ofthe 1980’s and 1990’s; potentially changing the relative importance of different aspectsof investment risk. Given most transport privatisation and associated literature andexperience are based on “downswing” conditions, reviewing these based on “upswing”conditions could be timely.Though focussed on profit maximisation (through risk management), betterunderstanding of changing risks should result in more efficient use of capital by private,public and aid agency sectors alike.DissFinal Page 1 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)1.2 Applicability and HypothesisThe Dissertation focuses on East Asia which is again emerging as a “powerhouse” ofeconomic growth, with commensurately strong demand for transport anticipated. TheWorld Bank (2003a) notes resurgent private sector involvement in infrastructureprovision since the 1980’s, with substantial tollway activity in East Asia (US$34 billionduring 1990-2001 into 149 projects). Although activity slowed following the AsianFinancial Crisis (AFC), by 2001 it returned to 1995 levels. Yepes (2004) expectshighways to be the second biggest infrastructure investment sector in East Asia during2006-2010. In addition to providing profit opportunities, there is evidence that projectscould facilitate substantial economic growth in poorer economies, as well as “tiger”economies (Corbett et al, 2006).However, besides a potential legacy of over-investment prior to the AFC (Di Bona,2002) suppressing the attractiveness of certain new projects, following 20 years ofdeclining interest rates and price inflation, it appears that they are now rising (Faber,2002). Arguably this is connected with an upturn in the long-wave business cycle(Kondratieff, 1926). Thus, the specific hypothesis is: “There is a significant change in the nature and extent of project finance risks for private stakeholders in East Asian toll roads during a period of increasing price inflation and interest rates”DissFinal Page 2 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)1.3 Geographic ScopeEast Asia is a large, diverse region, including some of the World’s richest and poorestsocieties, with differing political and legal systems and levels of economic openness.This Dissertation is concerned with its developing economies, which are likely tobenefit as: manufacturing hubs for the world; markets in their own right; and/or, naturalresource providers. It is in such economies that transport infrastructure demand growthmay be most marked.Whilst the literature review is deliberately broad, and the questionnaire survey relativelyso, the main focus is on inter-urban toll roads. Countries are included based on being: Sufficiently large (geographically) to accommodate inter-urban tolled highways; Developing economies; and, Countries where the author has at least some project experience.The countries thus considered are: Cambodia, China1, Indonesia, Laos, Malaysia,Myanmar, Philippines, Thailand and Vietnam; highlighted in Figure 1.A.Appendix 1 gives key demographic and economic data on these countries and a fewothers for benchmarking purposes. Appendix 2 gives headline transport statistics.Whilst countries such as China are anticipated to continue requiring and attractinginvestment in roads, increased scope for PPP is expected in other countries also.1 Being Mainland China, i.e. excluding Hong Kong SAR, Macau SAR and TaiwanDissFinal Page 3 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Mongolia N.Korea S.Korea Japan CHINA Hong Kong LAOSMYANMAR PHILIPPINES VIETNAMTHAILAND Brunei CAMBODIA MALAYSIA Singapore INDONESIA Timor-LesteSource of base map: Google EarthTM 2Figure 1.A: Map of East Asia2 Study Area countries in red on yellow text. Other countries/ territories in black on grey text.DissFinal Page 4 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)1.4 Research Approach and Dissertation StructureThe outline research approach is presented in Figure 1.B; also giving relevant Chapternumbers. 1. Introduction and Hypothesis Including definition of geographic scope 2. Literature Review 3. Environmental Analysis Including a priori evaluation Including country economics and and analysis thereof tollway market potential 4. Questionnaire Survey Analysis of respondent perceptions against findings of Literature Review and Environmental Analysis 5. Risk Simulation Modelling Quantitative testing of impacts of different economic assumptions and evaluation of relative importance of different risks, incorporating findings of Chapters 2, 3 & 4 6. Discussion and Conclusions Collating, comparing and summarising findings from Chapters 2, 3, 4 & 5. Evaluation of initial hypothesis and identifying areas for possible future investigation.Figure 1.B: Research ApproachDissFinal Page 5 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2. Literature Review2.1 Historical Perspective and Basic TaxonomyPrivate transport infrastructure financing and operation dates back to at least the 19thCentury, including railways (e.g. UK and USA) and the Suez Canal. IPFA (2006) notesfollowing the First World War government resumed most infrastructure provision,financing projects from public debt; subsequently developing countries followed thispractice, borrowing from development agencies (e.g. WB, ADB).By the 1980’s, government debt constrained public financing of schemes, especiallygiven high interest rates; yet economic and demographic forces continued to demandinfrastructure. Thus was private involvement reborn.There is much overlapping taxonomy regarding types of project privatisation. Guislainand Kerf (1995) note a continuum of options for private sector involvement, fromsupply and service contracts through leasing (wherein management of a built project islet to the private sector in exchange for a revenue-share and/or up-front payment) toBuild-Own/Operate-Transfer (BOT, BOOT) and Build-Own-Operate (BOO); wherein,the project is constructed then operated by the private concessionaire either in perpetuity(BOO) or for a fixed period (BOT). Other forms include Design-Build-Finance-Operate(DBFO) wherein the prospective concessionaire undertakes the design as well as buildof the project, often being wholly responsible for financing.DissFinal Page 6 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.2 Economic Benefits of Transport Infrastructure DevelopmentWhilst SACTRA (1994) questioned the benefits of additional trunk roads in developedeconomies with built-out highway networks, in developing economies new highwaysoften facilitate economic development. Christensen and Mertner (2004) showedCambodia’s factory gate price advantage over China for garments negated by transportcosts: China FOB prices are lower than Cambodia’s. Di Bona (2005) notedrehabilitation of Cambodia’s road networks transformed traffic levels and patterns;subsequent quantification estimated nationwide road traffic levels increased 83.6%above trend following the rehabilitation-to-date of roughly half of the trunk roadnetwork3 (Corbett et al, 2006, p.A2-99). The benefits of transport infrastructure indeveloping countries can be attested by increasing development aid for same (Luu,2006).In economic terms, rehabilitation greatly reduces generalised costs of travel (e.g. time,fuel, vehicular wear-and-tear and hence fares/ tariffs). Buchanan (1999) recommendsgovernments only approve projects yielding a given socio-economic return, beforedetermining likely profitability.Klein et al (1996) note privatisation appears to increase implementation costs, partiallydue to private sector participation bringing true costs to light. It also increases fundsavailable for development.3 83.6% estimated statistically, with traffic growth attributable directly to economic growth excluded.DissFinal Page 7 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.3 East Asian Transport Infrastructure Privatisation TrendsDeveloping countries’ transport infrastructure privatisation began in earnest in the1980’s, primarily with Malaysian, Mexican and Thai toll roads (WB, 2003a, p.126).During 1990-2001, East Asia was the second largest market, attracting US$56 billionprivate investment (41% of global total) into 229 projects (Ibid., p.135), particularly tollroads: US$34 billion into 149 projects (Ibid., pp25-26 & p.143). By 2001, China hadattracted more private investment than any other country (US$23.6 billion), andMalaysia the most per capita (US$582) (Ibid., p.136). Whilst activity slowed after the1997 Asian Financial Crisis (AFC), by 2001 it returned to 1995 levels (Ibid., p.2). Table2.1 illustrates substantial anticipated future expenditure (from Yepes, 2004); highwaysare anticipated to require the second most investment of any infrastructure category.Table 2.1: Investment and Maintenance Needs in East Asia, 2006-2010 (US$ million) (percent of GDP) Investment Maintenance Total Investment Maintenance TotalElectricity 63,446 25,744 89,190 2.4 1.0 3.4Telecoms 13,800 10,371 24,171 0.5 0.4 0.9Highways 23,175 10,926 34,102 0.9 0.4 1.3Railways 1,170 1,598 2,768 0.0 0.1 0.1 Water 2,571 5,228 7,799 0.1 0.2 0.3Sanitation 2,887 4,131 7,017 0.1 0.2 0.3 Total 107,049 57,998 165,047 4.0 2.3 6.3Buchanan (1999) notes the Malaysian boom in BOT highways followed the perceivedsuccess of the North-South Highway (PLUS) concession in 1988, through which privatefinance overcame public sector constraints and took-on risk, bringing private sectorskills and incentives to infrastructure operation. However, he believes PLUS appearedprofitable only because Government handed over 225km of existing expressway withtolling rights.DissFinal Page 8 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)In China several Provincial Governments established corporations for expresswaydevelopment. Soon after completing a flagship expressway, the company would belisted with revenues raised used to acquire or develop additional highways4. Thisrelatively rapid listing contrasts with experience elsewhere (see Willumsen and Russell,1998). Meanwhile, most foreign-invested BOT or leasing projects were Joint Ventures(JV) with government retaining equity in the operating company.Elsewhere in Asia, BOT concessions were the norm, though often undertaken by listedfirms. Operators occasionally issue bonds, although this practice is more widespread inthe Americas.2.4 Financial Valuation2.4.1 NPV and IRRThe decision to pursue a project and on what terms are primarily questions of projectvaluation and risk. Higson (1995, pp.60-61) notes project value may be defined via NetPresent Value (NPV) or Internal Rate of Return (IRR). NPV values future cashflows as: n Ct NPV   1  r t (1) t 0Where: Ct is net cashflow in period t r is the discount rate (equivalent to opportunity cost of capital) n is the number of periods covering the concession periodIRR expresses scheme value in terms of a percentage return on capital invested, beingthe discount rate at which NPV is exactly nought:4 See Appendix 3 for examples.DissFinal Page 9 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) n Ct NPV   0 t 0 1  R t (2)Ct can include social benefits of the scheme (see Section 2.2), as well as social costs(e.g. displacement, environmental degradation etc; not covered in this Dissertation)when used for social analysis.The Fisher-Hirshleifer theorem (ibid, pp.66-67) states firms should undertake projects ifreturn is greater than investors’ required return. Highways require substantial up-frontinvestment and traffic flows often take a few years to build-up to “break even” levels;attractiveness is greatly affected by timing of revenue receipts and the discount rate, aswell as by initial investment size.Investors treat own target FIRR as strictly confidential; so no directly citeable values areavailable. However, from the Author’s experience corroborated by off-the-recordconversations with fellow practitioners, a target FIRR of 16% p.a. is the usual thresholdrequired. This includes a modest risk premium (see 2.4.2); for particularly high riskprojects, or when capital is more expensive, FIRR would increase accordingly.2.4.2 CAPM and WACCThe above assumes certainty regarding all project aspects, including: demand, priceinflation for inputs, selling price, construction cost and time, operating period, implicitassumption of no sovereignty risks etc; yet uncertainty bedevils these parameters. TheCapital Asset Pricing Model (CAPM; ibid., p.123) suggests the return on a risky projectrj is: rj  ri   j (rm  ri ) (3)where: ri is the return on riskless borrowing/ lendingDissFinal Page 10 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) rm is the return on the money market as a wholeThe risk premium for j is a proportion βj of overall market risk-premium, as follows:  jm j  (4) m 2Required return can also be calculated as Weighted Average Cost of Capital (WACC;ibid, p.279): E MV  K e   DMV  K d  WACC  E MV  DMV  (5)Where: EMV is total market value of equity employed DMV is total market value of debt employed Ke is cost of equity, given by (6) Kd is cost of debt, given by (7)  Dividend  Ke     ExpectedDi videndGrowth (6)  Share Pr ice     Debenture Pr ice (%ofFaceValu e)   1  TaxRate  InterestRa te Kd    (7)  From (3) and (7) the Fisher-Hirshleifer theorem can be restated as pursue projects if: ri   j (rm  ri )  EMV  K e   DMV  K d  EMV  DMV  (8)2.4.3 Treatment of Price InflationOften (especially in transport scheme appraisal) a constant inflation rate is assumed withcalculations based in real prices (akin to zero price inflation throughout). Such priceneutrality simplifies calculations; however, it does preclude analysis of price-risksassociated with individual project inputs and outputs.DissFinal Page 11 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.4.4 Problems with CAPM and WACCβj might theoretically be known for existing highways, but is unknown for new projects.There may be insufficient local data to determine  m . β is intended for fully diversified 2investors, rather than appraising a scheme in isolation. Higson (ibid., p.136) notesCAPM assumes: (i) perfect markets, without taxes and transaction costs, full, freely available information and no-one with price-making power; (ii) investors are rational, risk-averse, wealth-maximising, with homogenous expectations of the future; (iii) assets are marketable and infinitely divisible, with normally distributed returns; and, (iv) there is a risk-free asset for comparison.Yet transaction costs can be substantial (professional fees, cross-border know-how, etc);information is imperfect and expectations are heterogeneous. Given skill-sets required,infrastructure investors are unlikely to be highly diversified. Highway projects’ sizemakes them relatively illiquid. There may be no risk-free asset: money is only risk-freeif possible depreciation/ price inflation is ignored.Lumby (1983) notes unless a project is financed with the same capital structure as thefirm itself (unlikely), WACC changes once the project is undertaken. Furthermore,WACC assumes constant cashflows and that project systematic risk to equal that of thecompany’s existing projects; both highly unlikely.DissFinal Page 12 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Ormerod (2005, p.173) notes whilst CAPM requires a normal probability distribution inderivative markets, they exhibit power-law behaviour; this discrepancy caused the 1998collapse of Long Term Capital Management. Whilst CAPM supports currencydiversification (e.g. in borrowing), Beaverstock and Doel (2001) note such borrowingcollapsed Steady Safe (an Indonesian taxi and bus firm) and in turn PeregrineInvestment Bank.2.4.5 Financial RatiosA number of financial ratios may be used to evaluate likely project performance andrisk. Given the capital-intensity of highway construction, coupled with typically longlead-times for demand build-up (see 2.10.4), financial ratios may not always be asrelevant to ex ante project valuation.Return on Capital Employed5 is likely to be poor for early years of a concession (unlessthe project is highly geared). Likewise, Gross Profit Margin, Profit On Sales, Expensesas Percent of Turnover, Sales to Capital Employed, Sales to Fixed Assets and AssetTurnover all typically take many years to build-up to levels normally deemed acceptablein many other businesses5.Some of the above ratios might be improved by heavy borrowing, but such borrowingand resultant debt-servicing increases the importance of Working Capital Requirements,the Current Ratio and the Debt Service Coverage Ratio5. Standard & Poor’s relies onInterest Cover (debt-service coverage) as the primary quantitative measure of a project’sfinancial strength (Rigby and Penrose, 2001, p.28).5 See Appendix 4 for definitions of these financial ratios.DissFinal Page 13 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.5 Project Risk AnalysisRigby and Penrose (2001) identify a pyramidal five-level framework for credit rating,which can be taken as a proxy for overall project investor risk, shown in Figure 2.A. Credit Credit Enhancement Enhancement Force Majeure Risk Institutional Risk Sovereign Risk Project-Level RisksFigure 2.A: Standard & Poor’s Risk PyramidProject-level risks comprise six broad elements, namely: Contractual foundations Technology, construction and operations: both pre-construction (e.g. construction delay/ quality issues) and post-construction (e.g. Operations and Maintenance) Competitive position of project within its market: including industry fundamentals, project’s competitive advantage/ likely market share, threats of new entrants, etc Legal structure, including choice of legal jurisdictionDissFinal Page 14 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Counterparty risks: e.g. extent to which JV partners can contribute equity if/when debt funding exhausted, reliability of suppliers, political risk guarantees, etc Cashflow and financial risks: in addition to expected cashflow, ability to cope with interest rate, inflation, foreign exchange, liquidity and funding risksGeorge et al (2004) note the uncertainty inherent in start-up tollways requires flexiblefinancing approaches. Willumsen and Russell (1998) illustrate project-level risks asshown in Figure 2.B. Predominating traffic/ revenue risks are discussed in Section 2.10. Construction Delay Change Orders Risk (nominal) Construction Costs Ramp Up Traffic & Revenue O&M er -2 -1 10 0 1 2 3 4 5 ov d an H YearFigure 2.B: Transport Concession RisksSovereign and institutional risks are concerned primarily with the project’s country:ratings usually constrained by government’s debt servicing/ foreign currency record,reflecting risks of currency conversion and overseas transfer. Institutional factorsDissFinal Page 15 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)include business and legal institutions, which are often weak/ nascent in developingcountries, with concepts of property rights and commercial law not fully developed,potentially leaving creditors/ investors exposed. La Porta et al (1997) found investorrights in developing countries though limited, are generally better under common lawthan civil law (especially French civil law, which often has weak enforcement).Force Majeure risks includes “Acts of God” (floods, earthquakes, etc) as well as civildisturbances, strikes, changes of law. Rigby and Penrose (2001) note toll roads aretypically less affected/ can return to normal service more quickly.Credit Enhancement refers to insuring/ re-insuring specific risks. However, litigationintrinsic in such claims can delay payment by years, so mitigation may be limited.2.6 The Kondratieff WaveOrthodox economics assumes given policies produce similar results at all times;Ormerod (1999, pp.96-102) notes experience contradicts this, due to periodic exogenousshocks. Others postulate cycles responding to exogenous shocks. But to some cycleadherents, such “exogenous” shocks are mostly endogenous. Schumpeter (1939)consolidated others’ preceding work, specifying three inter-related cycles: Kitchin (1923): based on fluctuations in business inventories (39+/– months) Juglar (1863): based on business investment in plant and equipment (7-11 years) Kondratieff (1926): based on development of new technologies/ sectors and impact of their adoption on socio-economic conditions (48-60 years; a.k.a. “K-Wave”)The K-Wave postulates periodic “Creative Destruction” (Schumpeter, 1950, Chap.VII)intrinsic to industrial-capitalism. Not all cycle proponents accept the K-Wave:DissFinal Page 16 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Kindleberger (1996, p.13) calls it “possibly… dubious and elusive.” There is alsodebate on periodicity. Whilst Schumpeter believed one K-Wave contained three JuglarCycles, each comprising in turn three Kitchin Cycles, Faber (2002, p.110) notesKondratieff never postulated precise periodicity.Kondratieff’s empirical work identified a number of patterns within each cycle. Furtheranalysis by Schumpeter (1939), summarised by Faber (2002, pp.116-138) notes: Before and during the beginning of Upswings there are profound changes in industrial techniques (based on new technologies) and/or involvement of new countries in the global economy and/or development of new transport technologies. Social upheavals and international conflict are more likely during Upswings. Agricultural prices decrease during downswings; industrial prices hold steady or fall slightly. During upswings, commodity price increases can create broader price inflation. Interest rates also follow this cycle. As appears to have been the case in recent years (see Section 3.7). Upswings are characterised by brevity of depressions and intensity of booms; the opposite being true during downswings.There are separate transitional phases at peaks and troughs, usually brief in relation toUpswing and Downswing phases and largely ignored in the context of this Dissertation.Appendix 5 presents K-Waves since 1787. Maddison (1995) estimated real global GDPper capita rose 2.90% p.a. from the 1950s-1970s (K-Wave upswing); but declined to1.11% p.a. until the 1990’s (K-Wave downswing).DissFinal Page 17 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.7 Kuznets Cycle, Kuznets Curve and S-CurvesKuznets (1930) identified a 15-25 year building construction cycle, concurring withSchumpeter that innovation drives growth endogenously to the economic cycle. He alsopostulated the Kuznets Curve (1955), plotting economic development against incomeinequality: inequality increasing in the early stages of economic development,plateauing then diminishing. Inequality can be measured using the Gini Coefficient(Gini, 1912): 0 denoting perfect equality and 100 perfect inequality (one person has allwealth).This implies few might afford cars or tolls in the early phases of growth, but aseconomies develop, tolls become substantially more affordable. Coupled with demandsaturation, this suggests an “S-Curve”, akin to the innovation/ adoption curve (Rogers,1962). Figure 2.C shows this inter-relationship between a Kuznets Curve and S-Curve,based on normal distribution. Norm al Density/ Kuznets Curve Cum ulative Norm al/ S-CurveFigure 2.C: Kuznets Curve and S-CurveDissFinal Page 18 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.8 Infrastructure Development, Cycles and CrisesInfrastructure may facilitate Upswings, but its short-term impact may triggerDownswings, fostering “Creative Destruction” (Schumpeter, 1950): purging oldmethods/ technologies for improved methods/ infrastructure to drive Upswings.Lawrence (1999) argues major skyscraper completions are cyclical, precedingrecessions. But do build-out peaks precipitate recessions, or are they “peaks” due tosubsequent demand failure, uncorrelated with preceding build-out (as espoused byKrugman, 2000)?Di Bona (2002) analyses Thailand6, where the Baht’s flotation triggered the AFC.Figure 2.D7 shows impressive real GDP growth until 1996, when close correlation withM2 broke. Continued M2 growth refutes Krugman’s attribution of the AFC to demandfailure, which ignored structural causes. 200 300 Real GDP (1991=100) 180 260 M2 (1991=100) 160 220 140 180 120 140 100 100 1991 1992 1993 1994 1995 1996 1997 1998 1999 Real GDP M2Figure 2.D: Indexed Thai Real GDP and M2, 1991-19996 Much of these Thai analyses originally presented in Di Bona, R.F. (2002) Surviving Bahtulism7 Raw data from APEC (www.apec.org); analysis my own.DissFinal Page 19 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Before the AFC, Thailand enjoyed a virtuous economic development cycle: increasedwealth boosted investment returns, attracting further investment. Keynesian multiplier-accelerator effects boosted growth, encouraging further development. Adaptiveexpectations of investment returns fuelled excessive capital works and otherinvestments. Bangkok planned several new residential and business hubs, which couldnot all be viable simultaneously: eventually supply outpaced demand.The Baht’s July 1997 flotation coincided with doubts regarding the sustainability ofThailand’s growth. Its depreciation (Figure 2.E8) ballooned offshore-financed corporatedebt. Ensuing capital flight intensified the crisis. Long infrastructure lead-times meantthere was still supply-in-waiting; many projects were stalled or abandoned. Figure 2.F9shows GFCF collapsing with no noticeable rebound by 2001. 0.05 0.045 0.04 USD per THB 0.035 0.03 0.025 0.02 0.015 0.01 0.005 0 4 4 5 5 6 6 7 7 8 8 9 9 0 0 1 1 99 99 99 99 99 99 99 99 99 99 99 99 00 00 00 00 n -1 l-1 n-1 l-1 n-1 l-1 n-1 l-1 n-1 l-1 n-1 l-1 n-2 l-2 n-2 l-2 Ja Ju Ja J u Ja J u Ja J u Ja J u Ja Ju Ja Ju Ja J uFigure 2.E: Baht-US$ Exchange Rate 1994-20018 Source data: www.fx.sauder.ubc.ca9 Source data: www.nesdb.go.th and www.fx.sauder.ubc.caDissFinal Page 20 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Hayek (1933) argues artificially low interest rates breed over-investment, precipitatingcrises with debt- and investment-overhangs delaying recovery. Faber (2002, pp.192-193) argues global liquidity injections following the 1995 Mexican crisis fuelled furtherAsian speculative growth, delaying but ultimately amplifying and prolonging the AFC. 14,000 12,000 10,000 million USD (1988 prices) 8,000 6,000 4,000 2,000 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 Gross Fixed Capital Formation Private Construction Government Construction Land Development Construction And Land DevelopmentFigure 2.F: Dollarised Thai GFCF 1994-2001Faber (2002, p.69) notes cycles are “particularly violent in the case of emergingeconomies, emerging industries and emerging companies, which grow and evolverapidly and are, therefore, capital-hungry.” Transport infrastructure construction isespecially capital-intensive.DissFinal Page 21 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Although infrastructure and utilities are often seen as defensive investments, Forsgren etal (1999) argue toll road performance is cyclical, noting with reference to China (notgenerally regarded as badly hit): Challenging business climate with (official) economic growth down to 7% p.a. Delayed construction of connector roads and reduced commerce reducing traffic growth (and occasionally traffic declines) Debt service coverage (operating revenues) short of base projections Growing doubts as to willingness and ability of local partners to pay minimum income guarantees to toll companies (note: these were abolished by decree in 2002) Increased refinancing and foreign exchange risks Periodic toll increases required to meet projections, yet approval process is opaque Problems with toll collection/ leakage Credit ratings deteriorating due to reduced credit quality of counterpartiesIn Indonesia, the rapid devaluation of the Rupiah in 1997, compounded by rapidlyincreasing fuel prices, massive economic and political uncertainty and civil unrest,substantially reduced Jakarta Intra Urban Tollroad traffic volumes (Ibid.).Such patterns are not new. Despite railways driving America’s economic developmentin the 19th Century, Faber (2002, pp.55-63) notes they exhibited cyclical booms andcrises. Moreover, historically overseas investors are often latecomers, repeatedly buyingpeaks to sell-out in the immediate aftermath of crisis.DissFinal Page 22 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.9 Transport ModellingCorbett and Di Bona (2006) note transport models provide inter alia: assessment ofdemand-side project risks; evaluation of alternative projects against one another; and,forecasts of economic and financial returns, for use in project valuation. Traditional“Four Stage” models (elucidated in Ortúzar and Willumsen, 1994) are outlined inAppendix 7; but such models are data hungry so simplifications are common. Theirapplicability to tollways has been questioned (Willumsen and Russell, 1998).Usually the modelled area is divided into spatial zones. Traditionally, traffic to/ fromeach zone is estimated based on land-use and corresponding trip generation rates.However, given sparseness of robust land use data in developing countries, econometricmodels of traffic levels are often used. Whilst Khan and Willumsen (1986) fitted S-curve models to vehicle ownership and usage, often historical traffic counts areregressed on corresponding income data to estimate income elasticities of trafficdemand, defined as:   t1  t 0      t t   2    T   0 1   y  T  Y   y1  y0  (9)       y0  y1   2    Where: to,t1 are traffic levels in periods 0 and 1 y1,y0 are income (GDP) levels in periods 0 and 1As elasticities might not hold over time forecast values are adjusted, based either on S-curves or a conservative assumption of gradually declining elasticities, taking implicitaccount of longer-term demand saturation or improved logistic efficiency (decreasedlorry empty-running). Though these ignore vehicle ownership/ usage costs, Pindyck andRubinfeld (1981, pp.396-398) note Hymans’s (1970) model of USA vehicle ownershipDissFinal Page 23 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)shows such factors have short-term impacts, income-ownership relationshipspredominating thereafter.In developing countries tollway appraisals, driver interview surveys scaled using trafficcounts are often used to obtain trip patterns. Effects of other modes (e.g. rail) arecommonly omitted; impacts might be insignificant, or data unavailable.In order to determine vehicle routeing, a variety of approaches are possible, including:Network Assignment Modelling: Where the network is complex (roads parallel andperpendicular to the toll-road significantly affecting patronage), network assignmentmodels should be used. In addition to interzonal trip matrices, the road network is coded(e.g. length, capacity, tolls and relationships between speed and congestion). Aniterative assignment process is used, with link speeds recalculated to reflect congestion.Typically forecasts are prepared for a base year, opening year and at 5 or 10-yearintervals thereafter, with intermediate years interpolated. Such models are calibrated byadjusting network coding and often using maximum entropy matrix estimation (see VanZuylen and Willumsen, 1980) to better match traffic counts.Logit-Based Corridor Modelling: A spreadsheet-based approach to model a corridor,typically with one competing route (e.g. with no/ lower tolls and lower speeds). Trafficis allocated between routes based on a logit function; (10) shows an absolute logit curvefor forecasting a new road’s traffic. For existing toll-roads incremental logit modelsmay be preferred, shown in (11). Commonly κ and λ would be estimated based onprevious studies (ideally existing toll-roads). Richardson (2004) notes a general biasagainst using toll roads (κ<0). Forecasts may be prepared for selected years(intermediate years interpolated) or for all years. Whilst congestion levels do notfeedback, increasing incomes make tolls more affordable.DissFinal Page 24 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 1 PijXt  ,       GC ij ,t  GC ijX,t L  (10) 1 eWhere: PijXt is the share of trips i→j in period t using the expressway, PijXt  PijLt  1 , , , r GCij ,t is the generalised cost for trip ij by route r (X=expressway, L=local road), comprising equivalenced time and monetary elements in year t κ, λ are calibrated parameters   1        PijXt    GCij ,t GCijX,t    L IPij ,t  Obij ,t 0   X   Obij ,t 0   X X , X 1  e    Pij ,t 0    (11)    1      1  e  GCij ,t 0 GCij ,t 0    L X  Where: ObijX,t 0 is the base year observed expressway market share for trips ij PijXt , is forecast expressway share in year t (absolute logit); t=0 is base year2.10 Traffic Risks and Forecasting IssuesBain and Wilkins (2002) analyse toll-traffic uncertainty and traffic forecast error,showing strong inter-correlation. Average initial year traffic was 70% of forecastoverall, 82% in lender-commissioned projections and 66% when commissioned byothers, suggesting commissioning party influence on forecasts: debt-financiersrelatively more concerned with down-side risk than equity-holders. Their Traffic RiskIndex (shown in Appendix 8) compares low and high risk factors for toll roads andtraffic forecasts in general.Whilst initial year errors might be due to ramp-up (see 2.10.4), which Streeter andMcManus (1999) reckon can last 3-5 years, Bain and Polakovic (2005) note optimismbias is “constant through Years 2 to 5” as shown in Table 2.2, signalling other errorsDissFinal Page 25 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)(discussed below). They also note drastic differences in forecasts by different parties forthe same projects, based in part on very different assumptions.Table 2.2: Bain and Polakovic Forecast Performance Statistics Operating Year Mean Actual/Forecast Traffic Standard Deviation 1 0.77 0.26 2 0.78 0.23 3 0.79 0.22 4 0.80 0.24 5 0.79 0.252.10.1 Toll Sensitivity and the Value of TimeExcepting “shadow tolling” (operator reimbursed based on patronage instead of user-tolling), willingness-to-pay tolls is critical. Typically choice is between a slow, cheaproad and a fast toll-road; time and money equivalenced using the behavioural Value ofTime (VOT) to give “generalised cost.” Whilst higher tolls are usually preferred (see2.10.4) sometimes they are too high (Wong and Moy, 2004). The price elasticity oftollway demand is:   q1  q0     Q   q0 q1   2   D  p    P  p1  p0  (14)       p0  p1   2    Where: ΔQ is change in traffic ΔP is change in price (toll) q1,q0 are traffic after and before toll change respectively p1,p0 are new and old tolls respectivelyDissFinal Page 26 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Figure 2.G shows the relationship between demand, revenue and η. When tolls arebeneath the revenue maximising level (i.e. p<Prm) 0   D  1 , toll increases boost prevenue; when p>Prm  D  1 (toll increases decrease revenue).  D  1 when p=Prm. p pWillumsen and Russell (1998) note in developing countries Stated Preference surveys toestimate  D and VOT are scarce and of uncertain quality. Reference is often made to pprevious studies, factored for income levels. But the income elasticity of VOT, VOT is ycomplicated: as income increases, VOT rises (“income effect”), as does expenditure onother products/ services (“substitution effect”) and possibly savings too (“savingseffect”), implying VOT  1 . In developed economies, Wardman (1998) suggests yVOT  0.49 ; Gunn and Sheldon (2001) advocate 0.35  VOT  0.7 . Cross-sectional y yanalysis between developing countries suggests VOT  1 yet time-series analysis within ya country VOT  1 to growth VOT thereafter10. y Revenue Maximisation -η Demand Total Revenue η= −1 Prm Price→Figure 2.G: Demand, Revenue and Price Elasticity of Demand10 Confidential source used in absence of public source.DissFinal Page 27 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Goods vehicles are of particular concern. Bain and Wilkins (2002) note in developingcountries long-distance tolls often exceed drivers’ wages, giving incentive to useuntolled routes (pocketing bosses’ toll money). Some studies (e.g. ADB, 2003) havefailed to establish any VOT for goods vehicles.2.10.2 Competing Routes and Link RoadsContractual guarantees theoretically limit competing routes’ development, presupposingthe contracting branch of government is willing and able to enforce such guaranteesacross multiple government layers.Jiangsu Expressway circumvented this risk by acquiring rights to highways parallel totheir flagship Shanghai-Nanjing Expressway and so manage (and toll) traffic on bothroutes. However, when GZI Transport listed in 1997, it was assumed that the ferryparallel to the (then) soon-to-open Humen Bridge would cease operation. But beingoperated by a different local government, operation continued with fares undercuttingbridge tolls, attracting substantial goods vehicle volumes from the Humen Bridge.Even when concessionaires gets first refusal at planned parallel routes, overinvestmentmay result in excess infrastructure relative to traffic levels. Buchanan (1999) notes inMalaysia those identifying schemes can often proceed (subject to financing) withoutdue diligence of impacts on existing BOT’s.Though more important for urban projects, provision of adequate link roads is alsoimportant. Congested approaches/ exits can result in “hurry up and wait” (Bain andWilkins, 2002), reducing tollways’ attractiveness.DissFinal Page 28 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.10.3 Toll Increases and Revenue GuaranteesContracts typically allow periodic price-indexed toll increases, or at a percentage ofprice inflation. However, Forsgren et al (1999) note toll increase approval processes areoften opaque and beset with delay. Bain and Wilkins (2002) note tariff escalation isoften politicised, especially where there is little previous “tolling culture.” Sometimessocial unrest follows tolls’ imposition (Orosz, 1998) or toll increases, especially duringeconomic downturns (Dizon, 2002).Some contracts give revenue guarantees to operators, underwritten by government.However, China’s 2002 State Council directive scrapped such revenue guaranteesoverriding contract provisions, leading to New World Development divesting from 13toll roads and bridges (Chan, 2003).Whilst non-toll revenues may be generated (e.g. service stations, advertising), Streeteret al (2004) note their contribution is usually dwarfed by toll revenues.2.10.4 Ramp-UpBain and Wilkins (2002) define ramp-up as information lag for users unfamiliar with anew highway and general reluctance to pay tolls (see Richardson, 2004 for experimentalevidence). Streeter and McManus (1999) reckon on 3-5 years’ ramp-up and note this isoften underestimated in traffic forecasts.Bain and Wilkins (2002) note ramp-up experience tends to cluster to extremes: either oflimited duration (even exceeding forecast traffic levels) or lagging for a long duration,maybe never “catching up”, particularly for projects with a high Traffic Risk Index (seeAppendix 8). They derived revenue-adjustment factors as per Table 2.3 for use infinancial stress-tests.DissFinal Page 29 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Table 2.3: Bain and Wilkins Ramp-Up Revenue-Adjustment Profiles Forecasts Lenders Others commissioned by Traffic Risk Low Average High Low Average High Year 1 revenue -10% -20% -30% -20% -35% -55% adjustment Ramp-up duration 2 5 8 2 5 8 (years) Eventual catch-up 100% 95% 90% 100% 90% 80%2.10.5 Operating CostsIn addition to tolls, many models also apply distance-based monetary Vehicle OperatingCosts (VOC) reflecting fuel, maintenance, depreciation, etc. Whilst economic values forthese parameters are derivable, accurate behavioural values are often elusive. In practicethey may be used to reflect certain advantages of higher quality roads, whereon wear-and-tear may be less and where smoother flow may yield fuel savings. However, theseare typically applied as fixed values with respect to distance and road-type, rather thanfeeding-back modelled forecast speeds. Where there are larger VOC savings from anexpressway ceteris paribus there is more scope for higher tolls. However, there is anissue as to who pays these costs (driver or employer).2.10.6 Toll LeakageSome vehicles use a facility without paying, either legitimately (e.g. certain governmentor military vehicles) or illegitimately. There may be theft by toll-collectors and fraud byadministrators. Forsgren et al (1999) note toll leakage can be as high as 20% ofrevenues. Sometimes computerised toll collection and auditing can restrain losses, buton lower volume routes the cost of such measures might outweigh savings.DissFinal Page 30 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.10.7 Induced TrafficWhen a new highway significantly reduces transport costs or relieves congestion, it mayresult in additional (induced) traffic. Corbett et al (2006, p.A2-99) report substantial,rapid induction on Cambodia’s roads following rehabilitation. On green-field sites, itmay also over time enable expanded development, generating further traffic demand.However, Willumsen and Russell (1998) note the difficulty of reliably forecasting sucheffects; Bain and Polakovic (2005) report the prevalence of significant errors in inducedtraffic forecasts.2.10.8 AnnualisationBain and Wilkins’ (2002) Traffic Risk Index shows projects with seasonal flow patternstend to be riskier. For inter-urban highways a “typical” day is usually modelled, withresults factored-up to annual forecasts. Thus seasonal changes might not be captured:forecasts represent an expansion of one part of the annual pattern. Even when AnnualAverage Daily Total (AADT) traffic is modelled, larger seasonal variations equate tolarger total variance between modelled day and actual day across the year.For those projects where modelled hours are considered, mathematically the problemincreases, given further factoring from a “typical” hour (or perhaps AM peak and PMpeak) to a “typical” day. Conversely, when modelling a day, future congestion in peakperiods and its impact on effective daily capacities may be under-estimated.DissFinal Page 31 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.10.9 Economic EffectsEconomic risks feed through many elements of traffic forecasts: Overall travel demand (e.g. car ownership and usage, freight volumes, extent of traffic induction) Willingness-to-pay tolls and try tollways (affordability; ramp-up extent and duration) Toll leakage (incentive for malfeasance) Over-investment increasing likelihood of competing routes being built/ upgradedEconomic cycles affect most aspects of the economy and decision-making, includingevaluation assumptions adopted. Transport consultants define economic growthscenarios either under guidance or instruction of commissioning parties. Whenexpectations are high more projects are evaluated, so proportionally more projects arelikely to founder on downturn (and be blamed on transport forecasts). This may createcynicism regarding tollway investments extending into the early economic recovery,resulting in under-investment in some areas, thence over-investment as returns onoperating (and newly opened) highways exceed expectations, thus creating a new“error of optimism” (Pigou, 1920).Luu (2006) and Gomez and Jomo (1999) cite governments in Vietnam and Malaysiapotentially over-expanding transport infrastructure development.DissFinal Page 32 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.11 Construction, Operations and MaintenanceConstruction cost overruns and delay (deferred/ lost revenue) may imperil initial debtrepayments. Rigby (1999) notes using engineering, procurement and construction (EPC)contractors’ reputations to proxy technical risk is both commonplace and erroneous:construction risks are often inadequately assessed. Based on UK experience, Flyvbjergand COWI (2004) recommend highway construction cost estimates be uplifted 15% if a50% chance of overrun/ delay is acceptable, or by 32% if 20% chance acceptable.Ruster (1996) notes construction cost overruns, delays and defects can be largelymitigated by liquidated damages, performance bonds, warranties, contingency fundsand insurance. As revenue losses are rarely disputed during delay/ overrun arbitrations,the focus of this Dissertation remains on demand-side risks. However, when thecontractor is the concessionaire, such risks should be analysed. Similarly, operationsand maintenance (O&M) risks should also be considered.Table 2.4 shows estimated costs for new expressways in China and Vietnam. Whilstcosts are dependent on terrain, design standards and local labour and material costs,there is significant difference between HHI costs and others (ADB potential projects),unlikely wholly attributable to differences in local prices, or the difference betweenDual-2 and Dual-3 standard. A distance-weighted average of US$4.633m per km ofDual-2 was derived, to be used in Chapter 5’s simulation model.There is a trade-off between construction and subsequent operations and maintenancecosts. The latter also affected by periodic major maintenance (e.g. immediately beforeconcession handback). Literature review found little agreement as to how to gauge suchcosts, and whether they should be related to construction or traffic flow/ revenue. Table2.5 shows some public domain values; some confidential sources suggested using 6% ofDissFinal Page 33 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)initial construction cost. In practice, there is likely to be a fixed element which could betaken as proportional to construction costs, plus a variable element proportional totraffic/ revenue. For simulations in Chapter 5, it is recommended to adopt 2% ofconstruction cost (fixed) plus 3% of toll revenue (variable).Table 2.4: Estimated Expressway Construction Costs Cost Cost per km Expressway Length (US$ million) (US$ million) Source Guangzhou E-S-W Ring 38 km, HHI (2003, pp US$542m 14.263 Road, China Dual 3 104-108) Phase 1 West, Guangzhou, 14.7 km, HHI (2003, pp US$207m 14.082 China Dual 3 114-118)Hanoi – Lao Cai Expressway, 260km, Corbett, et al US$915m 3.519 Vietnam Dual 2 (2006, p.VIII-4)Nanning – Baise Expressway, 189km, US$600m 3.175 China Dual 2 Corbett, et al Bien Hoa – Vung Tau 90km, (2006, p.VIII-5) US$680m 7.556 Expressway, Vietnam Dual 2 Dau Giay – Lien Khoung 189km, Corbett, et al US$600m 3.175 Expressway, Vietnam Dual 2 (2006, p.VIII-9) Hanoi – Haiphong 100km, US$410m 4.100 Expressway, Vietnam Dual 2 Da Nang – Quang Ngai 140km, US$700m 5.000 Corbett, et al Expressway, Vietnam Dual 2 (2006, p.IX-4)Saigon – Long Thanh – Dau 55km, US$350m 6.364 Day Expressway, Vietnam Dual 2/ 3 Hanoi Ring Road, Vietnam 65km US$600m 9.231 Total 1,140.7km US$5,604 4.913 Total (assuming Dual 2 throughout) 4.633Table 2.5: Operations and Maintenance Costs O&M as % of O&M as % of Construction Conservative Highway (Mean) Revenue (Mean) Source(s) Hefei-Nanjing Expressway 2.3% to 8.4% SWHK (1996a, (134km, Dual-2) (4.2%) 1996b)Shanghai-Nanjing Expressway 3.2% to 17.5% SWHK (1997a, (254km, Dual-2) (6.1%) 1997b) 1.3% to 7.8% Guangzhou E-S-W Ring PBA (2003) (2.9%) 2.1% to 5.1% Phase I West PBA (2003) (3.1%)DissFinal Page 34 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.12 Other ConsiderationsUsually concessions are tendered. Ormerod (2005, pp.94-98) notes even with relativelyfew pre-qualified firms (with technical, managerial, local and financial capability),oligopolistic Nash equilibria are elusive. Possible strategic stakes, asymmetricinformation, future expectations and track-records complicate bidding.Given major highways’ perceived importance, local reputation/ guanxi11 may beimportant. In Malaysia, Buchanan (1999) reports prospective concessionaires able toidentify then pursue projects uncontested. Gomez and Jomo (1999) observe well-connected businesses getting lucrative contracts in exchange for undertaking lesslucrative ones (possibly in other sectors). Whilst such arrangements distort markets,they sometimes enable achievement of specific national targets.Sometimes projects are pursued for local political rather than economic reasons. ADB etal (2005, p.92) note “pork barrelling” is prevalent in the Philippines, with an estimated22.5% of the public works’ budget over 1997-2001 allocated to these (Manasan, 2004).Government coordination is an issue in China, where local government officials’performance is correlated with the amount of GFCF in infrastructure generated,including FDI (ADB et al, 2005, p.102); WB (2005) argues this creates a danger ofover-investment. Whilst decentralisation is predicated on increasing responsiveness, alack of suitable local experience contributed to Mexico’s US$13bn 1989-94 toll roadprogramme amassing US$5.5bn in non-performing non-recourse loans (Irwin, 1999).11 See Appendix 6 for detailed definition.DissFinal Page 35 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Whilst expecting the same “game rules” as in the West is unreasonable, corruption is aconcern. Azfar et al (2000) estimate Philippine public sector corruption at 20%-40%.ADB et al (2005, p.116) note public/ highway works there often trigger unofficialpayments to each government tier involved. WB (2003b, 2004b) observes similarproblems in Indonesia, costing up to 30% of procurement budgets. Data in Appendix 9show corruption is widely perceived as a problem, both within the region and byTransparency International (2004); only Malaysia is (just) outside the “widespreadcorruption” definition.Brinkman (2003) and Kilsby (2004) identify other forecasting issues, such as models’opaqueness, lack of resources to properly forecast, plus psychological and ethicalfactors, overlapping to an extent with some of the “technical” issues above. Thisincludes modellers deluding themselves as to the infallibility and neutrality of theirforecasts, which are more often flawed and biased.DissFinal Page 36 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2.13 Summary of Key IssuesThis Literature Review identified a number of evaluation metrics which may be used(e.g. NPV, FIRR). Numerous project risks were also identified, whose importance mayvary between countries and projects (with some risks correlated), which may besummarised under the following headings: Macro-Economic Risks: including institutional, sovereign and broad economic risks. Market Risks: primarily concerning scheme attractiveness and riskiness. Forecasting Risks: pertaining to uncertainty and transport modelling practice.Stakeholder attitudes to many of these risks (and the utility of evaluation criteria) can betested by questionnaire surveys (Chapter 4), in terms of how often such risks areconsidered, whether they are deemed important and in the case of certain economicparameters, whether they are expected to increase or decrease in the near- to medium-term. Many risks may also be tested quantitatively by risk simulation modelling(Chapter 5). The factors and proposed testing methods are indicated in Table 2.6.Certain risks are beyond this Dissertation’s remit (e.g. bidding strategy) or not readilytestable by either questionnaire or risk simulation.Addressing the hypothesis, excepting the use of interest rates in financial analysis, littleliterature emphasised any importance of either price inflation or interest rates ontollways. Are they unimportant? Or is this merely symptomatic of most literature beingbased on Kondratieff downswing conditions? They are therefore included in the keyrisks to be considered in both the questionnaire surveys and risk simulation.DissFinal Page 37 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Table 2.6: Summary of Key Risks and Issues For Testing ByRisk Type Questionnaire Risk SimulationMacro-Economic Risks Country’s political and legal systems  Exchange risks: exchange rate and cash repatriation  Interest rates   Price inflation   Economic growth and business cycles   Income (in)equality  Tolling culture  Corruption Market Risks Road’s social/ economic benefits  Construction time/ threat of over-run   Construction cost/ threat of over-run   Operation & maintenance costs   Contractual foundations  Threat of competing routes   Ramp-up: size and length   Toll affordability   Enforceability of toll increases   Minimum income guarantees  Toll leakage   Truckers using free routes, pocketing boss’s toll money Guanxi  Connecting roads: access/ egress  Forecasting Risks Frequency of Over- and Under-Forecasting   Ramp up: length & size   Toll affordability   Sensitivity of traffic levels to GDP growth   Overall sensitivity of project traffic to tolls  Sensitivity of trucks/ large vehicles to tolls   Toll sensitivity to changes in income  Data availability/ quality for model calibration  Data availability/ quality for forecasting  Reliability of transport modelling process  Induced traffic Forecasters pressured by clients to adjust numbers  Treatment of connecting and competing routes Evaluation Criteria Use of financial metrics, e.g. NPV, Financial IRR  Project’s social cost/ benefit and which metrics used  Do counterparties mitigate or add to project risk? Other Risks Force Majeure “Pork Barrelling” Bidding StrategyDissFinal Page 38 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3. Environmental Analysis3.1 Introduction and PESTLE AnalysisWhilst Literature Review concentrated on generic project risks, environmental analysisis used to gauge potential risk and opportunity by country for toll roads.East Asia is too diverse for meaningful 5 Forces analysis (Porter, 1980); each projectjustifying its own framework. However, PESTLE analysis can identify externaldynamics affecting the market. Table 3.1 summarises key points (full analysis inAppendix 10), showing a growing desire overall for inter-urban transport. The keydriving-force is economics; however, political/ legal constraints include corruption.Table 3.1: Highlights of PESTLE Analysis Element Description Stability concerns in many countries, though not always deterring Political infrastructure investment Economies generally growing relatively rapidly, although wealth Economic levels varied. Generally much/ growing inter-urban travel, in parallel with rapid urbanisation. Demand suppressed in some cases by poor infrastructure. Social Some countries have developed foreign private financing more than others. In general, the scope for this sector’s contribution is acknowledged, but deep-seated nationalism can restrict foreign equity shares, sometimes creating management control issues.Technological Tolling is largely manual, excepting a few major routes. Legal A wide variety of legal systems, but with corruption often rife. Economic development predominates over environmentalEnvironmental considerationsDissFinal Page 39 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3.2 Political, Legal and Stakeholder IssuesCheong (1999) notes all countries experienced “maximum government” since 1945:military (Indonesia, Myanmar, Thailand12), emergency (Malaysia, Philippines) orcommunist rule (Cambodia, China, Laos, Vietnam). Such potential for centralisationremains either through current maximum government or switching from “rule of law"to “rule by law.” Risks might be compounded by multi-tiered government withoverlapping authority and a lack of transparency. Stakeholder mapping can illustrateopportunities and risks. Figure 3.A maps typical post-opening stakeholders13. Development Agencies (excepting Rest of Competing project donors/ Government Projects lenders) Government Equity (Concerned Dept.) Holders Concession Lenders Users Staff Suppliers Society Supplying Industries (including consultants, contractors, etc)Figure 3.A: Typical Concession Stakeholder Map12 Thailand reverting to military rule during the preparation of this Dissertation. Although to date, littleresultant impact appears to have been made on economic sentiment.13 Author’s own work.DissFinal Page 40 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Prior to award there are numerous potential concessionaires with equity/ debt financiers(and in some cases possibly conflicting units within government allied with certainbidders over others). During construction, suppliers would be more central.In addition to competition with other routes/ concessions, there may be conflict betweendifferent government departments having larger perceived stakes in other projects,either through governmental equity involvement or guanxi. There is also a trade-offbetween users and society; concession terms negotiated with government determine theextent of user subsidisation/ penalisation. Likewise, there may be conflicts betweenequity holders and government.The above illustrates possible conflicting/ coinciding interests, which should be mappedfor each project individually. Equally, factors’ impacts may change: following the AFC,connections with the Suharto family (previously critical to success in Indonesia) becamea business liability (Forsgren et al, 1999, p.152).La Porta et al (1997) found shareholder protection, good accounting standards and ruleof law strongly negatively correlated with concentration of ownership, suggesting suchfacets are important for good operation of capital markets to facilitate infrastructurefinancing.DissFinal Page 41 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3.3 Economic RecoverySection 2.8 discussed the causes and impacts to 2001 of the AFC in Thailand, showingover-investment precipitated the currency crisis and investment-downturn. However,there is evidence that economies (and construction as a proxy for infrastructureinvestment in general) have recently picked-up, with currencies largely stabilised.Figures 3.B14 and 3.C14 show a pick-up in GFCF since 2002 (in Baht) or 2003 (in US$),although construction continued to decline as a proportion of GFCF. However, privatesector construction has grown year-on-year since 2001/2002 (Baht/US$ respectively).Whilst Figures 3.D15 and 3.E15 show the collapse in GFCF relative to M2 and GDP asindexed to 1995 in Baht and US$ respectively. However, Figure 3.F15 shows a recoveryin GFCF in recent years (indexed to 2000); GFCF appears relatively income elastic,dipping lower than GDP in 2001, thereafter growing more rapidly. This suggests anupturn in GFCF, likely to increase construction spending and possibly tollways.Figure 3.G16 shows declines in a number of currencies following the AFC; only theChinese RMB was unscathed due to its pegging to the US$. Whilst time-series data onother currencies were not available, Vietnamese Đong, Cambodian Riel, MyanmarKyats (free-market rate) and especially Lao Kip all depreciated substantially over thisperiod also. However, Figure 3.H16 shows that since January 2001 currencies have beenbroadly stable; Myanmar Kyats (not shown) are the exception, continuing to devalue onthe free-market.14 Raw data from: NESDB (2006) and www.fx.sauder.ubc.ca15 Raw data from: www.bot.or.th and www.fx.sauder.ubc.ca16 Raw data from: www.fx.sauder.ubc.caDissFinal Page 42 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 350,000 60% 300,000 50% Construction as % of GFCF 250,000 Million Baht (1988 prices) 40% 200,000 30% 150,000 20% 100,000 10% 50,000 0 0% 93 94 95 96 97 98 99 00 01 02 03 04 05 06 19 19 19 19 19 19 19 20 20 20 20 20 20 20 Gross Fixed Capital Formation Private Construction Government Construction Construction as % of GFCFFigure 3.B: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) 14,000 60% 12,000 50% Construction as % of GFCF 10,000 Million USD (1988 prices) 40% 8,000 30% 6,000 20% 4,000 10% 2,000 0 0% 93 94 95 96 97 98 99 00 01 02 03 04 05 06 19 19 19 19 19 19 19 20 20 20 20 20 20 20 Gross Fixed Capital Formation Private Construction Government Construction Construction as % of GFCFFigure 3.C: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) in US$DissFinal Page 43 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 200 175 150 Index (1995=100) 125 100 75 50 25 0 1994 1996 1998 2000 2002 2004 GFCF GDP M2Figure 3.D: Thai GFCF, GDP and M2 in Baht, Indexed to 1995 200 175 150 Index (1995=100) 125 100 75 50 25 0 1994 1996 1998 2000 2002 2004 GFCF GDP M2Figure 3.E: Thai GFCF, GDP and M2 in US$, Indexed to 1995 150 125 Index (2000=100) 100 75 50 2000 2001 2002 2003 2004 GFCF GDP M2Figure 3.F: Thai GFCF, GDP and M2 in US$, Indexed to 2000DissFinal Page 44 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 120 Indexed Value vs. USD (Jan96=100) 100 80 60 40 20 0 Jan-94 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Chinese RMB Indonesian Rupiah Malaysia Ringgit Philippine Peso Thai BahtFigure 3.G: Currency Performance since 1994 120 Indexed Value vs. USD (Jan01=100) 100 80 60 40 20 0 Jan-01 Jul-01 Jan-02 Jul-02 Jan-03 Jul-03 Jan-04 Jul-04 Jan-05 Jul-05 Jan-06 Chinese RMB Indonesian Rupiah Malaysia Ringgit Philippine Peso Thai BahtFigure 3.H: Currency Performance since 2001DissFinal Page 45 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3.4 Vehicle OwnershipKhan and Willumsen (1986) note correlation between car ownership and roadspace indeveloping countries: statistically one proxying the other. ADB et al (2005, p.3) suggestthe following broad correlation between GDP and roadspace:  km _ of _ PavedRoads   LandArea _ in _ km2   0.5  0.5 ln US $ofGDPperCapita  ln   (15)  However, no goodness-of-fit is given (graphical presentation suggests low R2).Appendix 11 details a series of regressions undertaken using data in Appendices 1 and2, comprising fits on the Study Area 9 countries, plus 5 others for benchmarking. Thesesuggest S-curve relationships for paved roads, railway and airports in terms ofkilometrages/ number of airports per km2 or per capita. Figures 3.I and 3.J showequations fitted for roads per capita and per km2 respectively, with respect to GDP percapita, suggesting substantial road build-out/ vehicle ownership growth are likely aseconomies grow. These also suggest clustering as follows: Relatively developed networks, in countries with significant prior experience of transport infrastructure privatisation: China, Indonesia, Malaysia and Thailand; Relatively undeveloped networks, also correlating to a relative lack of infrastructure privatisation: Cambodia, Laos and Myanmar; and, Intermediate countries: with some problematic experience of privatisation (Philippines) or nascent interest in privatisation (Vietnam).DissFinal Page 46 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Population per km of Paved Road 10 MM 9 KH Ln(Population per km of Paved Road) PH VN 8 LA 7 ID TH CN MX KR 6 MY 5 UK PO US 4 3 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.)Figure 3.I: Relationship between Wealth and Roads Per Capita km 2 per km of Paved Road 6 MM 5 KH Ln(km2 per km of Paved Road) 4 LA 3 MX VN PH ID TH 2 CN MY 1 US KR 0 PO UK -1 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.)Figure 3.J: Relationship between Wealth and Road DensityDissFinal Page 47 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3.5 Traffic Performance of Existing Toll RoadsFor further analysis of income’s effect on traffic volumes, time series econometricanalyses were undertaken on data available as shown in Appendix 12, namely: Guangzhou-Shenzhen Superhighway, Guangdong (Hopewell Highway); Jiangsu Section of the Shanghai-Nanjing Expressway, (Jiangsu Expressway); and, Shanghai-Hangzhou-Ningbo Expressway (Zhejiang Expressway)Summary income elasticities (with respect to real GDP growth) are shown in Table 3.2.Whilst some caution is advised in interpretation as data cover different time periods andtoll changes are not considered, the overall income elasticity of expressway traffic isremarkably similar in all three instances; despite differences in vehicle ownershipsensitivities: although Guangdong Province is relatively more developed and thus mightbe higher-up the S-curve (ownership growth smoothing off), this does not explain thedifference in vehicle ownership growth between Jiangsu and Zhejiang.Table 3.2: Vehicle, Trip and Expressway Patronage Income Elasticities Guangdong Province/ Jiangsu Province/ Zhejiang Province/ Guangzhou-Shenzhen Shanghai-Nanjing Shanghai-Hangzhou-Income Elasticity of: Superhighway Expressway Ningbo ExpresswayVehicle Ownership 1.02 1.41 1.78Passenger-km 0.61 0.69 0.37Passenger Trip Length 0.14 0.23 0.03Freight MT-km 0.44 0.33 0.82Freight Trip Length 0.68 0.30 0.37Expressway Traffic 1.39 1.43 1.54 Car/Small n/a 1.14 1.70 Small/Medium n/a 1.54 1.05 Medium/Large n/a 1.35 1.34 Large/Heavy n/a 2.46 3.10Expressway Revenue 1.38 n/a 2.11Data from: 1995-2004 1997-2003 1998-2003DissFinal Page 48 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Whilst the largest vehicle categories’ expressway patronage is most responsive to GDPgrowth, they constitute a small proportion of traffic as shown in Figures 3.K and 3.L forShanghai-Nanjing and Shanghai-Hangzhou-Ningbo Expressways respectively.These analyses show the difference (and consequent risk) even between three leadingcoastal provinces in China, highlighting the importance of local factors for any project.However, they also show that inter-urban tollways can perform well with respect toGDP, even in a country with relatively well developed highway networks relative to therest of the region (see Section 3.4) and thus may make an attractive investment.Shenzhen Expressway (2006), Jiangsu Expressway (2006, p.143) and ZhejiangExpressway (2006, p.7) levy quite similar tolls on interurban highways 17. Cars are tolledat RMB0.40-0.60 per km (US$0.05-0.075 per km); and trucks (up 10 tonnes) atRMB1.00-2.40 per km (US$0.12-0.30 per km), though on most routes at aroundUS$0.15 per km. Thus representative values of US$0.06 and US$0.15 per km for carsand trucks could be adopted for the simulation modelling in Chapter 5.17 Other companies cited in this section typically report total toll revenue, not broken down by vehicleclass and with toll rates not readily available.DissFinal Page 49 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 35,000 Distance-Weighted Average Vehicles per Day 30,000 25,000 20,000 15,000 10,000 5,000 0 1997 1998 1999 2000 2001 2002 2003 Car Small Medium Large+HeavyFigure 3.K: Traffic Growth on Shanghai-Nanjing Expressway 35,000 Distance-Weighted Average Vehicles per Day 30,000 25,000 20,000 15,000 10,000 5,000 0 1998 1999 2000 2001 2002 2003 Revenue Small Medium Large HeavyFigure 3.L: Traffic Growth on Shanghai-Hangzhou-Ningbo ExpresswayDissFinal Page 50 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3.6 Opportunities and ThreatsSection 3.5 illustrated that different projects in the same country may performdifferently. Hence, it is not feasible to present Strengths and Weaknesses analysis forthe Study Area as a whole. Nevertheless a broad-brush Opportunities and Threatsanalysis can summarise key potential macro-level downside and upside risks, based onthe clusters suggested in Section 3.4:3.6.1 Cambodia, Laos and MyanmarOpportunities include potential for substantial growth in car usage and expansion ofhighway networks. Given their pressing development needs, favourable/ flexiblecontract terms might be possible, possibly with partial funding from aid agencies(subject to sanctions in case of Myanmar). Geographically these countries link thestronger regional economies: Thai-Vietnamese land transport either via Laos orCambodia; Sino-Thai trade via Laos or Myanmar; Myanmar offering land-linkagebetween East Asia and South Asia.A key threat is that current poverty may lengthen ramp-up and limit toll affordabilityand traffic levels. Significant sovereign and institutional risks persist, together withcorruption.These countries are thus quite risky.3.6.2 Philippines and VietnamHigh capacity trunk highway networks are largely undeveloped, meaning attractiveroutes remain to be developed. Economic growth suggests that tolls might beaffordable. Vietnam has signalled intent to open-up to expanded FDI, whilst thePhilippines is the most culturally westernised country in the sample.DissFinal Page 51 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)However, legal protection for investors remains weak and the extent of tollingaffordability is uncertain. Corruption remains a concern.These countries are fairly risky.3.6.3 China, Indonesia, Malaysia and ThailandThese countries have relatively strong economies with strong prospects, both in export-oriented manufacturing and commodity markets. Furthermore, they have sizeabledomestic economies possibly providing some resilience to international economicfactors. They also have strong track records in attracting FDI into transportinfrastructure, including reasonable legal systems (as compared to other countries in theregion). Tolls are relatively easily afforded by many drivers.However, these countries may risk over-investment in certain regions (as befell all barChina in the AFC). There remains some legal/ institutional risk, as well as corruption ora need to have business networks (e.g. guanxi). It might be argued that many of themost attractive projects have already been built.DissFinal Page 52 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3.7 Postulated Position on K-WaveBased on a peak in interest rates in most leading economies in the early 1980’s, alongwith a peak in commodity prices (especially gold) and inflation rates, the lastdownwave begun around 1980/1981. Equally, the bottoming out of many commodityprices at the end of the 1990’s suggest an upswing began around the same time (more-or-less coinciding with the NASDAQ peaking in 2000). Recent increases in US interestrates and strong commodity markets support this assertion. Regarding inflation, Faber(2003, p.10) notes that in the classical definition of inflation (increased money supply),low interest rates and easy credit now available in many countries, but specifically theUSA are evidence of inflation; price indices are likely to accelerate. Prolonged low realinterest rates since the K-Wave bottom are likely to yield gold prices over-and-abovewhat would normally be expected in the early stages of the upswing (Faber, 2005,2006). Figure 3.M plots US interest rates and nominal gold price, with a simplified K-Wave. The recent surge in gold prices is also shown.A few transport planner-economists (e.g. Kilsby, 2006a, 2006b) have recentlypostulated and examined the implications of significant fuel price increases; thoughsuch work is not yet widespread.Assuming the K-Wave exists, an upswing has likely begun. Given S-curves of vehicleownership and the above analyses on road build-out, this suggests an upturn ininvestment prospects, also coinciding more-or-less with the Kuznets cycle ready torebound (based on roughly a half-cycle elapsed since AFC). Private sector projectfinancing developed since the 1980’s (during a Downswing); will the Upswing changethe rules of the game?DissFinal Page 53 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 18 800 16 700 14 Treasury Bill Interest Rates (%) 600 Gold Price (USD per oz) 12 500 10 400 8 300 6 200 4 2 100 0 0 50 55 60 65 70 75 80 85 90 95 00 05 19 19 19 19 19 19 19 19 19 19 20 20 10 Year T-Bill 3 Month T-Bill Gold(USD/oz) K-WaveFigure 3.M: Interest Rates, Nominal Gold Price and Kondratieff WaveDissFinal Page 54 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4. Questionnaire Survey4.1 PurposeThe Literature Review (Chapter 2) presented a number of project risks associated withstart-up toll road projects, as well as critiques of transport planners’ failures to take dueconsideration of these. The Environmental Analysis (Chapter 3) intimated the potentialfor toll roads in Study Area countries. In order to test both literature and environmentalanalyses, a questionnaire survey was undertaken to test practitioners’ experience andperceptions regarding: Their scope of project experience; Relative weightings of difference macro- and micro-level project risks; Data availability and quality; Accuracy of forecasts and which metrics are employed to test risk; Market outlook in the nine Study Area countries; and, Expectations for economic parameters.In addition to comparing respondent attitudes against the findings of the literaturereview and environmental analysis, expectations were measured to help define aforecast scenario for risk simulation testing (see Chapter 5). Differences in attitudesbetween different project stakeholders/ professional groups were also evaluated, both totest how different stakeholders perceive risks and to identify gaps between transportplanners’ performance and others’ expectations of them.DissFinal Page 55 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.2 Design Concept and Sample SelectionThe questionnaire was designed to afford a relatively broad sample of opinion; variousadvisory professions were sampled. Respondents were also asked to state the extent oftheir working experience, the proportion of this spent in relevant fields and theirgeographic experience. Questions covered economic, legal, engineering andconnectivity risks, as well as (for those with modelling experience) an investigation intothe reliability of transport demand forecasts, attempting to identify where practitionersfeel their art is weakest.Given the relative obscurity of business cycle theory even amongst economists, onlyone question relates directly to the use of business cycles, though others testexpectations regarding price inflation and other economic variables. In order to preventcomparison with especially turbulent periods (e.g. AFC and NASDAQ topping-out),expectation comparisons were between the last 5 and next 10 years.Sampling was done via the author’s personal contacts, extracting contact details fromliterature reviewed, using a number of internet-based newsgroups (“yahoogroups”), plusreview of professional databases (e.g. www.legal500.com for legal professionals). Assuggested by “sub-tribalism” (Morris, 1971), the best response rate was from thoseknown to the author and those in the same primary field (transport planning), so thesample skewed towards transport planners/ economists. Such people also accounted formost of the optional (text) responses on broader issues.Given that some questions were designed specifically for transport planners this was nota problem (those without such experience being screened out of such questions, thoughnot the rest of the survey, as described below).DissFinal Page 56 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.3 Questionnaire Design and Survey ExecutionIn order to expedite survey diffusion and result collation, the internet-basedwww.surveymonkey.com was employed, allowing easy questionnaire disseminationand automatic result collation.Piloting occurred in September 2006, followed shortly thereafter by the main survey(into October 2006). Appendix 13 shows the final questionnaire design, withobservations on the Pilot in Appendix 14 (also detailing actions taken to revise thequestionnaire to incorporate pilot feedback).Approximately 40 respondents started the survey but dropped-out after just a fewquestions. These responses were excluded from the analysis. In a number of cases,respondents did not give answers to each question, but nonetheless gave answers tomany questions. Under such circumstances a “not sure” response was assumed foromitted answers. And when evaluating answers, such “not sure” responses weretypically excluded, such that analysis would concentrate on stated opinions only. A totalof 162 responses were considered as valid for analysis (though due to “not sure” andomitted answers, this number was often lower for specific questions).Data returns are in Appendix 15. As the first 6 questions concerned respondent identity(confidential), these are not included herein. The following sections set-out and analyseresponses by headings under which questions were grouped.DissFinal Page 57 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.4 The Survey SampleThe first question determined in which sector(s) respondents had experience, based on14 categories, with multiple answers permitted. The 162 responses were aggregated intosix categories as shown in Table 4.1, based on which different stakeholders’ attitudescould be examined (as shown later). The relative proportions are also shown in Figure4.A (note: many respondents worked in multiple sectors).Table 4.1: Aggregated Respondent Experience Categories Group Components Number Expressway Developer/ Operator/ Equity Investor Lawyer/ Attorney/ Solicitor Financial, Legal, Private Sector Lender 29 Operator Investment Banker Ratings Agency Accountant/ Valuer Insurer Transport Planner/ Transport Planning Consultant 98 Economist Economist Civil/ Structural/ Pavement/ Highway Engineer/ Architect 37 Engineer/ Architect Government/ Aid Government 43 Agency Aid Agency Academic Academic 22 Other Other 24The largest group was transport planners (for reasons explained in 4.2 above). The totalsample was relatively experienced (20.6 years mean working experience), as shown inFigure 4.B. The sample’s working experience cross-tabulated with years of experiencein Table 4.2 shows the average respondent has spent over 10 years on transportinfrastructure projects and over 7 in developing countries. Although there were perhapsnot a great many respondents in the Financial/ Legal/ Operator category, respondentsdid include a number of very senior figures within this category, including key decision-makers/ backers of private infrastructure schemes.DissFinal Page 58 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 100 80 60 40 20 0 Financial, Transport Engineer, Government, Academic Others Legal, Planner, Architect Aid Agency Operator EconomistFigure 4.A: Respondents by Experience Type 1 to 4 5 to 9 7% 6% 30+ 26% 10 to 19 31% 20 to 29 30%Figure 4.B: Respondents by Years of ExperienceDissFinal Page 59 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Table 4.2: Respondents’ Mean Years’ Experience in Various Fields Average Sample’s Years per Total Years’ Project Type Respondent Experience Transport infrastructure projects 10.66 1,642 All infrastructure projects (transport & non-transport) 13.13 2,022 Projects in developing economies 7.26 1,119 Tolled highway projects (urban and/or rural, anywhere) 2.57 396 Rural or inter-urban tolled highway projects 1.70 262 Rural/ inter-urban tolled highways in developing economies 1.12 173Figure 4.C shows experience by global region; 102 (65%) having worked in East Asia,broken-down by country in Figure 4.D. Table 4.3 gives sectoral experience by StudyArea countries, showing substantial numbers with China experience (69 respondents),through to few with Myanmar experience (5 respondents). As a significant proportion ofthe sample have experience within East Asia and a familiarity with developingeconomies, the sample appears suitable for analysis. Respondents with Experience in this Region 0 20 40 60 80 100 120 North America Latin America/ Caribbean Western Europe Eastern Europe Africa Middle East Central Asia South Asia East Asia Oceania/ Australasia OtherFigure 4.C: Respondents’ Global ExperienceDissFinal Page 60 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Respondents Experience by Country 0 10 20 30 40 50 60 70 Brunei Cambodia China Hong Kong Indonesia Japan North Korea South Korea Laos Macau Malaysia Mongolia Myanmar Philippines Singapore Taiwan Thailand Timor-Leste VietnamFigure 4.D: Respondents with Experience in East AsiaTable 4.3: Respondents with Experience in Study Area Other Other Non- Anything in Tolled Transport Infrastructure Infrastructure this Highways Projects Projects Projects CountryCambodia 1 18 9 5 20 China 38 49 29 25 69Indonesia 16 30 9 10 43 Laos 2 14 8 7 20 Malaysia 19 31 12 13 42Myanmar 0 4 1 0 5Philippines 19 30 13 13 44 Thailand 22 39 18 12 50 Vietnam 8 21 8 14 36DissFinal Page 61 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.5 Tollway AppraisalFigure 4.E shows respondents’ attitudes to macro-level risks (5 signifying critical and 1unimportant18; mean being the red bar and standard deviation the black line). Findingsare broadly consistent with literature, with sovereign and institutional risks (politicaland legal) predominating, followed by economic factors which, as shown previouslydrive much of a project’s likely success (e.g. Sections 2.10.1 and 3.4). Incomeinequality and toll familiarity were not deemed important. Overall there was neutralopinion towards corruption, currency risks, price inflation and interest rates. The lattertwo possibly due to adaptive expectations from recent lows in both; business cycleswere also deemed unimportant. The literature review found very little written transportliterature concerning business cycles; though business cycle economists (e.g. Faber,2002) cite transport infrastructure as integral to cycles.Table 4.4 shows rankings by respondent groups (as defined in Table 4.1). There is notmuch difference between groups, though transport planners are less concerned aboutcorruption than other groups; possibly as neither parties to the concession proper nor toconstruction, it affects them less. Figure 4.E showed a relatively large variance forcorruption; based on “mean plus one standard deviation”, corruption ranks third overall.Perceptions of project-level risks are shown in Figure 4.F and Table 4.5. Whilst legal/contractual foundations generally score highly, there is greater difference of opinionbetween groups. Financial/ legal/ operators are relatively more concerned withminimum income guarantees and toll affordability; and less concerned withconstruction time, construction and running costs, relative to most other groups; i.e.18 Values transposed from raw data in Appendix 15.DissFinal Page 62 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)they are more sensitive to revenues relative to costs than others. Ramp-up is rankedbottom by all, despite its potential impact on early project cashflow. 5 4 3 2 1 0 th ty li ty tion ts t ion m i sks m es s iari row rofi ycl e yste yst e Rat qua rup nfla cy R m il gP ic G sC al S al S n)E rest Cor ce I l Fa at in ren ines nom i tic Leg e (I Inte Pri Cur Tol atri B us Pol om Eco Rep IncFigure 4.E: Attitudes to Macro-Level RisksTable 4.4: Rankings of Macro-Level Risks by Respondent Category Financial, Transport Legal, Planner, Engineer, Government, All Operators Economist Architect Aid Agency Academic Other Political System 1 1 1 1 1 1 1 Legal System 2 2 2 2 2 2 3 Economic Growth 3 5 3 6 3 5 2 Corruption 4 3 5 3 4 4 4Repatriating Profits 5 4 4 5 5 3 6 Currency Risks 6 6 7 4 8 7 5 Price Inflation 7 8 7 8 5 9 6 Interest Rates 8 6 6 7 7 6 9 Business Cycles 9 9 9 9 11 10 8 Toll Familiarity 10 10 10 10 9 8 10Income (In)Equality 11 11 11 11 10 11 11DissFinal Page 63 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 5 4 3 2 1 0 les) ) p ns tees i tes sts me st f it s cles e ility t es g th a nx pu kag a tio n co ro u e co eh ic n ti rou en e ran l en Ra m ehi ea b Gu l lea und ct io an c ct io er v ing ic b ge v gu a ting sio n orc Tol l fo st ru pet ten st ru no m (o th (lar enf nec me ces tua Co m a in Co n Co n nco Co n Co n ase eco lity lity rac &m mi cre a bi a bi ia l/ ont i mu l in t in g ford ford So c c al/ Tol Min era l af l af Leg Op Tol TolFigure 4.F: Attitudes to Project-Level RisksTable 4.5: Rankings of Project-Level Risks by Respondent Category Financial, Transport Legal, Planner, Engineer, Government, All Operators Economist Architect Aid Agency Academic Other Legal/ contractual foundations 1 2 1 1 3 5 5 Construction cost 2 3 4 2 2 2 2 Competing routes 3 3 2 2 8 1 8 Toll increase enforceability 4 1 3 5 4 3 8 Social/ economic benefits 5 14 7 4 1 10 5 Construction time 6 7 5 7 6 4 1 Concession length 7 11 6 11 5 5 4 Operating & maintenance costs 8 12 11 9 7 13 3Toll affordability (large vehicles) 9 7 9 12 12 12 11 Connecting routes 10 7 8 14 9 5 13Toll affordability (other vehicles) 11 5 10 10 13 8 14 Guanxi 12 13 13 8 10 9 10 Minimum income guarantees 13 6 12 12 11 14 7 Toll leakage 14 10 14 6 14 11 12 Ramp up 15 15 15 15 15 15 15DissFinal Page 64 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.6 Transport Modelling IssuesConfined to those with modelling experience, Figures 4.G shows respondents’experience of data availability, for model calibration and forecasting, with 5 signifyingalways and 1 never19. This shows little difference between calibration and forecast dataavailability and reliability, that reliability is typically slightly worse than availability,but that data are generally more available and reliable in developed countries, as wouldbe expected. 5.00 4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 Developed Developed Developing Developing countries; countries; countries; countries; sufficient data reliable data sufficient data reliable data Calibration ForecastFigure 4.G: Data Availability and Reliability19 Values transposed from raw data in Appendix 15.DissFinal Page 65 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Figure 4.H presents attitudes to different transport model types (defined in Section 2.9),with 5 representing Strongly Agree and 1 Strongly Disagree 20. Four stage andassignment models are both seen as slightly reliable, with spreadsheets marginally lessso. On balance no model type is seen as too data hungry, simplistic or complicated (bydegree). Interestingly, four stage models are seen as least inappropriate for tollways(contrasting with Willumsen and Russell, 1998), though they are perceived as “blackboxes” (echoing much literature). There is little difference in perceived suitability fordeveloping economies. 