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 associated

with private investment in start-up

toll road projects in Developing East

Asian Economies?

Richard F. Di Bona

ID No.: 1005661

Dissertation submitted in partial fulfilment of the

requirements of Master of Business Administration

2006

Dissertation Richard F. DI BONA

Henley Management College (1005661)

DissFinal i December 2006

ACKNOWLEDGEMENTS

I am indebted to many for assistance and advice given during the preparation of this

Dissertation. Firstly, to my supervisor, David Parker; also to all the staff of the Henley

Hong Kong office, and to Ken Bull in Henley.

Within transport planning and associated professions, there are simply too many people

to thank individually. I believe I have learnt something from almost everyone I have

worked with over the last 14 years, who afforded me the opportunity to work across a

fascinating mix of countries. Over the last couple of years I have picked the brains of

many colleagues and clients, past and present; and due to frequent commercial

sensitivity, many comments and discussions have been on an anonymous basis. Many

also acted as disseminators of my questionnaire and as “sounding boards” to discuss

ideas and informally corroborate “ball park” figures used in the Monte Carlo risk

simulations.

I should also like to thank Consolidated Consultants in Amman, for their assistance with

printing the Dissertation.

Finally and most importantly, I must thank my wife Mariles for her moral support

throughout the course of my MBA studies and our daughter Vanessa (for helping me

take my mind off of my studies for essential relaxation).



DissFinal ii December 2006

DECLARATION

I confirm that this Dissertation is my own original work. It is submitted in partial

fulfilment of the requirements of Master of Business Administration in the Faculty of

Business Administration of Henley Management College. The work has not been

submitted before for any other degree or examination in any other university.



DissFinal iii December 2006

ABSTRACT

Since the 1980’s there has been a resurgence in private sector involvement in

infrastructure, especially in tolled highways, including in developing economies

(Malaysia, Mexico and Thailand were early adopters). Activity expanded during the

1990’s across much of Latin America and East Asia, the latter region being where the

author has worked extensively. Following a slowdown in the aftermath of the 1997

Asian Financial Crisis, activity has recently picked-up again.

The 1980’s and 1990’s were characterised by generally declining price inflation and

interest rates; whereas now there is evidence of them increasing. Based on the

Kondratieff Wave (long-term business cycle; a.k.a. “K-Wave”), price inflation and

interest rates could be expected to trend upwards significantly over the coming 10-15

years. This Dissertation seeks to determine whether this will significantly change the

nature 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 infrastructure

privatisation trends (globally and in East Asia), illustrating likely future demand.

Financial valuation methods are reviewed, suggesting that whilst FIRR and NPV can be

used, the upfront capital-intensity of toll roads makes annual ratios such as Return on

Capital Employed less relevant to ex ante project evaluation. Generic project risks are



DissFinal iv December 2006

then investigated, showing that most project-risks are “front-loaded” on toll roads. The

Kondratieff Wave is then introduced and its potential applicability discussed, followed

by Kuznets’ work on both infrastructure development cycles and development

economics. The implications of cycles on over-investment are then discussed, with

specific emphasis on the genesis and aftermath of the 1997 Asian Financial Crisis.

Transport modelling theory is presented, followed by discussion of traffic risks and

forecasting issues, resulting variously from uncertainty, institutional risks and

methodological weaknesses, but also demonstrating the primacy of economic growth on

outturn performance. Construction risks are also considered, followed by a brief

discussion of other issues (primarily related to governance and business norms).

Forecasts of toll road demand and construction cost have often been unreliable, with

serial underestimation of cost and overestimation of demand.

Environmental analyses of the East Asian countries studied are then presented, using

PESTLE and stakeholder analysis. Focussing on Thailand (for consistency with the

Literature Review’s analysis of the Asian Financial Crisis), recent economic

performance is assessed, suggesting that recovery is underway. Potential growth in

vehicle ownership and the demand for roadspace is then considered, benchmarking the

studied countries against more developed economies; this shows substantial up-side

potential. The performance of a number of Chinese expressways is then examined. The

opportunities and threats facing the studied countries are discussed, grouping the

countries into three categories corresponding to risk-versus-potential characteristics.

Finally, analysis of gold price and treasury bill rates are used to postulate the current

global economy’s position on the K-Wave, showing that it is likely in the early stages of

an upswing.



DissFinal v December 2006

Next, practitioner perceptions, expectations and experience were tested using a

questionnaire survey (which generated over 160 responses; respondents having a mean

of 20.6 years’ working experience). These showed that legal and political factors were

deemed most significant; but once detailed evaluation (i.e. modelling) commences,

economic factors predominate. As expected, data were perceived as less available and

reliable in developing economies. However, no strong preferences regarding the choice

of modelling method were shown; rather that the approach should be tailored to each

project in turn. Under-forecasting demand seemed rare and over-forecasting it relatively

common, in line with Literature Review findings. There was evidence of transport

modellers being pressured by clients to adjust forecasts. There was also evidence that

many forecasters do not appreciate differences between equity- and debt-side evaluation

requirements. NPV and FIRR are both widely used in evaluation. Based on perceptions

of individual countries’ prospective toll road markets, the country categorisations

proposed in the environmental analysis were broadly supported (with the exception of

Indonesia being seen more bearishly by respondents). Interestingly, respondents seemed

to generally expect many symptoms of the K-Wave upswing, in terms of rising interest

rates and price inflation. However, they were not that convinced of the impacts of these

parameters on forecast performance.

Consequently, Monte Carlo risk simulation modelling was employed to quantitatively

test likely impacts of different risk elements. The model comprised traffic/ revenue

forecasts 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 economic

scenarios, namely:



DissFinal vi December 2006

“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 more

variable than in the Conventional Case. Meanwhile, results from the Kondratieff Case

appeared quite volatile, tending to support theory. Furthermore, interest rates were

shown to become substantially more important to overall risk as they rise; and price

inflation may also increase in importance. Under Kondratieff Case conditions, if

economic growth outstrips the impacts of rising price inflation and interest rates, then

projected returns can be quite significant.