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Are reliable Too data hungry Too simplistic Too complicated Not suitable for tollways Not suitable for developing economies Too much of a black box Cannot provide meaningful outputs Four Stage Assignment SpreadsheetFigure 4.H: Attitudes to Transport Model Types20 Values transposed from raw data in Appendix 15.DissFinal Page 66 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.7 Forecast Performance and Evaluation CriteriaFigure 4.I presents perceptions of forecast performance. It appears fairly rare for outturntraffic to significantly exceed forecasts. All groups experienced significantoverforecasting (consistent with literature); whilst Financial/ Legal/ Operators havestrongest experience of this, Transport Planners/ Economists cite the next strongestexperience of overforecasting.It is acknowledged that clients can pressure transport consultants (as per Brinkman,2003), yet there is only weak acceptance of forecasts being different between equity anddebt perspectives. This is surprising; one pilot respondent noted (by follow-up email), itwould be “utterly wrong" if equity- and debt-side forecasts were the same, given thedifferent risk/ reward profiles of either side.This raises concern as to practitioners’ and users’ understanding of forecasting. It mayreinforce Brinkman’s (2003) assertion of forecasters being self-deceived as to thesupposed inscrutable neutrality of their models; and to systematic forecast errorsobserved by Bain and Polakovic (2005).Figure 4.J shows how often respondents’ consider various forecast outputs and otherfactors when appraising projects. Case study congestion is most often considered, thenbase/ central case traffic and revenue, then conservative forecasts, followed bycongestion on competing then feeder routes. Conservative forecasts are used more oftenthan optimistic ones (perhaps allaying some of the concerns regarding differencesbetween equity- and debt-side forecasts).Figure 4.K shows how often respondents’ consider various aspects of a project. NPVand FIRR are most commonly used, suggesting the primacy of financial returns overDissFinal Page 67 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)social returns. Despite the high ranking elsewhere of political and legal risks (see 4.5),sovereign/ institutional risks and counterparty risks are considered relatively inoften(possibly because pre-screening filters such risks). Portfolio correlation is consideredleast often. Some respondents also cited (by text entry) use of the Debt ServiceCoverage Ratio (both average and minimum), Payback Period and ROCE. 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 How often do projects significantly exceed forecast traffic/ revenue levels? How often do projects fall well short of forecast traffic/ revenue levels? How often do clients pressure transport planners to adjust forecasts? Are forecasts higher if for equity- rather than debt-side clients? Complete Sample Financial, Legal, Operator Transport Planner, Economist Engineer, Architect Government, Aid Agency Academic OtherFigure 4.I: Perceptions of Forecast PerformanceDissFinal Page 68 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 0.0 Never 1.0 Rarely 2.0 Sometimes 3.0 Usually 4.0 Always 5.0 Congestion on the Highway Studied Base/ Central Traffic Forecasts Base/ Central Revenue Forecasts Conservative/ Low Traffic Forecasts Congestion on Competing Routes Congestion on Feeder Routes Conservative/ Low Revenue Forecasts Optimistic/ High Traffic Forecasts Optimistic/ High Revenue ForecastsFigure 4.J: Which Forecast Outputs are Considered? 5.0 Always Usually 4.0 3.0S ometimes Rarely 2.0 Never 1.0 0.0 Net Present Financial Economic Social Cost/ Risk Counterparty Sovereign/ Value (NPV)Internal RateInternal Rate Benefit correlation risks Institutional of Return of Return Ratios versus other other (FIRR) (EIRR) projects in country/ portfolio legal risks NeverFigure 4.K: How Often Are Which Criteria Considered?DissFinal Page 69 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.8 Countries’ OutlooksRespondents’ expectations of the toll road market presented in Figure 4.L broadlyconcur with the categorisation of countries presented in Section 3.4 (countries colouredaccording to those categories), though Indonesia’s market is perceived as less developedthan that of the Philippines. Malaysia, China and Thailand have the most developedmarkets. Cambodia, Laos and Myanmar are perceived as being almost nascent onaverage (though with relatively large standard deviations, as shown by the black lines).Markets in Philippines and Vietnam, along with Indonesia are seen as nascent-to-developing.Whilst Figure 4.M indicates slight positive perceptions towards countries respondentshave worked in, showing a general positive bias by those with country experience.Differences are typically small (Indonesia and Malaysia having the largest); althoughCambodia and Myanmar are rated as nascent and Indonesia overtakes the Philippinesbeing rated developing-to-steady, by those with respective country experience.By respondent groups (Figure 4.N), there is usually little difference in perceptions.Notable exceptions being Academics with relatively positive views of Indonesia,Malaysia, Thailand and Vietnam and more bearish assessment of China, Laos andparticularly the Philippines. “Others” tend to be more bearish. Financial/ Legal/Operators are bearish relative to most others on Cambodia, Laos and Myanmar(consistent with their categorisation in Section 4.3), plus the Philippines and Thailand(possibly due to recent political problems in the Philippines and a Thai coup d’etatimmediately prior to the survey). Yet they are more bullish on Indonesia and Malaysia.DissFinal Page 70 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 6.0 O ver- Developed 5.5 5.0 Maturing 4.5 4.0 Steady 3.5 3.0 Developing 2.5 2.0 Nascent 1.5 1.0 No Market Malaysia China Thailand Philippines Indonesia Vietnam Cambodia Myanmar LaosFigure 4.L: Perceived Tollway Market Opportunities by Country No Market 1.0 Nascent 2.0 Developing 3.0 Steady 4.0 Maturing Over- 6.0 5.0 Developed Cambodia China Indonesia Laos Malaysia Myanmar Philippines Thailand Vietnam Full Sample Those With Country ExperienceFigure 4.M: Impact of Experience on Country PerceptionsDissFinal Page 71 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) No Market 1.0 1.5 Nascent 2.0 2.5 Developing 3.0 3.5 S teady 4.0 Over-Developed 4.5 5.0 Cambodia China Indonesia Laos Malaysia Myanmar Philippines Thailand Vietnam Complete Sample Financial, Legal, Operator Transport Planner, Economist Engineer, Architect Government, Aid Agency Academic OtherFigure 4.N: Country Perceptions by Respondent CategoryDissFinal Page 72 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.9 Economic OutlookFigure 4.O shows overall respondents anticipate substantially higher fuel prices in thefuture, as well as increased tolling acceptability and general price inflation. Interestrates, economic growth and exchange rate volatility are also expected to increase.Figure 4.P illustrates that there are no major differences of perception betweenrespondent groups. Perceptions are largely consistent with the economic outlook positedby the K-Wave as set out in Section 3.7, even if the perceived impacts of rising interestrates and price inflation are not deemed significant (Section 4.5) 5.0 S ignificant Increase 4.5 Increase to4.0 an Extent 3.5 3.0 No Change 2.5 Decrease to an 2.0 Extent 1.5 S ignificant Decrease 1.0 Fuel prices General price Interest rates Economic Exchange rate Tolling inflation growth volatility AcceptabilityFigure 4.O: Economic ExpectationsDissFinal Page 73 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Significant Decrease to Increase to Significant Decrease an Extent No Change an Extent Increase 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Fuel prices General price inflation Interest rates Economic growth Exchange rate volatility Tolling Acceptability Complete Sample Financial, Legal, Operator Transport Planner, Economist Engineer, Architect Government, Aid Agency Academic OtherFigure 4.P: Economic Expectations by Respondent GroupDissFinal Page 74 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)4.10 Other CommentsA variety of text comments were also received. With regards transport modelling, thekey theme was of tailoring models to local conditions and of the potential validity of allapproaches, subject to circumstances (e.g. data availability, timescale, phase of projectlife-cycle, client requirements, etc.)In all 71 respondents requested information on findings (23 on the survey, 48 onbroader research), perhaps intimating that this research is of perceived importance.4.11 Key Conclusions from the Questionnaire SurveyThere was a significant response rate from transport planners and economists, with alower number of responses from other groups. Nevertheless, it was deemed that therewere sufficient data to analyse different stakeholder perceptions (using groupings inTable 4.1). With mean working experience of 20.6 years, the sample has substantialexperience; the “average” respondent has just over one year’s experience in rural tolledhighways in developing countries. The sample is thus deemed sufficient for thepurposes of this Dissertation.There is perceived primacy of legal and political factors on viability; though oncemodelling commences, economic factors predominate. Business cycles, toll familiarityand income inequality are deemed slightly unimportant.Data quality and availability is deemed better in developed economies, as expected.There is no strong preference between four-stage, assignment and spreadsheet models.Rather each model should be tailored for specific conditions. Four-stage models areperceived as fairly reliable, but also as opaque.DissFinal Page 75 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Whilst under-forecasting appears relatively rare, over-forecasting happens much moreoften. There is some acknowledgement of transport planners adjusting forecasts to meetclients’ expectations. There appears a fundamental misunderstanding of the purposes ofequity- and debt-side forecasts; this based on only weak acceptance of differencesbetween forecasts for either side (as the author suspected a priori).NPV is the most often-used evaluation criterion, followed by FIRR, then economicmetrics. Counterparty risks and risk correlation versus other projects are used morerarely.Country categorisation in Section 3.4 is broadly supported, but with Indonesia seen asless advanced than posited in Section 3.4. On average, Malaysia is seen as steady-to-maturing; Thailand and China as developing-to-steady; Philippines, Indonesia andVietnam as nascent-to-developing; and Cambodia, Myanmar and Laos as sub-nascent.However, those with Indonesia experience rate the country as developing-to-steady; andthose with local experience regard Cambodia and Myanmar as nascent.There is a reasonable acceptance of symptoms of a K-Wave upswing, in terms ofincreasing price inflation (especially fuel prices), interest rates and to a lesser extent,economic growth. Tolling acceptability is predicted to increase. However, respondentsdid not deem the impacts of rising interest rates and price inflation to be significant.DissFinal Page 76 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)5. Risk Simulation Modelling5.1 IntroductionChapter 4 presented inter alia perceptions of different risks, as well as respondents’expectations of economic parameters. For example, on average respondents believedinflation and interest rates would increase, but would have little impact on project risk.Monte Carlo simulation is used to quantitatively estimate the relative importance ofdifferent risks, to test whether respondents might have underrated such risks. Althougheach project has specific locational and institutional risks, such are excluded herethrough use of a simplified, fictional case; the aim being to concentrate on the relativeimportance of broad risks irrespective of particular locational context.Three economic simulation scenarios are defined as follows: “Conventional Case” of interest rates and price inflation similar to recent values; “Respondents’ Case” based on questionnaire results (see 4.9) with increased fuel prices and some increase in general price inflation and interest rates; and, “Kondratieff Case” assuming an upswing with more substantial increases in price inflation, interest rates and also increased economic growth.DissFinal Page 77 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)5.2 The Case Study and Its ParameterisationThe case study is necessarily contrived to give a distribution of outcomes suitable foranalysing the relative importance of different risk elements; and in particular to test thehypothesis of the impact of increasing price inflation and interest rates on project risk.As far as practicable, parameters were taken from public sources (and whereappropriate, influenced by questionnaire responses); inevitably some use had to bemade of confidential sources. Finally, values were adjusted (primarily base traveldemand) to guarantee the distribution of financial outcomes outlined above, togetherwith a “base case” showing FIRR≈16% (see 5.4), corresponding to a typically requiredinvestment threshold (see 2.4.1).The case study network topology is shown in Figure 5.A, constituting six zones (for triporigins and destinations) and eight links, including the fictional tolled highway. Thelengths and freeflow speeds of each link are shown in Table 5.1. Though the number oflanes on some “local roads” may seem high, they proxy for multiple alternative routes.Assumed trip distribution is shown in Table 5.2 (in terms of trip total percentages oneach origin-destination movement); shaded cells correspond to movements that couldpotentially use the tollway. Both the total number of trips and road capacities areincluded amongst the simulation variables; all of which are shown in Appendix 16. 27parameters were common for all three Cases (Conventional, Respondents’ andKondratieff). A further 6 parameters had values specified for each Case differently,though for each Monte Carlo iteration, the values were inter-related.DissFinal Page 78 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Zone 1 Zone 6A G H Start of End of Highway HighwayB F Zone 2 Zone 3 Zone 4 Zone 5 C D EFigure 5.A: Case Study Notional NetworkTable 5.1: Basic Link Characteristics of Case Study Network Road Freeflow Speed (kph) Section Length Road Lanes per Small Large (“Link”) (km) Standard Direction Vehicles Vehicles A 10 Local Road 3 70 60 B 3 Local Road 4 70 60 C 15 Local Road 3 70 60 D 25 Local Road 3 70 60 E 10 Local Road 2 70 60 F 2 Local Road 4 70 60 G 10 Local Road 3 70 60 H 40 Tollway 2 120 100Table 5.2: Assumed Trip Distribution (% by O-D Pair) To Zone 1 2 3 4 5 6 Total 1 3% 2% 4% 4% 5% 18% 2 3% 5% 3% 3% 3% 17% From Zone 3 2% 5% 5% 3% 3% 18% 4 4% 3% 5% 3% 2% 17% 5 4% 3% 3% 3% 2% 15% 6 5% 3% 3% 2% 2% 15% Total 18% 17% 18% 17% 15% 15% 100%DissFinal Page 79 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Sections 2.10 and 2.11 gave evidence of serially overoptimistic errors in tollwayappraisal and transport models, supported by questionnaire findings (Section 4.7) withoutturn parameters more often being low than high. Thus it was not felt reasonable toadopt symmetric probability distributions for all parameters either side of a notional“mean”, “median” or “base” value in many instances. Rather a modal average wasspecified, together with a probability of below-modal value21; a different standarddeviation either side of this modal value was also applied in many cases. Finally, topreclude unrealistic outliers, a minimum and maximum was specified in each case;usually being twice the standard deviation.Usually transport models specify different parameters for different years over theforecast horizon (e.g. gradually declining economic growth rates); such detail wasdeemed superfluous for this exercise. It might be argued that a key risk of the K-Waveupswing pertains to those seeking downstream refinancing (typically ever pricier asopposed to cheaper during a downswing); however, the impacts of inter alia differentinterest rates are tested across the three scenarios and via simulation.Thus a single economic growth rate, together with a single elasticity across time in eachcase (though specified separately for small and large vehicles) was adopted. The use ofMonte Carlo techniques ought anyway to proxy for such uncertainty. Moreover, itpermits the analysis of economic growth,  T ,  D or VOT per se, which would not be y p yreadily feasible if such parameters changed across the forecast horizon. The relativeimportance of unforeseen changes in these parameters can be gauged to an extent fromanalysing changes in project value due to changes in these parameters. More detailed21 And by implication probability of above-modal values.DissFinal Page 80 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)analysis might be appropriate for a specific case study, wherein parameter sets tailoredto specific local conditions would be used in lieu of generic values and ranges, such asthose befitting and employed by this Dissertation. Such simplification also permits theconsideration of a wider range of forecast parameters. Furthermore, in the case ofinterest rates, it is reasonable to assume prospective concessionaires would size andacquire debt based on current interest rates (e.g. through issuance of bonds) and thatchanges to interest rates will primarily affect bridging loans or overdrafts requireddownstream (i.e. unbudgeted when deciding whether to proceed and on financestructuring).The Respondents’ and Kondratieff Cases adjusted Conventional Case values, withKondratieff Case and price inflation and interest rates greater than or equal toRespondents’ Case and these at least as great as Conventional values. Vehicle OperatingCosts in Kondratieff and Respondents’ cases were the same, as respondents largelypredict significantly higher fuel prices, in line with Kondratieff-based forecasts. Allrandom parameters are specified in Appendix 16.Fixed parameters are summarised in Appendix 17. The modelled concession length was30 years (modelled as 120 quarters), including construction time.DissFinal Page 81 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)5.3 Methodology10,000 iterations of Monte Carlo simulation were employed, using the followingmethodology:5.3.1 Defining Random ParametersAs stated above, one set of parameters were determined to be applied across all years,i.e. economic growth, price inflation rates, interest rates and elasticities were assumedconstant across all years; though different between Conventional, Respondents’ andKondratieff cases and with different values for each iteration. Distributions used areshown in Appendix 16.5.3.2 Applying Parameters to Derive Variables for Each QuarterQuarterly values of all cost indices, as well as value of time and trip matrix (demand)size were defined, based on progressive growthing in line with inflation from initial(“Quarter 0”) values through to Quarter 120. The parameters and equations used aregiven in Appendix 18. In the case of toll rates where increases were not uniform, butrather at certain intervals, assumed toll rates were kept fixed, being updated to thecorrect theoretical toll rate every x intervals (x= number of quarters between increases).5.3.3 Traffic Assignment for Each QuarterFor each quarter, two-class assignment (small and large vehicles) was performed usinga 10-iteration incremental loading (to take account of congestion) and logit equations toapportion loads on each iteration between expressway-using and non-expressway paths.Speeds were initially set to freeflow values. Generalised costs were then determinedbased on these speeds and on each iteration 1/10 of the matrices were assigned, beingsplit between expressway-using and non-expressway routes using a logit relationship,DissFinal Page 82 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)given in Appendix 18. Link loadings were then increased as appropriate based on thesplit of this traffic (10% of the matrix further split by the logit curve). Speeds wererevised based on the new loadings, using volume/capacity-to-speed relationships givenin Figure 5.B. The process was repeated until 10 iterations had been completed. 120 100 80 Speed (kph) 60 40 20 0 0 0.5 1 1.5 2 2.5 Volume/Capacity Ratio Tollway (Small Vehicles) Tollway (Large Vehicles) Local Road (Small Vehicles) Local Road (Large Vehicles)Figure 5.B: Volume/Capacity-to-Speed Relationships5.3.4 Financial AnalysisHaving obtained loadings of small and large vehicles on the expressway link for eachquarter, financial analysis followed.For any quarters prior to completion of the expressway, flows were set to zero, and theappropriate quarterly construction costs were accrued. For subsequent quartersfollowing opening (where revenues were generated), ramp-up was applied; which wasassumed to be linear from its first to final quarter. For later quarters, any flows over theexpressway’s capacity were capped-off, in line with industry standard practice andDissFinal Page 83 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)initial revenues were calculated. Variable operations & maintenance costs weresubtracted from these on a percentage basis as were losses from toll leakage. This gavea net revenue from which fixed operations & maintenance costs were subtracted.Any “surplus” revenues were used first to pay-off extra debts incurred, first paying offinterest and then principal. Any remaining surplus paid-off initial debts and interest.Any residual revenues (once all debts paid off) were taken as positive cashflow. For anyquarter without positive cashflow, additional interest payments and debt requirements(at the extra debt rate) were calculated and subtracted from the financial position22.Based on the resultant cashflow profile, financial analyses were performed, comprisingFIRR, payback period and NPV at various interest rates. For purposes of comparisonbetween cases and iterations, FIRR was used. Also, any run where FIRR≤0% or therewas no payback within 120 quarters was deemed to constitute “financial failure” (i.e.bankruptcy). Comparative probabilities of “failure” were also used to compare betweenruns (see 5.5).5.4 Comparison of Cases under “Base Run”As stated in 5.2, an initial “base” run was undertaken using modal values for eachparameter usually randomised. The objective being to ensure a realistic return on thebase scheme and to provide an ample spread of performance (i.e. a meaningful but notoverwhelming prevalence of “failure”); also to enable an initial comparison between thethree cases.22 Initially, debt was sized at 10% more than the envisaged construction cost providing a small bufferagainst interest rates.DissFinal Page 84 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)The results are shown in Table 5.3. This suggests that despite delayed Payback (due tohigher price inflation and interest rates) the Kondratieff case may give superior returnsthan the Conventional case, but that the Respondents’ case would yield the best returns;perhaps indicating more optimism amongst practitioners (transport planners being thelargest respondent group, as per Table 4.1). However, the risky nature of forecastingmeans that simulation results should also be investigated.Table 5.3: Comparison of “Base” Runs between Cases Case Conventional Respondents’ Kondratieff FIRR 16.83% 17.88% 16.95%Payback Period (years) 10.728 10.676 12.090 NPV (at 16%) $17,910,017 $45,944,246 $27,524,7255.5 Comparison of Simulation Results between CasesThough the results in 5.4 suggest that the Kondratieff case might be more beneficial toinvestors than the Conventional case (based on recent past experience), does thistranslate into less risk? Equally, is the apparent optimism of the Respondents’ caseconsistent over risk-testing also? Do the higher price inflation and interest rates inherentin the Kondratieff case (and to a lesser extent in the Respondents’ case) increaseriskiness when tested using Monte Carlo risk simulation techniques?Summary results from the 10,000 simulations for each case are shown in Table 5.4, withcumulative probability distributions of FIRR, payback and NPV (at 16%) shown inFigure 5.C, 5.D and 5.E respectively. Comparing Table 5.4 against Table 5.3, meanFIRR’s are greater, yet payback periods are longer except in the Kondratieff case(though this excludes 13 instances where there is no payback within 30 years). For bothFIRR and payback, the general pattern of Respondents’ Case being the most optimisticDissFinal Page 85 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)and the Conventional Case the most pessimistic holds. At a 16% discount rate,Respondents’ Case NPV is similar to the “base” run, but mean NPV is substantiallyhigher in the Conventional Case, though still less than in the Respondents’ Case.However, mean Kondratieff NPV is actually negative, despite mean FIRR of 17.62%;Figure 5.E shows that average NPV is lowered due to a significant number of largenegative NPV’s. In all cases, Kondratieff standard deviations are the greatest andConventional standard deviations the smallest. Furthermore, 12.5% of Kondratieff runsresulted in “failure” (i.e. negative FIRR or no payback); substantially greater than 1.1%of Respondents’ runs and 0.6% of Conventional runs. This suggests immediately thatnotwithstanding its superior mean values, the Respondents’ case is riskier than theConventional case; however, the Kondratieff case is substantially riskier still. The nextsection analyses the impacts of different risk elements, underlying these results.Table 5.4: Summary Results from Simulation Runs Case Metric Statistic Conventional Respondents’ Kondratieff FIRR Mean 17.20% 17.99% 17.62% Minimum* 0.01% 0.49% 0.11% Maximum 28.75% 29.26% 30.14% Standard Deviation 3.74% 3.77% 4.11% Payback Mean 10.83 10.85 11.60 Period Minimum 6.44 6.31 6.53 (years)† Maximum 29.98 29.23 29.66 Standard Deviation 2.32 2.41 2.73 NPV Mean $28.1 $46.4 -$37.7 at 16% Minimum -$377.8 -$936.0 -$5,600.3 ($ million) Maximum $296.6 $357.3 $465.0 Standard Deviation $80.3 $96.3 $308.9 Financial # of Cases 55 114 1250 Failure % of Cases 0.6% 1.1% 12.5%Note: * FIRR’s were not calculable once beneath 0%; hence 0% minimum value in all cases. † Excludes 13 instances under Kondratieff case where no payback obtained.DissFinal Page 86 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 100% 80% Cumulative % 60% 40% 20% 0% 5% 10% 15% 20% 25% Conventional Respondents KondratieffFigure 5.C: Cumulative Probability Distribution of FIRR (excluding FIRR<0%) 100% 80% Cumulative % 60% 40% 20% 0% 8 10 12 14 16 18 20 22 Conventional Respondents KondratieffFigure 5.D: Cumulative Probability Distribution of Payback Period (years) 100% 80% Cumulative % 60% 40% 20% 0% -500 -400 -300 -200 -100 0 100 200 300 400 Conventional Respondents KondratieffFigure 5.E: Cumulative Probability Distribution of NPV at 16% ($m)DissFinal Page 87 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)5.6 Analysis of Individual RisksAppendix 19 shows the impact of different simulation variables on FIRR and theprobability of financial failure. As some variables have a large but poorly correlatedeffect on FIRR, whilst others have a smaller but better correlated effect, Appendix 20takes the data from Appendix 19 and assesses importance as follows: The range (maximum less minimum) is calculated The range is multiplied by the linear regression equation’s coefficient 23 to determine the impact on FIRR; an absolute value is taken The impact is multiplied by the linear regression equation’s R2 to weight impact by strength-of-relationshipThe simulation variables were then grouped by category, so as to determine the relativeimportance of such risk categories, so as to avoid distortions due to the number ofvariables tested within each category. The categories were then ranked for each of thethree cases, as shown in Table 5.5. The Case Study appears to give very low importanceto the Value of Time, but this is likely a consequence of the nature of networkmodelled. What is more critical in the context of this Dissertation is the change inimpacts and rankings between cases. Excepting Vehicle Operating Costs and TollLeakage (the latter in the Respondents’ case), all parameters increase their impact onFIRR between Conventional and Respondents’ and Respondents’ and Kondratieffcases, indicating increased forecast risk volatility overall.23 i.e. coefficient rather than constant from linear regression in Appendix 19. Goodness-of-fit betweenlinear and log-linear equations was very similar, so for simplicity the linear regressions were used here.DissFinal Page 88 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Table 5.5: Rankings of Risk Categories’ Importance by Case Conventional Respondents Kondratieff Risk Group Impact*R2 Ranking Impact*R2 Ranking Impact*R2 Ranking Road Capacities 0.30% 11 0.53% 10 1.35% 9 Construction Cost & 5.68% 4 6.40% 4 8.75% 5 Duration All O&M Costs 0.59% 9 1.02% 8 1.98% 8 Value of Time & Its 0.00% 13 0.03% 13 0.05% 13 Income ElasticityVehicle Operating Costs 0.30% 10 0.19% 11 0.27% 12 Demand (Initial & 12.13% 1 12.69% 2 16.05% 2 Income Elasticity) Toll Revenue Leakage 0.95% 8 0.94% 9 1.20% 10Ramp-Up: Amplitude & 2.14% 6 2.30% 7 3.66% 7 DurationLogit Model Parameters 0.09% 12 0.12% 12 0.32% 11Toll Escalation Rate and 1.73% 7 2.56% 6 4.69% 6 Frequency GDP Growth 10.84% 2 12.62% 3 15.01% 3 Price Inflation 5.12% 5 5.91% 5 11.67% 4 Interest Rates 8.78% 3 15.73% 1 47.38% 1Interest rates increase markedly in importance in both Respondents’ and Kondratieffcases; this may be slightly overstated as initial interest rates feed into interest rates onany extra (subsequent) borrowings. However, in the Kondratieff case both sets ofinterest rates would individually outrank all other categories; illustrating the exponentialincrease in their impact as they rise, thus signifying that interest rates increase markedlyin importance in times of high (or increasing) interest rates. Demand ranks as mostimportant in the Conventional case and remains second only to interest rates in the othercases, followed by GDP growth (which itself permeates many other parameters asexplained in Sections 2.10.1 and 3.4). Price inflation and construction costs/ durationare 4th and 5th most important (precise ranking case-dependent). The importance of tollescalation rates and frequency of increases also increases in the Respondents’ andespecially the Kondratieff case, as might be expected: with price inflation increasing,the impact of delayed or incomplete adjustments increases. Indeed the impact of priceinflation does not appear simple. General price inflation accounts for most price impact,DissFinal Page 89 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)but is positively correlated with outturn performance; likely because it decreases the realvalue of initial debt and is abated to an extent by toll increases, so downside risksassociated with increased price inflation are statistically associated more strongly withtoll escalation-associated variables.Such inter-relationships between simulation system variables often occur; isolation ofindividual variables’ impacts is not always possible (Pindyck and Rubinfeld, 1981).With regards the Hypothesis, this suggests the impacts of increasing price inflation andinterest rates on various project risk elements are not always wholly linear; rather, morecomplex system-wide interactions are possible.DissFinal Page 90 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)5.7 Discussion of ResultsIn terms of the Hypothesis, the importance and riskiness of interest rates increasemarkedly as they increase. In the Case Study in both Respondents’ and Kondratieffcases interest rates become the most important determinant of project FIRR. The extentof this effect might be exaggerated due to the interrelation between “current” and“downstream” interest rates and by high gearing of the Case Study. However, as interestrates increase the risk-free return on cash (via bank deposits) would also increase andwith it investors’ required project returns. Hence the overall trend remains reasonable.Coupled with increasing price inflation, the impact of variance in almost all forecastvariables/ risks increases. Given the inter-relationships between forecasting parameters,impacts are not always linear; rather different parameters affect one another’s impacts(e.g. toll escalation rate’s and frequency’s importance affected by price inflation). Thissuggests that greater caution should be exercised by stakeholders when evaluatingschemes under such circumstances; and more investigation of risk be undertaken.Adopting K-Wave Theory, the increase in price inflation and moreover interest ratesshould be weighed against possible increased economic growth (also potentiallyaffecting initial demand). In fact, given that the K-Wave appears to still be in the earlystages of upswing, the impacts of increasing interest rates may not be too substantial atpresent, assuming investors can obtain fixed-rate debt (e.g. through bond issuance).However, as interest rates increase it will be ever more onerous raise extra finance.Excepting occasional short-term decreases in interest rates, it may no longer beadvisable to refinance projects down-stream; rather, sufficient fixed-rate debt should beacquired at project outset.DissFinal Page 91 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)6. Discussion and Conclusions6.1 IntroductionHaving undertaken literature review, environmental analysis, questionnaire surveys andrisk simulation modelling, this section discusses and summarises findings with a view toanswering the question posed by the Dissertation’s title and evaluating the hypothesis: What are the key risks associated with private investment in start-up toll road projects in Developing East Asian Economies?; and, There is a significant change in the nature and extent of project finance risks for private stakeholders in East Asian toll roads during a period of increasing price inflation and interest rates.In addition to reviewing evaluation criteria (discussed in 6.2), the literature reviewidentified three broad risk categories, each of which are discussed in turn,: Macro-economic risks, including regional risks and opportunities and evaluation of broader economic trends (Section 6.3); Market risks, including determination of scheme attractiveness (Section 6.4); and, Forecasting risks (Section 6.5).Section 6.6 examines the extent to which the market is anticipating change. Section 6.7makes observations and recommendations for both transport planners and projectfinance. Finally, formal evaluation of the hypothesis is performed (Section 6.8).Inevitably there is a certain degree of overlap between sections.DissFinal Page 92 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)6.2 Evaluation Criteria and Implications of the Time-Nature of RiskInvestors tend to focus on a scheme’s NPV and FIRR. The questionnaire surveyrevealed that NPV is used slightly more often than FIRR. Economic criteria are usedfairly often. Ratings agencies and lenders also consider Debt Service Coverage Ratio.The questionnaire survey found that sovereign and institutional risks, counterparty risksand risk correlation versus other portfolio projects are considered less often. However,this may be attributable to such criteria being used to “screen out” projects before fulldue diligence proceeds (bringing more appraisers, e.g. transport planners, into theprocess).The capital-intensive nature of infrastructure projects and in particular their dependenceon heavy up-front investment means that many standard financial ratios, e.g. Return onCapital Employed, Gross Profit Margin are unlikely to be that reliable in early years ofa project. Faber (2002, p.69) notes that returns are likely to be volatile in capital-hungryprojects, especially in emerging economies and “emerging companies” (as start-uptollways could be defined). Within transport planning, Willumsen and Russell (1998)showed schematically how risks are front-loaded to projects, reducing over time;corroborating the preceding statements.Given the inherent riskiness of such projects, it can be concluded that this dissertation’sinvestigation of risk is directly relevant to many aspects of the tollway industry and mayalso add value to other infrastructure investment sectors.DissFinal Page 93 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)6.3 Macro-Level Risks and OpportunitiesUse of private finance in infrastructure increased since the 1980’s in both developed anddeveloping economies; toll roads being one of the recipients of such investment,especially in developing countries. Although activity slowed in the aftermath of theAsian Financial Crisis (at least partially attributable to previous over-investment),recently it picked-up again and has begun spreading into some of the poorest countriesin the region (e.g. Cambodia; also noted by survey respondents with local experience(Figure 4.M)). This in parallel with economic recovery in recent years (Section 2.8discussed over-investment and Section 3.3 evidence of economic rebound).Developing countries are inherently riskier than developed ones, with weaker rule-of-law, increased corruption, poorer toll affordability and often more volatile economicgrowth rates, coupled with increased incidence of social and political upheaval.Questionnaire survey respondents ranked the political system, legal system, ease ofprofit repatriation, corruption and currency risks amongst the top six macro-level risks;all of which are predominantly developing country-risks.Weighed against the risks are the opportunities of investing in economies withpotentially explosive mobility growth. Based on Khan and Willumsen (1986)’sequivalencing of roadspace and vehicle ownership, regression analyses (in Section 3.4)identified potential for rapid demand growth; with the greatest ultimate growth potentialbeing amongst the poorest countries.Survey respondents rated the market in the next 10 years in Malaysian as steady-to-maturing; China and Thailand as developing-to-steady; Philippines, Indonesia andVietnam as nascent-to-developing. Cambodia, Myanmar and Laos were rated not-yet-nascent, signalling that whilst there may be significant percentage growth in vehicleDissFinal Page 94 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)ownership, tolling affordability and absolute market size may be insufficient to providea viable near-term tollroad market; though those with respective local experience ratedCambodia and Myanmar as nascent and Indonesia as developing-to-steady (Figure4.M). This suggests Malaysian investments should be seen as the least risky(commanding a smaller CAPM risk premium), with those in Cambodia, Myanmar andLaos as the riskiest (requiring a higher forecast return to proceed).Whilst interest rates do not feature greatly in transport planning literature, they are veryimportant in project finance: with investment risk increasing substantially as interestrates escalate; this corroborated by the risk simulation modelling (see 5.6 in particular).Price inflation can affect both construction and operating/ maintenance costs and theimpacts of delayed toll escalation or toll increases at under the rate of price inflation(discussed in more detail in Section 6.4). Neither price inflation nor interest rates wereseen as especially important in project risk analysis by most questionnaire respondents.However, economic growth was recognised as very important to project performance(ranked by survey respondents behind only political and legal systems). Economicgrowth feeds through many aspects of market and forecasting risks (both discussedbelow). Survey respondents expected increasing price inflation and interest rates (andespecially fuel prices), yet their importance was not rated that highly.To test these expectations, risk simulation modelling was undertaken based on threeeconomic scenarios. The first (“Conventional Case”) assumed similar trends to thoseexperienced in recent years; the second (“Respondents’ Case”) incorporatedrespondents’ expectations of higher fuel prices and slightly higher interest rates andeconomic growth; the third case (“Kondratieff Case”) was based on an upswing in theK-Wave (Kondratieff, 1926), resulting in markedly higher general price inflation andDissFinal Page 95 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)interest rates, as well as higher economic growth. Based on FIRR, the Respondents’Case generally gave the most optimistic returns, with mean Kondratieff Case returnsalso higher than those from the Conventional Case. However, 12.5% of KondratieffCase runs resulted in project failure (i.e. no payback and/or negative FIRR), versus0.6% in the Conventional Case and 1.1% in the Respondents’ Case.Given that economic growth correlates positively with FIRR, the impacts of increasedprice inflation and interest rates, where these outstrip economic growth would appear tohave a significant negative impact on project performance. Furthermore, the apparentvolatility of the Kondratieff Case would tend to support Faber’s (2002) assertionregarding the riskiness of the K-Wave Upswing (higher mean returns, but with aninherent danger of short-term reversals which can lead to bankruptcy), based on hisanalysis of the 19th Century American railroad industry.However, it also appears that the K-Wave upswing is being facilitated by the economicemergence of East Asia and that this could drive demand for transport infrastructure.With a period of roughly half-the-length of the Kuznets Cycle elapsed since the AsianFinancial Crisis, a further driver of infrastructure growth in the region could be posited.6.4 Market RisksRigby and Penrose (2001) define project-level risks as the most critical. Though thisDissertation concentrates on demand-side risks, construction cost and delay areimportant, affecting early-year performance (Willumsen and Russell, 1998). There isevidence of serial-underestimation of these costs (Flyvberg and COWI, 2004), whichare very significant to project performance: respondents ranked construction cost assecond only to contractual foundations (Table 4.5), with construction time ranked sixthout of fifteen project-risk criteria. The risk simulation model showed construction costDissFinal Page 96 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)and time as fourth or fifth most significant risk category out of thirteen (dependent onforecast case).Questionnaire respondents ranked toll increase enforceability as fourth most important(though most important to the Financial/ Legal/ Operators group; see Table 4.5).Although the risk simulation model ranked toll increase frequency and amount as sixthor seventh most important, the impact of toll increases is predicted to increase by 48%between the Conventional and Respondents’ Cases and by 170% between Conventionaland Kondratieff Cases (see Table 5.5)24. This risk is linked to contract enforceability(institutional risk). Minimum income guarantees ranked only thirteenth overall, butfinancial/ legal/ operators ranked them sixth; arguably because they are not of primaryconcern to those designing infrastructure (e.g. engineers) or determining demand(transport planners/ economists), whilst they are potentially critical to financiers.Survey respondents ranked competing routes as the third biggest market risk. The risksimulation did not consider competing routes (beyond an existing local road) as suchimpacts are very location-specific. It can be a contractual/ institutional issue, pertainingto the enforceability of agreements with governments to not approve competing routes.There may be correlation between the incidence of competing routes and over-investment, as witnessed prior to the AFC; meaning this risk may be partially cyclical,related to business confidence (and expectations of surplus demand requiring additionalroutes). Conversely, an absence of good connecting routes can subtract from projectperformance. Questionnaire respondents asserted that they usually consider congestionlevels on both their study routes and competing and feeder routes (Figure 4.J).24 Increases quantified by the percentage difference in “Impact*R2” between cases in Table 5.5.DissFinal Page 97 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Whilst survey respondents did not rank toll affordability highly amongst market risks, itlikely underpinned the low ratings of poor countries’ tollway market prospects(discussed in 6.3). A number of other market risks correlate with institutional risks (e.g.minimum income guarantees, importance of guanxi, “pork barrelling”)6.5 Forecasting RisksStart-up tollways do not have business track records for analysis. Every aspect of theirperformance has to be predicted (cost and revenue). However, transport plannersgenerally over-forecast demand and revenue (Bain and Polakovic, 2005); and the moreuncertain the environment the poorer the forecasting record (Bain and Wilkins, 2002).Questionnaire respondents broadly concurred with these assertions (Figure 4.I). Theyalso deemed the availability and reliability of data poorer in developing countries. Suchenvironments are typically more economically volatile, further compounding risk.Those with experience of using or developing transport models did not hold any modelform as significantly inherently better or worse than others (Figure 4.H). Whilst it wasacknowledged that clients sometimes pressure transport planners to manipulateforecasts (Figure 4.I; corroborating Brinkman, 2003), there was an ambivalent attitudeto whether equity-side forecasts should be higher than debt-side forecasts, suggestingmany forecasters do not understand the requirements of different financial perspectives.Economic growth underpins demand forecasting parameters. In addition to uncertaintyregarding economic forecasts, both price sensitivities (e.g. Value of Time (VOT)) andincome elasticities of traffic growth and of VOT are rarely known. Even wherehistorical data are available, given S-curve relationships (Sections 2.7, 2.9 and 3.4) andthe impact of locational specifics on any given project, uncertainty is inherent over-and-above that concerning economic growth per se.DissFinal Page 98 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Whilst Section 3.4 postulated cross-sectional income-car ownership relationships, inthemselves they are not sufficient to derive risk-free time-series sensitivities; Section3.5 highlighted significant time-series differences within a single country (China). Gunnand Sheldon (2001) advocate income elasticity of VOT in the range 0.35-0.70; despitethe implications of adopting 0.35 versus 0.70, there is not strong consensus as to whatvalues to use. There is even evidence of zero-VOT in some instances (ADB, 2003).More broadly, there is often a general bias against paying tolls (Richardson, 2004).The risk simulation model also illustrated the importance of GDP growth rates (Table5.5), with the demand-level (itself driven by GDP) ranking first or second mostsignificant on financial outcome of the case study.For new roads, induced traffic may result and this may be substantial, boosting localeconomic growth (Corbett et al, 2006). However, forecasting induced traffic is besetwith substantial error (Willumsen and Russell, 1998; Bain and Polakovic, 2005).Ramp-up also presents problems for forecasters. Although it may be a near-term risk, itmay affect a project’s ability to meet early debt repayments (Streeter and McManus,1999; Bain and Wilkins, 2002). However, it was ranked as the least-important marketrisk by all questionnaire respondent groups (Table 4.5).Due to practical limitations of scope, this dissertation did not investigate the directimpacts of interest rates on travel demand (compound errors through the economiclinkages determining disposable income for car purchase, discretionary travel, etcprohibited such analyses). However, should interest rates rise substantially it isreasonable to postulate reduced car purchases (often loan-financed) and travel (as anincreased portion of income is used to service existing debts).DissFinal Page 99 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)6.6 Is the Market Anticipating a Change in the Rules-of-the-Game?Economic literature and forecasts are notorious for differences of opinion. Even whenusing the same frameworks and assumptions, there can be substantial variance inforecasts. Conversely, despite the centrality of economics to demand forecasting, thereis a general lack of transport planning literature on economic development scenarios(excepting the “assumptions” sections of individual project reports). Kilsby (2006a,2006b) is an exception, positing “peak oil” driving fuel price increases25.The questionnaire survey (Figure 4.O) showed that respondents anticipate change,especially increasing fuel prices. Tolling acceptability was also expected to increase toan extent, followed by general price inflation, economic growth, interest rates andexchange rate volatility. Thus it might be argued that this is evidence of acceptance ofprinciples underlying the K-Wave and that the relatively weak acceptance of such trends(excepting fuel prices) is even consistent with the early stages of a change in directionof the K-Wave (before adaptive expectations have completely caught-up with thequalitative change of the K-Wave). However, given that the K-Wave is far fromcommon acceptance this perhaps states the case too strongly. Nevertheless, it doessuggest that some economic change is anticipated; and there was relatively littledisagreement between different stakeholders (Figure 4.P).Figure 4.E showed that whilst economic growth was seen as third-most-importantmacro-level risk, price inflation and interest rates were ranked only seventh and eighthrespectively (out of 11), though they still rated as “important.” Although the risksimulation case study is simplistic based on a shift from the “Conventional Case” to theDissFinal Page 100 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)“Respondents’ Case” (based on survey responses), it suggests that interest rates willbecome approximately 80% more important, with GDP growth and price inflationbecoming approximately 15% more significant (using “Impact*R2” in Table 5.526).However, based on the Kondratieff Case, interest rates become substantially moreimportant still and price inflation more than doubles in importance, amid morewidespread forecast volatility.6.7 What Lessons for Practitioners?Although practitioners appear to accept the likelihood of increased price inflation andinterest rates, comparing the outcomes of risk simulation modelling betweenConventional, Respondents’ and Kondratieff Cases, it appears that optimism-biaspersists. This despite a track-record of demand overforecasting.Uncertainty is inherent in forecasting, particularly for start-up facilities and indeveloping economies. Given this, reliance on base/ central case forecasts can bemisleading. Whilst full-blown Monte Carlo testing of traditional assignment models (letalone four-stage models) may not be practical, it is advisable that risk simulation testingbe undertaken on traffic and revenue forecasts; as well as to cost forecasts. This mightbe achieved through use of simplified forecasting models in spreadsheets, with keyvalues and sensitivities derived from orthodox traffic assignment models. If interestrates and price inflation escalate, the impacts of variance in these variables will25 “Peak oil” theorists hold that global oil production has either already or will shortly start to decline asreserves diminish. Consequently, given increasing global demand, oil prices are held to sharply escalate.26 “Impact” being defined as the difference in FIRR brought about between lower-bound and upper-boundof the parameter in question (e.g. initial interest rate; see Section 5.6). The percentages quoted refer to“Impact*R2” in the Respondents’ Case divided by the equivalent number in the Conventional Case.DissFinal Page 101 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)permeate many aspects of forecasting models (further compounded should escalatingeconomic growth in line with the K-Wave also be assumed).Inevitably this will require additional resources. However, the costs of projectevaluation pale in comparison with infrastructure costs. Those commissioning transportplanners should also resist pressuring their consultants from boosting forecasts; theenvironment is likely to get riskier, so distorted forecasts will result in an increased rateof project financial failure.Furthermore, it appears that many fail to appreciate the difference between equity- anddebt-side perspectives. Simply put, equity-side perspectives seek the mean value of aproject; whilst debt-side perspectives concentrate on all downside-risk elements. Withincreasing interest rates, the implications of mis-structuring finance will escalate;downstream re-financing tending to get more expensive (versus experience in the1980’s and 1990’s when interest rates generally decreased). This also suggests thatfixed-rate debt should be arranged wherever possible (e.g. bonds) and that prospectiveoperators ought to err on the side of taking-on extra up-front debt at lower rates, ratherthan risking downstream re-financing at a premium to initial rates. (Though not takingso much debt as to incur excessive debt servicing requirements.)Finally, it is suggested that further research is undertaken into the economic linkagesunderlying many transport models, with particular emphasis on developing countries.DissFinal Page 102 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)6.8 Conclusions: Evaluation of HypothesisDeveloping countries face more uncertainty than developed countries. Their increasedeconomic growth potential offset against economic volatility, corruption, sovereign andinstitutional issues and in particular contract enforceability. Yet East Asian economiesare increasingly important to global trade, as manufacturing centres and commodityproducers. Historical experience would support the view that such trends would drivedemand for transport infrastructure, including tollways. Given funding constraints,private participation is likely to remain important. However, performance is likely to bevolatile; this based on historical experience (e.g. 19th Century American railroads) andthe results of risk simulation modelling, with most forecast parameters exertingincreased impact on FIRR (Chapter 5).In addition to general forecast uncertainty, the following risks should be highlighted(based on Conventional Case risk simulations): Base demand (i.e. whether there is sufficient traffic congestion to drive demand); Economic growth (which is likely to be volatile); Interest rates (for financing); Construction costs and duration; and, Price inflation.The specific hypothesis is “There is a significant change in the nature and extent ofproject finance risks for private stakeholders in East Asian toll roads during a period ofincreasing price inflation and interest rates.” Practitioners generally held that bothprice inflation and interest rates would increase to an extent, though were less certain asDissFinal Page 103 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)to the significance of such increases. The risk simulations showed both were stronglycorrelated with project FIRR. Assuming that projects are substantially debt-financed,then increasing interest rates would markedly affect outturn performance. Meanwhile,price inflation will affect construction and operating and maintenance costs, as well asincreasing the impact of delayed toll increases (and increases beneath price inflationrates). Risk simulation showed that rising price inflation and especially interest rates arelikely to substantially increase their importance relative to other project-level risks.However, should fixed-rate debt be available (e.g. bonds) then risk can be offset (risingprice inflation decreasing the real debt burden) and subsequent increases in interest ratesare less important (so long as re-financing is not necessary).Furthermore, increasing price inflation and interest rates could be associated withaccelerating economic growth; though the K-Wave posits this, acceptance of the K-Wave is not necessarily required to accept the linkage between economic growth, priceinflation and interest rates. And economic growth is strongly positively correlated withproject performance, permeating most aspects of demand forecasting and potentiallymitigating some of the impacts of rising interest rates.Indeed if one accepts the K-Wave upswing scenario, then notwithstanding likelyperiodic reversals, economic prospects for East Asia are likely good. Furthermore, thereis a window of opportunity to set-up projects to take advantage of this growth beforeprice inflation and interest rates escalate markedly.In conclusion, rising price inflation and interest rates do appear likely to change thenature and extent of project finance risks for private stakeholders in East Asian tollroads.DissFinal Page 104 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) REFERENCES: LITERATUREThe following are referenced either in the Main Text (above) or in the Appendices:Anhui Expressway Ltd (1996) Prospectus for Placing and New Issue, Crosby CapitalMarkets (Asia) Limited and CEF Capital Limited, Hong Kong, 31st October 1996Asian Development Bank (ADB) (2003), Transit Fee and Tolling for Routes 3 and 9,Lao PDR, TA-3348 Final Report, Manila, April 2003Asian Development Bank (ADB), Japan Bank for International Cooperation andDevelopment (JBIC), and World Bank (WB), (2005) Connecting East Asia: A NewFramework for Infrastructure, Advance Edition, Manila, Washington, D.C. and Tokyo,May 2005Azfar, O., Gurgur, T., Kähkönen, S., Lanyi, A. and Meagher, P. (2000)“Decentralization and Governance: An Empirical Investigation of Public ServiceDelivery in the Philippines” IRIS Center, University of Maryland, College Park and TheWorld Bank, Washington D.C.Bain, R. and Polakovic, L, (2005) “Traffic Forecasting Risk Study Update 2005:Through Ramp-Up and Beyond”, Commentary, Standard & Poor’s, London, 25 August2005Bain, R. and Wilkins, M. (2002) “Credit Implications of Traffic Risk in Start-Up TollFacilities”, Infrastructure Finance, Standard & Poor’s, London, September 2002Beaverstock, J.V. and Doel, M.A. (2001) “Unfolding the Spatial Architecture of theEast Asian Financial Crisis: The Organizational Response of Global Investment Banks”,Geoforum, 32(1), pp.15-32DissFinal Page 105 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Brinkman, P.A. (2003) The Ethical Challenges and Professional Responses of TravelDemand Forecasters, PhD Dissertation, University of California at BerkeleyBuchanan, C.P. (1999) “The Role of BOT in the Highway Sector – Experience fromMalaysia and Elsewhere”, PIARC 21st World Roads Congress, Kuala Lumpur, October1999Central Intelligence Agency (CIA) (2005/2006) The World Factbook,www.odci.gov/cia/publications/factbook/index.htmlChan, C. (2003) “New World Divests in Mainland Toll Roads and Bridges”, SouthChina Morning Post, Hong Kong, 17 November 2003Cheong, Y.M. (1999) “The Political Structures of Independent States”, in Tarling, N.(Ed.) The Cambridge History of Southeast Asia, Volume Two Part Two (From WorldWar II to the Present), Cambridge University Press, Cambridge, pp 59-138Christensen, P. and Mertner, J. (1994), Assessment of Modal Competitiveness andTraffic Potential of a Rehabilitated Railway in Cambodia, Final Report, COWI Consult,Copenhagen and Asian Development Bank, ManilaCorbett, V.F. and Di Bona, R.F. (2006) “Discussion of Future Development,Management and Use of the GMS Transport Model”, Workshop on ADB TA6195-REG: GMS Transport Sector Strategy Study, Vientiane, Laos, 21 March 2006Corbett, V.F., Winston, B.P., Whittle, J.D., Mansell, P., Di Bona, R.F. and Husband, D.,(2006) GMS Transport Sector Strategy Study, Final Report, T.A. No. 6195-REG,PADECO Co. Ltd., Tokyo and Asian Development Bank, Manila, May 2006DissFinal Page 106 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Di Bona, R.F. (2002) Surviving Bahtulism: A Financial and Economic Appraisal ofScott Wilson Kirkpatrick Thailand Ltd, Managing Financial Resources assignment,Henley Management College.Di Bona, R.F. (2005) “GMS Transport Model”, Workshop on Draft Final Report ofADB TA 6195-REG: GMS Transport Sector Strategy Study, Ho Chi Minh City,Vietnam, 8-9 December 2005Dizon, N. (2002) “Laguna Folk Sue Skyway Managers for Breach of Contract”, SunStar, Manila, 19 January 2002, http://www.newsflash.org/2002/01/ht/ht002241.htm(sighted July 2006)Faber, M. (2002) Tomorrow’s Gold: Asia’s Age of Discovery, CLSA Books, HongKong, 2002Faber, M. (2003) “The Financial Implications of Reflation”, The Gloom, Boom & DoomReport, Marc Faber Limited, Hong Kong, June 23, 2003Faber, M. (2005) “Why are Investment Markets Stalling?”, Market Commentary,http://www.gloomboomdoom.com/marketcoms/mcdownloads/060509.pdf, 2 September2005Faber, M. (2006) “Poor Mr. Bernanke”, Market Commentary, 9 May 2006,http://www.gloomboomdoom.com/marketcoms/mcdownloads/060509.pdfFlyvbjerg, B. and COWI (2004) Procedures for Dealing with Optimism Bias inTransport Planning – Guidance Document, The British Department of Transport,London, and COWI, Copenhagen, 10 June 2004DissFinal Page 107 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Forsgren, K., Hecht J.E., Doud, T. and Wilkins, M. (1999) “The Toll Road Sector:Smooth Conditions Overall, but Watch for Caution Flags”, Standard & Poor’sInfrastructure Finance, October 1999, pp149-157George, C., Trommer, S. and McDermott, M. (2004) “The Continuing Search for Bliss:Flexible Toll Road Structures”, Project Finance Special Report, Fitch Ratings, NewYork, 20 October 2004Gini, C. (1912) “Variabilità e Mutabilità”, Studi Economico-Giuridici, Anno III, Parte2, Facoltà di Giurisprudenza, Regia Università di Cagliari, 1912Gomez, E.T. and Jomo K.S. (1999) Malaysia’s Political Economy: Politics, Patronageand Profits, 2nd Edition, Cambridge University Press.Guangdong Provincial Bureau of Statistics (1998), Guangdong Statistical Yearbook1998, China Statistics Press, Beijing, 1 August 1998Guangdong Provincial Bureau of Statistics (2000), Guangdong Statistical Yearbook2000, China Statistics Press, Beijing, 1 August 2000Guangdong Provincial Bureau of Statistics (2003), Guangdong Statistical Yearbook2003, China Statistics Press, Beijing, 1 August 2003Guangdong Provincial Bureau of Statistics (2005), Guangdong Statistical Yearbook2005, China Statistics Press, Beijing, 1 August 2005Guislain, P. and Kerf, M. (1995) “Concessions – The Way to Privatize InfrastructureSector Monopolies”, Public Policy for the Private Sector, Note No. 59, The WorldBank, Washington, D.C.DissFinal Page 108 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Gunn, H. and Sheldon, R. (2001) “The Value of Time”, The Transport Economist,Volume 28 Number 1, Spring 2001Hayek, F.A., (1933) Monetary Theory and the Trade Cycle, London, 1933Higson, C. (1995) Business Finance, Second Edition, Butterworths, London.Hopewell Highway Infrastructure Limited (HHI) (2003) Global Offering Prospectus,Hong Kong, 28 July 2003Hymans, S.H. (1970) “Consumer Durable Spending: Explanation and Prediction”,Brookings Papers on Economic Activity, no. 2, pp. 173-199, 1970International Project Finance Association (IPFA) (2006), “About Project Finance”,http://www.ipfa.org/about_pf.shtml, sighted in June 2006Irwin, C. (1999) “Decentralizing Transportation Infrastructure”, in Standard & Poor’sInfrastructure Finance: Project Finance, Utilities and Concessions, October 1999,pp.146-148Jiangsu Expressway Co. Ltd. (1997), Prospectus for Placing and New Issue, HSBCInvestment Bank Asia, Hong Kong, 18 June 1997Jiangsu Expressway Co. Ltd. (2006) 2005 Annual Report, NanjingJiangsu Provincial Statistics Bureau (1999), Jiangsu Statistical Yearbook 1999, ChinaStatistics Press, Beijing, 1 July 1999Jiangsu Provincial Statistics Bureau (2002), Jiangsu Statistical Yearbook 2002, ChinaStatistics Press, Beijing, 1 July 2002DissFinal Page 109 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Jiangsu Provincial Statistics Bureau (2003), Jiangsu Statistical Yearbook 2003, ChinaStatistics Press, Beijing, 1 July 2003Jiangsu Provincial Statistics Bureau (2004), Jiangsu Statistical Yearbook 2004, ChinaStatistics Press, Beijing, 1 June 2004Juglar, C. (1863) Des Crises Commerciales et leur Retour Périodique en France, enAngleterre, et aux Etats-Units, Imprimerie de Veuve Berger-Levrault, Strasbourg, 1863Khan, A. and Willumsen, L.G. (1986) “Modelling Car Ownership and Use inDeveloping Countries”, Traffic Engineering and Control, 27(11), pp554-60.Kilsby, D. (2004) “Ethical Challenges and Professional Responses of Travel DemandForecasters – Review”, downloaded from www.kilsby.com.au (in October 2006)Kilsby, D. (2006a) “Australia’s Future Oil Supply and Alternative Transport Fuels”,Submission to Senate Inquiry into Australia’s Future Oil Supply and AlternativeTransport Fuels, National Committee on Transport, Engineers Australia, downloadedfrom www.kilsby.com.au (in October 2006)Kilsby, D. (2006b) “Peak Oil – It’s Coming (Fast)”, Article in Cityscape, July andAugust 2006, downloaded from www.kilsby.com.au (in October 2006)Kindleberger, C.P. (1996) Manias, Panics and Crashes: A History of Financial Crises,3rd Edition, John Wiley & Sons, 1 December 1996Kitchin, J. (1923) “Cycles and Trends in Economic Factors”, Review of EconomicStatistics, Volume 5, pp. 10-16DissFinal Page 110 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Klein, M., So, J. and Shin, B. (1996) “Transaction Costs in Private InfrastructureProjects – Are They Too High”, Public Policy for the Private Sector, Note No. 95, TheWorld Bank, Washington, D.C., October 1996Kondratieff, N.D. (1926) “Die Langen Wellen der Konjunktur”, Archiv fürSozialwissenschaft und Sozialpolitik, 1926, Volume 56, No. 3, pp. 573-609; translatedby W. F. Stolper and reprinted as “Long Waves of Economic Life”, The Review ofEconomics and Statistics, Volume 17, Number 6, November 1935, pp. 20-42Krugman, P. (2000) The Return of Depression Economics, Paperback Edition, Penguin,LondonKuznets, S. (1930) Secular Movements in Production and Prices: Their Nature andtheir Bearing upon Cyclical Fluctuations, Houghton Miffin, Boston and New York.Kuznets, S. (1955) “Toward a Theory of Economic Growth”, in Lekachman, R. (Ed.),National Policy for Economic Welfare at Home and Abroad, Doubleday, Garden City,New York.Lumby, S. (1983) “The Case Against WACC in Investment Appraisal”, Accountancy,September 1983La Porta, R., López-de-Silanes, F., Shleifer, A., and Vishny, R. (1997) “WhichCountries Give Investors the Best Protection?”, Public Policy for the Private Sector,Note No. 109, The World Bank, Washington, D.C., April 1997Lawrence, A. (1999) “The Curse Bites: Skyscraper Index Strikes”, Property Report,Dresdner Kleinwort Benson Research, 3 March 1999DissFinal Page 111 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Luu, H.S. (2006) “Financing Sources for Transport Development in Vietnam”, NationalGraduate Institute for Policy Studies, Tokyo, 26th August 2006,www.grips.ac.jp/vietnam/VDFTokyo/Doc/25LHSon26Aug06Slides.pdfMaddison, A. (1995) Monitoring the World Economy, 1820-1992, Washington, D.C.:Organization for Economic Cooperation and DevelopmentManasan, R.G. (2004) “Infrastructure and Decentralization”, Background Paper forPhilippines AAA, World Bank, Washington D.C.Markowitz, H.M. (1952), “Portfolio Selection”, The Journal of Finance, Volume VII,No. 1, March 1952, pp77-91Mendez, C. (2004) “DOTC to Block Tollway Rate Hike”, STAR, Manila, July 7, 2004,http://www.newsflash.org/2004/02/hl/hl100616.htmMontlake, S. (2005) “Cambodia’s Killing Fields Get Privatized”, Christian ScienceMonitor, 3 May 2005, http://www.csmonitor.com/2005/0503/p06s01-woap.htmlMorris, D. (1971) The Human Zoo, London: CorgiNational and Economic Social Development Board (NESDB) (2006) Gross DomesticProduct: Q1/ 2006, Bangkok, downloaded from www.nesdb.go.th (in June 2006)National Institute of Statistics (2004), Cambodia Inter-Censal Population Survey 2004,General Report, Ministry of Planning, Phnom Penh, November 2004Ormerod, P. (1999) Butterfly Economics, Far East Edition, London: Faber and Faber.Ormerod, P. (2005) Why Most Things Fail: Evolution, Extinction and Economics,London: Faber and Faber.DissFinal Page 112 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Orosz, C. (1998) “The Effects of Tolled Motorways in Hungary”, European TransportConference, Loughborough, 14-18 September 1998Ortúzar, J.d.D. and Willumsen, L.G. (1994) Modelling Transport, 2nd Edition, JohnWiley & Sons.Parsons Brinckerhoff (Asia) Ltd (PBA) (2003) “Traffic Consultant’s Report”, inHopewell Highway Infrastructure Limited (2003) Global Offering ProspectusPigou, A.C. (1920) The Economics of Welfare, London.Pindyck, R.S. and Rubinfeld, D.L. (1981) Econometric Models and EconomicForecasts, Second Edition, McGraw-Hill International Editions, Singapore.Porter, M. (1980) Competitive Strategy, New York Free Press.Richardson, A.J. (2004) “Estimating Individual Values of Time in Stated PreferenceSurveys”, presented at 26th Conference of Australian Institutes of Transport Research(CAITR), Melbourne, December 2004Rigby, P.N. (1999) “Project Finance: Technical Risk Criteria”, Standard & Poor’sInfrastructure Finance, October 1999, pp 40-45Rigby, P. and Penrose, J. (2001) “Project Finance Summary Debt Rating Criteria”, inStandard & Poor’s Project & Infrastructure Finance Review: Criteria and Commentary,October 2001, pp. 19-32Rogers, E. (1962) Diffusion of Innovations, 1962Ruster, J. (1996) “Mitigating Commercial Risks in Project Finance”, Public Policy forthe Private Sector, Note No. 69, The World Bank, Washington, D.C., February 1996DissFinal Page 113 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Schumpeter, J.A. (1939) Business Cycles: A Theoretical, Historical and StatisticalAnalysis of the Capitalist Process, Philadelphia, 1939Schumpeter, J.A. (1950) Capitalism, Socialism and Democracy, 3rd Edition, New York:Harper & Brothers, 1950Scott Wilson (Hong Kong) Ltd. and RJ Nairn and Partners Pty Ltd (SWHK) (1997a)“Letter from the Traffic Forecast Consultant”, in Jiangsu Expressway Co. Ltd.Prospectus for Placing and New Issue, pp167-182Scott Wilson (Hong Kong) Ltd. and RJ Nairn and Partners Pty Ltd (SWHK) (1997b)“Letter from the Operation Review Consultant”, in Jiangsu Expressway Co. Ltd.Prospectus for Placing and New Issue, pp183-190Scott Wilson Kirkpatrick (Hong Kong) Ltd (SWHK) (1996a) “Letter from the TrafficForecast Consultant”, in Anhui Expressway Ltd (1996) Prospectus for Placing and NewIssue, pp105-111Scott Wilson Kirkpatrick (Hong Kong) Ltd (SWHK) (1996b) “Letter from theOperation Review Consultant”, in Anhui Expressway Ltd (1996) Prospectus forPlacing and New Issue, pp112-115Shenzhen Expressway (2006), “Toll Roads & Bridges >> Toll Standards”, webpagefound October 2006 at www.sz-expressway.comStanding Advisory Committee on Trunk Roads Assessment (SACTRA) (1994), TrunkRoads and the Generation of Traffic, Her Majesty’s Stationery Office, LondonStreeter, W., Hermans, M., McCarthy, B., Stephenson, K., Monnier, L., Kolotas, P. andOsako, M. (2004) “Toll Road Securitizations: Wherein the Future Flows”, InternationalDissFinal Page 114 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Structured Finance/ Project Finance Special Report, Fitch Ratings, New York, 20January 2004.Streeter, W. and McManus, K. (1999) “Challenges of Start-Up Toll Roads”, ProjectFinance Special Report, Fitch ICBA, Duff & Phelps, New York.Thornton, M. (2003) “Skyscrapers and Business Cycles”, Working Paper, Von MisesInstitute, http://www.mises.org/workingpapers.asp, 29 May 2003Transparency International (2004) Annual Report, Berlin.United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP)(2005) Statistical Indicators for Asia and the Pacific 2005 Compendium, Volume XXXV,December 2005, http://www.unescap.org/stat/data/statind/pdf/index.asp#VolXXXV,(downloaded July 2006)Van Zuylen, H and Willumsen, L.G. (1980) “The Most Likely Trip Matrix EstimatedFrom Traffic Counts”, Transportation Research, 14B(3), pp. 281-293Wardman, M. (1998) “Review of Service Quality Valuations”, European TransportConference, Loughborough, 14-18 September 1998Willumsen, L. and Russell, C. (1998) “Reducing Revenue Risk”, European TransportConference, Loughborough, 14-18 September 1998Wong, M. and Moy, P. (2004) “Higher Tolls Prove a Drag on Western Tunnel”, SouthChina Morning Post, Hong Kong, 2 November 2004World Bank (WB) (2003a), Private Participation in Infrastructure: Trends inDeveloping Countries 1990-2001, Washington D.C.DissFinal Page 115 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)World Bank (WB), (2003b) “Combating Corruption in Indonesia”, World Bank PovertyReduction and Economic Management Unit Report 27246, Washington D.C., 2003World Bank (WB) (2004a), Civil Society Perceptions. Background for the InfrastructureFlagship http://www.worldbank.org/eapinfrastructureWorld Bank (WB), (2004b) Averting an Infrastructure Crisis – Indonesia. AFramework for Policy and Action, Washington D.C., 2004World Bank (WB), (2005) China Quarterly Update, Washington D.C., February 2005Yepes, T. (2004) “Expenditure on Infrastructure in East Asia Region, 2006-2010”,Paper for East Asia Pacific Infrastructure Flagship Study, Asian Development Bank(ADB), Manila, Japan Bank for International Cooperation (JBIC), Tokyo and TheWorld Bank (WB), Washington, D.C.Yuen, J. (2005) China on the March – Again: A Business Culture Perspective, MBADissertation, Henley Management College, Henley.Zhejiang Expressway Co. Ltd. (2006) 2005 Annual Report, HangzhouZhejiang Provincial Bureau of Statistics (1999), Zhejiang Statistical Yearbook 1999,China Statistics Press, Beijing, 1 July 1999Zhejiang Provincial Bureau of Statistics (2000), Zhejiang Statistical Yearbook 2000,China Statistics Press, Beijing, 1 August 2000Zhejiang Provincial Bureau of Statistics (2002), Zhejiang Statistical Yearbook 2002,China Statistics Press, Beijing, 1 July 2002Zhejiang Provincial Bureau of Statistics (2004), Zhejiang Statistical Yearbook 2004,China Statistics Press, Beijing, 1 July 2004DissFinal Page 116 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) REFERENCES: INTERNET RESOURCESThe following websites were used variously to obtain raw data, literature, figures,definitions, for dissemination of questionnaire surveys (as defined in parentheses):Asia-Pacific Economic Cooperation (for reports and economic statistics): www.apec.orgAsian Development Bank (for reports and economic statistics): www.adb.orgBank of Thailand (for economic statistics): www.bot.or.thCentral Intelligence Agency (CIA) World Factbook (for data on various countries):www.odci.gov/cia/publications/factbook/index.htmlChristian Science Monitor (for news article on Cambodia): www.csmonitor.comForeign and Commonwealth Office, UK (for data on various countries): www.fco.go.ukGoogleEarthTM (for map in Figure 1.A): earth.google.comHenley Management College (for survey dissemination in addition to structuralguidance on the Dissertation): www.henleymc.ac.ukHopewell Highway Infrastructure Limited (for expressway traffic and revenue data):www.hopewellhighway.comInternational Project Finance Association (IPFA) (for background information onhistory of project finance): www.ipfa.orgJiangsu Expressway Co. Ltd. (for expressway traffic data): www.jsexpressway.com(formerly www.jsexpressway.com.cn)Kilsby Australia (for articles): www.kilsby.com.auLegal500.com (for identifying suitable legal professionals for the survey):www.legal500.comDr. Marc Faber/ Gloom Boom Doom Report website (for reports and marketcommentaries): www.gloomboomdoom.com (note: in mid-2006 much of this contentbecame subscriber-only)DissFinal Page 117 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)National Economic and Social Development Board, Thailand (for economic statistics):www.nesdb.go.thNational Graduate Institute for Policy Studies, Tokyo, Japan (for papers):www.grips.ac.jpNewsflash.org (for a number of news articles on the Philippines): www.newsflash.orgPacific Exchange Rate Service (for historical exchange rate information):www.fx.sauder.ubc.caShenzhen Expressway Co. Ltd. (for toll rates): www.sz-expressway.comSurvey MonkeyTM (used for conducting the questionnaire survey):www.surveymonkey.comThe Urban Transport Institute (for articles): www.tuti.com.auUnited Nations Economic and Social Commission for Asia and the Pacific (foreconomic data and reports): www.unescap.orgVon Mises Institute (for working papers and reports): www.mises.orgWorld Bank (WB) (for economic data and reports): www.worldbank.orgWren Investment Advisers (for historical gold prices and interest rates):www.wrenresearch.com.au/downloads/index.htmYahoo! Newsgroups (for survey dissemination): EMME/2 users’ group: http://groups.yahoo.com/group/emme2users TransCAD users’ group: http://tech.groups.yahoo.com/group/transcadZhejiang Expressway Co. Ltd. (for expressway traffic and revenue data):www.zjec.com.cnDissFinal Page 118 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) APPENDICESAppendix 1: Headline Demographic and Economic Statistics ..................................... 120Appendix 2: Headline Transport Statistics ................................................................... 122Appendix 3: Examples of Listed Provincial Chinese Expressway Operators .............. 123Appendix 4: Typical Financial Ratios........................................................................... 124Appendix 5: Kondratieff Waves since 1787 ................................................................. 125Appendix 6: Definition of Guanxi ................................................................................ 126Appendix 7: Typical Structure of Four-Stage Transport Models ................................. 127Appendix 8: Traffic Risk Index .................................................................................... 128Appendix 9: Measures of Corruption and its Impact .................................................... 130Appendix 10: PESTLE Analysis ................................................................................... 131Appendix 11: Correlation between Wealth and Transport Networks ........................... 137Appendix 12: Expressway and Economic Index Calculations ..................................... 148Appendix 13: Survey Questionnaire: Question Specification and Logical Flow ......... 163Appendix 14: Amendments Made to Questionnaire Following Pilot Survey .............. 171Appendix 15: Questionnaire Responses........................................................................ 175Appendix 16: Risk Simulation Modelling: Simulation Parameters Employed ............ 199Appendix 17: Risk Simulation Modelling: Fixed Parameters ...................................... 203Appendix 18: Risk Simulation Modelling: Equations Employed ................................. 204Appendix 19: Risk Simulation Modelling: Results by Parameter ................................ 206Appendix 20: Risk Simulation Modelling: Comparison of Parameters’ Impacts ........ 245DissFinal Page 119 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 1: Headline Demographic and Economic StatisticsData extracted from CIA Factbook (http://www.odci.gov/cia/publications/factbook/index.html)in May 2006, unless indicated by footnote to the contrary. Each datum corresponds to the valuesgiven. In most cases, data are from 2005, although in some cases older data were quoted. Thedata collection method is not known, but could be expected to vary from country-to-countrybased on what statistics are available and what degree of estimation is required in each case.Data are given for the 9 countries examined as part of the Dissertation, as well as for five othercountries for comparison and benchmarking purposes. Land Population Population % Below Median Life Area (Estimated (Annual Poverty Age Expectancy Country (km2) July 2006) Change) Line (years) at Birth Cambodia 176,520 13,881,427 1.78% 40% 20.6 59.29 China 9,596,410 1,313,973,713 0.59% 10% 32.7 72.50 Indonesia 1,826,440 245,452,739 1.41% 16.7% 26.8 69.87 Laos 230,800 6,368,481 2.39% 40% 18.9 55.49 Malaysia 328,550 24,385,858 1.78% 8% 24.1 72.50 Myanmar 657,740 47,382,633 0.81% 25% 27.0 60.97 Philippines 298,170 89,468,677 1.80% 40% 22.5 70.21 Thailand 511,770 64,631,595 0.68% 10% 31.9 72.25 Vietnam 325,360 84,402,966 1.02% 19.5% 25.9 70.85 South Korea 98,190 48,846,823 0.42% 15% 35.2 77.04 Poland 304,465 38,536,869 -0.05% 17% 37.0 74.97 Mexico 1,923,040 107,449,525 1.16% 40% 25.3 75.41 UK 241,590 60,609,153 0.28% 17% 39.3 78.54 USA 9,161,923 298,444,215 0.91% 12% 36.5 77.85 GDP (billion US$) Real GDP Proportion of GDP by Sector Gross Fixed Official Growth Agri- Investment Country PPP Exchange Rate Rate culture Industry Services as % of GDP Cambodia 29.89 4.791 6.0% 35.0% 30.0% 35.0% 22.8% China 8182 1790 9.3% 14.4% 53.1% 32.5% 43.6% Indonesia 901.7 270 5.4% 14.7% 30.6% 54.6% 21.5% Laos 11.92 2.541 7.2% 48.6% 25.9% 25.5% n/a Malaysia 248.7 121.2 5.2% 7.2% 33.3% 59.5% 20.3% Myanmar 76.36 8.042 1.5% 54.6% 13.0% 32.4% 11.5% Philippines 451.3 90.3 4.6% 14.8% 31.7% 53.5% 16.3% Thailand 545.8 177.2 4.4% 9.3% 45.1% 45.6% 31.7% Vietnam 253.2 44.66 8.4% 20.9% 41.0% 38.1% 38.7% South Korea 965.3 801.2 3.9% 3.7% 40.1% 56.3% 28.9% Poland 489.8 242.7 3.5% 2.8% 31.7% 65.5% 18.5% Mexico 1068 699.5 3.0% 4.0% 26.5% 69.5% 21.1% UK 1869 2218 1.7% 1.1% 26.0% 72.9% 16.3% USA 12410 12470 3.5% 1.0% 20.7% 78.3% 16.8%DissFinal Page 120 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Public Debt as Annual GDP per % of GDP/ Consumer capita Labour Gini Index External Debt Price (PPP Force Unemploy- on Family Country (US$ billion)27 Inflation method) (millions) ment (%) Income Cambodia 0.8 4.3% 2,200 7 7.1%28 40.0% China 28.8% 1.9% 6,300 791.4 4.2%29 44.0% Indonesia 52.6% 10.4% 3,700 94.2 10.9% 34.3% Laos 2.49 9.4% 1,900 2.8 5.7% 37.0% Malaysia 48.3% 2.9% 10,400 10.67 3.6% 49.2% Myanmar 6.97 25.0% 1,600 27.75 5.0% n/a Philippines 77.4% 7.9% 5,100 36.73 12.2% 46.6% Thailand 35.9% 4.8% 8,300 35.36 1.4% 51.1% Vietnam 75.5% 8.4% 3,000 44.39 5.5% 36.1% South Korea 30.1% 2.6% 20,400 23.53 3.7% 35.8% Poland 47.3% 2.1% 12,700 17.1 18.3% 34.1% Mexico 39.1% 3.3% 10,100 43.4 3.6%30 54.6% UK 42.2% 2.2% 30,900 30.07 4.7% 36.8% USA 64.7% 3.2% 42,000 149.3 5.1% 45.0%27 Left aligned numbers refer to Public Debt as % of GDP. Right Aligned numbers refer to External Debtin US$billion. Data were available in one or the other format, but not for both, in each case.28 Source: National Institute of Statistics (2004, p. xiv) (A figure of 2.5% unemployment in Cambodiawas quoted in the CIA Factbook, based on a 2000 estimate, which appeared very low to the Author.Hence, an alternative source was sought for this datum.)29 4.2% official registered unemployment in urban areas in 2004; substantial unemployment andunderemployment in rural areas; an official Chinese journal estimated overall unemployment (includingrural areas) for 2003 at 20%. (This note taken from CIA Factbook.)30 Mexico has 3.6% unemployment plus underemployment of perhaps 25% (2005 est.). (This note takenfrom CIA Factbook.)DissFinal Page 121 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 2: Headline Transport Statistics Airports Roadway (km) With Paved Railway Waterway Total Runway (km) Total Paved (km) Notes Cambodia 20 6 602 12,323 1,996 2,400 (1) China 489 389 71,898 1,809,829 1,447,682 123,964 (9) Indonesia 668 161 6,458 368,360 213,649 21,579 Laos 44 9 0 32,620 4,590 4,600 (2), (9) Malaysia 117 37 1,890 71,814 55,943 7,200 Myanmar 84 19 3,955 27,000 3,200 12,800 (9) Philippines 256 83 897 200,037 19,804 3,219 (3) Thailand 108 65 4,071 57,403 56,542 4,000 (4), (9) Vietnam 28 23 2,600 94,354 23,589 17,702 (5) South Korea 108 70 3,472 97,252 75,641 1,608 (6) Poland 123 84 23,852 423,997 295,356 3,997 Mexico 1,832 227 17,634 349,038 116,928 2,900 UK 471 334 17,274 387,674 387,674 3,200 (7) USA 14,893 5,120 227,736 6,407,637 4,164,964 41,009 (8)Primary data source: CIA (2005-2006) The World FactbookNotes: (1) Estimate of length of Cambodias paved roads from 2000. Since this time there has been rehabilitation of many key routes within the country. (2) Additional 2,897km of waterways in Laos seasonally navigable by craft with draft up to 0.5m. (3) Philippine waterways limited to vessels with draft under 1.5m. (4) 3,701km of Thailands waterways are restricted to vessels with draft up to 0.9m. (5) 5,000km of Vietnams waterways restricted to vessels with upto 1.8m draft. The apparently large length of waterways is largely attributable to the Red River Delta in the north and the Mekong Delta in the south. Roadway statistics taken from ADB et al (2005) Connecting East Asia. (6) South Koreas waterways mostly navigable only by small craft. (7) Only 620km of UKs waterways used for commerce. (8) Only 19,312km of USAs waterways used for commerce. These figures include 3,769km shared with Canada. (9) In light of the qualification given above (8), it is believed that the length of shared waterways (i.e. defining borders) are included under both countries concerned in each instance. Within East Asia, this would primarily affect the Mekong which defines substantial portions of the Laos-Thailand border, as well as borders between China, Laos and Myanmar.DissFinal Page 122 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 3: Examples of Listed Provincial Chinese Expressway Operators Company Year Bourse Anhui Expressway 1996 Hong Kong Guangdong Provincial Expressway 1996 Shenzhen 1997 Hong Kong Jiangsu Expressway 1999 Shanghai Shandong Infrastructure 2000 Shanghai Sichuan Expressway 1997 Hong Kong 1997 Hong Kong Zhejiang Expressway 2000 LondonDissFinal Page 123 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 4: Typical Financial Ratios PROFITABILITY RATIOS Return on Capital Employed Profit(Loss) Before Interest and Tax (ROCE) Total Assets less Current Liabilities Gross Profit Margin (GPM) Gross Profit Turnover Profit on Sales (POS) Profit(Loss) Before Interest and Tax Turnover Expenses as Percentage of Expenses Turnover (EPT) Turnover Sales to Capital Employed (SCE) . Turnover . Total Assets less Current Liabilities Sales to Fixed Assets (SFA) . Turnover . Fixed Assets Sales to Working Capital . Turnover . Net Current Assets LIQUIDITY/ WORKING CAPITAL MANAGEMENT RATIOS Working Capital Requirement Current Assets less Current Liabilities (WCR) Current Ratio . Current Assets . Current Liabilities Asset Turnover . Turnover . Total Assets less Current Liabilities Interest Cover/ Debt Service Profit(Loss) Before Interest and Tax Coverage Ratio Interest PayableDissFinal Page 124 December 2006

Panic of 1819 Panic of 1837 Panic of 1873 Crash 1929 Crash 1973 Crash 1987 1815 1866 1921 1976 2033DissFinal Dissertation 1787 1842 1896 1949 2004 2058 Depression Hard Time Depression Depression Great Depression Henley Management College First Kondratieff Second Kondratieff Third Kondratieff Fourth Kondratieff Fifth Kondratieff 1787 - 1842 1842 - 1896 1896 - 1949 1949 - 2004 2004 - 2058 Source: Faber, M. (2002, p.120) Upswing: from late 1780s to period Upswing: from late 1840s to early Upswing: from early 1890s to period Upswing: from 1940s to 1970s Upswing: from 1995-2004 to period 1810-1817 1870s 1914-1920 2025-2035 Downswing: from period 1810-1817 to Downswing: from early 1870s to early Downswing: from beginning of 1914- Downswing: from late 1970s to early Downswing: from 2025-2035 to period late 1840s 1890s 1920s 2000s 2055-2065 Displacement Appendix 5: Kondratieff Waves since 1787 Canals Railroadisation of America Progress in: Electronics Opening of new markets, China,Page 125 Eastern Europe, Russia Roads Gold Discoveries in California and Electricity Aerospace Australia Telecommunications Bridges Communication, chemical and auto Consumerism industry Entry of America into world markets Information technology, etc Services including health care, leisure, etc Application of new inventions to manufacturing (Industrial Revolution) Richard F. DI BONA (1005661)December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 6: Definition of GuanxiStrictly speaking, guanxi applies to China and to relationships amongst Chinese, yetthere are parallels in other Asian cultures. The following definition of guanxi isreproduced with permission from Yuen (2005, pp.75-76):There is no direct English translation of the term guanxi, which would fully convey itsmeaning of connections or relationships defined by reciprocity and mutual obligationand underpinned by a sense of goodwill and personal affection. Guanxi is based onmutual trust and shared experiences. Guanxi is a manifestation of China’s Confucianheritage. Its origins can be traced back to ancient Chinese social customs, in whichreciprocity and mutual obligation were used to build and maintain interpersonalrelationships throughout society.Guanxi exists in various forms. These differ depending on the closeness of therelationship between the parties involved. Chinese see relationships as existing on oneof three levels, each denoting a differing social proximity.1. Jiaren denotes family members (including extended family members). These represent the closest possible relationships in Chinese society.2. Shuren denotes non-family members, with whom one shares a significant connection, including people from the same town or village. Relationships with shuren, although not as close as those with jiaren, are still important.3. Shengren denotes strangers, to whom there is greater wariness as there is initially no basis for mutual trust. Not until such trust has been established can strangers become shuren.Renqing is a crucial concept for both understanding and cultivating guanxi. This term isused to express the reciprocation of outstanding favours, which accrue through guanxi.DissFinal Page 126 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 7: Typical Structure of Four-Stage Transport Models Land Use/ Planning Data (or Economic Growth Data) First Stage: Trip Generation Trip Rates x Planning or Economic Data Output: Total Trips to/from Each Zone Costs to/from (by Trip Purpose and/or Vehicle Type and/or Time of Day) each zone Second Stage: Trip Distribution Linkage between Zones (e.g. by Purpose/ Vehicle Type) Output: Trip Patterns Zone-to-Zone Zone-to-Zone (by Trip Purpose and/or Vehicle Type and/or Time of Day) costs Third Stage: Mode Split Proportions of Trips by Transport Mode Output: Zone-to-Zone Trips by Mode costs by mode Zone-to-Zone Fourth Stage: Assignment Routeings between each zonal pair Output: Flows on each network link Travel costs for each zonal pair by mode Modelled flows on network ("demand forecast") and economic analyses, etcDissFinal Page 127 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 8: Traffic Risk IndexThis is the Traffic Risk Index to be applied to traffic forecasts, postulated by Bain andWilkins (2002). Its purpose is to provide guidance as to the likely reliability of trafficforecasts.Project Attributes Low High Tolling Regime Shadow tolls User-paid tolls Tolling Culture Toll roads well established; data No toll roads in country; on actual use available uncertainty over acceptance Tariff Escalation Flexible rate setting/ escalation All tariff hikes require regulatory formula; no government approval approval required Forecast Horizon Near-term forecasts required Long-term (30+ year) forecasts requiredToll-Facility Details Facility already open Facility at the very early stages of planning Estuarial crossings Dense, urban networks Radial corridors into urban areas Ring-roads/ beltways around urban areas Extension of existing road Green-field site Alignment: strong rationale Confused/ unclear road (including tolling points and objectives (not where people intersections) want to go) Alignment: strong economics Alignment: strong politics Clear understanding of future Many options for network highway network extensions exist Stand-alone (single) facility Reliance on other, proposed highway improvements Highly congested corridor Limited/ no congestion Few competing roads Many alternative routes Clear competitive advantage Weak competitive advantage Only highway competition Multimodal competition Good, high capacity connectors “Hurry-up-and-wait” (congested access/ egress routes) “Active” competition protections Autonomous authorities can do (e.g. traffic calming, truck bans) what they want Surveys/ data Easy to collect (laws exist) Difficult/ dangerous to collect collection Experienced surveyors No culture of data collection Up-to-date Historical information Locally-calibrated parameters Parameters imported from elsewhere (another country?) Existing zone framework (widely Develop zone framework from used) scratch Users: private Clear market segment(s) Unclear market segments Few, key origins and destinations Multiple origins and destinations Dominated by single journey Multiple journey purposes purpose (e.g. commute, airport) High income, time-sensitive Average/ low income market marketDissFinal Page 128 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Project Attributes Low High Tolls in line with existing Tolls higher than the norm – facilities extended ramp-up? Simple toll structure Complex toll structure (local discounts, frequent users, variable pricing, etc) Flat demand profile (time-of-day, Highly seasonal and/ or “peaky” day-of-week, etc) demand profileUsers: commercial Fleet operator pays toll Owner-driver pays toll Clear time and operating cost Unclear competitive advantage savings Simple route choice decision- Complicated route choice making decision-making Strong compliance with weight Overloading of trucks is restrictions commonplace Micro-economics Strong, stable, diversified local Weak/ transitional local/ national economy economy Strict land-use planning regime Weak planning controls/ enforcement Stable, predictable population Population forecast dependent on growth many, exogenous factors Traffic growth Driven by/ correlated with Reliance upon future factors, new existing, established and developments, structural changes, predictable factors etc High car ownership Low/ growing car ownershipDissFinal Page 129 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 9: Measures of Corruption and its ImpactNGO perceptions of corruption and its impact on infrastructure development in China,Indonesia, Japan, Philippines, Thailand and Vietnam are summarised below from WB(2004a): Agree / Disagree/ Serious Obstacle Not Serious Obstacle Extent to which corruption is an obstacle 95% 5% Extent to which potential for corruption 91% 4% should be taken into accountGovernment does not do enough to prevent 77% 23% corruption in infrastructure developmentThe following table presents data from Transparency International (2004) on corruption.A score of 10 indicates highly clean; scores below 5 indicate widespread corruption;and, scores below 3 indicate rampant corruption. Corruption is a problem in general.However, business is often carried out with those to whom one is “connected”; thiswould be seen as biased from a western perspective. It reinforces the need to “know thesystem” and “know the people” in initiating and operating a project Country Level of Corruption/ Transparency Indices Cambodia n/a China 3.4 Indonesia 2.0 Laos n/a Malaysia 5.0 Myanmar 1.7 Philippines 2.6 Thailand 3.6 Vietnam 2.6DissFinal Page 130 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 10: PESTLE AnalysisPOLITICAL Official Name Government Democracy; Cambodian People’s Party in notional Cambodia Kingdom of Cambodia coalition with FUNCINPEC China People’s Republic of China One-Party State (Chinese Communist Party) Republic with Parliamentary Elections (Golkar Indonesia Republic of Indonesia currently ruling) Lao People’s Democratic Laos One-Party State (Lao People’s Revolutionary Party) Republic Rotating Monarchy with democracy: rule by Barisan Malaysia Malaysia Nasional coalition; United Malays National Organisation (UMNO) is main party therein Military Junta (State Peace and Development Myanmar Union of Myanmar Council) Philippines Republic of the Philippines US-Style Presidential Republic (Lakas Party ruling) Transitional military rule (pending restoration of Thailand Kingdom of Thailand bicameral democracy), under a Monarchy Socialist Republic of Vietnam One-Party State (Communist Party of Vietnam) Vietnam 31POLITICAL Stability Foreign Relations Public order fragile; whilst Prime Minister Hun Sen seen Generally good and improving. PM close to Cambodia as a “strong man”, much of Vietnam. Major recipient of development aid. state apparatus relatively weak Generally stable. But increased Improving. Strong trade with most neighbours. China labour and social unrest in Whilst still receiving development aid, has some areas. expanded its own aid donations in the region. Unrest in outlying areas; Generally good, but seen by some as weak on Indonesia ongoing terrorist threat. Muslim Militants. Improving following chairing of ASEAN. Close Laos Sporadic rural banditry to Vietnam. Increasing cooperation with Thailand. Major recipient of development aid. Malaysia Generally stable Generally good. Economic sanctions by much of the West plus Unsettled; insurgencies in some Myanmar political pressure within ASEAN constrain areas; bombings in capital economic development and receipt of aid. High crime level; threat of Government seen as bulwark against terrorism, Philippines bombings and kidnappings; but reputation is hurt by ongoing allegations of ongoing Presidential crisis Presidential vote-rigging and cronyism. Generally good. Seen as the political and trade Generally stable, except in Thailand centre of Mainland S.E. Asia. Increasingly south (unrest and bombings) involved in development aid to its neighbours. Improving. Vietnam still acts as an influence in Vietnam Generally stable Laos and Cambodia. Still some tension with China, but ties improving. Stability concerns in many Summary countries, though not Most foreign relations improving, with possible and necessarily deterring exception of Myanmar. Comments infrastructure investment.31 Source: www.fco.gov.uk, 23 August 2005, supplemented by some of the author’s own observations.DissFinal Page 131 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) GDP (PPP method) Real GDP Proportion of GDP by Sector Gross Fixed GDP Per capita Growth Agri- Investment ECONOMIC (US$ bn) (US$) Rate culture Industry Services as % of GDP Cambodia 29.89 2,200 6.0% 35.0% 30.0% 35.0% 22.8% China 8182 6,300 9.3% 14.4% 53.1% 32.5% 43.6% Indonesia 901.7 3,700 5.4% 14.7% 30.6% 54.6% 21.5% Laos 11.92 1,900 7.2% 48.6% 25.9% 25.5% n/a Malaysia 248.7 10,400 5.2% 7.2% 33.3% 59.5% 20.3% Myanmar 76.36 1,600 1.5% 54.6% 13.0% 32.4% 11.5% Philippines 451.3 5,100 4.6% 14.8% 31.7% 53.5% 16.3% Thailand 545.8 8,300 4.4% 9.3% 45.1% 45.6% 31.7% Vietnam 253.2 3,000 8.4% 20.9% 41.0% 38.1% 38.7% Typically Excepting Myanmar, growing rapidly, substantial Summary albeit typically from a relatively low Varied. GFI as % of base. GDP.Source: CIA Factbook (http://www.odci.gov/cia/publications/factbook/index.html) in May 2006DissFinal Page 132 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Urbanisation32 SOCIAL Propensity to Travel 1990 2004 Inter-urban travel restricted by poor highway networks Cambodia 11.6% 15.0% and affordability. Increasing inter-urban travel. Some restriction in remote China 26.4% 41.8% areas due to poor networks. Indonesia 20.9% 45.0% On Java, substantial. Less in more remote areas. Laos 18.6% 21.0% Inter-urban travel often arduous. Malaysia 54.7% 60.0% Much inter-urban travel. Inter-urban travel often arduous and some areas Myanmar 24.8% 30.0% restricted.Philippines 48.6% 62.0% Much low cost inter-urban travel. Thailand 17.7% 31.0% Much inter-urban travel. Vietnam 19.5% 26.0% Growing inter-urban travel. Increasing urbanisation. Generally much/ growing inter-urban travel. Suppressed Summary Often quite dramatic. in some cases by poor transport networks. SOCIAL Attitudes to Foreign Private Sector Involvement in Infrastructure Provision In general, keen to attract foreign investment, although local partners required in Cambodia many investments. However, controversy over privatisation of Choeung Ek (Killing Fields) and associated toll-road33. Substantial involvement of private sector in toll road provision, especially by China overseas Chinese. Stock market listings and Bond Issues of State majority-owned operators. Local connections often critical. Revenue guarantees largely abolished. Pre-Asian Financial Crisis there was substantial activity in toll-road financing. Indonesia Activity once again picking-up, but local connections often critical. Sector not yet developed. However, State Railways of Thailand extending their Laos network into Laos (Nong Khai – Friendship Bridge – Vientiane Municipality). Much private sector involvement. However, concessions go mainly to well- Malaysia connected locals who may then raise finance from overseas34. Keen to attract foreign investment largely curtailed by sanctions and shareholder Myanmar activism. Some roads have been financed by domestic BOT arrangements, but concessions go to well-connected individuals, rather than FDI PPP. Whilst local partners are usually required, much infrastructure has been financed/Philippines developed by international companies. However, there have been problems enforcing toll increases, especially on foreign-invested projects35. Thailand Overseas investors long active in Thailand. Relatively few foreign private investments in Vietnamese toll-roads to date. Vietnam Vietnam is tipped by some to develop this sector quickly in coming years. Some countries have developed foreign private financing more than others. In general, the scope for this sector’s contribution is acknowledged, but deep-seated Summary nationalism can restrict foreign equity shares, sometimes creating management control issues.32 Source: UNESCAP (2005, p.3)33 See: Montlake, M (2005)34 See: Gomez and Jomo (1999)35 For example, the South Luzon Expressway has had many challenges and cancellations of toll increases.For recent example, see: Mendez (2004)DissFinal Page 133 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)TECHNO- Toll Collection SystemsLOGICALCambodia Manual collection systems Manual collection on minor routes. Increased usage of automatic systems on China major routes.Indonesia Primarily manual Laos Manual Malaysia Computerised systems on major routes, but substantial manual collection.Myanmar ManualPhilippines Primarily manual. Thailand Primarily manual Vietnam Manual Largely manual. Use of computerised and automated systems increasing, Summary typically on higher-volume routes in richer countries. Development of Manufacturing and Primary Industries Agriculture predominates, with some basic export-oriented industries. However, Cambodia export price competitiveness restrained by efficiency of transport networks. In coastal areas China is a world-leader in manufacturing. However, other parts China of China are yet to catch-up. Large producer and consumer of many commodities. Whilst a largely agricultural society, there is also substantial and growing Indonesia manufacturing, mining and timber industries. Primarily agricultural. Mining has been hindered by being land-locked with Laos under-developed internal transport networks. Malaysia Well developed manufacturing and primary industries. Agriculture predominates. Sanctions and shareholder pressure limit development Myanmar of export-oriented manufacturing. Commodity exploitation increasing (especially exports to China). Agriculture predominates through much of the country. However, some areasPhilippines (e.g. Subic, Metro Manila) have significant manufacturing industries. Roll-out elsewhere often hindered by transport networks. Whilst the north (and especially north-east) are predominantly agricultural, there Thailand is substantial manufacturing especially around Greater Bangkok/ Laem Chabang areas. Agriculture predominates. However, manufacturing is rapidly increasing. Given Vietnam Vietnam’s geography distances to sea are typically short. Manufacturing is relocating globally to East Asia. Some countries are also very Summary important as commodity producers.DissFinal Page 134 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) LEGAL Legal System36 primarily a civil law mixture of French-influenced codes from the United Nations Transitional Authority in Cambodia (UNTAC) period, royal decrees, Cambodia and acts of the legislature, with influences of customary law and remnants of communist legal theory; increasing influence of common law in recent years a complex amalgam of custom and statute, largely criminal law; rudimentary civil code in effect since 1 January 1987; new legal codes in effect since 1 China January 1980; continuing efforts are being made to improve civil, administrative, criminal, and commercial law based on Roman-Dutch law, substantially modified by indigenous concepts and Indonesia by new criminal procedures and election codes; has not accepted compulsory ICJ jurisdiction based on traditional customs, French legal norms and procedures, and socialist Laos practice based on English common law; judicial review of legislative acts in the Supreme Malaysia Court at request of supreme head of the federation; has not accepted compulsory ICJ jurisdiction Myanmar has not accepted compulsory ICJ jurisdiction based on Spanish and Anglo-American law; accepts compulsory ICJ jurisdiction,Philippines with reservations based on civil law system, with influences of common law; has not accepted Thailand compulsory ICJ jurisdiction Vietnam based on communist legal theory and French civil law systemSummary varied LEGAL Level of Corruption/ Transparency Indices37Cambodia n/a China 3.4Indonesia 2.0 Laos n/a Malaysia 5.0Myanmar 1.7Philippines 2.6 Thailand 3.6 Vietnam 2.6 Corruption a problem in general. However, business is often carried out with those to whom one is “connected”; this would be seen as biased from a western Summary perspective. It reinforces the need to “know the system” and “know the people” in initiating and operating a project.36 Quoted from: CIA (2005)37 Source: Transparency International (2004): A score of 10 indicates highly clean; scores below 5indicate widespread corruption; and, scores below 3 indicate rampant corruption. Comments under“Summary and Comments” are the author’s own.DissFinal Page 135 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)ENVIRON- Perceived Importance of Environmental Considerations MENTAL Economic development first priority. However, environmental considerationsCambodia increasingly considered with potential for eco-tourism. Moves towards forestry management, but sometimes hampered by poor local control. Increased attention being paid to environment. However, many Chinese cities China remain amongst the worst polluted in the world. Environmental efforts sometimes undermined by local failings in rule of law. In some areas environmental considerations increasingly important. However, Indonesia highly variable across this very large archipelago. Economic development first priority, with increasing attention to environmental Laos considerations, aided by development agencies’ involvement. However, e.g. Nam Theun II dam and hydroelectric project remains contentious. Whilst economic development remains a priority, environmental protection is Malaysia increasingly important, both in major cities and in indigenous areas. Economic development first priority. Environmental considerations very much Myanmar secondary to development of individual projects. Some aspects of environment increasingly important, especially regarding tourismPhilippines projects. However, environmental efforts often undermined by local failings in rule of law or ability to enforce. Environment increasingly important, especially in tourism areas. Environmental Thailand efforts sometimes undermined by local failings. Environment increasingly important, however standards not always applied equally Vietnam across the country. Economic development predominates over environmental considerations, but the Summary environment is of increasing importance, possibly correlating with extent of economic development. Geographic Considerations Currently the most convenient land route from Bangkok to southern Vietnam, evenCambodia with road rehabilitation yet to be completed (still ongoing). Varied topography. Very large and diverse geography. Southern and eastern coastal regions more China developed. Western hinterlands more mountainous. A massive archipelago with differing customs and levels of economic development. Indonesia However, as part of the “Ring of Fire” much of the country is mountainous. Dominated by Mekong River and in north and east by mountains. Offers shortest Laos crow-fly route between Thailand and Vietnam, but mountainous. Sparsely populated. Peninsular Malaysia is most economically advanced part of Study Area, with well developed north-south highways, though east coast and east-west routes less Malaysia developed. East Malaysia (Sarawak and Sabah, plus Labuan) relatively less developed; topology dominated by rivers and mountains. Very large and diverse topography. Underdeveloped transport networks. Myanmar Myanmar offers the most logical land-routes between South Asia and China and between South Asia and South-East Asia. Also important for Sino-Thai land-based trade. Varied archipelago. High-capacity trunk road network relatively limited outsidePhilippines Metro Manila and immediate environs. There are long-term plans (with Japanese funding) to link main islands with a series of bridges. Whilst topography is varied, trunk highway network is relatively well developed. Thailand Tollways concentrated around Greater Bangkok. A “long” country and relatively “thin” excepting Red River area in the north. Vietnam Typically north-south (between Hanoi and Ho Chi Minh City) there is a coastal route and a secondary inland route through mountains. Summary Varied. However, mountainous areas typically less penetrated.DissFinal Page 136 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 11: Correlation between Wealth and Transport NetworksThe econometrics presented in this Appendix, together with graphical representationsthereof are the author’s own work, using data presented in Appendices 1 and 2. Threesets of models were developed:(1) A “Full Sample” of 9 Study Area countries and 5 others (used for benchmarking).(2) Models on the 9 countries in the Study Area(3) Models on the 5 benchmarking countriesAn amalgamation of the last two sets of equations is presented in Section 3.4.1. Models on Full SampleCountries in Study Area Other Countries (1)KH Cambodia KR South KoreaCN China MX MexicoID Indonesia PO PolandLA Laos UK United KingdomMY Malaysia US United States of AmericaMM MyanmarPH PhilippinesTH ThailandVN VietnamNote: (1) Other countries presented for benchmarking and indication of trends given further economic development.DissFinal Page 137 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Structure of Equations FittedA series of equations were fitted. In each case GDP per capita was taken as theexplanatory variable. Two basic structural forms were fitted, as follows:(A) Dependent      GDPpc(B) Dependent     GDPpcThe following dependent variables were used: TotalPopul ation(1) PopAP Population per Airport NumberofAirports TotalLandArea(km2)(2) Km2AP Land Area per Airport NumberofAirports TotalPopul ation(3) PopRail Population per km of Railway Railways (km) TotalLandArea(km2)(4) Km2Rail Land Area per km of Railway Railways (km) TotalPopul ation(5) PopRoad Population per km of Paved Road PavedRoad (km) TotalLandArea(km2)(6) Km2Road Land Area per km of Paved Road PavedRoad (km)DissFinal Page 138 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Regression Results on All Countries (1A) PopAP = 994600 – 27.26 GDPpc R2 = 12.1% (343480) (21.20) (1B) PopAP = 738400 x 0.999928GDPpc R2 = 41.9% (0.3962) (0.000024) (2A) Km2AP = 7469 – 0.2170 GDPpc R2 = 23.1% (1853) (0.1144) (2B) Km2AP = 6404 x 0.999929GDPpc R2 = 55.2% (0.2984) (0.000018) (3A) PopRail = 32490 – 0.9155 GDPpc R2 = 18.3% (9786) (0.5824) (3B) PopRail = 28590 x 0.999926GDPpc R2 = 52.7% (0.3541) (0.000021) (4A) Km2Rail = 212.1 – 0.005871 GDPpc R2 = 44.3% (33.35) (0.001985) (4B) Km2Rail = 201.4 x 0.999936GDPpc R2 = 49.6% (0.3287) (0.000020) (5A) PopRoad = 4352 – 0.1521 GDPpc R2 = 20.6% (1397) (0.0862) (5B) PopRoad = 3175 x 0.999898GDPpc R2 = 63.7% (0.3591) (0.000022) (6A) Km2Road = 51.75 – 0.001889GDPpc R2 = 16.5% (19.88) (0.001227) (6B) Km2Road = 27.53 x 0.999900GDPpc R2 = 51.8% (0.4534) (0.000028)Standard errors associated with each parameter are shown in parenthesis beneath theparameter in question. As can be seen, in each instance equation form (B) gave a bettergoodness-of-fit than (A), in terms of R2. The fits obtained by the (B) series of equations,together with observed data are shown in the following graphs:DissFinal Page 139 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Population per Airport 15 VN CN 14 KH TH Ln(Population per Airport) MM 13 KR IDPH PO MY 12 LA UK 11 MX 10 US 9 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.) km 2 per Airport 10 CN VN KH 9 MM LA TH Ln(km2 per Airport) 8 ID MY PO 7 PH MX KR US UK 6 5 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.)DissFinal Page 140 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Population per km of Railw ay 12 PH 11 Ln(Population per km of Railway) ID VN 10 KH CN TH MY KR MM 9 MX UK 8 PO US 7 6 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.) Km 2 per km of Railw ay 6 PH KHID MM MY 5 CNTH Ln(km2 per km of Railway) VN MX 4 US KR 3 UK PO 2 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.)DissFinal Page 141 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Population per km of Paved Road 10 MM 9 KH Ln(Population per km of Paved Road) PH VN 8 LA 7 ID TH CN MX KR 6 MY 5 UK PO US 4 3 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.) km 2 per km of Paved Road 6 MM 5 KH Ln(km2 per km of Paved Road) 4 LA 3 MX VN PH ID TH 2 CN MY 1 US KR 0 PO UK -1 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.)DissFinal Page 142 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2. Models on the 9 Study Area CountriesCountries in Study AreaKH CambodiaCN ChinaID IndonesiaLA LaosMY MalaysiaMM MyanmarPH PhilippinesTH ThailandVN VietnamThe rationale of these analyses was to determine whether a different set of functionsexists within developing East Asian economies. The same equations were fitted as in(1).DissFinal Page 143 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Regression Results (1A) PopAP = 107000 – 23.65 GDPpc R2 = 0.4% (740700) (133.65) (1B) PopAP = 668300 x 0.999970GDPpc R2 = 0.8% (0.7005) (0.000126) (2A) Km2AP = 8425 – 0.2643 GDPpc R2 = 2.0% (3860) (0.6964) (2B) Km2AP = 7595 x 0.999925GDPpc R2 = 7.2% (0.5628) (0.000102) (3A) PopRail = 37830 – 1.239 GDPpc R2 = 1.7% (22240) (3.809) (3B) PopRail = 32287 x 0.999945GDPpc R2 = 5.8% (0.5253) (0.000090) (4A) Km2Rail = 247.2 – 0.008482 GDPpc R2 = 9.6% (62.23) (0.01066) (4B) Km2Rail = 232.6 x 0.999960GDPpc R2 = 9.4% (0.2984) (0.000051) (5A) PopRoad = 7947 – 0.8623 GDPpc R2 = 32.9% (2580) (0.4655) (5B) PopRoad = 7835 x 0.999731GDPpc R2 = 53.9% (0.5215) (0.000094) (6A) Km2Road = 104.5 – 0.01264GDPpc R2 = 34.4% (36.6) (0.00660) (6B) Km2Road = 89.04 x 0.999686GDPpc R2 = 58.3% (0.5554) (0.000100)Standard errors associated with each parameter are shown in parenthesis beneath theparameter in question. As can be seen, results for airports and railway are notstatistically meaningful, with R2 in all instances below 10%. Only the equationsregarding paved road give results which could be deemed meaningful; and as withAppendix 5, (B) form equations give better fits in terms of R2 than (A) equations. (5B)and (6B) are plotted below.DissFinal Page 144 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Population per km of Paved Road 10 MM Ln(Population per km of Paved Road) 9 KH PH VN 8 LA 7 ID TH CN MY 6 0 2,000 4,000 6,000 8,000 10,000 12,000 GDP Per Capita (USD p.a.) km 2 per km of Paved Road 6 MM 5 Ln(km2 per km of Paved Road) KH 4 LA 3 PH VN ID TH 2 CN MY 1 0 2,000 4,000 6,000 8,000 10,000 12,000 GDP Per Capita (USD p.a.)DissFinal Page 145 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3. Models on 5 Benchmarking Countries Other Countries (1) KR South Korea MX Mexico PO Poland UK United Kingdom US United States of AmericaNote: (1) Other countries presented for benchmarking and indication of trends given further economic development.Only roads data were regressed, as follows:Regression Results (5A) PopRoad = 804.6 – 0.01809 GDPpc R2 = 40.3% (330.5) (0.01268) (5B) PopRoad = 915.6 x 0.999943GDPpc R2 = 47.5% (0.9017) (0.000035) (6A) Km2Road = 10.41 – 0.0002622GDPpc R2 = 26.1% (6.644) (0.0002548) (6B) Km2Road = 4.765 x 0.999962GDPpc R2 = 15.7% (1.323) (0.000051)Standard errors associated with each parameter are shown in parenthesis beneath theparameter in question. As can be seen, whilst a relationship appears to hold for roadsper capita, geographic road density gives unsatisfactory results. This sample is notmeant to be necessarily significant, merely to give some guidance as to an S-curve forroad provision with respect to economic development, as presented in Section 3.4.DissFinal Page 146 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Population per km of Paved Road 10 9 Ln(Population per km of Paved Road) 8 7 MX KR 6 5 UK PO US 4 3 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.) km 2 per km of Paved Road 3 MX 2.5 2 Ln(km2 per km of Paved Road) 1.5 1 US 0.5 KR 0 PO -0.5 UK -1 0 10,000 20,000 30,000 40,000 50,000 GDP Per Capita (USD p.a.)DissFinal Page 147 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 12: Expressway and Economic Index CalculationsAppendix comprises: 1. Data on Guangdong Province/ Guangzhou-Shenzhen Superhighway 2. Data on Jiangsu Province/ Shanghai-Nanjing Expressway 3. Data on Zhejiang Province/ Shanghai-Hangzhou-Ningbo Expressway1. Data on Guangdong Province and Guangzhou-Shenzhen SuperhighwaySocio-economic data from Guangdong Statistical Yearbooks relating to Guangdong Province asa whole:GDSY98: Guangdong Provincial Bureau of Statistics (1998)GDSY00: Guangdong Provincial Bureau of Statistics (2000)GDSY03: Guangdong Provincial Bureau of Statistics (2003)GDSY05: Guangdong Provincial Bureau of Statistics (2005) Year GDP current GDP growth rate year- Implied price Real GDP price on-year inflation year-on- growth year-on- (100m RMB) (comparable price) year (%) year (%) 1995 5,733.97 14.9% 10.5% 14.7% 1996 6,519.14 10.7% 2.7% 10.5% 1997 7,315.51 10.6% 1.5% 10.4% 1998 7,919.12 10.2% -1.8% 10.1% 1999 8,464.31 9.5% -2.4% 9.3% 2000 9,662.23 10.8% 3.0% 10.5% 2001 10,647.71 9.6% 0.5% 9.5% 2002 11,735.64 11.4% -1.1% 11.3% 2003 13,625.87 14.3% 1.6% 14.2% 2004 16,039.46 14.2% 3.1% 14.1% Source GDSY05, p70 GDSY05, p72 derived derived Year Civil Vehicle Ownership Source 1995 1,147,348 GDSY98, p422 1996 1,163,339 GDSY98, p422 1997 1,234,317 GDSY98, p422 1998 1,355,074 GDSY00, p433 1999 1,437,963 GDSY00, p433 2000 1,729,054 GDSY03, p365 2001 1,919,150 GDSY03, p365 2002 2,308,875 GDSY03, p365 2003 2,579,592 GDSY05, p387 2004 3,054,025 GDSY05, p387DissFinal Page 148 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Year Highway Passenger Highway Passenger- Average Highway Trips (10,000 people) km (100 million) Passenger Trip Length (km) 1995 118,406 613.07 51.78 1996 117,815 626.60 53.19 1997 113,259 616.48 54.43 1998 121,795 630.65 51.78 1999 137,324 700.74 51.03 2000 148,945 780.74 52.42 2001 161,967 858.86 53.03 2002 171,197 945.16 55.21 2003 174,288 983.67 56.44 2004 183,012 1,076.06 58.80 Source GDSY05, p386 GDSY05, p386 derived Year Highway Freight Highway MT-km Average Freight Trip Transport (10,000 MT) (100 million) Length (km) 1995 68,884 352.45 107.0 1996 60,131 327.81 127.6 1997 62,728 341.68 127.1 1998 65,682 371.08 144.6 1999 70,626 426.7 171.7 2000 75,365 472.49 183.9 2001 86,555 522.89 197.8 2002 92,736 576.35 219.5 2003 97,806 614.01 224.0 2004 102,843 657.49 220.8 Source GDSY05, p386 GDSY05, p386 derivedTraffic and revenue data for Guangzhou-Shenzhen Superhighway from Hopewell HighwayInfrastructure (www.hopewellhighway.com): Year and Average Average Daily Year and Average Average Daily Month Daily Revenue Month Daily Revenue Traffic (thousand Traffic (thousand (vehicles) RMB) (vehicles) RMB) 1995_01 41,000 1500 1996_01 54,000 1892 1995_02 37,000 1450 1996_02 50,000 1799 1995_03 47,000 1712 1996_03 57,000 2044 1995_04 49,000 1814 1996_04 59,000 2118 1995_05 49,000 1769 1996_05 59,000 2106 1995_06 49,000 1765 1996_06 58,000 2067 1995_07 51,000 1842 1996_07 60,000 2137 1995_08 50,000 1869 1996_08 62,000 2218 1995_09 51,000 1876 1996_09 62,000 2468 1995_10 51,000 1853 1996_10 61,000 2597 1995_11 52,000 1845 1996_11 60,000 2543 1995_12 55,000 1959 1996_12 60,000 2547 1997_01 63,000 2737 1998_01 71,000 3359 1997_02 52,000 2348 1998_02 70,000 3323 1997_03 62,000 2733 1998_03 73,000 3431 1997_04 65,000 2832 1998_04 74,000 3554DissFinal Page 149 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Year and Average Average Daily Year and Average Average Daily Month Daily Revenue Month Daily Revenue Traffic (thousand Traffic (thousand (vehicles) RMB) (vehicles) RMB) 1997_05 64,000 2685 1998_05 73,000 3397 1997_06 62,000 2585 1998_06 72,000 3314 1997_07 64,000 2618 1998_07 73,000 3452 1997_08 71,000 2980 1998_08 74,000 3503 1997_09 72,000 3415 1998_09 76,000 3577 1997_10 72,000 3415 1998_10 76,000 3540 1997_11 71,000 3317 1998_11 76,000 3541 1997_12 72,000 3341 1998_12 76,000 3624 1999_01 78,000 3510 2000_01 100,000 4657 1999_02 77,000 3584 2000_02 86,000 4071 1999_03 83,000 3828 2000_03 101,000 4629 1999_04 87,000 3949 2000_04 105,000 4828 1999_05 82,000 3693 2000_05 103,000 4689 1999_06 87,000 3876 2000_06 102,000 4689 1999_07 87,000 4006 2000_07 106,000 4864 1999_08 88,000 4144 2000_08 111,000 5136 1999_09 91,000 4284 2000_09 112,000 5156 1999_10 94,000 4324 2000_10 105,000 4783 1999_11 93,000 4255 2000_11 104,000 4587 1999_12 94,000 4342 2000_12 105,000 4683 2001_01 100,000 4636 2002_01 122,000 4961 2001_02 104,000 4656 2002_02 114,000 5022 2001_03 112,000 4909 2002_03 129,000 5378 2001_04 113,000 4963 2002_04 134,000 5472 2001_05 111,000 4838 2002_05 126,000 5168 2001_06 112,000 4869 2002_06 125,000 5082 2001_07 115,000 5033 2002_07 136,000 5373 2001_08 123,000 5373 2002_08 146,000 5686 2001_09 127,000 5492 2002_09 149,000 5726 2001_10 121,000 5181 2002_10 149,000 5623 2001_11 122,000 5070 2002_11 151,000 5599 2001_12 121,000 4971 2002_12 157,000 5774 2003_01 168,000 6302 2004_01 169,000 6513 2003_02 149,000 5704 2004_02 181,000 6640 2003_03 173,000 6223 2004_03 192,000 6941 2003_04 166,000 6037 2004_04 202,000 7340 2003_05 150,000 5367 2004_05 191,000 6888 2003_06 172,000 5934 2004_06 201,000 7222 2003_07 185,000 6495 2004_07 216,000 7780 2003_08 189,000 6770 2004_08 221,000 7910 2003_09 196,000 7123 2004_09 229,000 8146 2003_10 186,000 6967 2004_10 221,000 7874 2003_11 182,000 6804 2004_11 224,000 7878 2003_12 192,000 7127 2004_12 226,000 7965DissFinal Page 150 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Annual Average Daily Traffic and Revenue thus derived: Year Average Daily Vehicles Average Daily Revenue 1995 48,575 1,773,216 1996 58,533 2,212,628 1997 65,929 2,920,529 1998 73,688 3,468,795 1999 86,800 3,985,011 2000 103,402 4,733,689 2001 115,137 5,000,984 2002 136,649 5,407,778 2003 175,849 6,409,405 2004 206,134 7,426,615Indices thus obtained from above data: Civil Vehicle Passenger Trip GDP Ownership Passenger-km Length 1995 100.0 100.0 100.0 100.0 1996 110.7 101.4 102.2 102.7 1997 122.4 107.6 100.6 105.1 1998 134.9 118.1 102.9 100.0 1999 147.7 125.3 114.3 98.6 2000 163.7 150.7 127.3 101.2 2001 179.4 167.3 140.1 102.4 2002 199.9 201.2 154.2 106.6 2003 228.4 224.8 160.4 109.0 2004 260.9 266.2 175.5 113.6 Freight Trip Superhighway Superhighway Freight MT-km Length Traffic Revenue 1995 100.0 100.0 100.0 100.0 1996 87.3 93.0 120.5 124.8 1997 91.1 96.9 135.7 164.7 1998 95.4 105.3 151.7 195.6 1999 102.5 121.1 178.7 224.7 2000 109.4 134.1 212.9 267.0 2001 125.7 148.4 237.0 282.0 2002 134.6 163.5 281.3 305.0 2003 142.0 174.2 362.0 361.5 2004 149.3 186.5 424.4 418.8Income elasticities thus calculated (1995 to 2004): Income elasticity of: Value Civil Vehicle Ownership 1.02 Passenger-km 0.61 Passenger Trip Length 0.14 Freight MT-km 0.44 Freight Trip Length 0.68 Superhighway Traffic 1.39 Superhighway Revenue 1.38DissFinal Page 151 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Graph of Guangdong Province/ Guangzhou-Shenzhen Superhighway data (indexed to1995): 450.0 400.0 350.0 300.0 Index (1995=100) 250.0 200.0 150.0 100.0 50.0 0.0 95 96 97 98 99 00 01 02 03 04 19 19 19 19 19 20 20 20 20 20 Year GDP Civil Vehicle Ow nership Passenger-km Passenger Trip Length Freight MT-km Freight Trip Length Superhighw ay Traffic Superhighw ay RevenueDissFinal Page 152 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)2. Jiangsu Province and Shanghai-Nanjing Expressway (Jiangsu Section)Socio-economic data from Jiangsu Statistical Yearbooks relating to Jiangsu Province as awhole:JSSY99: Jiangsu Provincial Statistics Bureau (1999)JSSY02: Jiangsu Provincial Statistics Bureau (2002)JSSY03: Jiangsu Provincial Statistics Bureau (2003)JSSY04: Jiangsu Provincial Statistics Bureau (2004) Year GDP current GDP growth rate year- Implied price Real GDP price on-year inflation year-on- growth year-on- (100m RMB) (comparable price) year (%) year (%) 1997 6,680.34 12.0% -0.7% 11.9% 1998 7,199.95 11.0% -2.9% 10.9% 1999 7,697.82 10.1% -2.9% 10.1% 2000 8,582.73 10.6% 0.8% 10.4% 2001 9,511.91 10.2% 0.6% 10.2% 2002 10,631.75 11.6% 0.2% 11.6% 2003 12,460.83 13.6% 3.2% 13.6% Source JSSY04, p61 JSSY04, p62 derived derived Year Civil Vehicle Ownership Source 1997 519,930 JSSY99, p228 1998 561,129 JSSY99, p228 1999 639,152 JSSY02, p241 2000 745,106 JSSY02, p241 2001 871,191 JSSY02, p241 2002 1,044,960 JSSY03, p288 2003 1,317,673 JSSY04, p288 Year Highway Passenger Highway Passenger- Average Highway Trips (10,000 people) km (100 million) Passenger Trip Length (km) 1997 88,826 504.05 56.7 1998 92,215 527.62 57.2 1999 95,564 554.04 58.0 2000 101,713 594.48 58.4 2001 105,105 682.25 64.9 2002 110,139 719.08 65.3 2003 118,046 774.11 65.6 Source JSSY04, p285 JSSY04, p285 derivedDissFinal Page 153 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Year Highway Freight Highway MT-km Average Freight Trip Transport (10,000 MT) (100 million) Length (km) 1997 52,441 681.42 122.2 1998 54,328 661.85 148.0 1999 54,803 704.01 148.1 2000 59,056 746.39 131.5 2001 59,058 757.58 150.9 2002 60,299 770.03 157.0 2003 64,321 995.34 156.5 Source JSSY04, p286 JSSY04, p286 derivedTraffic data for Jiangsu Section of Shanghai-Nanjing Expressway from Jiangsu Expressway Co.Ltd. (www.jsexpressway.com.cn): Year Average Daily Composition (%) and Traffic Month (vehicles) Car Small Medium Large Heavy1997_01 11,876 47.82% 21.92% 26.49% 3.47% 0.26%1997_02 9,325 50.62% 23.09% 23.73% 2.31% 0.24%1997_03 12,187 48.33% 24.72% 23.40% 3.21% 0.33%1997_04 12,725 50.04% 24.44% 21.70% 3.41% 0.40%1997_05 11,962 51.08% 24.48% 20.26% 3.71% 0.47%1997_06 11,001 51.47% 23.92% 19.96% 4.03% 0.62%1997_07 11,115 51.41% 23.12% 20.79% 4.08% 0.61%1997_08 12,047 50.61% 23.08% 21.52% 4.22% 0.56%1997_09 13,180 48.17% 23.59% 23.22% 4.50% 0.52%1997_10 13,245 49.08% 23.25% 22.47% 4.69% 0.51%1997_11 13,380 48.69% 23.09% 22.59% 5.13% 0.49%1997_12 13,200 48.87% 23.12% 22.57% 4.95% 0.49%1998_01 11,890 49.07% 22.80% 22.53% 5.09% 0.51%1998_02 12,225 44.79% 22.48% 26.53% 5.69% 0.51%1998_03 13,865 44.37% 23.25% 24.65% 7.11% 0.63%1998_04 15,387 44.99% 23.24% 24.40% 6.74% 0.63%1998_05 14,453 44.35% 23.20% 25.09% 6.80% 0.56%1998_06 13,530 44.58% 23.03% 24.83% 6.86% 0.70%1998_07 13,550 45.48% 23.01% 24.00% 6.78% 0.73%1998_08 13,642 45.47% 22.80% 24.22% 6.71% 0.80%1998_09 15,186 43.97% 23.16% 25.25% 6.74% 0.88%1998_10 14,890 45.80% 23.69% 23.46% 6.22% 0.83%1998_11 14,856 45.95% 23.38% 23.47% 6.42% 0.78%1998_12 14,027 45.84% 23.83% 23.00% 6.48% 0.85%1999_01 13,403 46.58% 23.05% 23.12% 6.46% 0.79%1999_02 13,680 48.90% 22.27% 23.71% 4.48% 0.64%1999_03 15,357 45.01% 23.66% 24.18% 6.30% 0.85%1999_04 16,547 46.23% 23.98% 23.56% 6.31% 0.93%1999_05 15,591 45.80% 23.90% 22.90% 6.42% 0.99%1999_06 14,749 46.04% 23.84% 22.39% 6.67% 1.06%1999_07 16,341 43.99% 23.41% 24.47% 6.92% 1.21%1999_08 16,822 45.30% 23.55% 22.56% 7.23% 1.36%1999_09 19,753 40.95% 23.64% 25.16% 8.77% 1.49%1999_10 18,366 45.01% 23.24% 22.80% 7.60% 1.35%1999_11 17,513 45.71% 24.61% 21.18% 7.23% 1.27%DissFinal Page 154 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Year Average Daily Composition (%) and Traffic Month (vehicles) Car Small Medium Large Heavy1999_12 16,571 46.12% 25.19% 19.99% 7.41% 1.29%2000_01 16,797 45.83% 24.75% 21.30% 6.91% 1.21%2000_02 14,733 47.39% 23.97% 22.29% 5.23% 1.12%2000_03 18,001 45.91% 25.55% 20.15% 6.99% 1.40%2000_04 19,184 45.69% 25.50% 20.18% 7.28% 1.35%2000_05 18,204 47.74% 25.66% 19.31% 6.26% 1.03%2000_06 17,254 46.19% 25.68% 19.62% 7.02% 1.49%2000_07 17,453 45.99% 25.44% 20.01% 7.03% 1.53%2000_08 18,432 46.35% 25.67% 19.51% 7.04% 1.43%2000_09 20,042 44.72% 26.08% 20.18% 7.56% 1.46%2000_10 19,033 46.47% 25.69% 19.57% 6.96% 1.31%2000_11 18,963 45.79% 25.97% 19.48% 7.39% 1.37%2000_12 18,837 45.34% 26.09% 19.42% 7.71% 1.44%2001_01 18,508 47.24% 24.58% 21.27% 5.75% 1.16%2001_02 18,818 43.23% 25.16% 23.17% 7.10% 1.34%2001_03 20,005 45.63% 25.92% 19.53% 7.52% 1.40%2001_04 20,938 45.72% 25.84% 19.68% 7.37% 1.39%2001_05 20,730 45.86% 25.44% 20.04% 7.28% 1.38%2001_06 20,340 43.77% 25.61% 20.91% 8.01% 1.70%2001_07 20,886 42.28% 25.45% 22.16% 8.41% 1.70%2001_08 21,579 43.29% 25.27% 21.76% 8.09% 1.59%2001_09 23,789 41.90% 25.42% 22.30% 8.66% 1.71%2001_10 22,159 43.79% 24.97% 21.59% 7.94% 1.72%2001_11 22,698 43.87% 25.01% 20.85% 8.40% 1.87%2001_12 21,618 43.83% 25.00% 20.64% 8.66% 1.87%2002_01 21,569 43.46% 24.60% 21.37% 8.65% 1.92%2002_02 23,405 44.17% 23.34% 25.02% 5.84% 1.64%2002_03 24,926 42.31% 24.50% 22.14% 8.88% 2.16%2002_04 25,922 42.56% 24.78% 21.28% 9.08% 2.29%2002_05 24,598 43.97% 24.43% 20.88% 8.55% 2.17%2002_06 23,608 41.73% 24.48% 21.74% 9.62% 2.42%2002_07 24,804 41.79% 24.56% 22.30% 9.25% 2.10%2002_08 26,047 41.32% 24.61% 22.39% 9.60% 2.07%2002_09 28,275 40.54% 25.07% 22.38% 9.84% 2.17%2002_10 27,347 42.83% 24.26% 21.93% 9.15% 1.82%2002_11 27,030 41.72% 24.66% 21.85% 9.80% 1.96%2002_12 26,668 41.18% 25.57% 21.25% 9.88% 2.12%2003_01 31,203 40.74% 25.12% 23.20% 8.81% 2.13%2003_02 27,926 40.28% 24.35% 25.83% 7.61% 1.94%2003_03 30,389 39.19% 26.15% 21.85% 10.40% 2.44%2003_04 28,402 42.41% 25.69% 20.48% 9.32% 2.10%2003_05 16,865 48.29% 24.50% 17.75% 8.24% 1.23%2003_06 26,161 41.89% 25.95% 19.11% 10.49% 2.56%2003_07 31,025 40.92% 26.25% 20.10% 10.23% 2.50%2003_08 33,998 39.55% 26.14% 21.09% 10.64% 2.59%2003_09 37,802 36.99% 26.69% 21.69% 11.73% 2.89%2003_10 36,517 40.65% 25.20% 20.62% 10.87% 2.66%2003_11 36,081 39.70% 25.89% 20.45% 11.24% 2.72%2003_12 35,927 39.07% 25.97% 20.61% 11.33% 3.02%DissFinal Page 155 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Annual Average Daily Traffic thus derived: Year Total Car Small Medium Large+Heavy 1997 12,120 6,013 2,847 2,715 546 1998 13,964 6,333 3,237 3,389 1,006 1999 16,249 7,356 3,853 3,735 1,306 2000 18,087 8,335 4,621 3,623 1,508 2001 21,013 9,278 5,318 4,444 1,974 2002 25,355 10,711 6,237 5,582 2,826 2003 31,039 12,536 7,988 6,563 3,952Indices thus obtained from above data: Civil Vehicle Passenger Trip GDP Ownership Passenger-km Length 1997 100.0 100.0 100.0 100.0 1998 111.0 107.9 104.7 100.8 1999 122.2 122.9 109.9 102.2 2000 135.2 143.3 117.9 103.0 2001 149.0 167.6 135.4 114.4 2002 166.2 201.0 142.7 115.1 2003 188.8 253.4 153.6 115.6 Freight Trip Expressway Freight MT-km Length Traffic 1997 100.0 100.0 100.0 1998 103.6 104.2 115.2 1999 104.5 105.4 134.1 2000 112.6 112.3 149.2 2001 112.6 112.3 173.4 2002 115.0 116.0 209.2 2003 122.7 120.3 256.1Income elasticities thus calculated (1997 to 2003): Income elasticity of: Value Civil Vehicle Ownership 1.41 Passenger-km 0.69 Passenger Trip Length 0.23 Freight MT-km 0.33 Freight Trip Length 0.30 Total Expressway Traffic 1.43 Expressway Cars 1.14 Expressway Small 1.54 Expressway Medium 1.35 Expressway Large+Heavy 2.46DissFinal Page 156 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Graph of Jiangsu Province/ Shanghai-Nanjing Expessway data (indexed to 1997): 300.0 250.0 200.0 Index (1997=100) 150.0 100.0 50.0 0.0 97 98 99 00 01 02 03 19 19 19 20 20 20 20 Year GDP Civil Vehicle Ow nership Passenger-km Passenger Trip Length Freight MT-km Freight Trip Length Jiangsu Expressw ay TrafficDissFinal Page 157 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)3. Zhejiang Province and Shanghai-Hangzhou-Ningbo ExpresswaySocio-economic data from Zhejiang Statistical Yearbooks relating to Zhejiang Province as awhole:ZJSY99: Zhejiang Provincial Bureau of Statistics (1999)ZJSY00: Zhejiang Provincial Bureau of Statistics (2000)ZJSY02: Zhejiang Provincial Bureau of Statistics (2002)ZJSY04: Zhejiang Provincial Bureau of Statistics (2004) Year GDP current GDP growth rate year- Implied price Real GDP price on-year inflation year-on- growth year-on- (100m RMB) (comparable price) year (%) year (%) 1998 4,988 10.10% -2.3% 10.0% 1999 5,365 10.00% -2.2% 10.0% 2000 6,036 11.01% 1.4% 10.9% 2001 6,748 10.50% 1.2% 10.5% 2002 7,796 12.50% 2.7% 12.4% 2003 9,395 14.40% 5.3% 14.4% Source ZJSY04, p24 ZJSY04, p26 derived derived Year Civil Vehicle Ownership Source 1998 478,297 ZJSY99, p395 1999 575,882 ZJSY00, p379 2000 680,586 ZJSY02, p415 2001 855,642 ZJSY02, p415 2002 1,078,311 ZJSY04, p445 2003 1,358,209 ZJSY04, p445 Year Highway Passenger Highway Passenger- Average Highway Trips (10,000 people) km (100 million) Passenger Trip Length (km) 1998 111,847 436.34 39.01 1999 111,771 433.50 38.78 2000 116,996 449.51 38.42 2001 126,008 479.53 38.06 2002 128,980 519.20 40.25 2003 133,968 531.63 39.68 Source ZJSY04, p447 ZJSY04, p448 derivedDissFinal Page 158 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Year Highway Freight Highway MT-km Average Freight Trip Transport (10,000 MT) (100 million) Length (km) 1998 45,338 257.11 193.2 1999 45,754 256.90 155.3 2000 55,008 280.02 156.3 2001 55,706 282.53 141.7 2002 63,532 293.60 119.5 2003 70,907 313.70 106.0 Source ZJSY04, p449 ZJSY04, p450 derivedTraffic and revenue data for Shanghai-Hangzhou-Ningbo Expressway from ZhejiangExpressway Co. Ltd. (www.zjec.com.cn): Average Composition (%) Daily Year Revenue Average and (thousand Daily Traffic 0-2T 2-5T 5-10T 10-20T >20T Month RMB) (vehicles) (Small) (Medium) (Large) (Heavy) (Heavy)1998_01 9,881 1,196.8 61.62% 26.64% 11.07% 0.53% 0.15%1998_02 9,683 1,229.5 56.82% 29.53% 13.03% 0.47% 0.16%1998_03 11,096 1,413.0 55.24% 32.09% 11.93% 0.62% 0.11%1998_04 12,159 1,528.7 56.23% 31.89% 11.14% 0.63% 0.11%1998_05 11,485 1,439.2 56.60% 31.48% 11.22% 0.58% 0.12%1998_06 11,264 1,381.4 59.25% 29.83% 10.23% 0.56% 0.13%1998_07 11,004 1,355.9 59.83% 29.14% 10.29% 0.59% 0.15%1998_08 11,115 1,365.8 60.65% 28.19% 10.42% 0.60% 0.13%1998_09 12,448 1,530.1 60.90% 28.18% 10.14% 0.64% 0.14%1998_10 12,710 1,567.6 58.71% 30.14% 10.40% 0.64% 0.09%1998_11 13,028 1,615.9 57.35% 31.45% 10.37% 0.71% 0.11%1998_12 12,347 1,641.5 59.61% 29.83% 9.65% 0.77% 0.13%1999_01 12,559 2,168.0 60.42% 29.05% 9.50% 0.88% 0.15%1999_02 11,688 1,985.5 66.56% 23.66% 8.89% 0.77% 0.12%1999_03 13,687 2,433.4 60.93% 28.56% 9.58% 0.80% 0.14%1999_04 15,062 2,641.1 62.20% 27.75% 9.11% 0.80% 0.14%1999_05 14,474 2,531.5 62.58% 27.42% 9.18% 0.69% 0.14%1999_06 14,066 2,437.7 63.56% 26.60% 8.89% 0.81% 0.14%1999_07 14,546 2,674.6 62.62% 27.29% 9.05% 0.89% 0.15%1999_08 15,204 2,790.7 62.73% 26.96% 9.24% 0.93% 0.14%1999_09 16,610 2,955.6 61.80% 27.61% 9.47% 1.00% 0.12%1999_10 17,012 3,128.0 63.09% 26.34% 9.52% 0.94% 0.12%1999_11 16,744 3,101.3 62.85% 26.61% 9.39% 1.01% 0.14%1999_12 16,386 3,024.9 63.57% 26.18% 9.10% 1.01% 0.14%2000_01 17,125 3,145.2 63.99% 25.62% 9.22% 1.04% 0.14%2000_02 13,853 2,486.8 67.39% 21.88% 9.68% 0.91% 0.14%2000_03 18,082 3,322.5 63.67% 25.73% 9.37% 1.10% 0.13%2000_04 19,458 3,647.0 63.94% 25.26% 9.56% 1.10% 0.14%2000_05 19,061 3,522.4 65.77% 23.74% 9.37% 1.02% 0.11%2000_06 17,496 3,231.9 66.17% 23.33% 9.24% 1.13% 0.12%2000_07 17,058 3,165.0 65.98% 23.28% 9.44% 1.17% 0.13%2000_08 17,738 3,274.2 66.51% 23.08% 9.20% 1.09% 0.11%2000_09 18,750 3,478.4 66.18% 23.36% 9.21% 1.15% 0.11%DissFinal Page 159 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Average Composition (%) Daily Year Revenue Average and (thousand Daily Traffic 0-2T 2-5T 5-10T 10-20T >20T Month RMB) (vehicles) (Small) (Medium) (Large) (Heavy) (Heavy)2000_10 18,300 3,360.7 67.17% 22.34% 9.24% 1.17% 0.09%2000_11 18,155 3,339.9 67.28% 22.25% 9.15% 1.23% 0.09%2000_12 17,990 3,305.5 67.64% 21.92% 9.07% 1.26% 0.10%2001_01 17,290 3,090.3 71.79% 18.69% 8.25% 1.17% 0.10%2001_02 18,450 3,394.1 68.47% 20.87% 9.31% 1.25% 0.10%2001_03 20,557 3,755.9 68.91% 20.67% 8.98% 1.35% 0.09%2001_04 20,993 3,830.8 69.37% 20.02% 9.20% 1.33% 0.09%2001_05 20,776 3,735.8 70.95% 18.79% 9.01% 1.18% 0.08%2001_06 19,962 3,759.0 65.65% 22.72% 10.13% 1.39% 0.11%2001_07 19,520 3,819.4 60.84% 26.52% 11.01% 1.53% 0.10%2001_08 21,172 4,141.9 60.64% 26.67% 10.99% 1.60% 0.10%2001_09 22,666 4,467.8 59.69% 27.63% 11.01% 1.57% 0.09%2001_10 21,887 4,234.2 61.79% 25.87% 10.65% 1.62% 0.07%2001_11 22,219 4,312.9 61.06% 26.22% 10.80% 1.83% 0.08%2001_12 21,525 4,154.8 61.15% 26.20% 10.59% 1.97% 0.08%2002_01 21,804 4,179.9 60.53% 26.93% 10.49% 1.97% 0.08%2002_02 20,952 3,805.8 65.02% 24.59% 8.74% 1.58% 0.08%2002_03 24,830 4,801.9 58.42% 28.42% 11.24% 1.82% 0.10%2002_04 25,541 4,876.1 60.94% 26.29% 10.78% 1.86% 0.14%2002_05 24,900 4,678.9 62.91% 24.69% 10.43% 1.85% 0.12%2002_06 24,044 4,593.2 61.54% 25.25% 10.99% 2.09% 0.13%2002_07 24,595 4,707.2 61.64% 25.19% 10.86% 2.18% 0.13%2002_08 26,203 5,058.8 60.89% 25.21% 11.26% 2.48% 0.16%2002_09 27,471 5,354.3 60.41% 25.26% 11.46% 2.69% 0.18%2002_10 27,094 5,150.3 62.49% 24.12% 10.82% 2.42% 0.15%2002_11 26,840 5,161.4 62.23% 24.43% 10.05% 3.18% 0.11%2002_12 26,048 4,965.9 63.33% 23.74% 9.52% 3.35% 0.07%2003_01 26,036 4,835.7 65.55% 22.48% 8.73% 3.18% 0.06%2003_02 23,240 4,221.3 67.09% 21.40% 8.83% 2.62% 0.06%2003_03 27,286 5,145.5 63.70% 23.32% 9.68% 3.25% 0.06%2003_04 27,003 5,082.0 63.72% 23.41% 9.48% 3.34% 0.06%2003_05 21,253 4,053.4 62.32% 24.30% 9.71% 3.59% 0.08%2003_06 26,471 4,876.4 66.11% 21.52% 9.09% 3.18% 0.10%2003_07 28,190 5,244.6 65.90% 21.44% 9.35% 3.20% 0.10%2003_08 29,405 5,488.5 65.60% 21.56% 9.49% 3.24% 0.11%2003_09 31,370 5,904.1 65.09% 21.75% 9.70% 3.34% 0.11%2003_10 32,198 5,994.5 65.80% 21.10% 9.86% 3.12% 0.12%2003_11 30,790 5,824.6 64.82% 21.55% 10.03% 3.46% 0.13%2003_12 31,712 6,013.6 65.20% 21.16% 9.41% 3.71% 0.14%DissFinal Page 160 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Annual Average Daily Traffic and Revenue thus derived: Revenue Year (RMB) Total Small Medium Large Heavy 1998 1,439,675 11,525 6,750 3,449 1,240 86 1999 2,660,135 14,854 9,312 4,015 1,375 152 2000 3,275,939 17,769 11,716 4,180 1,653 219 2001 3,893,291 20,593 13,345 4,853 2,068 327 2002 4,783,404 25,051 15,442 6,343 2,653 612 2003 5,229,751 27,930 18,189 6,150 2,646 944Indices thus obtained from above data: Civil Vehicle Passenger Trip GDP Ownership Passenger-km Length 1998 100.0 100.0 100.0 100.0 1999 110.0 120.4 99.3 99.4 2000 122.1 142.3 103.0 98.5 2001 134.9 178.9 109.9 97.5 2002 151.8 225.4 119.0 103.2 2003 173.6 284.0 121.8 101.7 Freight Trip Expressway Expressway Freight MT-km Length Traffic Revenue 1997 100.0 100.0 100.0 100.0 1998 100.9 99.9 128.9 184.8 1999 121.3 108.9 154.2 227.5 2000 122.9 109.9 178.7 270.4 2001 140.1 114.2 217.4 332.3 2002 156.4 122.0 242.3 363.3 2003 100.0 100.0 100.0 100.0Income elasticities thus calculated (1997 to 2003): Income elasticity of: Value Civil Vehicle Ownership 1.78 Passenger-km 0.37 Passenger Trip Length 0.03 Freight MT-km 0.82 Freight Trip Length 0.37 Zhejiang Expressway Traffic 1.54 Zhejiang Expressway Revenue 2.11 Expressway Traffic Small 1.70 Expressway Traffic Medium 1.05 Expressway Traffic Large 1.34 Expressway Traffic Heavy 3.10DissFinal Page 161 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Graph of Zhejiang Province/ Shanghai-Hangzhou-Ningbo Expressway data (indexed to1998): 400.0 350.0 300.0 Index (1998=100) 250.0 200.0 150.0 100.0 50.0 0.0 98 99 00 01 02 03 19 19 20 20 20 20 Year GDP Civil Vehicle Ow nership Passenger-km Passenger Trip Length Freight MT-km Freight Trip Length Zhejiang Expressw ay Traffic Zhejiang Expressw ay RevenueDissFinal Page 162 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 13: Survey Questionnaire: Question Specification and Logical Flowwww.SurveyMonkey.com was used to prepare the questionnaire and undertake thesurvey. The survey is reproduced page-by-page as follows:Page 1 IntroductionFor ALL RespondentsFirst of all, thank you for taking part in this survey. In most cases I hope this should take nomore than 10-15 minutes of your time.You are free to complete this survey on an anonymous basis. However, if you would bewilling to let me know a little bit more about you, you may like to complete some (or all) ofthe questions on this first page. But if you would rather remain anonymous, feel free to skipthese questions...Should you have any problems completing this survey, or wish to make a comment where abox for optional comments is not provided, please do email me at: rfdibona@yahoo.comFinally, if you have any colleagues who might be appropriate respondents to this survey,please feel free to pass them the survey details.1 Your name: (text response) Optional2 Your organisation: (text response) Optional3 Your position/ role: (text response) Optional4 Your email: (text response) Optional5 Telephone: (text response) Optional6 Would you be happy for me to contact you YES/ NO Optional directly for further discussions?DissFinal Page 163 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Page 2 Your Sector of ExpertiseFor ALL Respondents7 Please select which sectors you have worked in: (multiple choice from Mandatory menu)  Expressway developer/ operator/ equity investor  Lawyer/ Attorney/ Solicitor  Private Sector Lender (i.e. lending own/ employer’s money)  Investment Banker  Ratings Agency (e.g. Fitch, Moodys, Standard & Poor’s)  Accountant/ Valuer  Insurer  Transport Planning Consultant  Economist  Civil/ Structural/ Pavement/ Highway Engineer/ Architect  Government  Aid-agency (e.g. ADB, World Bank, JICA, etc)  Academic  Other (please specify)8 Approximately how many years’ working (text response) Mandatory experience do you have?9 What percentage of this time has been spent on: (rating scale) Mandatory (please answer for each row; as some categories One answer per row, from: overlap the total time across all rows may exceed 100%)  Transport infrastructure projects 0%  All infrastructure projects (transport & non-transport) 1%-10%  Projects in developing economies 11%-25% 26%-50%  Tolled highway projects (urban and/or rural, anywhere in world) 51%-75%  Rural or inter-urban tolled highway projects 76%-95%  Rural or inter-urban tolled highway projects in developing 96%-100% economiesDissFinal Page 164 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Page 3 Your International ExperienceFor ALL Respondents10 In which parts of the world have you worked (select all that apply) Optional on projects?  North America (USA/ Canada)  Central/ South America/ Caribbean  Western Europe  Eastern Europe  Africa  Middle East  Central Asia  South Asia  East Asia  Oceania/ Australasia  Other (please specify)11 Have you worked on projects in East Asia? (select all that apply) Optional Multiple answers per row, from:  Brunei  Tolled Highways  Cambodia  Other Transport Projects  Mainland China (i.e. excluding Hong  Other Infrastructure Projects Kong, Macau, Taiwan)  Non-Infrastructure Projects  Hong Kong  Indonesia  Japan  North Korea  South Korea  Laos  Macau  Malaysia  Mongolia  Myanmar (Burma)  Philippines  Singapore  Taiwan  Thailand  Timor-Leste (East Timor)  VietnamDissFinal Page 165 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Page 4 Socio-Economic RisksFor ALL Respondents12 Based on your overall experience in infrastructure (rating scale) Mandatory projects, with an emphasis on transport projects and particularly tolled highways (if applicable), please rate the importance of the following risks: One answer per row, from:  The prevailing political system and its stability  Critical  The legal system  Strong  Currency (exchange) risks impact  Ease of repatriating profits  Important  Interest rates  Limited  Price inflation impact  Income (in)equality  Not usually  Economic growth considered  Business cycles (as distinct from recent growth)  Not sure  Drivers’ familiarity with highway tolls  Corruption13 Similarly, please rate the importance of the following (rating scale) Mandatory to a project’s likely performance/ riskiness: One answer per row, from:  The project’s social/ economic benefits  Critical  Guanxi/ the importance of business connections  Strong impact  The project’s overall legal/ contractual foundations  Important  The length of the operating concession  Limited impact  Construction time/ risk of delayed opening  Not usually  Construction cost/ risk of cost over-run considered  Reliability of operating and maintenance cost estimates  Not sure  The enforceability of toll/ tariff increases  Minimum income guarantees and their enforceability  The threat of competing routes/ alternatives to the project  Standard of connecting routes  Toll affordability for large vehicles (e.g. large trucks/ goods vehicles)  Toll affordability for other vehicles  Toll leakage/ evasion  Ramp up length (i.e. the time taken for drivers to familiarise themselves with the benefits of a new tolled highway)14 Have you experience of using, undertaking or (choice) Mandatory reviewing traffic and/or revenue forecasts undertaken by transport consultants/ economists? (a) Yes, I have prepared forecasts myself (b) I have supervised forecasts made by my staff, but have not prepared them (c) Yes, I have reviewed or used forecasts undertaken by others, but have not prepared them (d) I have worked alongside transport consultants, but have not used their forecasts (e) No, none of the aboveDissFinal Page 166 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Page 5 Data Availability for ModellingOnly for those respondents replying (a), (b) or (c) to Question 14.15 From your experience, do you feel there are sufficient (rating scale) Mandatory data to successfully calibrate/ validate models? One answer per row, from:  In developed countries, there are sufficient data available  Always  In developed countries, data are reliable  Usually  In developing countries, there are sufficient data available  Sometimes  In developing countries, data are reliable  Rarely  Never  Not sure16 From your experience, do you feel there are sufficient (rating scale) Mandatory data to successfully prepare meaningful traffic and revenue forecasts? One answer per row, from:  In developed countries, there are sufficient data available  Always  In developed countries, data (e.g. land use/economic forecasts)  Usually are reliable  Sometimes  In developing countries, there are sufficient data available  Rarely  In developing countries, data (e.g. land use/economic  Never forecasts) are reliable  Not sure17 If you would like to be more specific regarding (text response) Optional particular problems with data collection, its quality, etc, either with regards specific parameters (e.g. traffic counts, Values of Time, GDP forecasts, etc) or with regards specific countries which are especially good or bad for data availability/ reliability, you may comment below:DissFinal Page 167 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Page 6 Transport Modelling IssuesOnly for those respondents replying (a), (b) or (c) to Question 14.18 Regarding the applicability of full four-stage models (rating scale) Mandatory (i.e. including trip generation, distribution and mode split in addition to assignment), please indicate how strongly you agree or disagree with the following statements: One answer per row, from:  Such models are reliable  Strongly  Such models are too data hungry to be relied upon agree  Such models are too complicated to be of worth  Agree  Such models are not suitable for toll-road work  Neutral  Economic uncertainties/ pace of change makes them irrelevant opinion in developing economies  Disagree  Such models are too much of a black box for non-specialists to  Strongly properly critique model outputs disagree  Not sure19 Regarding network assignment models (e.g. based on (rating scale) Mandatory traffic counts and/or Origin-Destination surveys), but NOT full four-stage models, please indicate how strongly you agree or disagree with the following statements: One answer per row, from:  Such models are reliable  Strongly  Such models are too data hungry to be relied upon agree  Such models are too simplistic to be relied upon  Agree  Such models are too complicated to be of worth  Neutral  Such models are not suitable for toll-road work opinion  Economic uncertainties/ pace of change makes them irrelevant  Disagree in developing economies  Strongly  Such models are too much of a black box for non-specialists to disagree properly critique model outputs  Not sure20 Regarding spreadsheet-based traffic/ revenue models, (rating scale) Mandatory please indicate how strongly you agree or disagree with the following statements: One answer per row, from:  Such models are reliable  Strongly  Such models are too data hungry to be relied upon agree  Such models are too simplified to be of worth  Agree  Such models are not suitable for toll-road work  Neutral  Economic uncertainties/ pace of change makes them irrelevant opinion in developing economies  Disagree  Such models are too simplistic to provide meaningful outputs  Strongly disagree  Not sure21 If you have any specific comments on issues with (text response) Optional developing/ calibrated/ forecasting with traffic and revenue models (of any type), please give your comments here:DissFinal Page 168 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Page 7 General Reliability of Traffic ForecastsOnly for those respondents replying (a), (b), (c) or (d) to Question 14.22 From your experience, please state how often you feel (rating scale) Mandatory each of the following statements is true: One answer per row, from:  How often do projects significantly exceed forecast traffic/  Very often revenue levels?  Quite often  How often do projects fall well short of forecast traffic/  Sometimes revenue levels?  Rarely  Do you believe that transport planners are pressured by clients  Never to adjust forecasts to meet their expectations?  Not sure  Do you believe that transport consultants’ forecasts are higher if they are engaged by equity- rather than debt-side clients?Page 8 Project Evaluation CriteriaFor ALL Respondents23 How often do you explicitly consider the following in (rating scale) Mandatory appraising tolled highways? One answer per row, from:  Traffic forecasts: Base and/or Central Case  Always  Traffic forecasts: Optimistic and/or High Case  Usually  Traffic forecasts: Conservative and/or Low Case  Sometimes  Revenue forecasts: Base and/or Central Case  Rarely  Revenue forecasts: Optimistic and/or High Case  Never  Revenue forecasts: Conservative and/or Low Case  Not sure  Congestion on competing/ alternative routes  Congestion on link-roads/ feeder routes  Capacity of the highway being considered24 How often do you explicitly consider the following (rating scale) Mandatory criteria in appraising infrastructure projects (with an emphasis on tolled highways if applicable)? One answer per row, from:  Net Present Value (NPV)  Always  Financial Internal Rate of Return (FIRR)  Usually  Economic Internal Rate of Return (EIRR; including social  Sometimes impacts)  Rarely  Social Cost/ Benefit Ratios  Never  Risk correlation versus other projects in company’s/ client’s  Not sure portfolio  Counterparty risks: can partners contribute equity/ debt  Sovereign/ Institutional other country/ legal risks25 If you use any financial ratios when appraising (text response) Optional projects, please state which ratios you normally use:DissFinal Page 169 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Page 9 Future ProspectsFor ALL Respondents26 How would you rate the potential for inter-urban (rating scale) Mandatory tolled highways over the next 10 years in each of the following countries: One answer per row, from:  Cambodia  Sector overdeveloped (few  Mainland China (i.e. excluding Hong Kong, prospects) Macau, Taiwan)  Maturing market (decline)  Indonesia  Already strong and likely to  Laos remain so (steady)  Malaysia  Fast developing (growing)  Myanmar (Burma)  Only just starting (nascent)  Philippines  Undeveloped and negligible  Thailand prospects (no market)  Vietnam  Not sure27 Comparing the next 10 years (2006-2016) with the (rating scale) Mandatory last 5 years (2001-2006), how do you feel each the following will change: One answer per row, from:  Fuel prices  Will be significantly greater  General price inflation  Will increase to an extent  Interest rates  No significant change  Economic growth  Will decrease to an extent  Exchange rate volatility  Will significantly decrease  Acceptability of road tolls and toll increases  Not sure28 If you believe that there will be any other significant (text response) Optional changes to factors affecting toll road performance, please state which factors and how you feel they will change below. Similarly, if you feel that patterns will be markedly different between certain economies, please explain below (citing which countries may have above-trend growth in which variables, and which countries you feel will have below-trend changes):Page 10 And finally…For ALL Respondents29 Finally, if you have any other comments you would (text response) Optional like to make, either about issues in project finance/ transport forecasting, or about this survey, please let me have your thoughts. Thank you.30 If you would like information about the survey (choose from Optional responses, once collated, or about my broader menu) research, please indicate below:  No, thank you.  Yes, regarding survey results.  Yes, regarding your research.DissFinal Page 170 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 14: Amendments Made to Questionnaire Following Pilot SurveyA number of useful comments were made during the piloting of the questionnaire.Some comments could be readily incorporated through amending the wording of aquestion. In other instances a multi-choice rather than single-choice response wasimplemented. In one case a question was split into two separate questions. The Questionnumber references given refer to the question numbers in the Final Survey (as shown inAppendix 13).Question 7: changed from single-choice to multi-choice, following feedback from thosewho have developed their career through consultancy and academia and/or the publicsector and/or aid agencies.Question 9: following a comment received, the question was clarified through theaddition of the text: “(please answer for each row; as some categories overlap the totaltime across all rows may exceed 100%)”Question 11: “Timor-Leste” changed to “Timor-Leste (East Timor)” to provide greaterclarity; this a result of the Author’s own review of questions.Questions 18 and 19: these were originally a single question, referring to networkassignment models and software. An initial comment was received via the pilot,pointing out that approaches ought to be largely independent of software platforms. Thiscomment initially suggested use of the term “Four Stage Model” in lieu of softwareplatforms. However, subsequent consideration and discussion with the originator of thespecific comment (by email) and with another respondent (by telephone) resulted in theinitial question being subdivided into two, as now shown. Section 2.9 cited practicaldifficulties of using Four Stage models; hence the subdivision was into a questionDissFinal Page 171 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)regarding Four Stage models, and a question regarding network assignment modelsoutwith Four Stage models.Question 22: one respondent pointed out in their comments that it would be “utterlywrong" if equity- and debt-side forecasts were the same, given the different risk/ rewardprofiles of the different perspectives. It was clarified that the purpose of this questionwas to ascertain the extent to which practitioners are aware of these differences.Question 23: one respondent picked-out the difference between Base and Central cases,citing Central as the most probable outcome (50% cumulative probability) with Baseusually lower than Central. However, from experience the two terms are often usedinterchangeably, hence “Base or Central” was replaced with “Base and/or Central”;similar changes in wording (i.e. “or” to “and/or”) were made to High/ Optimistic andLow/ Conservative.Question 24: one respondent claimed this question was ambiguous, with attention tofinancial returns, social returns or portfolio-based risk-spreading being determined bywhom one is working for. This ambiguity was at least semi-intentional; the aim being tosee how often any group considers which set of objectives. Depending on eventualsurvey returns, the intention being to see if one group are inherently more interested inone set of factors than another. (A priori, government and aid agencies ought to be moreinterested in social returns, the private sector in financial returns and possibly in risk-spreading also.)General Comments #1: one respondent requested the ability to review all questionsbefore answering. However, within the context of the software used (which includessome logical branching) this is not feasible. A possible work-around would be tocomplete the survey and then return to the beginning to revise answers. But it was feltDissFinal Page 172 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)that the survey introduction would get overly complicated to include this possibility inthe preamble.General Comments #2: following subsequent discussions with respondents, aside fromspecific question-related comments (given above) no major areas appeared to have beenmissed out.General Comments #3: the time take to complete appeared to have been around 15minutes by experienced transport planners/ modellers (the group likely to have thelongest response time as they need to answer all questions). This was felt to be a littlebit long by some. Whilst a paper-based approach might be quicker for respondents,dissemination and return of results would become an issue (and so would increase thelikely response time once printing off and faxing back, etc were included). Also, therewere no obvious candidate questions to be omitted. Thus, whilst a 10-minute responsetime might be preferable, it might not be attainable by those answering questions onmodelling. However, for those without hands-on modelling experience, a 10-minuteresponse ought to be feasible, hence the preamble was revised from “no more than 15minutes of your time” to “no more than 10-15 minutes of your time.”General Comments #4: one respondent suggested that a distinction between transportinfrastructure projects and infrastructure in general was not clear; that they cross-relateto a great extent. Indeed, this is one of the rationales of the survey and respondenttargeting, that there are often substantial similarities. However, where appropriate theintention remains for the respondent to concentrate on transport projects, should theyhave such experience (and broader infrastructure experience where they do not).Conversely, one respondent (non-transport planner) who successfully completed thesurvey commented that he was unable to comment further as he was not an expert in theDissFinal Page 173 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)transport field, notwithstanding his response to question 28 “Fuel cost to have a impacton efficiency in routing. Possible slow-down in road projects in more mature marketswhere mass transport may be considered more appropriate going forward.”General Comments #5: some respondents reported problems on pages with questionscontaining mandatory answers. As such, logical control on giving mandatory answerswas over-ridden (enabling mandatory answers to be skipped in theory). The exceptionwas Question 14, where an answer was required to determine which, if any of questions15 to 22 the respondent would be presented with.DissFinal Page 174 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 15: Questionnaire ResponsesQuestions 1 to 6Relate to respondent identity; no “results” presented due to confidentialityconsiderations.Question 7: Please select which sectors you have worked in (159 responses) Sector Number Percent Expressway developer/ operator/ equity investor 15 9.4% Lawyer/ Attorney/ Solicitor 5 3.1% Private Sector Lender (i.e. lending own/ employer’s money) 2 1.3% Investment Banker 10 6.3% Ratings Agency (e.g. Fitch, Moodys, Standard & Poor’s) 4 2.5% Accountant/ Valuer 3 1.9% Insurer 1 0.6% Transport Planning Consultant 95 59.7% Economist 22 13.8% Civil/ Structural/ Pavement/ Highway Engineer/ Architect 37 23.3% Government 37 23.3% Aid-agency (e.g. ADB, World Bank, JICA, etc) 9 5.7% Academic 22 13.8% Other 24 15.1%Total (as multiple selections possible, total may exceed 100%) 286 179.9%These sectors were then aggregated into 6 groups to permit meaningful analysis ofdifferent perceptions by stakeholder types, as follows: Aggregated Sectors Number Percent Financial, Legal, Operator 29 18.2% Transport Planner/ Economist 98 61.6% Civil/ Structural/ Pavement/ Highway Engineer/ Architect 37 23.3% Government/ Aid Agency 43 27.0% Academic 22 13.8% Other 24 15.1%Total (as multiple selections possible, total may exceed 100%) 253 159.1%Question 8: Approximately how many years’ working experience do you have?(162 responses) Number of Years Responses Percent 30 or more years 42 25.9% 20 to 29 years 49 30.2% 10 to 19 years 50 30.9% 5 to 9 years 10 6.2% 1 to 4 years 11 6.8% Mean number of years 20.6 Standard deviation 10.9DissFinal Page 175 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 9: What percentage of this time has been spent on… (162 responses) 0% 1-10% 11-25% 26-50% 51-75% 76-95% 96-100%Transport infrastructure 10 22 29 19 26 38 18 projectsAll infrastructure projects (transport & non- 5 14 18 21 30 45 29 transport) Projects in developing 42 28 24 21 20 17 10 economies Tolled highway projects (urban and/or rural, 49 54 30 18 6 4 1 anywhere in world) Rural or inter-urban 69 51 24 11 2 4 1 tolled highway projects Rural or inter-urbantolled highway projects in 88 50 13 7 1 2 1 developing economiesAssuming mid-range values (e.g. 5.5% for 1-10%) the following means and standarddeviations were calculated: Mean Standard Deviation Transport infrastructure projects 49.4% 35.8% All infrastructure projects (transport & non- 60.3% 34.9% transport) Projects in developing economies 31.3% 35.6% Tolled highway projects (urban and/or rural, 14.3% 22.5% anywhere in world) Rural or inter-urban tolled highway projects 10.3% 22.6% Rural or inter-urban tolled highway projects in 6.7% 21.1% developing economiesCombining the percentages of time spent on each kind of work with number of years ofworking experience, the following estimates were obtained of years per kind of work: Mean 30+ 20-29 10-19 5-9 1-4 0Transport infrastructure 10.66 6 21 38 29 58 7 projectsAll infrastructure projects (transport & non- 13.13 8 28 48 30 43 2 transport) Projects in developing 7.26 5 13 21 24 57 39 economies Tolled highway projects (urban and/or rural, 2.57 0 0 9 20 84 46 anywhere in world) Rural or inter-urban 1.70 0 0 4 10 79 66 tolled highway projects Rural or inter-urbantolled highway projects in 1.12 0 0 2 7 65 85 developing economiesDissFinal Page 176 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 10: In which parts of the world have you worked on projects? (150responses) North America (USA/ Canada) 36 22.8% Central/ South America/ Caribbean 34 21.5% Western Europe 80 50.6% Eastern Europe 36 22.8% Africa 40 25.3% Middle East 46 29.1% Central Asia 21 13.3% South Asia 67 42.4% East Asia 102 64.6% Oceania/ Australasia 47 29.7% Other (please specify) 6 3.8%Question 11: Have you worked on projects in East Asia? Other Other Non- Anything Tolled Transport Infrastructure Infrastructure in this Highways Projects Projects Projects Country* Brunei 0 1 1 3 5 Cambodia 1 18 9 5 20 China 38 49 29 25 69 Hong Kong 29 55 32 25 69 Indonesia 16 30 9 10 43 Japan 5 6 4 4 11 North Korea 0 4 2 1 4 South Korea 11 17 4 6 28 Laos 2 14 8 7 20 Macau 2 12 7 4 19 Malaysia 19 31 12 13 42 Mongolia 0 4 3 2 7 Myanmar 0 4 1 0 5 Philippines 19 30 13 13 44 Singapore 8 32 12 11 39 Taiwan 2 17 5 5 25 Thailand 22 39 18 12 50 Timor-Leste 0 2 0 0 2 Vietnam 8 21 8 14 36Note: * The last column can be smaller than the total of the previous four, as arespondent may have worked on a number of different kinds of project within onecountry/ territory.Note: Countries being explicitly considered under this Dissertation are shown in bold.DissFinal Page 177 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 12: Based on your overall experience in infrastructure projects, with anemphasis on transport projects and particularly tolled highways (if applicable),please rate the importance of the following risks:Excluding “Not Sure” and null responses: 1 2 3 4 5 Strong Limited Not Usually Critical Impact Important Impact ConsideredThe prevailing political system and 49 56 31 7 4 its stability The legal system 22 62 48 10 2 Currency (exchange) risks 9 30 64 24 12 Ease of repatriating profits 10 35 63 14 10 Interest rates 7 22 77 22 8 Price inflation 5 30 68 28 7 Income (in)equality 3 11 46 51 26 Economic growth 17 46 66 13 2 Business cycles (as distinct from 3 19 48 46 15 recent growth)Drivers’ familiarity with highway 2 22 44 47 22 tolls Corruption 23 31 42 21 17Using the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation The prevailing political system and its stability 147 2.05 0.99 The legal system 144 2.36 0.87 Currency (exchange) risks 139 3.00 1.00 Ease of repatriating profits 132 2.84 0.98 Interest rates 136 3.01 0.87 Price inflation 138 3.01 0.88 Income (in)equality 137 3.63 0.95 Economic growth 144 2.56 0.86 Business cycles (as distinct from recent growth) 131 3.39 0.95 Drivers’ familiarity with highway tolls 137 3.47 0.99 Corruption 134 2.84 1.25DissFinal Page 178 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)The mean scores from above were also compared with mean scores based on the 6aggregated experience sectors from Question 7, as shown below: All FLO TpEc E&A G&A Acad Oth Political system 2.05 1.59 2.08 2.00 2.13 1.78 2.24 Legal system 2.36 1.69 2.43 2.24 2.24 2.06 2.62 Currency risks 3.00 2.72 3.06 2.62 3.20 3.12 2.95 Repatriating profits 2.84 2.45 2.87 2.64 3.00 2.71 3.05 Interest rates 3.01 2.72 3.02 2.97 3.03 2.94 3.22 Price inflation 3.01 2.83 3.06 3.09 3.00 3.29 3.05 Income (in)equality 3.63 3.69 3.76 3.52 3.43 3.41 3.50 Economic growth 2.56 2.54 2.54 2.69 2.68 2.76 2.38 Business cycles 3.39 3.17 3.48 3.39 3.44 3.38 3.11 Toll familiarity 3.47 3.28 3.59 3.50 3.42 3.17 3.47 Corruption 2.84 2.34 2.99 2.44 2.76 2.75 2.80Key: FLO = Financial, Legal, Operator TpEc = Transport Planners and Economists E&A = Engineers and Architects G&A = Government and Aid Agencies Acad = Academics Oth = OthersQuestion 13: Similarly, please rate the importance of the following to a project’slikely performance/ riskiness:Excluding “Not Sure” and null responses: 1 2 3 4 5 Strong Limited Not Usually Critical Impact Important Impact Considered The project’s social/ economic 20 49 59 17 2 benefitsGuanxi/ the importance of business 9 40 59 25 3 connections The project’s overall legal/ 33 45 58 8 1 contractual foundations The length of the operating 15 46 62 14 2 concession Construction time risk 15 56 52 19 1Construction cost/ risk of over-run 22 58 53 9 2 Reliability of operating and 13 43 65 22 2 maintenance cost estimates The enforceability of toll/ tariff 27 45 45 12 8 increasesMinimum income guarantees and 14 35 57 19 9 their enforceability The threat of competing routes/ 26 47 51 14 5 alternatives to the project Standard of connecting routes 7 44 73 15 3Toll affordability for large vehicles 12 37 68 12 6Toll affordability for other vehicles 13 40 57 19 8 Toll leakage/ evasion 13 28 58 30 7 Ramp up length 3 27 53 39 12DissFinal Page 179 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Using the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation The project’s social/ economic benefits 147 2.54 0.91 Guanxi/ the importance of business connections 136 2.80 0.89The project’s overall legal/ contractual foundations 145 2.30 0.90 The length of the operating concession 139 2.58 0.86 Construction time risk 143 2.55 0.87 Construction cost/ risk of over-run 144 2.38 0.87 Operating and maintenance cost estimates 145 2.70 0.88 The enforceability of toll/ tariff increases 137 2.48 1.08 Enforceability of minimum income guarantees 134 2.81 1.03 The threat of competing routes 143 2.48 1.01 Standard of connecting routes 142 2.74 0.79 Toll affordability for large vehicles 135 2.73 0.91 Toll affordability for other vehicles 137 2.77 1.00 Toll leakage/ evasion 136 2.93 1.00 Ramp up length 134 3.22 0.94The mean scores from above were also compared with mean scores based on the 6aggregated experience sectors from Question 7, as shown below: All FLO TpEc E&A G&A Acad Oth Social/ economic benefits 2.54 2.71 2.63 2.38 2.24 2.58 2.57 Guanxi 2.80 2.68 2.90 2.73 2.91 2.50 2.80 Legal/ contractual 2.30 1.90 2.35 2.18 2.34 2.44 2.57 foundationsOperating concession length 2.58 2.48 2.62 2.81 2.62 2.44 2.55 Construction time 2.55 2.41 2.58 2.71 2.63 2.35 2.38 Construction cost 2.38 2.14 2.46 2.36 2.28 2.17 2.43Operating and maintenance 2.70 2.52 2.79 2.76 2.64 2.76 2.48 costs Toll/ tariff increase 2.48 1.83 2.44 2.45 2.57 2.19 2.70 enforceabilityMinimum income guarantee 2.81 2.31 2.81 2.83 2.94 2.88 2.67 enforceabilityThreat of competing routes 2.48 2.14 2.43 2.36 2.68 2.11 2.70 Standard of connecting 2.74 2.41 2.70 2.88 2.89 2.44 3.00 routesToll affordability for large 2.73 2.41 2.71 2.83 2.94 2.71 2.85 vehiclesToll affordability for other 2.77 2.21 2.79 2.80 2.94 2.47 3.05 vehicles Toll leakage/ evasion 2.93 2.45 2.98 2.70 3.06 2.65 2.95 Ramp up length 3.22 2.93 3.21 3.35 3.41 3.12 3.44DissFinal Page 180 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 14: Have you experience of using, undertaking or reviewing trafficand/or revenue forecasts undertaken by transport consultants/ economists? (156responses) Yes, I have prepared forecasts myself 61 39%I have supervised forecasts made by my staff, but have 12 8% not prepared them Yes, I have reviewed or used forecasts undertaken by 44 28% others, but have not prepared them I have worked alongside transport consultants, but 16 10% have not used their forecasts No, none of the above 23 15%Question 15: From your experience, do you feel there are sufficient data tosuccessfully calibrate/ validate models?Excluding “Not Sure” and null responses: 1 2 3 4 5 Always Usually Sometimes Rarely Never In developed countries, there are 9 65 29 6 1 sufficient data available In developed countries, data are 2 53 48 6 0 reliable In developing countries, there are 1 9 36 52 2 sufficient data available In developing countries, data are 1 8 37 47 6 reliableUsing the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation In developed countries, there are sufficient data 110 2.32 0.74 available In developed countries, data are reliable 109 2.53 0.63 In developing countries, there are sufficient data 100 3.45 0.73 available In developing countries, data are reliable 99 3.49 0.77DissFinal Page 181 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 16: From your experience, do you feel there are sufficient data tosuccessfully prepare meaningful traffic and revenue forecasts?Excluding “Not Sure” and null responses: 1 2 3 4 5 Always Usually Sometimes Rarely Never In developed countries, there are 8 69 28 4 0 sufficient data available In developed countries, data (e.g. land use/economic forecasts) are 6 55 41 4 1 reliable In developing countries, there are 1 11 51 34 2 sufficient data available In developing countries, data (e.g. land use/economic forecasts) are 1 9 42 43 4 reliableUsing the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation In developed countries, there are sufficient data 109 2.26 0.64 available In developed countries, data (e.g. land 107 2.43 0.70 use/economic forecasts) are reliable In developing countries, there are sufficient data 99 3.25 0.72 available In developing countries, data (e.g. land 99 3.40 0.75 use/economic forecasts) are reliableDissFinal Page 182 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 17: If you would like to be more specific regarding particular problemswith data collection, its quality, etc, either with regards specific parameters (e.g.traffic counts, Values of Time, GDP forecasts, etc) or with regards specificcountries which are especially good or bad for data availability/ reliability, youmay comment below: (27 responses) 1 In developing countries, each project is an unique experience. Even in cities where there are existing models the zonal system is often not refined (small) enough to prepare highly reliable forecasts at the micro level. 2 Developing country work normally requires one to develop his own data, adding to costs. Some countries are better than others. Malaysia, Singapore, South Korea, Hong Kong are particularly good. Indonesia, China, Vietnam are notoriously bad. Thailand is in between. 3 The consultancies with which I have worked have had repeated trouble in obtaining even the most basic data in The Peoples Republic of China 4 Most of my response relates to a tolled highway in Vietnam where the client was a Korean company seeking to get substantial land options in return. They were knowlingly going into a very risky market - until they went bankrupt. Existing data is often unreliable, but, with some effort, it is possible to collect reliable data 5 I have encountered several issues in this regard: - Validity of data - Institutional ability to keep data up to date - Local consultant capability - Excessively high combined requirement for local consultants on projects funded by IFIs or bilateralagencies, creating extremely difficult conditions for project implementation 6 Quality of work depends largely on time and budget made available to consultants to gather and build up meaningful databases of traffic info. 7 In my experience there is never sufficient reliable data to answer all the questions expected of the traffic & revenue forecasts. Hence there are many judgements required many of which are based more on gut feel than real local data. 8 My involvement in transport projects is from an equality/inclusion perspective and the impact of providing an inclusive transport system is never adequately considered - we do not even have a clear understanding of what needs to be measured, let alone how to measure it, in anything other than anecdotal terms. 9 Countries with structured and consistent methods and systems to collect data tend to provide more useful inputs versus those without. However, even with high quality inputs and successful calibration of models, the forecasted outputs in developed countries are unreliable for greenfield projects and also for established projects that seek to maximize revenue, therefore putting into question the validity of the models in the first place. 10 There are significant differences in these values provided by experts in those areas. 11 (In Australia) Land Use Data: Existing Population data is generally available and reliable, forecasts can be questionable. Employment data is generally either not available or unreliable. Traffic Count Data: Mostly available, if not available fairly easy to get/commission counts. 12 The problem in Australia (a developed contry) is the low resource base, the low value placed by public authorities on good data, and the patchy nature of what is availabe - eg very good on the journey to work, very bad on freight. 13 The Commitment of the Government to Infrastructure Building and the market reactions are important contributors . 14 different sources 15 I question the validity of SP surveys (i.e. VOT) in developing countries, as their economies are so much more volatile than in developed countries. 16 Diff reqmts for diff types of road. Inter-urban much easier than urban, upgrade existing roads much easier than new routes. Network assignment type models only really necessary for urban roads otherwise simple, transparent spreadsheets generally moreDissFinal Page 183 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) useful. Traffic counts are quick and cheap to conduct, shouldnt be a data problem. Economic growth forecasts are always highly uncertain. Values of time are dealt with v badly within models anyway, single VofT used when really broad spectrum of values which is crucial to understanding how tolls divert traffic 17 Private transit operators may not have as much information in written form, may have more flexible/relaxed operating policies, and may be reluctant to share data. 18 to develop Intelligent Transportation Sysytems in different cities, and develop good data fusion and relevant algorithms as well. 19 Greater consistency of approach established in developed countries (more agreement regarding methodology). 20 Value of time data is usually insufficient, particularly as different people, on different trips, can have widely divergent values of time. 21 With regard to transportation demand, in growing regions timeliness is critical; as is the need for panel data as the demographic profile of a region changes. The cost and complexity of implementing a successful travel survey sometimes prohibits having good base year calibration. The quality and reliability of land use and economic forecasts varies widely because those inputs are as often political as empirical. 22 trend data often lacking value of time data not often calibrated OD data suspect 23 Dont have too many issues here as you can generate the data yourself, although at a cost. 24 In our country, for example, data are not managed, consolidated and ussually is not easy to collect. GDP forecasts always are much higher than in reality. Traffic counts are carried out but for a limited number of days and convgerted to AADT. Value of time is rather not given importance as other economic activities generating revenue to road users in saved time is yet almost absent. 25 Often need due diligence and confirmatory studies. On Hong Kong - Guangzhou Superhighway, first appraisal (1982) done on moving observer traffic count basis - one pass in one direction Hong Kong - Macau via Guangzhou. Thereafter, five or six full studies carried out at behest of potential funders with road finally opening in 1995. 26 The critical issues are how long ahead the forecasts have to run. For a 3-5 year span, trend based reviews may be OK; but the critical point is how much of the profit depends on large growth after this period, because many countries have problems with longer term growth parameters. 27 Whether it is primary or secondary, data tends to be collected for the sake of it rather than with its usefulness for future planning in mind. Its therefore usually in a difficult format, inaccurate and never close to being comprehensive. Supplementing it with further data collection is a strenuous task because local enumerators do not understand the importance of rigour and accuracy.DissFinal Page 184 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 18: Regarding the applicability of full four-stage models (i.e. includingtrip generation, distribution and mode split in addition to assignment), pleaseindicate how strongly you agree or disagree with the following statements:Excluding “Not Sure” and null responses: 1 2 3 4 5 Strongly Strongly Agree Agree Neutral Disagree Disagree Such models are reliable 2 35 56 7 2Such models are too data hungry to 0 29 35 30 5 be relied uponSuch models are too complicated to 4 12 26 47 11 be of worth Such models are not suitable for 0 11 26 42 14 toll-road work Economic uncertainties/ pace of change makes them irrelevant in 2 20 31 35 6 developing economies Such models are too much of a black box for non-specialists to 10 37 22 27 6 properly critique model outputsUsing the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Such models are reliable 102 2.73 0.70Such models are too data hungry to be relied upon 99 3.11 0.89 Such models are too complicated to be of worth 100 3.49 0.97 Such models are not suitable for toll-road work 93 3.63 0.88 Economic uncertainties/ pace of change makes 94 3.24 0.93 them irrelevant in developing economiesSuch models are too much of a black box for non- 102 2.82 1.11 specialists to properly critique model outputsDissFinal Page 185 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 19: Regarding network assignment models (e.g. based on traffic countsand/or Origin-Destination surveys), but NOT full four-stage models, pleaseindicate how strongly you agree or disagree with the following statements:Excluding “Not Sure” and null responses: 1 2 3 4 5 Strongly Strongly Agree Agree Neutral Disagree Disagree Such models are reliable 2 39 47 9 2Such models are too data hungry to 1 10 37 37 9 be relied uponSuch models are too simplistic to be 0 21 33 37 5 relied uponSuch models are too complicated to 0 4 28 49 16 be of worth Such models are not suitable for 1 12 28 40 11 toll-road work Economic uncertainties/ pace of change makes them irrelevant in 3 13 30 38 7 developing economies Such models are too much of a black box for non-specialists to 4 19 30 36 10 properly critique model outputsUsing the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Such models are reliable 99 2.70 0.74Such models are too data hungry to be relied upon 94 3.46 0.85 Such models are too simplistic to be relied upon 96 3.27 0.86 Such models are too complicated to be of worth 97 3.79 0.76 Such models are not suitable for toll-road work 92 3.52 0.90 Economic uncertainties/ pace of change makes 91 3.36 0.93 them irrelevant in developing economiesSuch models are too much of a black box for non- 99 3.29 1.02 specialists to properly critique model outputsDissFinal Page 186 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 20: Regarding spreadsheet-based traffic/ revenue models, please indicatehow strongly you agree or disagree with the following statements:Excluding “Not Sure” and null responses: 1 2 3 4 5 Strongly Strongly Agree Agree Neutral Disagree Disagree Such models are reliable 0 24 53 16 4Such models are too data hungry to 1 2 27 48 16 be relied upon Such models are too simplified to 3 12 39 34 5 be of worth Such models are not suitable for 2 10 38 32 5 toll-road work Economic uncertainties/ pace of change makes them irrelevant in 1 14 32 36 5 developing economies Such models are too simplistic to 5 10 40 39 3 provide meaningful outputsUsing the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Such models are reliable 97 3.00 0.76Such models are too data hungry to be relied upon 94 3.81 0.78 Such models are too simplified to be of worth 93 3.28 0.87 Such models are not suitable for toll-road work 87 3.32 0.84 Economic uncertainties/ pace of change makes 88 3.34 0.85 them irrelevant in developing economies Such models are too simplistic to provide 97 3.26 0.88 meaningful outputsQuestion 21: If you have any specific comments on issues with developing/calibrated/ forecasting with traffic and revenue models (of any type), please giveyour comments here: (20 responses) 1 As one goes through the planning cycle, different models and levels of disaggregation should be used. Thus, spreadsheets and, for example, EMME/2 have their own roles. 2 Depends on the particular project, availability of data, or circumstances. More robust data justifies more complexity and higher confidence in the result 3 Its a matter of horses for courses. 4 Stage models are the most appropriate approach, particularly when development means demands are changing rapidly. They are important, even if uncertainties mean that all they can produce are a number of wide- ranging forecasts. Cheaper assignment and spreadsheet models have their place in initial stages or for reality checks on the more complex models. I suppose they might even be sufficient in themselves - if the case is so strong that detailed assessment of the demand is not required, but then you are missing out on opportunities to optimise the scheme. 4 I feel that each of these approached are valid, and worthy of application based on need. It is possible to have a model of the 3rd type that is rigorous and reliable in producing meaningful outputs..DissFinal Page 187 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 5 Proper market segmentation and application of realistic willingness to pay diversion curves is the essential, however implemented - spreadsheet or assignment / route choice model. 6 There is a place for each of the model types. All models have strengths and weaknesses. It is important that this is recognised (regardless of the model type/form) when considering the outputs. 7 Again depends on location and status of prior work undertaken. Journey Time and Traffic counts are essential. OD becomes essential if no reliable matrices available. Best approach if relevant is to use pre-existing 4Stage model to provide strategic inputs to a refined highway assignment model. 8 I have built some very complex models using spreadsheets that included the land use and assignment. These models included the nodes of the paths from the origins to destinations. There was an ability to allocate proportions of the trips to be split and loaded onto those paths. This was not for a toll road, but for an inner city development. These additional trips were added to the background traffic (that data was available from surveys). The advantage is that all the outputs can be analysed by everyone and that the inputs can be varied and thus agreements can be reached on the assumptions and outputs. 9 I assume that the models are being applied by someone who knows what they are doing! 10 Static assignment models have shown their limitations in congested areas, where toll roads would have the greatest success. Dynamic traffic assignment techniques will have to be adapted to toll reality but the industry is only starting in this area. 11 see comments on previous page re inter-urban vs urban routes 12 I have no experience of models identified in q.19 and q.20 13 Unfortunately for advocates of simple aggregate models, reality is disaggregate, and the differences a fine levels of detail really do matter. Four-step models can be quite useful if they are treated as tools, not Delphic oracles, and the coming tour-based models show great promise of being significantly better. 14 These questions are difficult to answer in the scale provided because, of course, a well developed 4-stage travel model that is calibrated and validated and run based on reasonable land use forecasts, iterated properly to a point of reasonable reliability, and interpreted by qualified and experience professionals can be very useful to evaluate alternatives and impacts of a proposed changes. The same is true for spreadsheet models--- there are good ones and bad ones. The spreadsheet itself can be used to implement simple models that do a great job, or complex models that do a very poor job, or the inverse. These can only be evaluated on a case-by-case basis, and in relation to the purpose to which the model or forecast is being applied. 15 cut yr clothe... 16 Model should be compatible with the local capability and shall not call for institutional support from outsider foreever 17 Any information is useful and as such big four stage models can help inform the analyst; particualrly in the urban areas. Spreadsheet models are most relevant for inter- urban toll roads where route choice is limited. In austrlaia most assignment models willnow be breaking tiem into different categories of moving and delay effectively recognising that not to do so impairs calibration and accuracy of forecasts. 18 The issue is that there is risk, and models should address risk. In fact they often do not. Data exoistence, quanity etc are to some extent not the core problem, but the fundamental failure to recognise the scale of uncertainty and the imperative to get to grips with it. No models alone do this adequately, without reality-chackign against comparable projects whose charactyeristics are documented. 19 Each of the 3 types has their uses, and many of the responses are it depends. Relevant factor are timescale and budget, availability of data, green-field new corridor or existing facilities, length of period unde examination.etc.DissFinal Page 188 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 20 As with all modelling the structure, theory, format and software used, etc is of less relevance than the data and assumptions used in their creation. If you get the latter right, then you can produce reliable forecasts for simple scenarios without going near a computer.Question 22: From your experience, please state how often you feel each of thefollowing statements is true:Excluding “Not Sure” and null responses: 1 2 3 4 5 Very Quite Often Often Sometimes Rarely NeverHow often do projects significantly exceed forecast traffic/ revenue 5 18 43 48 0 levels? How often do projects fall well short of forecast traffic/ revenue 10 51 42 12 0 levels? Do you believe that transportplanners are pressured by clients to 18 43 45 12 3 adjust forecasts to meet their expectations? Do you believe that transportconsultants’ forecasts are higher if 10 22 44 14 3they are engaged by equity- rather than debt-side clients?Using the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean DeviationHow often do projects significantly exceed forecast 114 3.18 0.85 traffic/ revenue levels? How often do projects fall well short of forecast 115 2.49 0.80 traffic/ revenue levels? Do you believe that transport planners are pressured by clients to adjust forecasts to meet 121 2.50 0.95 their expectations?Do you believe that transport consultants’ forecasts are higher if they are engaged by equity- rather 93 2.76 0.94 than debt-side clients?DissFinal Page 189 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)The mean scores from above were also compared with mean scores based on the 6aggregated experience sectors from Question 7, as shown below: All FLO TpEc E&A G&A Acad Oth Significantly exceed 3.18 3.42 3.25 3.17 3.00 3.13 2.70 forecast traffic/ revenueFall well short of forecast 2.49 2.23 2.45 2.67 2.69 2.40 2.40 traffic/ revenue Are transport planners pressured by clients to 2.50 2.32 2.41 2.56 2.38 2.40 2.55 adjust forecasts? Are forecasts are higher for equity- rather than 2.76 2.86 2.64 3.14 2.86 2.69 2.63 debt-side clients?Key: FLO = Financial, Legal, Operator TpEc = Transport Planners and Economists E&A = Engineers and Architects G&A = Government and Aid Agencies Acad = Academics Oth = OthersQuestion 23: How often do you explicitly consider the following in appraisingtolled highways?Excluding “Not Sure” and null responses: 1 2 3 4 5 Always Usually Sometimes Rarely Never Traffic forecasts: Base and/or 53 31 15 1 5 Central CaseTraffic forecasts: Optimistic and/or 22 37 24 16 6 High Case Traffic forecasts: Conservative 35 44 15 6 5 and/or Low Case Revenue forecasts: Base and/or 46 31 16 3 6 Central Case Revenue forecasts: Optimistic 21 34 22 16 8 and/or High Case Revenue forecasts: Conservative 32 41 16 7 6 and/or Low Case Congestion on competing/ 37 38 16 6 7 alternative routes Congestion on link-roads/ feeder 37 37 20 4 8 routes Capacity of the highway being 57 33 10 1 5 consideredDissFinal Page 190 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Using the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Traffic forecasts: Base and/or Central Case 105 1.80 1.04 Traffic forecasts: Optimistic and/or High Case 105 2.50 1.15Traffic forecasts: Conservative and/or Low Case 105 2.07 1.06 Revenue forecasts: Base and/or Central Case 102 1.94 1.12 Revenue forecasts: Optimistic and/or High Case 101 2.56 1.21Revenue forecasts: Conservative and/or Low Case 102 2.16 1.12 Congestion on competing/ alternative routes 104 2.12 1.15 Congestion on link-roads/ feeder routes 106 2.14 1.16 Capacity of the highway being considered 106 1.72 1.01Question 24: How often do you explicitly consider the following criteria inappraising infrastructure projects (with an emphasis on tolled highways ifapplicable)?Excluding “Not Sure” and null responses: 1 2 3 4 5 Always Usually Sometimes Rarely Never Net Present Value (NPV) 52 40 19 2 4Financial Internal Rate of Return 39 43 19 6 7 (FIRR)Economic Internal Rate of Return 28 35 27 14 10(EIRR; including social impacts) Social Cost/ Benefit Ratios 25 31 34 20 9 Risk correlation versus other projects in company’s/ client’s 19 12 31 27 20 portfolioCounterparty risks: can partners 17 23 19 23 21 contribute equity/ debt Sovereign/ Institutional other 26 22 20 20 20 country/ legal risksUsing the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Net Present Value (NPV) 117 1.85 0.98 Financial Internal Rate of Return (FIRR) 114 2.11 1.12 Economic Internal Rate of Return (EIRR; 114 2.50 1.23 including social impacts) Social Cost/ Benefit Ratios 119 2.64 1.20 Risk correlation versus other projects in 109 3.16 1.33 company’s/ client’s portfolio Counterparty risks: can partners contribute 103 3.08 1.38 equity/ debt Sovereign/ Institutional other country/ legal risks 108 2.87 1.44DissFinal Page 191 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 25: If you use any financial ratios when appraising projects, please statewhich ratios you normally use? (20 responses) 1 Debt Service Coverage Ratio in addition to the above, and perhaps other depending on the debt instrument. 2 The use and importance give to NPV, EIRR and B/C ratio vary with the client. 3 I just do the traffic and revenue forecasts 4 Financial rate of return: 10% 5 EIRR +15% 6 EIRR >= 12% 7 commercially sensitive. 8 I have not been directly involved in this work - I have been near it, but I have not actually done it. 9 Debt Service Cover Ratio (Average and Min) Loan Life COver Ratio Project Life Cover Ratio Initial Debt/Equity Ratio 10 Question 24 & 25: These are tasks of other colleagues of the team 11 Accounting and cashflow payback period, return on capital investment 12 Cost/Benefit ratios where costs are generally limited to financial or easily-monetized values (for example, construction cost, relocation cost, operation cost), and benefits are limited to mobility/accessibility measures such as PMT. 13 Question 24 is difficult to answer if applied to both tolled and non-tolled projects because they are so different. A typical non-tolled public project doenst really have to meet a threshold for economic performance; and the risk profile is usually on on developed in relation to the construction cost. Privately or publicly financed toll projects have to go through a more rigorous process to justify a bond issue for initial construction. So it is diffucult to answer the questions in 24 for both tolled & non- tolled; it would be better to have two questions or answer it as either/or. 14 cash yields, IRR, NPV, DSCRs, payback periods 15 Benefit to country/Client verses to the consessionnaire 16 Only FIRR 17 Debt service cover ratio 18 FIRR 19 The main sources of financial risk in major transport infrastructure projects are : 1. construction cost overruns induced by, for instance, government, client, management, contractor or accident; 2. increased financing costs, caused by changes in interest and exchange rates and by delays; and 3 lower than expected revenues, produced by changes in traffic volumes and in payments per unit of traffic. From an analytical point of view, it is expedient to identify the following types of risk of relevance to both a financial and an economic perspective. i) project-specific risks ii) market risks iii) sector-policy risks iv) capital-market risks. When appaising projects in the case of toll roads on occassion government may need to make up the difference between the private capital injection and the total investment cost, if the roads are to be built. Typically , this has been done by providing land for free, or on the basis of deferred payments, namely by sharing or dedicating toll revenues from other roads (for example, Bangkok Second Stage expressway, Sydney Harbour Tunnel & Dartford Bridge), or for direct grants or subsidies. 20 IRR Rate of return thresholdsDissFinal Page 192 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 26: How would you rate the potential for inter-urban tolled highwaysover the next 10 years in each of the following countries:Excluding “Not Sure” and null responses: 1 2 3 4 5 6 Over- No developed Maturing Steady Developing Nascent Market Cambodia 2 0 0 6 40 18 China 1 12 37 39 4 1 Indonesia 2 4 9 21 24 7 Laos 1 0 0 3 42 28 Malaysia 6 25 27 8 4 1 Myanmar 2 0 1 5 35 29 Philippines 2 4 11 21 22 4 Thailand 3 11 24 18 10 2 Vietnam 2 0 3 28 28 5Using the values 1 to 6 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Cambodia 66 5.06 1.08 China 94 3.38 0.84 Indonesia 67 4.22 1.24 Laos 74 5.28 0.94 Malaysia 71 2.75 1.05 Myanmar 72 5.19 1.25 Philippines 64 4.08 1.17 Thailand 68 3.40 1.16 Vietnam 66 4.44 0.96The mean scores from above were also compared with mean scores based on the 6aggregated experience sectors from Question 7, as shown below: All FLO TpEc E&A G&A Acad Oth Cambodia 5.06 5.31 5.14 5.10 5.00 5.00 5.29 China 3.38 3.45 3.35 3.54 3.57 3.50 3.63 Indonesia 4.22 4.07 4.00 4.52 4.06 3.25 4.88 Laos 5.28 5.44 5.41 5.30 5.30 5.50 5.35 Malaysia 2.75 2.60 2.66 2.53 2.68 2.20 3.08 Myanmar 5.19 5.31 5.22 5.33 5.29 5.33 5.27 Philippines 4.08 4.54 3.97 4.14 3.88 5.00 4.45 Thailand 3.40 3.81 3.18 3.67 3.37 2.60 4.00 Vietnam 4.44 4.53 4.50 4.57 4.20 3.40 4.54Key: FLO = Financial, Legal, Operator TpEc = Transport Planners and Economists E&A = Engineers and Architects G&A = Government and Aid Agencies Acad = Academics Oth = OthersDissFinal Page 193 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Repeating the analysis for Question 26, but this time including responses ONLY fromthose with experience in the country in question (from answer to Question 11) and onceagain excluding any “Not Sure” responses: 1 2 3 4 5 6 Over- No developed Maturing Steady Developing Nascent Market Cambodia 1 0 0 0 13 3 China 0 10 26 20 1 0 Indonesia 1 3 5 7 8 0 Laos 0 0 0 2 10 7 Malaysia 2 15 10 2 0 0 Myanmar 0 0 1 0 2 2 Philippines 0 1 5 11 8 0 Thailand 0 7 12 10 4 1 Vietnam 0 0 1 15 8 2Using the values 1 to 6 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Cambodia 17 4.94 1.16 China 57 3.21 0.74 Indonesia 24 3.75 1.16 Laos 19 5.26 0.80 Malaysia 29 2.41 0.72 Myanmar 5 5.00 1.41 Philippines 25 4.04 0.82 Thailand 34 3.41 1.05 Vietnam 26 4.42 0.72Comparing the mean of all respondents who expressed an opinion with the sub-set ofthose with experience in the country: All Respondents Those with Country Difference (A) Experience (B) (A – B) Cambodia 5.06 4.94 0.12 China 3.38 3.21 0.17 Indonesia 4.22 3.75 0.47 Laos 5.28 5.26 0.02 Malaysia 2.75 2.41 0.33 Myanmar 5.19 5.00 0.19 Philippines 4.08 4.04 0.04 Thailand 3.40 3.41 -0.01 Vietnam 4.44 4.42 0.02DissFinal Page 194 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 27: Comparing the next 10 years (2006-2016) with the last 5 years (2001-2006), how do you feel each the following will change:Excluding “Not Sure” and null responses: 1 2 3 4 5 Significant Increase to Decrease to Significant Increase an Extent No Change an Extent Decrease Fuel prices 49 70 2 2 0 General price 4 74 42 1 0 inflation Interest rates 7 41 63 5 0 Economic growth 5 60 41 17 0 Exchange rate 6 43 52 8 1 volatility Acceptability of road tolls and toll 22 63 28 2 2 increasesUsing the values 1 to 5 (as per column headings above), the mean and standarddeviation of responses as follows: Standard Responses Mean Deviation Fuel prices 123 1.65 0.60 General price inflation 121 2.33 0.55 Interest rates 116 2.57 0.67 Economic growth 123 2.57 0.78 Exchange rate volatility 110 2.59 0.74 Acceptability of road tolls and toll increases 117 2.14 0.79The mean scores from above were also compared with mean scores based on the 6aggregated experience sectors from Question 7, as shown below: All FLO TpEc E&A G&A Acad Oth Fuel prices 1.65 1.72 1.61 1.81 1.55 1.53 1.78 General price inflation 2.33 2.38 2.35 2.25 2.21 2.31 2.26 Interest rates 2.57 2.58 2.59 2.65 2.48 2.64 2.79 Economic growth 2.57 2.63 2.67 2.44 2.57 2.57 2.42 Exchange rate volatility 2.59 2.61 2.60 2.50 2.69 2.67 2.68Acceptability of road tolls and toll increases 2.14 2.42 2.14 2.29 2.36 2.36 2.11Key: FLO = Financial, Legal, Operator TpEc = Transport Planners and Economists E&A = Engineers and Architects G&A = Government and Aid Agencies Acad = Academics Oth = OthersDissFinal Page 195 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Question 28: If you believe that there will be any other significant changes tofactors affecting toll road performance, please state which factors and how you feelthey will change below. Similarly, if you feel that patterns will be markedlydifferent between certain economies, please explain below (citing which countriesmay have above-trend growth in which variables, and which countries you feel willhave below-trend changes): (19 responses) 1 Lack of resources and capacity to build and manage infrastructure. 2 Fuel cost to have a impact on efficiency in routing. Possible slow-down in road projects in more mature markets where mass transport may be considered more appropriate going forward. 3 Critical is the preceived friendliness of the government to private sector involvement in the BOT type projects. This varies with time. China will be near the bottom of the list even though they have an number of toll roads. The ability to adjust the tolls is likewise important since there is always political pressure not to allow changes even if clearly stated in the concession agreement. 4 Experience in Indonesia (albeit not directly involving toll road acceptability) indicates that professional drivers avoid them in order to avoid the need to pay tolls, even though this may mean sitting in traffic queues for hours at a time. 5 Not able to respond 6 General willingness to pay for new facilities using new technology Globalisation of inductrial production 7 Dependence on surface/road based freight logistics system. Needs for punctual delivery of goods. if they are high, the performance of tolll road network will be positive. 8 Re Q 27 - it is not clear which part of the world you are asking about. 9 where do you want the invoice to be sent? 10 I am wondering if there will be any land use changes resulting from the higher fuel=private transportation costs. 11 Increased congestion in urban areas will push drivers onto toll roads - especially if compounded by wieght limits (eg, against big trucks) and strictly enforced speed limits / traffic calming in towns/villages. 12 Chinese economic growth will slow down because of resource constraints. Other regional economies will probably follow China. 13 Toll road use is closely tied to government tax policy. Low tax approaches put financial pressure on public infrastructure providers, which in turn pushes user fee approaches such as toll roads. If low-tax trends continue, toll road projects will increase. 14 Institutions, legal systems, social and political volatility and corruption are critical issues. 15 increased difference between urban / inter-urban / bridges. In China, urban toll roads are becoming less acceptable on traffic management grounds e.g. Shanghai / SZ have removed tolls. GZ tolls the ring roads but not the arterials - counterproductive. 16 Willingness to Pay & Ability to Pay on the part of road users and its relationship with the level of service provision could be one of the important issue. This is much eminent in developing economies with difficulty in utilizing saved time out of use of toll road. 17 Use of managed lanes with variable pricing may increase the divide of use between weathly and middle-lower class. 18 The major problem is that the IFIs have decided to institute tolling on roads to guarantee sustainable maintenance. In SE Asia, fine. In Africa, very difficult. No culture of paying tolls and IFI insistence of imposing tolls on roads with neglible traffic levels. I was recently asked to design a tolling framework for a road in central Africa characterized by traffic flows uniformly below 200 vpd.DissFinal Page 196 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 19 Existing statistics indicate that there is difficulty to obtain reliable traffic forecasts. To support the statement the actual traffic as percentage of forecast traffic, opening year: Project Actual traffic as % of forecast traffic (opening year) Channel tunnel, UK, France 18% Third Dartford Crossing, UK 115% Pont de Normandie, France 120%Question 29: Finally, if you have any other comments you would like to make,either about issues in project finance/ transport forecasting, or about this survey,please let me have your thoughts. Thank you. (30esponses) 1 I do not have a strong background in the stated field. 2 See above. Please note the time need to complete it far exceeded the time indicated in the introductory para. 3 Cool Survey Dude! My invoice is in the mail. 4 Corruption levels, strong political will, reliable legal framework are the most important factors in succeeding in developing economies. If corruption exists, it must be quantifiable. 5 Good luck with your project. It is a pity that I have no direct involvement in toll road projects. 6 As far as constructing a good survey is concerned, the initial few pages were off putting as it seemed suspiciously like fishing for a recruitment agency. 7 The survey focuses on aspects of projects that up to now I have not often come across in my work. I would be interested in the outcome. 8 Hi Richard, I have answered the questions based on most of my experience in the Pacific and limited experience in Australia, and almost no toll road experience. 9 We are in worthy sector to improve quality of life. 10 Questionnaire a bit too long. 11 one good idea when designing surveys is to tell the respondent how many pages/questions are in the survey from the start, or some kind of progress i.e. 10% done 20% done is useful 12 Regarding the study sponsor, debt versus equity, clearly debt sponsors have an incentive to be conservative given their sole interest in getting repaid. However, they do not have to be right on the upside. Equity investors have to be right on the upside, so their investors get a reasonable return, and the downside, so creditors get paid. At the end of the day, the diligence of the project sponsor, be it private or public sector, in getting the best sense of the range of possibilities for project performance is the best indicator of forecast accuracy. 13 The important issue, in my opinion, is contract. In my country, some State Governments are pursuing a legal battle trying to break contracts alleging public interest. They claim the toll is to high, although determined by social survey and negotiated. Without a solid contract all other considerations are secondary. The importance increases with governmente instability and lack of proper policy. 14 I think that this is a very worthwhile project and I hope that others appreciate it as well. 15 Richard, I have not answered a number of the questions as they relate to toll roads and I have no experience of these. Sorry I could not be of more help. Mike 16 Contingent valuation (CV) methods could be useful in assessing, for example, drivers willingness to pay toll fees (eg, bench-marking against numeraire such as prevailing price of petrol/litre). 17 Theres a review of the pressure on forecasters in the archive of my web site www.kilsby.com.au - see entry for 02/04. 18 I dont know how much help I have been - it is all a bit tangential to my experience! Good luck. 19 No relevant experience for 24. 25 20 You are free enough to contact us in case of need.DissFinal Page 197 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) 21 My experience of toll road projects is limited, as is my direct involvement in the use of transport models. My responses to the questions in this survey reflect what I believe, but are based on my limited knowledge of the subject, and may therefore be of little value for the survey. In any case, please use with caution! 22 Good luck! 23 richard Am I the only one to reply? cheers! Tim 24 Soon there could be a blurring of the difference between highway tolling and congestion pricing, as traffic congestion gets worse in many urban areas. 25 I have personally carried out viability study for a potential road project through BOT model. However, the level of traffic yet sems not adequate to generate adequate revenue for private party to invest ~ 7 bl NRs. Also the present political chaos and formation of a permanent regulatory body is needed to ensure potential investor that the consession contract shall be respected by public sector for usually long concession period. Your survey questionnaire are well prepared, however, I feel that its analyses and outcome is rather oriented towards endorsing the private financing of the road project. 26 You have clearly thought about some important issues! Im sorry that I dont have more time to take more of an interest in your research. In any case, Im now out of toll road forecasting and doing congestion charging instead. 27 Thanks for the opportunity. I had difficulty in answering Q12 as the objective of the question is (in my view) not sufficiently clear. 28 Consider the following publication for your literature review; Megaprojects and Risk by B.Flybjerg, N.Bruzelius & W.Rothengatter. Pub. Cambridge Press. ISBN 0 521 00946 4 Fraqnce & Spain have had the longest and most extensive experience of building private roads financed by tolls. 29 Take some of this lot with a pinch of salt because my exposure to transport planning in the developing world has been minimal since 2000. Looks like an interseting project though. 30 I dont think I am a suitable respondee for this survey - I have no involvement in road transport projects nor in any projects in Asia.Question 30: If you would like information about the survey responses, oncecollated, or about my broader research, please indicate below: (80 responses) No, thank you 9 Yes, regarding survey results 23 Yes, regarding your research 48DissFinal Page 198 December 2006