What the above implies is that the nature and extent of project finance risks for private

stakeholders 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 interest

rates increase significantly, these risks can be mitigated. Price inflation subsequent to

the issuing of bonds would also serve to decrease the real value of debt outstanding. But

downstream refinancing is likely to prove increasingly costly (versus experience during

the 1980s and 1990s when cheaper refinancing was often available as a consequence of



DissFinal vii December 2006

declining interest rates). In summary, therefore the hypothesis is broadly supported by

evidence.

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 facilities



DissFinal viii December 2006

TABLE OF CONTENTS

1. Introduction ............................................................................................................. 1

1.1 Terms of Reference/ Personal Development............................................................ 1

1.2 Applicability and Hypothesis ................................................................................... 2

1.3 Geographic Scope .................................................................................................... 3

1.4 Research Approach and Dissertation Structure ...................................................... 5

2. Literature Review.................................................................................................... 6

2.1 Historical Perspective and Basic Taxonomy ........................................................... 6

2.2 Economic Benefits of Transport Infrastructure Development ................................. 7

2.3 East Asian Transport Infrastructure Privatisation Trends ...................................... 8

2.4 Financial Valuation ................................................................................................. 9

2.5 Project Risk Analysis ............................................................................................. 14

2.6 The Kondratieff Wave ............................................................................................ 16

2.7 Kuznets Cycle, Kuznets Curve and S-Curves ........................................................ 18

2.8 Infrastructure Development, Cycles and Crises .................................................... 19

2.9 Transport Modelling .............................................................................................. 23

2.10 Traffic Risks and Forecasting Issues ..................................................................... 25

2.11 Construction, Operations and Maintenance.......................................................... 33

2.12 Other Considerations ............................................................................................ 35

2.13 Summary of Key Issues .......................................................................................... 37

3. Environmental Analysis ....................................................................................... 39

3.1 Introduction and PESTLE Analysis ....................................................................... 39

3.2 Political, Legal and Stakeholder Issues................................................................. 40

3.3 Economic Recovery ............................................................................................... 42

3.4 Vehicle Ownership ................................................................................................. 46

3.5 Traffic Performance of Existing Toll Roads .......................................................... 48

3.6 Opportunities and Threats ..................................................................................... 51

3.7 Postulated Position on K-Wave ............................................................................. 53

4. Questionnaire Survey ........................................................................................... 55

4.1 Purpose .................................................................................................................. 55

4.2 Design Concept and Sample Selection .................................................................. 56

4.3 Questionnaire Design and Survey Execution ........................................................ 57

4.4 The Survey Sample ................................................................................................. 58

4.5 Tollway Appraisal .................................................................................................. 62

4.6 Transport Modelling Issues ................................................................................... 65

4.7 Forecast Performance and Evaluation Criteria .................................................... 67

4.8 Countries’ Outlooks ............................................................................................... 70

4.9 Economic Outlook ................................................................................................. 73

4.10 Other Comments .................................................................................................... 75

4.11 Key Conclusions from the Questionnaire Survey .................................................. 75

5. Risk Simulation Modelling ................................................................................... 77

5.1 Introduction ........................................................................................................... 77

5.2 The Case Study and Its Parameterisation ............................................................. 78

5.3 Methodology .......................................................................................................... 82

5.4 Comparison of Cases under “Base Run” .............................................................. 84

5.5 Comparison of Simulation Results between Cases ................................................ 85



DissFinal ix December 2006

5.6 Analysis of Individual Risks ................................................................................... 88

5.7 Discussion of Results ............................................................................................. 91

6. Discussion and Conclusions .................................................................................. 92

6.1 Introduction ........................................................................................................... 92

6.2 Evaluation Criteria and Implications of the Time-Nature of Risk ........................ 93

6.3 Macro-Level Risks and Opportunities ................................................................... 94

6.4 Market Risks .......................................................................................................... 96

6.5 Forecasting Risks................................................................................................... 98

6.6 Is the Market Anticipating a Change in the Rules-of-the-Game? ....................... 100

6.7 What Lessons for Practitioners? ......................................................................... 101

6.8 Conclusions: Evaluation of Hypothesis ............................................................... 103

References: Literature ................................................................................................ 105

References: Internet Resources ................................................................................. 117

Appendices ................................................................................................................... 119

LIST OF TABLES

Table 2.1: Investment and Maintenance Needs in East Asia, 2006-2010 ......................... 8

Table 2.2: Bain and Polakovic Forecast Performance Statistics ..................................... 26

Table 2.3: Bain and Wilkins Ramp-Up Revenue-Adjustment Profiles .......................... 30

Table 2.4: Estimated Expressway Construction Costs .................................................... 34

Table 2.5: Operations and Maintenance Costs ................................................................ 34

Table 2.6: Summary of Key Risks and Issues ................................................................ 38

Table 3.1: Highlights of PESTLE Analysis .................................................................... 39

Table 3.2: Vehicle, Trip and Expressway Patronage Income Elasticities....................... 48

Table 4.1: Aggregated Respondent Experience Categories ............................................ 58

Table 4.2: Respondents’ Mean Years’ Experience in Various Fields ............................ 60

Table 4.3: Respondents with Experience in Study Area ................................................. 61

Table 4.4: Rankings of Macro-Level Risks by Respondent Category ............................ 63

Table 4.5: Rankings of Project-Level Risks by Respondent Category ........................... 64

Table 5.1: Basic Link Characteristics of Case Study Network ....................................... 79

Table 5.2: Assumed Trip Distribution (% by O-D Pair) ................................................. 79

Table 5.3: Comparison of “Base” Runs between Cases ................................................. 85

Table 5.4: Summary Results from Simulation Runs ....................................................... 86