Chance of Standard Standard Modal Smaller Deviation Minimum Deviation MaximumDissFinal Case Variable Value Value (Low Values) Value (High Values) Value Source/ Justification Dissertation Conventional Expressway Capacity/ From other studies (corroborated Respondents 20,000 50% 2,000 20,000 2,000 28,000 Lane (pcu’s per day) by other transport planners) Kondratieff Conventional Local Road Capacity/ From other studies (corroborated Respondents 10,000 50% 1,000 8,000 1,000 12,000 Henley Management College Lane (pcu’s per day) by other transport planners) Kondratieff Conventional $4,633,000 per km * 40 km; cost Respondents Construction Cost ($)* $185.32m 25% 5% 90% 15% 30% overrun more likely than Kondratieff underrun (see 2.11) Conventional time overrun more likely than Construction Duration Respondents 10 20% 1 8 2 14 underrun (see 2.11); specified in (in Quarters) Kondratieff whole quarters (i.e. integers) Conventional Operations & From other studies (corroboratedPage 199 Respondents Maintenance Fixed Costs 2% 50% 0.50% 0.10% 1% 4% by other transport planners) Kondratieff (% of Construction Cost) Conventional Operations & From other studies (corroborated Respondents Maintenance Variable 3% 50% 1% 1% 1% 5% by other transport planners) Kondratieff Costs (% of Revenues) Conventional Base Value of Time for From other studies (corroborated Respondents $4 50% $1 $2 $1 $6 Small Vehicles ($/hour) by other transport planners) Kondratieff Conventional Base Value of Time for From other studies (corroborated Respondents $3 50% $1 $1 $1 $5 Large Vehicles ($/hour) by other transport planners) Kondratieff Conventional Income Elasticity of Value Respondents 0.5 50% 0.15 0.2 0.15 0.8 see 2.10.1 Appendix 16: Risk Simulation Modelling: Simulation Parameters Employed of Time (Small Vehicles) Kondratieff Conventional Income Elasticity of Value Richard F. DI BONA (1005661)December 2006 Respondents 0.5 50% 0.15 0.2 0.15 0.8 see 2.10.1 of Time (Large Vehicles) Kondratieff

Chance of Standard Standard Modal Smaller Deviation Minimum Deviation MaximumDissFinal Case Variable Value Value (Low Values) Value (High Values) Value Source/ Justification Dissertation Conventional Base Small Vehicle From other studies (corroborated Respondents Operating Costs on $0.06 50% $0.015 $0.03 $0.015 $0.09 by other transport planners) Kondratieff Expressways ($/km) Conventional Base Large Vehicle From other studies (corroborated Respondents Operating Costs on $0.10 50% $0.025 $0.05 $0.025 $0.15 Henley Management College by other transport planners) Kondratieff Expressways ($/km) Conventional Small Vehicle Local Road From other studies (corroborated Respondents Operating Cost (relative to 1.5 50% 0.25 1.0 0.15 2.0 by other transport planners) Kondratieff expressway) Conventional Large Vehicle Local Road From other studies (corroborated Respondents Operating Cost (relative to 2.0 50% 0.25 1.5 0.25 2.5 by other transport planners) Kondratieff expressway) Conventional arbitrary to reflect possible error Factor for Base Small Respondents 100% 50% 15% 70% 15% 130% range in initial surveys on a "realPage 200 Vehicle Demand Matrix Kondratieff world" project Conventional arbitrary to reflect possible error Factor for Base Large Respondents 100% 50% 15% 70% 15% 130% range in initial surveys on a "real Vehicle Demand Matrix Kondratieff world" project Conventional Income Elasticity of Small Respondents 1.25 50% 0.20 0.85 0.20 1.65 see 3.5 Vehicle Traffic Kondratieff Conventional see 3.5 (smaller value to allow for Income Elasticity of Large Respondents 1.10 50% 0.20 0.70 0.20 1.50 larger trucks and coaches and Vehicle Traffic Kondratieff increased efficiency) Conventional From other studies (corroborated Respondents Toll Revenue Leakage (%) 10% 50% 2.50% 5% 5% 20% by other transport planners) Kondratieff Richard F. DI BONA (1005661)December 2006