Table 5.5: Rankings of Risk Categories’ Importance by Case ....................................... 89

LIST OF FIGURES

Figure 1.A: Map of East Asia ........................................................................................... 4

Figure 1.B: Research Approach ........................................................................................ 5

Figure 2.A: Standard & Poor’s Risk Pyramid ................................................................. 14

Figure 2.B: Transport Concession Risks ......................................................................... 15



DissFinal x December 2006

Figure 2.C: Kuznets Curve and S-Curve......................................................................... 18

Figure 2.D: Indexed Thai Real GDP and M2, 1991-1999 .............................................. 19

Figure 2.E: Baht-US$ Exchange Rate 1994-2001 .......................................................... 20

Figure 2.F: Dollarised Thai GFCF 1994-2001 ................................................................ 21

Figure 2.G: Demand, Revenue and Price Elasticity of Demand ..................................... 27

Figure 3.A: Typical Concession Stakeholder Map ......................................................... 40

Figure 3.B: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) .................... 43

Figure 3.C: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) in US$ ....... 43

Figure 3.D: Thai GFCF, GDP and M2 in Baht, Indexed to 1995 ................................... 44

Figure 3.E: Thai GFCF, GDP and M2 in US$, Indexed to 1995 .................................... 44

Figure 3.F: Thai GFCF, GDP and M2 in US$, Indexed to 2000 .................................... 44

Figure 3.G: Currency Performance since 1994 ............................................................... 45

Figure 3.H: Currency Performance since 2001 ............................................................... 45

Figure 3.I: Relationship between Wealth and Roads Per Capita .................................... 47

Figure 3.J: Relationship between Wealth and Road Density .......................................... 47

Figure 3.K: Traffic Growth on Shanghai-Nanjing Expressway ...................................... 50

Figure 3.L: Traffic Growth on Shanghai-Hangzhou-Ningbo Expressway ..................... 50

Figure 3.M: Interest Rates, Nominal Gold Price and Kondratieff Wave ........................ 54

Figure 4.A: Respondents by Experience Type ................................................................ 59

Figure 4.B: Respondents by Years of Experience .......................................................... 59

Figure 4.C: Respondents’ Global Experience ................................................................. 60

Figure 4.D: Respondents with Experience in East Asia ................................................. 61

Figure 4.E: Attitudes to Macro-Level Risks ................................................................... 63

Figure 4.F: Attitudes to Project-Level Risks .................................................................. 64

Figure 4.G: Data Availability and Reliability ................................................................. 65

Figure 4.H: Attitudes to Transport Model Types ............................................................ 66

Figure 4.I: Perceptions of Forecast Performance ............................................................ 68

Figure 4.J: Which Forecast Outputs are Considered? ..................................................... 69

Figure 4.K: How Often Are Which Criteria Considered?............................................... 69

Figure 4.L: Perceived Tollway Market Opportunities by Country ................................. 71

Figure 4.M: Impact of Experience on Country Perceptions ........................................... 71

Figure 4.N: Country Perceptions by Respondent Category ............................................ 72

Figure 4.O: Economic Expectations ............................................................................... 73

Figure 4.P: Economic Expectations by Respondent Group ............................................ 74

Figure 5.A: Case Study Notional Network ..................................................................... 79

Figure 5.B: Volume/Capacity-to-Speed Relationships ................................................... 83

Figure 5.C: Cumulative Probability Distribution of FIRR (excluding FIRR<0%) ......... 87

Figure 5.D: Cumulative Probability Distribution of Payback Period (years) ................. 87

Figure 5.E: Cumulative Probability Distribution of NPV at 16% ($m) .......................... 87



DissFinal xi December 2006

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 Page 1 December 2006

1. Introduction

1.1 Terms of Reference/ Personal Development

For 14 years, I have worked in transport planning, economics and demand forecasting

across 20 countries/territories, mostly on transport infrastructure scheme appraisal, often

for privatisation, and usually in East Asia (covering rich, “tiger” and poor economies).

One reason for pursuing the MBA, the Business Finance Elective and this Dissertation

topic 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, most

notably business cycles, leading me to the Kondratieff Wave. This postulates a cycle of

48-60 years duration; comprising inter alia phases of increasing interest rates and

commodity prices followed by decreases in same. Given recent increases in Federal

Reserve interest rates and commodity prices, Kondratieff theorists posit a

commencement of an “upswing” phase, qualitatively different from the “downswing” of

the 1980’s and 1990’s; potentially changing the relative importance of different aspects

of investment risk. Given most transport privatisation and associated literature and

experience are based on “downswing” conditions, reviewing these based on “upswing”

conditions could be timely.

Though focussed on profit maximisation (through risk management), better

understanding of changing risks should result in more efficient use of capital by private,

public and aid agency sectors alike.




1.2 Applicability and Hypothesis

The Dissertation focuses on East Asia which is again emerging as a “powerhouse” of

economic growth, with commensurately strong demand for transport anticipated. The

World Bank (2003a) notes resurgent private sector involvement in infrastructure

provision since the 1980’s, with substantial tollway activity in East Asia (US$34 billion

during 1990-2001 into 149 projects). Although activity slowed following the Asian

Financial Crisis (AFC), by 2001 it returned to 1995 levels. Yepes (2004) expects

highways to be the second biggest infrastructure investment sector in East Asia during

2006-2010. In addition to providing profit opportunities, there is evidence that projects

could 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 of

declining 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


inflation and interest rates”




1.3 Geographic Scope

East Asia is a large, diverse region, including some of the World’s richest and poorest

societies, with differing political and legal systems and levels of economic openness.

This Dissertation is concerned with its developing economies, which are likely to

benefit as: manufacturing hubs for the world; markets in their own right; and/or, natural

resource providers. It is in such economies that transport infrastructure demand growth

may be most marked.