Chance of Standard Standard Modal Smaller Deviation Minimum Deviation MaximumDissFinal Case Variable Value Value (Low Values) Value (High Values) Value Source/ Justification Dissertation Conventional Initial Amplitude of Ramp- Respondents 40% 50% 10% 20% 20% 80% see 2.10.4 Up (% traffic decrease) Kondratieff Conventional Ramp-Up Duration Respondents 8 40% 2 4 5 20 see 2.10.4 Henley Management College (in Quarters) Kondratieff Conventional Small Vehicles Tolling see 2.9 (and corroborated by other Respondents 10 50% 5 0 5 20 Penalty (minutes) transport planners) Kondratieff Conventional Large Vehicles Tolling see 2.9 (and corroborated by other Respondents 15 50% 5 5 5 25 Penalty (minutes) transport planners) Kondratieff Conventional Small Vehicle Routeing Respondents Sensitivity "Lambda" for 0.05 50% 0.0125 0.25 0.0125 0.75 arbitrary valuePage 201 Kondratieff Logit Sub-Model Conventional Large Vehicle Routeing Respondents Sensitivity "Lambda" for 0.05 50% 0.0125 0.25 0.0125 0.75 arbitrary value Kondratieff Logit Sub-Model Conventional Toll Escalation Rate (% of Respondents Retail Price Index 90% 50% 15% 60% 5% 100% see 2.10.3 Kondratieff Inflation) Conventional Quarters Between Toll Respondents 12 40% 2 8 4 20 see 2.10.3 Increases Kondratieff Richard F. DI BONA (1005661)December 2006