Whilst the literature review is deliberately broad, and the questionnaire survey relatively

so, 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 few

others for benchmarking purposes. Appendix 2 gives headline transport statistics.

Whilst countries such as China are anticipated to continue requiring and attracting

investment 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 Taiwan




Source of base map: Google EarthTM 2

Figure 1.A: Map of East Asia

2 Study Area countries in red on yellow text. Other countries/ territories in black on grey text.

MALAYSIA

INDONESIA

Brunei

Singapore

Hong Kong

THAILAND

MYANMAR

CHINA

LAOS

VIETNAM

CAMBODIA

Japan S.Korea

N.Korea

Mongolia

Timor-Leste

PHILIPPINES




1.4 Research Approach and Dissertation Structure

The outline research approach is presented in Figure 1.B; also giving relevant Chapter

numbers.

Figure 1.B: Research Approach

1. Introduction and Hypothesis Including definition of geographic scope

2. Literature Review Including a priori evaluation

and analysis thereof

3. Environmental Analysis Including country economics and

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.




2. Literature Review

2.1 Historical Perspective and Basic Taxonomy

Private transport infrastructure financing and operation dates back to at least the 19th

Century, including railways (e.g. UK and USA) and the Suez Canal. IPFA (2006) notes

following the First World War government resumed most infrastructure provision,

financing projects from public debt; subsequently developing countries followed this

practice, borrowing from development agencies (e.g. WB, ADB).

By the 1980’s, government debt constrained public financing of schemes, especially

given high interest rates; yet economic and demographic forces continued to demand

infrastructure. Thus was private involvement reborn.

There is much overlapping taxonomy regarding types of project privatisation. Guislain

and Kerf (1995) note a continuum of options for private sector involvement, from

supply and service contracts through leasing (wherein management of a built project is

let to the private sector in exchange for a revenue-share and/or up-front payment) to

Build-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 build

of the project, often being wholly responsible for financing.




2.2 Economic Benefits of Transport Infrastructure Development

Whilst SACTRA (1994) questioned the benefits of additional trunk roads in developed

economies with built-out highway networks, in developing economies new highways

often facilitate economic development. Christensen and Mertner (2004) showed

Cambodia’s factory gate price advantage over China for garments negated by transport

costs: China FOB prices are lower than Cambodia’s. Di Bona (2005) noted

rehabilitation 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 road

network3 (Corbett et al, 2006, p.A2-99). The benefits of transport infrastructure in

developing 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) recommends

governments only approve projects yielding a given socio-economic return, before

determining likely profitability.

Klein et al (1996) note privatisation appears to increase implementation costs, partially

due to private sector participation bringing true costs to light. It also increases funds

available for development.

3 83.6% estimated statistically, with traffic growth attributable directly to economic growth excluded.




2.3 East Asian Transport Infrastructure Privatisation Trends

Developing countries’ transport infrastructure privatisation began in earnest in the

1980’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 billion

private investment (41% of global total) into 229 projects (Ibid., p.135), particularly toll

roads: US$34 billion into 149 projects (Ibid., pp25-26 & p.143). By 2001, China had

attracted more private investment than any other country (US$23.6 billion), and

Malaysia the most per capita (US$582) (Ibid., p.136). Whilst activity slowed after the

1997 Asian Financial Crisis (AFC), by 2001 it returned to 1995 levels (Ibid., p.2). Table

2.1 illustrates substantial anticipated future expenditure (from Yepes, 2004); highways

are 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 Total

Electricity 63,446 25,744 89,190 2.4 1.0 3.4

Telecoms 13,800 10,371 24,171 0.5 0.4 0.9

Highways 23,175 10,926 34,102 0.9 0.4 1.3

Railways 1,170 1,598 2,768 0.0 0.1 0.1

Water 2,571 5,228 7,799 0.1 0.2 0.3

Sanitation 2,887 4,131 7,017 0.1 0.2 0.3

Total 107,049 57,998 165,047 4.0 2.3 6.3

Buchanan (1999) notes the Malaysian boom in BOT highways followed the perceived

success of the North-South Highway (PLUS) concession in 1988, through which private

finance overcame public sector constraints and took-on risk, bringing private sector

skills and incentives to infrastructure operation. However, he believes PLUS appeared

profitable only because Government handed over 225km of existing expressway with

tolling rights.




In China several Provincial Governments established corporations for expressway

development. Soon after completing a flagship expressway, the company would be

listed with revenues raised used to acquire or develop additional highways4. This

relatively 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 listed

firms. Operators occasionally issue bonds, although this practice is more widespread in

the Americas.

2.4 Financial Valuation

2.4.1 NPV and IRR

The decision to pursue a project and on what terms are primarily questions of project

valuation and risk. Higson (1995, pp.60-61) notes project value may be defined via Net

Present Value (NPV) or Internal Rate of Return (IRR). NPV values future cashflows as:

n

tt

t

r

CNPV

0 1 (1)

Where: 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 period

IRR expresses scheme value in terms of a percentage return on capital invested, being

the discount rate at which NPV is exactly nought:

4 See Appendix 3 for examples.




010

n

tt

t

R

CNPV (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 if

return is greater than investors’ required return. Highways require substantial up-front

investment 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, as

well as by initial investment size.

Investors treat own target FIRR as strictly confidential; so no directly citeable values are

available. However, from the Author’s experience corroborated by off-the-record

conversations with fellow practitioners, a target FIRR of 16% p.a. is the usual threshold

required. This includes a modest risk premium (see 2.4.2); for particularly high risk

projects, or when capital is more expensive, FIRR would increase accordingly.

2.4.2 CAPM and WACC

The above assumes certainty regarding all project aspects, including: demand, price

inflation for inputs, selling price, construction cost and time, operating period, implicit

assumption of no sovereignty risks etc; yet uncertainty bedevils these parameters. The

Capital Asset Pricing Model (CAPM; ibid., p.123) suggests the return on a risky project

rj is:

)( imjij rrrr (3)

where: ri is the return on riskless borrowing/ lending




rm is the return on the money market as a whole

The risk premium for j is a proportion βj of overall market risk-premium, as follows:

2

m

jm

j

(4)

Required return can also be calculated as Weighted Average Cost of Capital (WACC;

ibid, p.279):

MVMV

dMVeMV

DE

KDKEWACC

(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)

thvidendGrowExpectedDiiceShare

DividendKe

Pr (6)

TaxRateeofFaceValuiceDebenture

teInterestRaKd

1

)(%Pr (7)

From (3) and (7) the Fisher-Hirshleifer theorem can be restated as pursue projects if:

MVMV

dMVeMVimji

DE

KDKErrr

)( (8)

2.4.3 Treatment of Price Inflation

Often (especially in transport scheme appraisal) a constant inflation rate is assumed with

calculations based in real prices (akin to zero price inflation throughout). Such price

neutrality simplifies calculations; however, it does preclude analysis of price-risks

associated with individual project inputs and outputs.




2.4.4 Problems with CAPM and WACC

βj might theoretically be known for existing highways, but is unknown for new projects.

There may be insufficient local data to determine 2

m . β is intended for fully diversified

investors, rather than appraising a scheme in isolation. Higson (ibid., p.136) notes

CAPM 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’ size

makes them relatively illiquid. There may be no risk-free asset: money is only risk-free

if possible depreciation/ price inflation is ignored.

Lumby (1983) notes unless a project is financed with the same capital structure as the

firm itself (unlikely), WACC changes once the project is undertaken. Furthermore,

WACC assumes constant cashflows and that project systematic risk to equal that of the

company’s existing projects; both highly unlikely.




Ormerod (2005, p.173) notes whilst CAPM requires a normal probability distribution in

derivative markets, they exhibit power-law behaviour; this discrepancy caused the 1998

collapse of Long Term Capital Management. Whilst CAPM supports currency

diversification (e.g. in borrowing), Beaverstock and Doel (2001) note such borrowing

collapsed Steady Safe (an Indonesian taxi and bus firm) and in turn Peregrine

Investment Bank.

2.4.5 Financial Ratios

A number of financial ratios may be used to evaluate likely project performance and

risk. Given the capital-intensity of highway construction, coupled with typically long

lead-times for demand build-up (see 2.10.4), financial ratios may not always be as

relevant to ex ante project valuation.

Return on Capital Employed5 is likely to be poor for early years of a concession (unless

the project is highly geared). Likewise, Gross Profit Margin, Profit On Sales, Expenses

as Percent of Turnover, Sales to Capital Employed, Sales to Fixed Assets and Asset

Turnover all typically take many years to build-up to levels normally deemed acceptable

in many other businesses5.

Some of the above ratios might be improved by heavy borrowing, but such borrowing

and resultant debt-servicing increases the importance of Working Capital Requirements,

the Current Ratio and the Debt Service Coverage Ratio5. Standard & Poor’s relies on

Interest Cover (debt-service coverage) as the primary quantitative measure of a project’s

financial strength (Rigby and Penrose, 2001, p.28).

5 See Appendix 4 for definitions of these financial ratios.




2.5 Project Risk Analysis

Rigby 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.

Figure 2.A: Standard & Poor’s Risk Pyramid

Project-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 jurisdiction

Force Majeure Risk

Credit

Enhancement

Institutional Risk

Sovereign Risk

Project-Level Risks

Force Majeure Risk

Credit

Enhancement

Institutional Risk

Sovereign Risk

Project-Level Risks

Force Majeure Risk

Credit

Enhancement

Institutional Risk

Sovereign Risk

Project-Level Risks




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 risks

George et al (2004) note the uncertainty inherent in start-up tollways requires flexible

financing approaches. Willumsen and Russell (1998) illustrate project-level risks as

shown in Figure 2.B. Predominating traffic/ revenue risks are discussed in Section 2.10.

Figure 2.B: Transport Concession Risks

Sovereign 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 factors

O&M

Traffic &

Revenue

Ramp Up

Construction

Costs

Construction

Delay

Change Orders

-2 -1 0 1 2 3 4 5 10

Han

dove

r

Year

Ris

k (

no

min

al)




include business and legal institutions, which are often weak/ nascent in developing

countries, with concepts of property rights and commercial law not fully developed,

potentially leaving creditors/ investors exposed. La Porta et al (1997) found investor

rights in developing countries though limited, are generally better under common law

than civil law (especially French civil law, which often has weak enforcement).

Force Majeure risks includes “Acts of God” (floods, earthquakes, etc) as well as civil

disturbances, strikes, changes of law. Rigby and Penrose (2001) note toll roads are

typically less affected/ can return to normal service more quickly.

Credit Enhancement refers to insuring/ re-insuring specific risks. However, litigation

intrinsic in such claims can delay payment by years, so mitigation may be limited.

2.6 The Kondratieff Wave

Orthodox economics assumes given policies produce similar results at all times;

Ormerod (1999, pp.96-102) notes experience contradicts this, due to periodic exogenous

shocks. Others postulate cycles responding to exogenous shocks. But to some cycle

adherents, 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:




Kindleberger (1996, p.13) calls it “possibly… dubious and elusive.” There is also

debate on periodicity. Whilst Schumpeter believed one K-Wave contained three Juglar

Cycles, each comprising in turn three Kitchin Cycles, Faber (2002, p.110) notes

Kondratieff never postulated precise periodicity.

Kondratieff’s empirical work identified a number of patterns within each cycle. Further

analysis 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 to

Upswing and Downswing phases and largely ignored in the context of this Dissertation.

Appendix 5 presents K-Waves since 1787. Maddison (1995) estimated real global GDP

per capita rose 2.90% p.a. from the 1950s-1970s (K-Wave upswing); but declined to

1.11% p.a. until the 1990’s (K-Wave downswing).