Chance of Standard StandardDissFinal Modal Smaller Deviation Minimum Deviation Maximum Dissertation Case Variable Value Value (Low Values) Value (High Values) Value Source/ Justification Conventional 6% 50% 2% 2% 2% 10% base for other cases also Respondents GDP Growth (% p.a.) +1% 50% 0.5% +0% 0.5% +2% in addition to Conventional value Kondratieff +1% 50% 0.5% +0% 0.5% +2.5% in addition to Respondents value Conventional 2.5% 50% 1% 0.5% 1% 4.5% base for other cases also Vehicle Operating Cost Henley Management College Respondents +2.5% 50% 1.5% +0% 1.5% +5.5% in addition to Conventional value Price Inflation (% p.a.) Kondratieff +0% n/a n/a n/a n/a n/a same as Respondents value Conventional Construction, Operations 2.5% 50% 1% 0.5% 1% 4.5% base for other cases also Respondents & Maintenance Cost +0.75% 50% 0.25% +0.25% 0.25% +1.25% in addition to Conventional value Kondratieff Inflation (% p.a.) +1.0% 50% 0.50% +0% 0.50% +2.5% in addition to Respondents value Conventional 2.5% 50% 1% 0.5% 1% 4.5% base for other cases also General Price Inflation (% Respondents +0.75% 50% 0.25% +0.25% 0.25% +1.25% in addition to Conventional value p.a.) Kondratieff +1.0% 50% 0.50% +0% 0.50% +2.5% in addition to Respondents valuePage 202 Conventional Interest Rates for Initial 5% 50% 1% 3% 1% 7% base for other cases also Respondents Debt (% p.a.) +1% 50% 0.50% +0% 0.50% +2% in addition to Conventional value Kondratieff based on construction costs +2% 50% 1% +0% 1% +4% in addition to Respondents value Conventional Interest Rates for Extra +2% 50% 1% +0% 1% +4% in addition to initial debt rate Respondents Debt (% p.a.) +2% 50% 1% +0% 1% +4% in addition to initial debt rate Kondratieff for subsequent cash shortfalls +2% 50% 1% +0% 1% +4% in addition to initial debt rate Richard F. DI BONA (1005661)December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 17: Risk Simulation Modelling: Fixed ParametersConcession Length30 years (120 quarters), including construction time (commencing in quarter 1)Network Road LengthsAs described in Section 5.2Pcu Factors(to equivalence different vehicle types to a common unit for congestion analysis)Small Vehicles = 1Large Vehicles = 2Speeds by Road CapacityAs shown in Figure 5.B.DissFinal Page 203 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 18: Risk Simulation Modelling: Equations EmployedGrowing Prices in Line with Appropriate Inflation Rate Applies To: Using: Construction Cost per Quarter (charged Construction, Operations & Maintenance during construction period only) Cost Inflation Operations & Maintenance Fixed Costs Construction, Operations & Maintenance (charged following start of operations) Cost InflationVehicle Operating Costs for Expressways Vehicle Operating Cost Price Inflation and Local Roads, for Small and Large Vehicles (used in path-building on assignment) Toll Rates (toll rates updated every X General Price Inflation * Toll Escalation quarters following start of operations, Factorwhere X is the number of quarters between toll increases)PRICE q  PRICE q1  1  InflationR ate 4  1    Growthing Trip Matrices Applies To: Using: Small Vehicle Matrix (trips in each cell) GDP Growth Rate and Small Vehicle Income Elasticity of Traffic Large Vehicle Matrix (trips in each cell) GDP Growth Rate and Large Vehicle Income Elasticity of TrafficTRIPS q  TRIPS q1  1  GDPGrowthRate  Elasticity 4  1    Growthing Value of Time Applies To: Using: Small Vehicle Value of Time ($/hour) General Price Inflation, GDP Growth Rate and Small Vehicle Income Elasticity of Value of Time Large Vehicle Value of Time ($/hour) General Price Inflation, GDP Growth Rate and Small Vehicle Income Elasticity of Value of Time    VOTq  VOTq 1  1  Inflation  4   1  GDPGrowthRate  Elasticity  4  1 1    DissFinal Page 204 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Generalised Costs of Routes Using and Not Using ExpresswayTravel time for each link: LinkLength(km)MinutesOfT ime  60  Speed (kph)Monetary cost for each link:CostInDollars  LinkLength(km)  VehicleOpe ratingCostsInDollarsGeneralised Time for each link: CostInDollarsGeneralisedTime  MinutesOfT ime  60  ValueOfTim e($ / hour )Total Generalised Time for non-expressway route;NonExpresswayTime   GeneralisedTime forl1...numberofli nks lTotal Generalised Time for expressway route; DollarsOfToll  60ExpresswayTime    GeneralisedTimel ValueOfTim e($ / hour ) forl1...numberoflinksShare of expressway trips (logit relationship): 1ExpresswayShare  1  e Lambda NonExpresswayTimeExpresswayTimeTollingPenaltyDissFinal Page 205 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 19: Risk Simulation Modelling: Results by Parameter Variable: Capacity per Expressway Lane (pcus) Minimum: 20,000 Distribution of Variable 100% 6% Maximum: 28,000 Mean: 23,995 80% 5% 4% Monte Carlo Settings: 60% Modal Value 24,000 3% % < Modal 50% 40% 2% SD (<Modal) 2,000 20% SD (>Modal) 2,000 1% 0% 0% 0 0 0 0 0 0 0 0 0 , 00 , 00 , 00 , 00 , 00 , 00 , 00 , 00 , 00 20 21 22 23 24 25 26 27 28 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.14863 + 0.000001 * Variable 16% R2= 0.400 15% 14% FIRR = 0.14985 * 1.000005 ^ Variable 13% 12% R2= 0.398 11% 10% 0 0 0 0 0 0 0 0 0 , 00 , 00 , 00 , 00 , 00 , 00 , 00 , 00 , 00 Respondents Case 20 21 22 23 24 25 26 27 28 Conventional Respondents Kondatrieff FIRR = 0.14586 + 0.000001 * Variable R2= 0.505 Chance of Failure by Variable Value and Forecast Case 18% 16% FIRR = 0.14832 * 1.000007 ^ Variable 14% R2= 0.500 12% 10% 8% Kondratieff Case 6% 4% FIRR = 0.09223 + 0.000003 * Variable 2% R2= 0.656 0% 0 0 0 0 0 0 0 0 0 , 00 , 00 , 00 , 00 , 00 , 00 , 00 , 00 , 00 FIRR = 0.10269 * 1.000017 ^ Variable 20 21 22 23 24 25 26 27 28 Conventional Respondents Kondatrieff R2= 0.654DissFinal Page 206 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Capacity per Local Lane (pcus) Minimum: 8,000 Distribution of Variable 100% 6% Maximum: 12,000 Mean: 9,994 80% 5% 4% Monte Carlo Settings: 60% Modal Value 10,000 3% 40% % < Modal 50% 2% SD (<Modal) 1,000 20% 1% SD (>Modal) 1,000 0% 0% 0 00 5 00 0 00 5 00 00 00 00 00 00 8, 8, 9, 9, ,0 ,5 ,0 ,5 ,0 10 10 11 11 12 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17226 + -1.24E-07 * Variable 16% R2= 0.004 15% 14% FIRR = 0.17223 * 0.999999 ^ Variable 13% 12% R2= 0.004 11% 10% 0 00 5 00 0 00 5 00 00 00 00 50 0 00 0 8, 8, 9, 9, ,0 ,5 ,0 , , Respondents Case 10 10 11 11 12 Conventional Respondents Kondatrieff FIRR = 0.17887 + -1.04E-07 * Variable R2= 0.002 Chance of Failure by Variable Value and Forecast Case 18% 16% FIRR = 0.17885 * 0.999999 ^ Variable 14% R2= 0.002 12% 10% 8% Kondratieff Case 6% 4% FIRR = 0.16067 + -6.81E-07 * Variable 2% R2= 0.032 0% 0 00 5 00 0 00 5 00 00 00 00 50 0 00 0 8, 8, 9, 9, ,0 ,5 ,0 , , FIRR = 0.16086 * 0.999996 ^ Variable 10 10 11 11 12 Conventional Respondents Kondatrieff R2= 0.033DissFinal Page 207 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Total Construction Cost (Base Year million$) Minimum: 166.79 million Distribution of Variable 100% 6% Maximum: 240.92 million Mean: 199.77 million 80% 5% 4% Monte Carlo Settings: 60% Modal Value 185.3 USDm* 3% % < Modal 25% 40% 2% SD (<Modal) 5% USDm** 20% SD (>Modal) 15% USDm** 1% Note: * USD4.633 per km 0% 0% Note: ** SD as % of Modal Value 7. 5 7. 5 7. 5 7. 5 7. 5 7. 5 7. 5 7. 5 16 17 18 19 20 21 22 23 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 22% Conventional Case 20% 18% FIRR = 0.2952068 + -0.00062 * Variable R2= 0.960 16% 14% FIRR = 0.3574268 * 0.996323 ^ Variable R2= 0.955 12% 10% 5 5 5 5 5 5 5 5 7. 7. 7. 7. 7. 7. 7. 7. Respondents Case 16 17 18 19 20 21 22 23 Conventional Respondents Kondatrieff FIRR = 0.3122438 + -0.00067 * Variable R2= 0.961 Chance of Failure by Variable Value and Forecast Case 30% FIRR = 0.3837237 * 0.996162 ^ Variable 25% R2= 0.955 20% 15% Kondratieff Case 10% FIRR = 0.3220731 + -0.00083 * Variable 5% R2= 0.941 0% 5 5 5 5 5 5 5 5 7. 7. 7. 7. 7. 7. 7. 7. FIRR = 0.4713168 * 0.994423 ^ Variable 16 17 18 19 20 21 22 23 Conventional Respondents Kondatrieff R2= 0.929DissFinal Page 208 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Construction Duration (Quarters) Minimum: 8 Distribution of Variable 100% 30% Maximum: 14 Mean: 11.08 25% 80% Monte Carlo Settings: 20% 60% Modal Value 10 15% % < Modal 20% 40% SD (<Modal) 1 10% SD (>Modal) 2 20% 5% 0% 0% 8 9 10 11 12 13 14 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.19502 + -0.00217 * Variable 16% R2= 0.989 15% 14% FIRR = 0.19672 * 0.987384 ^ Variable 13% R2= 0.989 12% 11% 10% Respondents Case 8 9 10 11 12 13 14 Conventional Respondents Kondatrieff FIRR = 0.20836 + -0.00276 * Variable R2= 0.988 20% Chance of Failure by Variable Value and Forecast Case 18% FIRR = 0.21105 * 0.984577 ^ Variable 16% R2= 0.987 14% 12% 10% Kondratieff Case 8% 6% FIRR = 0.21041 + -0.00504 * Variable 4% R2= 0.968 2% 0% FIRR = 0.2212 * 0.967949 ^ Variable 8 9 10 11 12 13 14 Conventional Respondents Kondatrieff R2= 0.969DissFinal Page 209 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Fixed Annual Operations and Maintenance Costs (as % of Construction Cost) Minimum: 0.1% Distribution of Variable 100% 18% Maximum: 4.0% 16% Mean: 2.2% 80% 14% Monte Carlo Settings: 12% 60% Modal Value 2% * 10% % < Modal 50% 8% 40% SD (<Modal) 0.5% 6% SD (>Modal) 1.0% 4% 20% Note: * as % of initial construction cost 2% 0% 0% 0.1% 1.1% 2.1% 3.1% 4.1% Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.18088 + -0.46747 * Variable 16% R2= 0.170 15% 14% FIRR = 0.17939 * 0.088485 ^ Variable 13% R2= 0.128 12% 11% 10% Respondents Case 0.1% 1.1% 2.1% 3.1% 4.1% Conventional Respondents Kondatrieff FIRR = 0.19135 + -0.58727 * Variable R2= 0.329 Chance of Failure by Variable Value and Forecast Case 25% FIRR = 0.19064 * 0.044776 ^ Variable 20% R2= 0.316 15% Kondratieff Case 10% FIRR = 0.17735 + -0.96474 * Variable 5% R2= 0.504 0% FIRR = 0.17678 * 0.002841 ^ Variable 0.1% 1.1% 2.1% 3.1% 4.1% Conventional Respondents Kondatrieff R2= 0.536DissFinal Page 210 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Variable Operations and Maintenance Costs (as % of Revenues) Minimum: 1.0% Distribution of Variable 100% 9% Maximum: 5.0% 8% Mean: 3.0% 80% 7% Monte Carlo Settings: 6% 60% Modal Value 3% * 5% % < Modal 50% 4% 40% SD (<Modal) 1% 3% SD (>Modal) 1% 2% 20% Note: * as % of revenues 1% 0% 0% 1.00% 1.80% 2.60% 3.40% 4.20% 5.00% Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17554 + -0.15022 * Variable 16% R2= 0.468 15% 14% FIRR = 0.17557 * 0.415625 ^ Variable 13% R2= 0.468 12% 11% 10% Respondents Case 1.00% 1.80% 2.60% 3.40% 4.20% 5.00% Conventional Respondents Kondatrieff FIRR = 0.1825 + -0.15471 * Variable R2= 0.434 20% Chance of Failure by Variable Value and Forecast Case 18% FIRR = 0.18253 * 0.419220 ^ Variable 16% R2= 0.434 14% 12% 10% Kondratieff Case 8% 6% FIRR = 0.15889 + -0.15932 * Variable 4% R2= 0.159 2% 0% FIRR = 0.1588 * 0.361829 ^ Variable 1.00% 1.80% 2.60% 3.40% 4.20% 5.00% Conventional Respondents Kondatrieff R2= 0.154DissFinal Page 211 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Base Value of Time ($/hr) for Small Vehicles Minimum: 2.00 $/hour Distribution of Variable 100% 9% Maximum: 6.00 $/hour 8% Mean: 4.00 $/hour 80% 7% Monte Carlo Settings: 6% 60% Modal Value 4 5% % < Modal 50% 4% 40% SD (<Modal) 1 3% SD (>Modal) 1 2% 20% 1% 0% 0% 2.0 2.8 3.6 4.4 5.2 6.0 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17092 + 0.000061 * Variable 16% R2= 0.001 15% 14% FIRR = 0.17089 * 1.000384 ^ Variable 13% R2= 0.001 12% 11% 10% Respondents Case 2.0 2.8 3.6 4.4 5.2 6.0 Conventional Respondents Kondatrieff FIRR = 0.17837 + -0.00005 * Variable R2= 0.001 Chance of Failure by Variable Value and Forecast Case 18% 16% FIRR = 0.17835 * 0.999741 ^ Variable 14% R2= 0.001 12% 10% Kondratieff Case 8% 6% FIRR = 0.15427 + 0.000071 * Variable 4% R2= 0.0004 2% 0% FIRR = 0.15419 * 1.000505 ^ Variable 2.0 2.8 3.6 4.4 5.2 6.0 Conventional Respondents Kondatrieff R2= 0.0005DissFinal Page 212 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Base Value of Time ($/hr) for Large Vehicles Minimum: 1.00 $/hour Distribution of Variable 100% 9% Maximum: 5.00 $/hour 8% Mean: 3.01 $/hour 80% 7% Monte Carlo Settings: 6% 60% Modal Value 3 5% % < Modal 50% 4% 40% SD (<Modal) 1 3% SD (>Modal) 1 2% 20% 1% 0% 0% 1.0 1.8 2.6 3.4 4.2 5.0 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17087 + 0.00004 * Variable 16% R2= 0.001 15% 14% FIRR = 0.17086 * 1.000212 ^ Variable 13% R2= 0.001 12% 11% 10% Respondents Case 1.0 1.8 2.6 3.4 4.2 5.0 Conventional Respondents Kondatrieff FIRR = 0.17881 + -0.00031 * Variable R2= 0.032 Chance of Failure by Variable Value and Forecast Case 18% 16% FIRR = 0.1788 * 0.998246 ^ Variable 14% R2= 0.032 12% 10% Kondratieff Case 8% 6% FIRR = 0.15591 + -0.00055 * Variable 4% R2= 0.032 2% 0% FIRR = 0.15588 * 0.996423 ^ Variable 1.0 1.8 2.6 3.4 4.2 5.0 Conventional Respondents Kondatrieff R2= 0.032DissFinal Page 213 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Income Elasticity of Value of Time (Small Vehicles) Minimum: 0.20 Distribution of Variable 100% 14% Maximum: 0.80 Mean: 0.50 12% 80% 10% Monte Carlo Settings: 60% Modal Value 0.5 8% % < Modal 50% 6% 40% SD (<Modal) 0.15 4% SD (>Modal) 0.15 20% 2% 0% 0% 0.20 0.40 0.60 0.80 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.1719 + -0.00074 * Variable 16% R2= 0.007 15% 14% FIRR = 0.17187 * 0.995842 ^ Variable 13% R2= 0.007 12% 11% 10% Respondents Case 0.20 0.40 0.60 0.80 Conventional Respondents Kondatrieff FIRR = 0.17828 + 0.000159 * Variable R2= 0.0003 Chance of Failure by Variable Value and Forecast Case 16% 14% FIRR = 0.17825 * 1.001120 ^ Variable R2= 0.0004 12% 10% 8% Kondratieff Case 6% FIRR = 0.15602 + -0.00228 * Variable 4% R2= 0.027 2% 0% FIRR = 0.15593 * 0.986271 ^ Variable 0.20 0.40 0.60 0.80 Conventional Respondents Kondatrieff R2= 0.024DissFinal Page 214 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Income Elasticity of Value of Time (Large Vehicles) Minimum: 0.20 Distribution of Variable 100% 16% Maximum: 0.80 Mean: 0.50 14% 80% 12% Monte Carlo Settings: 10% 60% Modal Value 0.5 8% % < Modal 50% 40% 6% SD (<Modal) 0.15 SD (>Modal) 0.15 4% 20% 2% 0% 0% 0.20 0.40 0.60 0.80 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17136 + -0.00137 * Variable 16% R2= 0.031 15% 14% FIRR = 0.17135 * 0.991996 ^ Variable 13% R2= 0.032 12% 11% 10% Respondents Case 0.20 0.40 0.60 0.80 Conventional Respondents Kondatrieff FIRR = 0.17885 + -0.00275 * Variable R2= 0.136 Chance of Failure by Variable Value and Forecast Case 18% 16% FIRR = 0.17885 * 0.984583 ^ Variable 14% R2= 0.136 12% 10% Kondratieff Case 8% 6% FIRR = 0.15044 + 0.00561 * Variable 4% R2= 0.118 2% 0% FIRR = 0.15042 * 1.037364 ^ Variable 0.20 0.40 0.60 0.80 Conventional Respondents Kondatrieff R2= 0.120DissFinal Page 215 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Base Expressway Vehicle Operating Costs ($/km) for Small Vehicles Minimum: 0.03 Distribution of Variable 100% 14% Maximum: 0.09 Mean: 0.06 12% 80% 10% Monte Carlo Settings: 60% Modal Value 0.06 8% % < Modal 50% 6% 40% SD (<Modal) 0.015 4% SD (>Modal) 0.015 20% 2% 0% 0% 0.030 0.050 0.070 0.090 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.16901 + 0.033129 * Variable 16% R2= 0.130 15% 14% FIRR = 0.169 * 1.214432 ^ Variable 13% R2= 0.131 12% 11% 10% Respondents Case 0.030 0.050 0.070 0.090 Conventional Respondents Kondatrieff FIRR = 0.17596 + 0.030190 * Variable R2= 0.102 Chance of Failure by Variable Value and Forecast Case 16% 14% FIRR = 0.17595 * 1.185655 ^ Variable R2= 0.102 12% 10% 8% Kondratieff Case 6% FIRR = 0.1515 + 0.041592 * Variable 4% R2= 0.067 2% 0% FIRR = 0.15152 * 1.305841 ^ Variable 0.030 0.050 0.070 0.090 Conventional Respondents Kondatrieff R2= 0.066DissFinal Page 216 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Base Expressway Vehicle Operating Costs ($/km) for Large Vehicles Minimum: 0.05 Distribution of Variable 100% 9% Maximum: 0.15 8% Mean: 0.10 80% 7% Monte Carlo Settings: 6% 60% Modal Value 0.1 5% % < Modal 50% 4% 40% SD (<Modal) 0.025 3% SD (>Modal) 0.025 2% 20% 1% 0% 0% 0.050 0.070 0.090 0.110 0.130 0.150 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.16677 + 0.037643 * Variable 16% R2= 0.359 15% 14% FIRR = 0.1668 * 1.247046 ^ Variable 13% R2= 0.361 12% 11% 10% Respondents Case 0.050 0.070 0.090 0.110 0.130 0.150 Conventional Respondents Kondatrieff FIRR = 0.17422 + 0.031820 * Variable R2= 0.251 Chance of Failure by Variable Value and Forecast Case 20% 18% FIRR = 0.17424 * 1.196431 ^ Variable 16% R2= 0.251 14% 12% 10% Kondratieff Case 8% 6% FIRR = 0.15235 + 0.01331 * Variable 4% R2= 0.013 2% 0% FIRR = 0.15238 * 1.085659 ^ Variable 0.050 0.070 0.090 0.110 0.130 0.150 Conventional Respondents Kondatrieff R2= 0.012DissFinal Page 217 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Local Roads VOC Multiplier for Small Vehicles (Expressway VOC * Factor) Minimum: 1.00 Distribution of Variable 100% 9% Maximum: 2.00 8% Mean: 1.50 80% 7% 6% Monte Carlo Settings: 60% 5% Modal Value 1.5 4% % < Modal 50% 40% 3% SD (<Modal) 0.25 SD (>Modal) 0.25 20% 2% 1% 0% 0% 0 00 2 00 40 0 60 0 80 0 00 0 1. 1. 1. 1. 1. 2. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.16777 + 0.002163 * Variable 16% R2= 0.141 15% 14% FIRR = 0.1678 * 1.012712 ^ Variable 13% R2= 0.141 12% 11% 10% 0 00 2 00 4 00 6 00 8 00 0 00 Respondents Case 1. 1. 1. 1. 1. 2. Conventional Respondents Kondatrieff FIRR = 0.17628 + 0.001067 * Variable R2= 0.038 Chance of Failure by Variable Value and Forecast Case 16% 14% FIRR = 0.1763 * 1.005942 ^ Variable 12% R2= 0.037 10% 8% Kondratieff Case 6% 4% FIRR = 0.15063 + 0.002226 * Variable 2% R2= 0.040 0% 0 00 2 00 4 00 6 00 8 00 0 00 FIRR = 0.15068 * 1.014337 ^ Variable 1. 1. 1. 1. 1. 2. Conventional Respondents Kondatrieff R2= 0.039DissFinal Page 218 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Local Roads VOC Multiplier for Large Vehicles (Expressway VOC * Factor) Minimum: 1.50 Distribution of Variable 100% 9% Maximum: 2.50 8% Mean: 2.00 80% 7% 6% Monte Carlo Settings: 60% 5% Modal Value 2 4% % < Modal 50% 40% 3% SD (<Modal) 0.25 SD (>Modal) 0.25 20% 2% 1% 0% 0% 5 00 7 00 90 0 10 0 30 0 50 0 1. 1. 1. 2. 2. 2. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.16216 + 0.004343 * Variable 16% R2= 0.258 15% 14% FIRR = 0.16234 * 1.025811 ^ Variable 13% R2= 0.258 12% 11% 10% 5 00 7 00 9 00 1 00 3 00 5 00 Respondents Case 1. 1. 1. 2. 2. 2. Conventional Respondents Kondatrieff FIRR = 0.16966 + 0.003963 * Variable R2= 0.229 Chance of Failure by Variable Value and Forecast Case 18% 16% FIRR = 0.1698 * 1.022631 ^ Variable 14% R2= 0.228 12% 10% 8% Kondratieff Case 6% 4% FIRR = 0.13897 + 0.00737 * Variable 2% R2= 0.335 0% 5 00 7 00 9 00 1 00 3 00 5 00 FIRR = 0.13942 * 1.049863 ^ Variable 1. 1. 1. 2. 2. 2. Conventional Respondents Kondatrieff R2= 0.335DissFinal Page 219 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Base Traffic Factor (Small Vehicles) to Expand/Contract Initial Demand Minimum: 0.70 Distribution of Variable 100% 14% Maximum: 1.30 Mean: 1.00 12% 80% 10% Monte Carlo Settings: 60% 8% Modal Value 1 % < Modal 50% 40% 6% SD (<Modal) 0.15 4% SD (>Modal) 0.15 20% 2% 0% 0% 7 00 90 0 1 00 30 0 0. 0. 1. 1. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.10093 + 0.068177 * Variable 16% R2= 0.994 15% 14% FIRR = 0.11246 * 1.499343 ^ Variable 13% R2= 0.991 12% 11% 10% 7 00 9 00 1 00 3 00 Respondents Case 0. 0. 1. 1. Conventional Respondents Kondatrieff FIRR = 0.10692 + 0.068921 * Variable R2= 0.996 Chance of Failure by Variable Value and Forecast Case 25% FIRR = 0.11823 * 1.483236 ^ Variable 20% R2= 0.993 15% Kondratieff Case 10% FIRR = 0.06407 + 0.087354 * Variable 5% R2= 0.989 0% 7 00 9 00 1 00 3 00 FIRR = 0.08385 * 1.795023 ^ Variable 0. 0. 1. 1. Conventional Respondents Kondatrieff R2= 0.980DissFinal Page 220 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Base Traffic Factor (Large Vehicles) to Expand/Contract Initial Demand Minimum: 0.70 Distribution of Variable 100% 14% Maximum: 1.30 Mean: 1.00 12% 80% 10% Monte Carlo Settings: 60% 8% Modal Value 1 % < Modal 50% 40% 6% SD (<Modal) 0.15 4% SD (>Modal) 0.15 20% 2% 0% 0% 7 00 90 0 1 00 30 0 0. 0. 1. 1. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.10347 + 0.066009 * Variable 16% R2= 0.975 15% 14% FIRR = 0.1144 * 1.477431 ^ Variable 13% R2= 0.973 12% 11% 10% 7 00 9 00 1 00 3 00 Respondents Case 0. 0. 1. 1. Conventional Respondents Kondatrieff FIRR = 0.10919 + 0.067103 * Variable R2= 0.970 Chance of Failure by Variable Value and Forecast Case 25% FIRR = 0.1201 * 1.464043 ^ Variable 20% R2= 0.968 15% Kondratieff Case 10% FIRR = 0.07515 + 0.07755 * Variable 5% R2= 0.968 0% 7 00 9 00 1 00 3 00 FIRR = 0.09152 * 1.660623 ^ Variable 0. 0. 1. 1. Conventional Respondents Kondatrieff R2= 0.970DissFinal Page 221 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Income Elasticity of Traffic (Small Vehicles) Minimum: 0.85 Distribution of Variable 100% 12% Maximum: 1.65 Mean: 1.25 80% 10% 8% Monte Carlo Settings: 60% Modal Value 1.25 6% % < Modal 50% 40% 4% SD (<Modal) 0.2 SD (>Modal) 0.2 20% 2% 0% 0% 9 00 1 00 30 0 50 0 0. 1. 1. 1. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.13436 + 0.028688 * Variable 16% R2= 0.950 15% 14% FIRR = 0.13782 * 1.183292 ^ Variable 13% R2= 0.949 12% 11% 10% 9 00 1 00 3 00 5 00 Respondents Case 0. 1. 1. 1. Conventional Respondents Kondatrieff FIRR = 0.1383 + 0.030938 * Variable R2= 0.960 Chance of Failure by Variable Value and Forecast Case 25% FIRR = 0.14214 * 1.190834 ^ Variable 20% R2= 0.957 15% Kondratieff Case 10% FIRR = 0.09821 + 0.043622 * Variable 5% R2= 0.921 0% 9 00 1 00 3 00 5 00 FIRR = 0.10639 * 1.332811 ^ Variable 0. 1. 1. 1. Conventional Respondents Kondatrieff R2= 0.910DissFinal Page 222 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Income Elasticity of Traffic (Large Vehicles) Minimum: 0.70 Distribution of Variable 100% 12% Maximum: 1.50 Mean: 1.10 80% 10% 8% Monte Carlo Settings: 60% Modal Value 1.1 6% % < Modal 50% 40% 4% SD (<Modal) 0.2 SD (>Modal) 0.2 20% 2% 0% 0% 7 00 9 00 10 0 30 0 50 0 0. 0. 1. 1. 1. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.14173 + 0.026119 * Variable 16% R2= 0.967 15% 14% FIRR = 0.14394 * 1.165400 ^ Variable 13% R2= 0.969 12% 11% 10% 7 00 9 00 1 00 3 00 5 00 Respondents Case 0. 0. 1. 1. 1. Conventional Respondents Kondatrieff FIRR = 0.14477 + 0.029486 * Variable R2= 0.969 Chance of Failure by Variable Value and Forecast Case 20% 18% FIRR = 0.14747 * 1.180804 ^ Variable 16% R2= 0.971 14% 12% 10% Kondratieff Case 8% 6% FIRR = 0.10672 + 0.04219 * Variable 4% 2% R2= 0.933 0% 7 00 9 00 1 00 3 00 5 00 FIRR = 0.11279 * 1.317465 ^ Variable 0. 0. 1. 1. 1. Conventional Respondents Kondatrieff R2= 0.933DissFinal Page 223 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Toll Revenue Leakage (%) Minimum: 5% Distribution of Variable 100% 16% Maximum: 20% 14% Mean: 11% 80% 12% Monte Carlo Settings: 60% 10% Modal Value 10% 8% % < Modal 50% 40% 6% SD (<Modal) 2.5% 4% SD (>Modal) 5% 20% 2% 0% 0% 5% 9% 13 % 17 % 21 % Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.18011 + -0.07629 * Variable 16% R2= 0.833 15% 14% FIRR = 0.18033 * 0.639348 ^ Variable 13% R2= 0.833 12% 11% 10% Respondents Case 5% 9% 13 % 17 % 21 % Conventional Respondents Kondatrieff FIRR = 0.18704 + -0.07777 * Variable R2= 0.805 18% Chance of Failure by Variable Value and Forecast Case 16% FIRR = 0.18728 * 0.644238 ^ Variable 14% R2= 0.801 12% 10% 8% Kondratieff Case 6% FIRR = 0.16656 + -0.10153 * Variable 4% R2= 0.788 2% 0% FIRR = 0.16707 * 0.514851 ^ Variable 5% 9% 13 % 17 % 21 % Conventional Respondents Kondatrieff R2= 0.787DissFinal Page 224 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Initial Amplitude of Ramp-Up (%) Minimum: 20% Distribution of Variable 100% 20% Maximum: 80% 18% Mean: 44% 80% 16% 14% Monte Carlo Settings: 60% 12% Modal Value 40% 10% % < Modal 50% 40% 8% SD (<Modal) 10% 6% SD (>Modal) 20% 20% 4% 2% 0% 0% 20% 40% 60% 80% Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17962 + -0.01845 * Variable 16% R2= 0.861 15% 14% FIRR = 0.17983 * 0.897345 ^ Variable 13% R2= 0.862 12% 11% 10% % % % % .0 .0 .0 .0 Respondents Case 20 40 60 80 Conventional Respondents Kondatrieff FIRR = 0.18678 + -0.01924 * Variable R2= 0.866 Chance of Failure by Variable Value and Forecast Case 20% 18% FIRR = 0.18701 * 0.896982 ^ Variable 16% R2= 0.867 14% 12% 10% Kondratieff Case 8% 6% FIRR = 0.16915 + -0.03187 * Variable 4% 2% R2= 0.826 0% % % % % .0 .0 .0 .0 FIRR = 0.16997 * 0.810841 ^ Variable 20 40 60 80 Conventional Respondents Kondatrieff R2= 0.818DissFinal Page 225 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Ramp-Up Duration (Quarters) Minimum: 4 Distribution of Variable 100% 16% Maximum: 20 14% Mean: 9.79 80% 12% Monte Carlo Settings: 60% 10% Modal Value 8 8% % < Modal 40% 40% 6% SD (<Modal) 2 SD (>Modal) 5 4% 20% 2% 0% 0% 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.18002 + -0.00094 * Variable 16% R2= 0.792 15% 14% FIRR = 0.18042 * 0.994412 ^ Variable 13% R2= 0.785 12% 11% 10% Respondents Case 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Conventional Respondents Kondatrieff FIRR = 0.18777 + -0.00103 * Variable R2= 0.788 20% Chance of Failure by Variable Value and Forecast Case 18% FIRR = 0.18822 * 0.994098 ^ Variable 16% R2= 0.780 14% 12% 10% Kondratieff Case 8% 6% FIRR = 0.1693 + -0.00156 * Variable 4% R2= 0.834 2% 0% FIRR = 0.17053 * 0.989561 ^ Variable 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Conventional Respondents Kondatrieff R2= 0.822DissFinal Page 226 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Small Vehicles Tolling Penalty (Minutes) Minimum: 0 Distribution of Variable 100% 10% Maximum: 20 9% Mean: 10 80% 8% 7% Monte Carlo Settings: 60% 6% Modal Value 10 5% % < Modal 50% 40% 4% SD (<Modal) 5 3% SD (>Modal) 5 20% 2% 1% 0% 0% 0 4 8 12 16 20 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17277 + -0.00015 * Variable 16% R2= 0.240 15% 14% FIRR = 0.17275 * 0.999112 ^ Variable 13% R2= 0.238 12% 11% 10% Respondents Case 0 4 8 12 16 20 Conventional Respondents Kondatrieff FIRR = 0.17982 + -0.00017 * Variable R2= 0.291 16% Chance of Failure by Variable Value and Forecast Case 14% FIRR = 0.1798 * 0.999047 ^ Variable R2= 0.290 12% 10% 8% Kondratieff Case 6% 4% FIRR = 0.15837 + -0.00035 * Variable R2= 0.417 2% 0% FIRR = 0.15833 * 0.997740 ^ Variable 0 4 8 12 16 20 Conventional Respondents Kondatrieff R2= 0.422DissFinal Page 227 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Large Vehicles Tolling Penalty (Minutes) Minimum: 5 Distribution of Variable 100% 9% Maximum: 25 8% Mean: 15 80% 7% Monte Carlo Settings: 6% 60% Modal Value 15 5% % < Modal 50% 40% 4% SD (<Modal) 5 3% SD (>Modal) 5 20% 2% 1% 0% 0% 5 9 13 17 21 25 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17198 + -0.00006 * Variable 16% R2= 0.042 15% 14% FIRR = 0.17195 * 0.999631 ^ Variable 13% R2= 0.040 12% 11% 10% Respondents Case 5 9 13 17 21 25 Conventional Respondents Kondatrieff FIRR = 0.17833 + -0.00003 * Variable R2= 0.007 18% Chance of Failure by Variable Value and Forecast Case 16% FIRR = 0.1783 * 0.999852 ^ Variable 14% R2= 0.007 12% 10% Kondratieff Case 8% 6% FIRR = 0.15341 + 0.00003 * Variable 4% R2= 0.005 2% 0% FIRR = 0.1534 * 1.000199 ^ Variable 5 9 13 17 21 25 Conventional Respondents Kondatrieff R2= 0.004DissFinal Page 228 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Routeing Sensitivity ("Lambda") for Small Vehicles Minimum: 0.025 Distribution of Variable 100% 18% Maximum: 0.075 16% Mean: 0.050 80% 14% 12% Monte Carlo Settings: 60% 10% Modal Value 0.05 8% % < Modal 50% 40% 6% SD (<Modal) 0.0125 SD (>Modal) 0.0125 20% 4% 2% 0% 0% 0 25 04 5 06 5 0. 0. 0. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.16919 + 0.02656 * Variable 16% R2= 0.040 15% 14% FIRR = 0.16917 * 1.170635 ^ Variable 13% R2= 0.040 12% 11% 10% 0 25 0 45 0 65 Respondents Case 0. 0. 0. Conventional Respondents Kondatrieff FIRR = 0.17557 + 0.03624 * Variable R2= 0.084 Chance of Failure by Variable Value and Forecast Case 16% 14% FIRR = 0.17553 * 1.231156 ^ Variable 12% R2= 0.086 10% 8% Kondratieff Case 6% 4% FIRR = 0.15357 + 0.00169 * Variable 2% R2= 0.0001 0% 0 25 0 45 0 65 FIRR = 0.15348 * 1.018592 ^ Variable 0. 0. 0. Conventional Respondents Kondatrieff R2= 0.0003DissFinal Page 229 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Routeing Sensitivity ("Lambda") for Large Vehicles Minimum: 0.025 Distribution of Variable 100% 18% Maximum: 0.075 16% Mean: 0.050 80% 14% 12% Monte Carlo Settings: 60% 10% Modal Value 0.05 8% % < Modal 50% 40% 6% SD (<Modal) 0.0125 SD (>Modal) 0.0125 20% 4% 2% 0% 0% 0 25 04 5 06 5 0. 0. 0. Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17011 + 0.01230 * Variable 16% R2= 0.030 15% 14% FIRR = 0.1701 * 1.075460 ^ Variable 13% R2= 0.030 12% 11% 10% 0 25 0 45 0 65 Respondents Case 0. 0. 0. Conventional Respondents Kondatrieff FIRR = 0.1773 + 0.00671 * Variable R2= 0.010 Chance of Failure by Variable Value and Forecast Case 16% 14% FIRR = 0.17729 * 1.038873 ^ Variable 12% R2= 0.010 10% 8% Kondratieff Case 6% 4% FIRR = 0.15144 + 0.04419 * Variable 2% R2= 0.112 0% 0 25 0 45 0 65 FIRR = 0.15142 * 1.336206 ^ Variable 0. 0. 0. Conventional Respondents Kondatrieff R2= 0.114DissFinal Page 230 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Toll Escalation Rate (% of RPI Inflation) Minimum: 60% Distribution of Variable 100% 20% Maximum: 100% 18% Mean: 86% 80% 16% 14% Monte Carlo Settings: 60% 12% Modal Value 90% 10% % < Modal 50% 40% 8% SD (<Modal) 15% 6% SD (>Modal) 5% 20% 4% 2% 0% 0% 62.5% 72.5% 82.5% 92.5% Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.13572 + 0.04035 * Variable 16% R2= 0.861 15% 14% FIRR = 0.13862 * 1.270939 ^ Variable 13% R2= 0.856 12% 11% 10% Respondents Case 62.5% 72.5% 82.5% 92.5% Conventional Respondents Kondatrieff FIRR = 0.13126 + 0.05336 * Variable R2= 0.917 Chance of Failure by Variable Value and Forecast Case 25% FIRR = 0.13604 * 1.358413 ^ Variable 20% R2= 0.912 15% Kondratieff Case 10% FIRR = 0.07763 + 0.08760 * Variable 5% R2= 0.964 0% FIRR = 0.09193 * 1.803628 ^ Variable 62.5% 72.5% 82.5% 92.5% Conventional Respondents Kondatrieff R2= 0.960DissFinal Page 231 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Quarters between Toll Increases Minimum: 8 Distribution of Variable 100% 16% Maximum: 20 14% Mean: 13.29 80% 12% Monte Carlo Settings: 60% 10% Modal Value 12 8% % < Modal 40% 40% 6% SD (<Modal) 2 SD (>Modal) 4 4% 20% 2% 0% 0% 8 9 10 11 12 13 14 15 16 17 18 19 20 Cumulative (Left Axis) Probability Density (Right Axis) Regression Analysis: FIRR on Variable Mean FIRR by Variable Value and Forecast Case 20% 19% Conventional Case 18% 17% FIRR = 0.17765 + -0.00051 * Variable 16% R2= 0.565 15% 14% FIRR = 0.17781 * 0.997028 ^ Variable 13% R2= 0.561 12% 11% 10% Respondents Case 8 9 10 11 12 13 14 15 16 17 18 19 20 Conventional Respondents Kondatrieff FIRR = 0.18744 + -0.00073 * Variable R2= 0.690 16% Chance of Failure by Variable Value and Forecast Case 14% FIRR = 0.18773 * 0.995904 ^ Variable R2= 0.687 12% 10% 8% Kondratieff Case 6% FIRR = 0.17121 + -0.00129 * Variable 4% R2= 0.850 2% 0% FIRR = 0.17223 * 0.991630 ^ Variable 8 9 10 11 12 13 14 15 16 17 18 19 20 Conventional Respondents Kondatrieff R2= 0.851DissFinal Page 232 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: GDP Growth (% p.a.) Page 1 of 2 Conventional Distribution of Variable: Conventional 100% 12% Minimum: 2% Maximum: 10% 80% 10% Mean: 6% 8% 60% Modal Value 6% 6% % < Modal 50% 40% 4% SD (<Modal) 2% 20% SD (>Modal) 2% 2% 0% 0% 2% 4% 6% 8% 10% Cumulative (Left Axis) Probability Density (Right Axis) Respondents Distribution of Variable: Respondents 100% 12% Minimum: 2% Maximum: 12% 80% 10% Mean: 7% 8% 60% Modal Value 1% added to Conventional 6% % < Modal 50% 40% 4% SD (<Modal) 0.5% 20% SD (>Modal) 0.5% 2% 0% 0% 2% 4% 6% 8% 10% 12% Cumulative (Left Axis) Probability Density (Right Axis) Kondratieff Distribution of Variable: Kondratieff 100% 10% Minimum: 2% 9% Maximum: 14% 80% 8% Mean: 8% 7% 60% 6% Modal Value 1% added to Respondents 5% % < Modal 50% 40% 4% SD (<Modal) 0.5% 3% SD (>Modal) 0.5% 20% 2% 1% 0% 0% 2% 4% 6% 8% 10% 12% 14% Cumulative (Left Axis) Probability Density (Right Axis) All Cases Distribution of Variable: All Cases 100% 10% Minimum: 2% 9% Maximum: 14% 80% 8% Mean: 7% 7% 60% 6% Modal Value n/a 5% % < Modal n/a 40% 4% SD (<Modal) n/a 3% SD (>Modal) n/a 20% 2% 1% 0% 0% 2% 4% 6% 8% 10% 12% 14% Cumulative (Left Axis) Probability Density (Right Axis)DissFinal Page 233 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: GDP Growth (% p.a.) Page 2 of 2 Regression Analysis: FIRR on Variable Conventional Case Mean FIRR by Variable Value and Forecast Case 25% FIRR = 0.07953 + 1.408371 * Variable R2= 0.962 20% FIRR = 0.0923 * 8818 ^ Variable R2= 0.908 15% 10% Respondents Case 5% FIRR = 0.07433 + 1.349827 * Variable R2= 0.935 0% 2% 4% 6% 8% 10% 12% 14% Conventional Respondents Kondatrieff All FIRR = 0.08764 * 6825 ^ Variable R2= 0.860 Kondratieff Case FIRR = 0.03262 + 1.384828 * Variable R2= 0.938 Chance of Failure by Variable Value and Forecast Case 60% FIRR = 0.05548 * 55033 ^ Variable 50% R2= 0.855 40% All Cases 30% 20% FIRR = 0.09232 + 0.94170 * Variable R2= 0.905 10% FIRR = 0.09977 * 472 ^ Variable 0% R2= 0.854 2% 4% 6% 8% 10% 12% 14% Conventional Respondents Kondatrieff AllDissFinal Page 234 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Price Inflation for Vehicle Operating Costs (% p.a.) Page 1 of 2 Conventional Distribution of Variable: Conventional 100% 25% Minimum: 0% Maximum: 5% 80% 20% Mean: 2% 60% 15% Modal Value 2.5% % < Modal 50% 40% 10% SD (<Modal) 1% SD (>Modal) 1% 20% 5% 0% 0% 0% 1% 2% 3% 4% Cumulative (Left Axis) Probability Density (Right Axis) Respondents Distribution of Variable: Respondents 100% 14% Minimum: 0% 12% Maximum: 11% 80% Mean: 5% 10% 60% Modal Value 2.5% added to Conventional 8% % < Modal 50% 40% 6% SD (<Modal) 1.5% 4% SD (>Modal) 1.5% 20% 2% 0% 0% 0% 2% 4% 6% 8% 10% Cumulative (Left Axis) Probability Density (Right Axis) Kondratieff Distribution of Variable: Kondratieff 100% 14% Minimum: 0% 12% Maximum: 11% 80% Mean: 5% 10% 60% Modal Value 0% same as Respondents 8% % < Modal n/a 40% 6% SD (<Modal) n/a 4% SD (>Modal) n/a 20% 2% 0% 0% 1% 3% 5% 7% 9% 11% 13% Cumulative (Left Axis) Probability Density (Right Axis) All Cases Distribution of Variable: All Cases 100% 12% Minimum: 0% Maximum: 11% 80% 10% Mean: 4% 8% 60% Modal Value n/a 6% % < Modal n/a 40% SD (<Modal) n/a 4% SD (>Modal) n/a 20% 2% 0% 0% 0% 2% 4% 6% 8% 10% 12% Cumulative (Left Axis) Probability Density (Right Axis)DissFinal Page 235 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Price Inflation for Vehicle Operating Costs (% p.a.) Page 2 of 2 Regression Analysis: FIRR on Variable Conventional Case Mean FIRR by Variable Value and Forecast Case 25% FIRR = 0.17127 + -0.00581 * Variable R2= 0.003 20% FIRR = 0.17126 * 0.967 ^ Variable R2= 0.003 15% 10% Respondents Case 5% FIRR = 0.1764 + 0.041954 * Variable R2= 0.090 0% 0% 2% 4% 6% 8% 10% 12% Conventional Respondents Kondatrieff All FIRR = 0.17628 * 1.274 ^ Variable R2= 0.097 Kondratieff Case FIRR = 0.15218 + 0.036530 * Variable R2= 0.031 Chance of Failure by Variable Value and Forecast Case 25% FIRR = 0.15202 * 1.271 ^ Variable 20% R2= 0.031 15% All Cases 10% FIRR = 0.16869 + -0.00970 * Variable R2= 0.009 5% FIRR = 0.16869 * 0.941 ^ Variable 0% R2= 0.010 0% 2% 4% 6% 8% 10% 12% Conventional Respondents Kondatrieff AllDissFinal Page 236 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Price Inflation for Construction and (Fixed) Operations & Maintenance Costs (% p.a.) Page 1 of 2 Conventional Distribution of Variable: Conventional 100% 25% Minimum: 0% Maximum: 5% 80% 20% Mean: 3% 60% 15% Modal Value 2.5% % < Modal 50% 40% 10% SD (<Modal) 1% SD (>Modal) 1% 20% 5% 0% 0% 0% 1% 2% 3% 4% Cumulative (Left Axis) Probability Density (Right Axis) Respondents Distribution of Variable: Respondents 100% 20% Minimum: 1% 18% Maximum: 6% 80% 16% Mean: 3% 14% 60% 12% Modal Value 0.75% added to Conventional 10% % < Modal 50% 40% 8% SD (<Modal) 0.25% 6% SD (>Modal) 0.25% 20% 4% 2% 0% 0% 0% 1% 2% 3% 4% 5% Cumulative (Left Axis) Probability Density (Right Axis) Kondratieff Distribution of Variable: Kondratieff 100% 18% Minimum: 1% 16% Maximum: 8% 80% 14% Mean: 4% 12% 60% Modal Value 1% added to Respondents 10% % < Modal 50% 40% 8% SD (<Modal) 0.5% 6% 20% 4% SD (>Modal) 0.5% 2% 0% 0% 1% 2% 3% 4% 5% 6% 7% 8% Cumulative (Left Axis) Probability Density (Right Axis) All Cases Distribution of Variable: All Cases 100% 16% Minimum: 0% 14% Maximum: 8% 80% 12% Mean: 3% 60% 10% Modal Value n/a 8% % < Modal n/a 40% 6% SD (<Modal) n/a 4% SD (>Modal) n/a 20% 2% 0% 0% 0% 1% 2% 3% 4% 5% 6% 7% Cumulative (Left Axis) Probability Density (Right Axis)DissFinal Page 237 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Price Inflation for Construction and (Fixed) Operations & Maintenance Costs (% p.a.) Page 2 of 2 Regression Analysis: FIRR on Variable Conventional Case Mean FIRR by Variable Value and Forecast Case 20% FIRR = 0.17777 + -0.25581 * Variable 18% R2= 0.923 16% 14% FIRR = 0.17789 * 0.2225 ^ Variable R2= 0.920 12% 10% 8% Respondents Case 6% 4% FIRR = 0.18585 + -0.22984 * Variable 2% R2= 0.780 0% 0% 1% 2% 3% 4% 5% 6% 7% Conventional Respondents Kondatrieff All FIRR = 0.186 * 0.2737 ^ Variable R2= 0.775 Kondratieff Case FIRR = 0.15982 + -0.06411 * Variable R2= 0.024 Chance of Failure by Variable Value and Forecast Case 20% 18% FIRR = 0.15966 * 0.6604 ^ Variable 16% R2= 0.025 14% 12% All Cases 10% 8% FIRR = 0.17853 + -0.31342 * Variable 6% R2= 0.503 4% 2% FIRR = 0.17888 * 0.1471 ^ Variable 0% R2= 0.488 0% 1% 2% 3% 4% 5% 6% 7% Conventional Respondents Kondatrieff AllDissFinal Page 238 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: General Price Inflation (% p.a.) Page 1 of 2 Conventional Distribution of Variable: Conventional 100% 25% Minimum: 0% Maximum: 5% 80% 20% Mean: 3% 60% 15% Modal Value 2.5% % < Modal 50% 40% 10% SD (<Modal) 1% 20% 5% SD (>Modal) 1% 0% 0% 0% 1% 2% 3% 4% Cumulative (Left Axis) Probability Density (Right Axis) Respondents Distribution of Variable: Respondents 100% 25% Minimum: 1% Maximum: 6% 80% 20% Mean: 3% 60% 15% Modal Value 0.75% added to Conventional % < Modal 50% 40% 10% SD (<Modal) 0.25% 20% 5% SD (>Modal) 0.25% 0% 0% 0% 1% 2% 3% 4% 5% Cumulative (Left Axis) Probability Density (Right Axis) Kondratieff Distribution of Variable: Kondratieff 100% 20% Minimum: 1% 18% Maximum: 8% 80% 16% Mean: 4% 14% 60% 12% Modal Value 1% added to Respondents 10% % < Modal 50% 40% 8% SD (<Modal) 0.5% 6% SD (>Modal) 0.5% 20% 4% 2% 0% 0% 1% 2% 3% 4% 5% 6% 7% 8% Cumulative (Left Axis) Probability Density (Right Axis) All Cases Distribution of Variable: All Cases 100% 18% Minimum: 0% 16% Maximum: 8% 80% 14% Mean: 3% 12% 60% Modal Value n/a 10% % < Modal n/a 40% 8% SD (<Modal) n/a 6% 20% 4% SD (>Modal) n/a 2% 0% 0% 0% 1% 2% 3% 4% 5% 6% 7% Cumulative (Left Axis) Probability Density (Right Axis)DissFinal Page 239 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: General Price Inflation (% p.a.) Page 2 of 2 Regression Analysis: FIRR on Variable Conventional Case Mean FIRR by Variable Value and Forecast Case 25% FIRR = 0.14125 + 1.072175 * Variable R2= 0.973 20% FIRR = 0.1429 * 554 ^ Variable R2= 0.967 15% 10% Respondents Case 5% FIRR = 0.13892 + 1.069124 * Variable R2= 0.931 0% 0% 1% 2% 3% 4% 5% 6% 7% Conventional Respondents Kondatrieff All FIRR = 0.14147 * 468 ^ Variable R2= 0.929 Kondratieff Case FIRR = 0.06937 + 1.802280 * Variable R2= 0.945 Chance of Failure by Variable Value and Forecast Case 60% FIRR = 0.07889 * 642111 ^ Variable 50% R2= 0.835 40% All Cases 30% 20% FIRR = 0.14241 + 0.71620 * Variable R2= 0.914 10% FIRR = 0.14446 * 61 ^ Variable 0% R2= 0.931 0% 2% 4% 6% Conventional Respondents Kondatrieff AllDissFinal Page 240 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Interest Rates for Initial Debt (% p.a.) Page 1 of 2 Conventional Distribution of Variable: Conventional 100% 20% Minimum: 3% 18% Maximum: 7% 80% 16% Mean: 5% 14% 60% 12% Modal Value 5% 10% % < Modal 50% 40% 8% SD (<Modal) 1% 6% SD (>Modal) 1% 20% 4% 2% 0% 0% 3% 4% 5% 6% 7% Cumulative (Left Axis) Probability Density (Right Axis) Respondents Distribution of Variable: Respondents 100% 20% Minimum: 3% 18% Maximum: 9% 80% 16% Mean: 6% 14% 60% 12% Modal Value 1% added to Conventional 10% % < Modal 50% 40% 8% SD (<Modal) 0.5% 6% SD (>Modal) 0.5% 20% 4% 2% 0% 0% 3% 4% 5% 6% 7% 8% 9% Cumulative (Left Axis) Probability Density (Right Axis) Kondratieff Distribution of Variable: Kondratieff 100% 14% Minimum: 3% 12% Maximum: 13% 80% Mean: 8% 10% 60% Modal Value 2% added to Respondents 8% % < Modal 50% 40% 6% SD (<Modal) 1% 4% SD (>Modal) 1% 20% 2% 0% 0% 3% 5% 7% 9% 11% Cumulative (Left Axis) Probability Density (Right Axis) All Cases Distribution of Variable: All Cases 100% 14% Minimum: 3% 12% Maximum: 13% 80% Mean: 6% 10% 60% Modal Value n/a 8% % < Modal n/a 40% 6% SD (<Modal) n/a 4% SD (>Modal) n/a 20% 2% 0% 0% 3% 5% 7% 9% 11% 13% Cumulative (Left Axis) Probability Density (Right Axis)DissFinal Page 241 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Interest Rates for Initial Debt (% p.a.) Page 2 of 2 Regression Analysis: FIRR on Variable Conventional Case Mean FIRR by Variable Value and Forecast Case 30% FIRR = 0.23277 + -1.19631 * Variable R2= 0.970 25% FIRR = 0.24606 * 8.E-04 ^ Variable 20% R2= 0.959 15% 10% Respondents Case 5% FIRR = 0.27426 + -1.68129 * Variable R2= 0.848 0% 3% 5% 7% 9% 11% 13% Conventional Respondents Kondatrieff All FIRR = 0.33297 * 1.E-05 ^ Variable R2= 0.752 Kondratieff Case FIRR = 0.36567 + -2.78234 * Variable R2= 0.936 Chance of Failure by Variable Value and Forecast Case 100% 90% FIRR = 1.16299 * 1.E-13 ^ Variable 80% R2= 0.681 70% 60% All Cases 50% 40% FIRR = 0.29774 + -2.14979 * Variable 30% R2= 0.853 20% 10% FIRR = 0.70788 * 1.74E-11 ^ Variable 0% R2= 0.603 3% 5% 7% 9% 11% 13% Conventional Respondents Kondatrieff AllDissFinal Page 242 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Interest Rates for Extra Debt (% p.a.) Page 1 of 2 Conventional Distribution of Variable: Conventional 100% 30% Minimum: 3% Maximum: 11% 80% 25% Mean: 7% 20% 60% Modal Value 2% added to Initial Interest Rate 15% % < Modal 50% 40% SD (<Modal) 1% 10% SD (>Modal) 1% 20% 5% 0% 0% 3% 5% 7% 9% 11% Cumulative (Left Axis) Probability Density (Right Axis) Respondents Distribution of Variable: Respondents 100% 30% Minimum: 3% Maximum: 13% 80% 25% Mean: 8% 20% 60% Modal Value 2% added to Initial Interest Rate 15% % < Modal 50% 40% SD (<Modal) 1% 10% SD (>Modal) 1% 20% 5% 0% 0% 3% 5% 7% 9% 11% 13% Cumulative (Left Axis) Probability Density (Right Axis) Kondratieff Distribution of Variable: Kondratieff 100% 25% Minimum: 4% Maximum: 16% 80% 20% Mean: 10% 60% 15% Modal Value 2% added to Initial Interest Rate % < Modal 50% 40% 10% SD (<Modal) 1% SD (>Modal) 1% 20% 5% 0% 0% 4% 6% 8% 10% 12% 14% 16% Cumulative (Left Axis) Probability Density (Right Axis) All Cases Distribution of Variable: All Cases 100% 25% Minimum: 3% Maximum: 16% 80% 20% Mean: 8% 60% 15% Modal Value n/a % < Modal n/a 40% 10% SD (<Modal) n/a SD (>Modal) n/a 20% 5% 0% 0% 3% 5% 7% 9% 11% 13% 15% 17% Cumulative (Left Axis) Probability Density (Right Axis)DissFinal Page 243 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Variable: Interest Rates for Extra Debt (% p.a.) Page 2 of 2 Regression Analysis: FIRR on Variable Conventional Case Mean FIRR by Variable Value and Forecast Case 30% FIRR = 0.22287 + -0.65067 * Variable R2= 0.796 25% FIRR = 0.22774 * 2.53E-02 ^ Variable 20% R2= 0.799 15% 10% Respondents Case 5% FIRR = 0.26441 + -1.16853 * Variable R2= 0.631 0% 3% 7% 11% 15% Conventional Respondents Kondatrieff All FIRR = 0.3541 * 6.37E-05 ^ Variable R2= 0.492 Kondratieff Case FIRR = 0.33942 + -1.94655 * Variable R2= 0.945 Chance of Failure by Variable Value and Forecast Case 100% 90% FIRR = 0.5787 * 2.75E-07 ^ Variable 80% R2= 0.767 70% 60% All Cases 50% 40% FIRR = 0.28635 + -1.56141 * Variable 30% R2= 0.882 20% 10% FIRR = 0.39919 * 5.59E-06 ^ Variable 0% R2= 0.686 3% 7% 11% 15% Conventional Respondents Kondatrieff AllDissFinal Page 244 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661)Appendix 20: Risk Simulation Modelling: Comparison of Parameters’ Impacts Parameter Conventional Respondents KondratieffRoad Capacities Range 8,000 8,000 8,000 Impact 0.74% 1.06% 2.05% Capacity per R2 40% 51% 66%Expressway Lane (pcus) Impact*R2 0.30% 0.53% 1.34% +/- Impact +ve +ve +ve Range 4,000 4,000 4,000 Impact 0.05% 0.04% 0.27%Capacity per Local Road R2 0% 0% 3% Lane (pcus) Impact*R2 0.00% 0.00% 0.01% +/- Impact -ve -ve -ve Sum of Impact * R2 0.30% 0.53% 1.35%Construction Cost & Duration Range 74 74 74 Impact 4.58% 4.96% 6.19% Construction Cost R2 96% 96% 94% (Base Year $m) Impact*R2 4.39% 4.76% 5.82% +/- Impact -ve -ve -ve Range 6 6 6 Impact 1.30% 1.66% 3.03% Construction Duration R2 99% 99% 97% (Quarters) Impact*R2 1.29% 1.64% 2.93% +/- Impact -ve -ve -ve Sum of Impact * R2 5.68% 6.40% 8.75%All O&M Costs Range 4% 4% 4% Fixed O&M Costs Impact 1.81% 2.27% 3.73% (as % of Construction R2 17% 33% 50% Costs) Impact*R2 0.31% 0.75% 1.88% +/- Impact -ve -ve -ve Range 4% 4% 4% Impact 0.60% 0.62% 0.64% Variable O&M Costs R2 47% 43% 16% (as % of Revenue) Impact*R2 0.28% 0.27% 0.10% +/- Impact -ve -ve -ve Sum of Impact * R2 0.59% 1.02% 1.98%Value of Time & Its Income Elasticity Range 4 4 4 Impact 0.02% 0.02% 0.03% Small Vehicle VOT R2 0% 0% 0% ($/hr) Impact*R2 0.00% 0.00% 0.00% +/- Impact +ve -ve +ve Range 4 4 4 Impact 0.01% 0.13% 0.22% Large Vehicle VOT R2 0% 3% 3% ($/hr) Impact*R2 0.00% 0.00% 0.01% +/- Impact +ve -ve -ve Range 1 1 1 Income Elasticity of Impact 0.04% 0.01% 0.14% VOT (Small Vehicles) R2 1% 0% 3%DissFinal Page 245 December 2006