2.7 Kuznets Cycle, Kuznets Curve and S-Curves

Kuznets (1930) identified a 15-25 year building construction cycle, concurring with

Schumpeter that innovation drives growth endogenously to the economic cycle. He also

postulated the Kuznets Curve (1955), plotting economic development against income

inequality: 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 all

wealth).

This implies few might afford cars or tolls in the early phases of growth, but as

economies develop, tolls become substantially more affordable. Coupled with demand

saturation, 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.

Normal Density/ Kuznets Curve

Cumulative Normal/ S-Curve

Figure 2.C: Kuznets Curve and S-Curve




2.8 Infrastructure Development, Cycles and Crises

Infrastructure may facilitate Upswings, but its short-term impact may trigger

Downswings, fostering “Creative Destruction” (Schumpeter, 1950): purging old

methods/ technologies for improved methods/ infrastructure to drive Upswings.

Lawrence (1999) argues major skyscraper completions are cyclical, preceding

recessions. But do build-out peaks precipitate recessions, or are they “peaks” due to

subsequent demand failure, uncorrelated with preceding build-out (as espoused by

Krugman, 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 with

M2 broke. Continued M2 growth refutes Krugman’s attribution of the AFC to demand

failure, which ignored structural causes.

Figure 2.D: Indexed Thai Real GDP and M2, 1991-1999

6 Much of these Thai analyses originally presented in Di Bona, R.F. (2002) Surviving Bahtulism

7 Raw data from APEC (www.apec.org); analysis my own.

100

120

140

160

180

200

1991 1992 1993 1994 1995 1996 1997 1998 1999

Re

al G

DP

(1

99

1=

10

0)

100

140

180

220

260

300

M2

(1

99

1=

10

0)

Real GDP M2




Before the AFC, Thailand enjoyed a virtuous economic development cycle: increased

wealth boosted investment returns, attracting further investment. Keynesian multiplier-

accelerator effects boosted growth, encouraging further development. Adaptive

expectations of investment returns fuelled excessive capital works and other

investments. Bangkok planned several new residential and business hubs, which could

not all be viable simultaneously: eventually supply outpaced demand.

The Baht’s July 1997 flotation coincided with doubts regarding the sustainability of

Thailand’s growth. Its depreciation (Figure 2.E8) ballooned offshore-financed corporate

debt. Ensuing capital flight intensified the crisis. Long infrastructure lead-times meant

there was still supply-in-waiting; many projects were stalled or abandoned. Figure 2.F9

shows GFCF collapsing with no noticeable rebound by 2001.

Figure 2.E: Baht-US$ Exchange Rate 1994-2001

8 Source data: www.fx.sauder.ubc.ca

9 Source data: www.nesdb.go.th and www.fx.sauder.ubc.ca

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

Jan-1

994

Jul-1

994

Jan-1

995

Jul-1

995

Jan-1

996

Jul-1

996

Jan-1

997

Jul-1

997

Jan-1

998

Jul-1

998

Jan-1

999

Jul-1

999

Jan-2

000

Jul-2

000

Jan-2

001

Jul-2

001

US

D p

er

TH

B

http://www.fx.sauder.ubc.ca/

http://www.nesdb.go.th/





Hayek (1933) argues artificially low interest rates breed over-investment, precipitating

crises with debt- and investment-overhangs delaying recovery. Faber (2002, pp.192-

193) argues global liquidity injections following the 1995 Mexican crisis fuelled further

Asian speculative growth, delaying but ultimately amplifying and prolonging the AFC.

Figure 2.F: Dollarised Thai GFCF 1994-2001

Faber (2002, p.69) notes cycles are “particularly violent in the case of emerging

economies, emerging industries and emerging companies, which grow and evolve

rapidly and are, therefore, capital-hungry.” Transport infrastructure construction is

especially capital-intensive.

0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

1994 1995 1996 1997 1998 1999 2000 2001 2002

millio

n U

SD

(1

988

pri

ces)

Gross Fixed Capital Formation Private Construction Government Construction

Land Development Construction And Land Development




Although infrastructure and utilities are often seen as defensive investments, Forsgren et

al (1999) argue toll road performance is cyclical, noting with reference to China (not

generally 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 counterparties

In Indonesia, the rapid devaluation of the Rupiah in 1997, compounded by rapidly

increasing 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 development

in the 19th

Century, Faber (2002, pp.55-63) notes they exhibited cyclical booms and

crises. Moreover, historically overseas investors are often latecomers, repeatedly buying

peaks to sell-out in the immediate aftermath of crisis.




2.9 Transport Modelling

Corbett and Di Bona (2006) note transport models provide inter alia: assessment of

demand-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 in

Appendix 7; but such models are data hungry so simplifications are common. Their

applicability to tollways has been questioned (Willumsen and Russell, 1998).

Usually the modelled area is divided into spatial zones. Traditionally, traffic to/ from

each zone is estimated based on land-use and corresponding trip generation rates.

However, given sparseness of robust land use data in developing countries, econometric

models of traffic levels are often used. Whilst Khan and Willumsen (1986) fitted S-

curve models to vehicle ownership and usage, often historical traffic counts are

regressed on corresponding income data to estimate income elasticities of traffic

demand, defined as:

2

2

10

01

10

01

yy

yy

tt

tt

Y

TT

y (9)

Where: to,t1 are traffic levels in periods 0 and 1

y1,y0 are income (GDP) levels in periods 0 and 1

As elasticities might not hold over time forecast values are adjusted, based either on S-

curves or a conservative assumption of gradually declining elasticities, taking implicit

account of longer-term demand saturation or improved logistic efficiency (decreased

lorry empty-running). Though these ignore vehicle ownership/ usage costs, Pindyck and

Rubinfeld (1981, pp.396-398) note Hymans’s (1970) model of USA vehicle ownership




shows such factors have short-term impacts, income-ownership relationships

predominating thereafter.

In developing countries tollway appraisals, driver interview surveys scaled using traffic

counts are often used to obtain trip patterns. Effects of other modes (e.g. rail) are

commonly 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 and

perpendicular to the toll-road significantly affecting patronage), network assignment

models 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). An

iterative 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-year

intervals thereafter, with intermediate years interpolated. Such models are calibrated by

adjusting network coding and often using maximum entropy matrix estimation (see Van

Zuylen 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). Traffic

is allocated between routes based on a logit function; (10) shows an absolute logit curve

for forecasting a new road’s traffic. For existing toll-roads incremental logit models

may be preferred, shown in (11). Commonly κ and λ would be estimated based on

previous studies (ideally existing toll-roads). Richardson (2004) notes a general bias

against using toll roads (κ<0). Forecasts may be prepared for selected years

(intermediate years interpolated) or for all years. Whilst congestion levels do not

feedback, increasing incomes make tolls more affordable.




Xtij

Ltij GCGC

X

tij

eP

,,1

1,

(10)

Where: X

tijP , is the share of trips i→j in period t using the expressway, 1,, L

tij

X

tij PP

r

tijGC , 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

Xtij

Ltij

Xtij

Ltij

GCGC

GCGCX

tijX

tij

X

tijX

tij

X

tij

e

eOb

P

PObIP

0,0,

,,

1

1

1

1

0,

0,

,

0,,

(11)

Where: X

tijOb 0, is the base year observed expressway market share for trips ij

X

tijP , is forecast expressway share in year t (absolute logit); t=0 is base year

2.10 Traffic Risks and Forecasting Issues

Bain and Wilkins (2002) analyse toll-traffic uncertainty and traffic forecast error,

showing strong inter-correlation. Average initial year traffic was 70% of forecast

overall, 82% in lender-commissioned projections and 66% when commissioned by

others, suggesting commissioning party influence on forecasts: debt-financiers

relatively more concerned with down-side risk than equity-holders. Their Traffic Risk

Index (shown in Appendix 8) compares low and high risk factors for toll roads and

traffic forecasts in general.

Whilst initial year errors might be due to ramp-up (see 2.10.4), which Streeter and

McManus (1999) reckon can last 3-5 years, Bain and Polakovic (2005) note optimism

bias is “constant through Years 2 to 5” as shown in Table 2.2, signalling other errors




(discussed below). They also note drastic differences in forecasts by different parties for

the 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 0.25

2.10.1 Toll Sensitivity and the Value of Time

Excepting “shadow tolling” (operator reimbursed based on patronage instead of user-

tolling), willingness-to-pay tolls is critical. Typically choice is between a slow, cheap

road and a fast toll-road; time and money equivalenced using the behavioural Value of

Time (VOT) to give “generalised cost.” Whilst higher tolls are usually preferred (see

2.10.4) sometimes they are too high (Wong and Moy, 2004). The price elasticity of

tollway demand is:

2

2

10

01

10

01

pp

pp

qq

qq

P

Qp

D (14)

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 respectively




Figure 2.G shows the relationship between demand, revenue and η. When tolls are

beneath the revenue maximising level (i.e. p<Prm) 10 p

D , toll increases boost

revenue; when p>Prm 1p

D (toll increases decrease revenue). 1p

D when p=Prm.

Willumsen and Russell (1998) note in developing countries Stated Preference surveys to

estimate p

D and VOT are scarce and of uncertain quality. Reference is often made to

previous studies, factored for income levels. But the income elasticity of VOT, y

VOT is

complicated: as income increases, VOT rises (“income effect”), as does expenditure on

other products/ services (“substitution effect”) and possibly savings too (“savings

effect”), implying 1y

VOT . In developed economies, Wardman (1998) suggests

49.0y

VOT ; Gunn and Sheldon (2001) advocate 7.035.0 y

VOT . Cross-sectional

analysis between developing countries suggests 1y

VOT yet time-series analysis within

a country 1y

VOT to growth VOT thereafter10

.

Figure 2.G: Demand, Revenue and Price Elasticity of Demand

10 Confidential source used in absence of public source.

Price→

Demand

Total

Revenue

-η

η= −1

Revenue

Maximisation

Prm




Goods vehicles are of particular concern. Bain and Wilkins (2002) note in developing

countries long-distance tolls often exceed drivers’ wages, giving incentive to use

untolled routes (pocketing bosses’ toll money). Some studies (e.g. ADB, 2003) have

failed to establish any VOT for goods vehicles.

2.10.2 Competing Routes and Link Roads

Contractual guarantees theoretically limit competing routes’ development, presupposing

the contracting branch of government is willing and able to enforce such guarantees

across multiple government layers.

Jiangsu Expressway circumvented this risk by acquiring rights to highways parallel to

their flagship Shanghai-Nanjing Expressway and so manage (and toll) traffic on both

routes. However, when GZI Transport listed in 1997, it was assumed that the ferry

parallel to the (then) soon-to-open Humen Bridge would cease operation. But being

operated by a different local government, operation continued with fares undercutting

bridge tolls, attracting substantial goods vehicle volumes from the Humen Bridge.

Even when concessionaires gets first refusal at planned parallel routes, overinvestment

may result in excess infrastructure relative to traffic levels. Buchanan (1999) notes in

Malaysia those identifying schemes can often proceed (subject to financing) without

due diligence of impacts on existing BOT’s.

Though more important for urban projects, provision of adequate link roads is also

important. Congested approaches/ exits can result in “hurry up and wait” (Bain and

Wilkins, 2002), reducing tollways’ attractiveness.