Dissertation Richard F. DI BONAHenley Management College (1005661) Parameter Conventional Respondents Kondratieff Impact*R2 0.00% 0.00% 0.00% +/- Impact -ve +ve -ve Range 1 1 1 Impact 0.08% 0.17% 0.34% Income Elasticity of R2 3% 14% 12% VOT (Large Vehicles) Impact*R2 0.00% 0.02% 0.04% +/- Impact -ve -ve +ve Sum of Impact * R2 0.00% 0.03% 0.05%Vehicle Operating Costs Range 0.06 0.06 0.06 Small Vehicle Impact 0.20% 0.18% 0.25% Expressway VOC R2 13% 10% 7% ($/km) Impact*R2 0.03% 0.02% 0.02% +/- Impact +ve +ve +ve Range 0.10 0.10 0.10 Large Vehicle Impact 0.38% 0.32% 0.13% Expressway VOC R2 36% 25% 1% ($/km) Impact*R2 0.14% 0.08% 0.00% +/- Impact +ve +ve +ve Range 1.0 1.0 1.0 VOC Multiplier (Small Impact 0.22% 0.11% 0.22% Vehicles on Local R2 14% 4% 4% Roads) Impact*R2 0.03% 0.00% 0.01% +/- Impact +ve +ve +ve Range 1.0 1.0 1.0 VOC Multiplier (Large Impact 0.43% 0.40% 0.74% Vehicles on Local R2 26% 23% 34% Roads) Impact*R2 0.11% 0.09% 0.25% +/- Impact +ve +ve +ve Sum of Impact * R2 0.30% 0.19% 0.27%Demand (Initial & Income Elasticity) Range 60% 60% 60% Impact 4.09% 4.14% 5.24% Small Vehicle Demand R2 99% 100% 99% Impact*R2 4.07% 4.12% 5.18% +/- Impact +ve +ve +ve Range 60% 60%

What Are The Key Risks Associated With Private Investment In Start Up Toll Road Projects In Developing East Asian Economies

by dibona on Apr 02, 2011

Accessibility

Categories

Upload Details

Usage Rights

1 Embed 1

Statistics

What Are The Key Risks Associated With Private Investment In Start Up Toll Road Projects In Developing East Asian Economies Document Transcript