2.10.3 Toll Increases and Revenue Guarantees

Contracts typically allow periodic price-indexed toll increases, or at a percentage of

price inflation. However, Forsgren et al (1999) note toll increase approval processes are

often opaque and beset with delay. Bain and Wilkins (2002) note tariff escalation is

often politicised, especially where there is little previous “tolling culture.” Sometimes

social unrest follows tolls’ imposition (Orosz, 1998) or toll increases, especially during

economic downturns (Dizon, 2002).

Some contracts give revenue guarantees to operators, underwritten by government.

However, China’s 2002 State Council directive scrapped such revenue guarantees

overriding contract provisions, leading to New World Development divesting from 13

toll roads and bridges (Chan, 2003).

Whilst non-toll revenues may be generated (e.g. service stations, advertising), Streeter

et al (2004) note their contribution is usually dwarfed by toll revenues.

2.10.4 Ramp-Up

Bain and Wilkins (2002) define ramp-up as information lag for users unfamiliar with a

new highway and general reluctance to pay tolls (see Richardson, 2004 for experimental

evidence). Streeter and McManus (1999) reckon on 3-5 years’ ramp-up and note this is

often underestimated in traffic forecasts.

Bain and Wilkins (2002) note ramp-up experience tends to cluster to extremes: either of

limited 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 (see

Appendix 8). They derived revenue-adjustment factors as per Table 2.3 for use in

financial stress-tests.




Table 2.3: Bain and Wilkins Ramp-Up Revenue-Adjustment Profiles

Forecasts

commissioned by Lenders Others

Traffic Risk Low Average High Low Average High

Year 1 revenue

adjustment -10% -20% -30% -20% -35% -55%

Ramp-up duration

(years) 2 5 8 2 5 8

Eventual catch-up 100% 95% 90% 100% 90% 80%

2.10.5 Operating Costs

In addition to tolls, many models also apply distance-based monetary Vehicle Operating

Costs (VOC) reflecting fuel, maintenance, depreciation, etc. Whilst economic values for

these parameters are derivable, accurate behavioural values are often elusive. In practice

they 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, these

are typically applied as fixed values with respect to distance and road-type, rather than

feeding-back modelled forecast speeds. Where there are larger VOC savings from an

expressway ceteris paribus there is more scope for higher tolls. However, there is an

issue as to who pays these costs (driver or employer).

2.10.6 Toll Leakage

Some vehicles use a facility without paying, either legitimately (e.g. certain government

or military vehicles) or illegitimately. There may be theft by toll-collectors and fraud by

administrators. Forsgren et al (1999) note toll leakage can be as high as 20% of

revenues. Sometimes computerised toll collection and auditing can restrain losses, but

on lower volume routes the cost of such measures might outweigh savings.




2.10.7 Induced Traffic

When a new highway significantly reduces transport costs or relieves congestion, it may

result 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, it

may also over time enable expanded development, generating further traffic demand.

However, Willumsen and Russell (1998) note the difficulty of reliably forecasting such

effects; Bain and Polakovic (2005) report the prevalence of significant errors in induced

traffic forecasts.

2.10.8 Annualisation

Bain and Wilkins’ (2002) Traffic Risk Index shows projects with seasonal flow patterns

tend to be riskier. For inter-urban highways a “typical” day is usually modelled, with

results 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 Annual

Average Daily Total (AADT) traffic is modelled, larger seasonal variations equate to

larger total variance between modelled day and actual day across the year.

For those projects where modelled hours are considered, mathematically the problem

increases, given further factoring from a “typical” hour (or perhaps AM peak and PM

peak) to a “typical” day. Conversely, when modelling a day, future congestion in peak

periods and its impact on effective daily capacities may be under-estimated.




2.10.9 Economic Effects

Economic 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/ upgraded

Economic cycles affect most aspects of the economy and decision-making, including

evaluation assumptions adopted. Transport consultants define economic growth

scenarios either under guidance or instruction of commissioning parties. When

expectations are high more projects are evaluated, so proportionally more projects are

likely to founder on downturn (and be blamed on transport forecasts). This may create

cynicism regarding tollway investments extending into the early economic recovery,

resulting in under-investment in some areas, thence over-investment as returns on

operating (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 Malaysia

potentially over-expanding transport infrastructure development.




2.11 Construction, Operations and Maintenance

Construction cost overruns and delay (deferred/ lost revenue) may imperil initial debt

repayments. 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, Flyvbjerg

and COWI (2004) recommend highway construction cost estimates be uplifted 15% if a

50% chance of overrun/ delay is acceptable, or by 32% if 20% chance acceptable.

Ruster (1996) notes construction cost overruns, delays and defects can be largely

mitigated by liquidated damages, performance bonds, warranties, contingency funds

and insurance. As revenue losses are rarely disputed during delay/ overrun arbitrations,

the focus of this Dissertation remains on demand-side risks. However, when the

contractor is the concessionaire, such risks should be analysed. Similarly, operations

and maintenance (O&M) risks should also be considered.

Table 2.4 shows estimated costs for new expressways in China and Vietnam. Whilst

costs 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 between

Dual-2 and Dual-3 standard. A distance-weighted average of US$4.633m per km of

Dual-2 was derived, to be used in Chapter 5’s simulation model.

There is a trade-off between construction and subsequent operations and maintenance

costs. The latter also affected by periodic major maintenance (e.g. immediately before

concession handback). Literature review found little agreement as to how to gauge such

costs, and whether they should be related to construction or traffic flow/ revenue. Table

2.5 shows some public domain values; some confidential sources suggested using 6% of




initial construction cost. In practice, there is likely to be a fixed element which could be

taken as proportional to construction costs, plus a variable element proportional to

traffic/ revenue. For simulations in Chapter 5, it is recommended to adopt 2% of

construction cost (fixed) plus 3% of toll revenue (variable).

Table 2.4: Estimated Expressway Construction Costs

Expressway Length

Cost

(US$ million)

Cost per km

(US$ million) Source

Guangzhou E-S-W Ring

Road, China

38 km,

Dual 3 US$542m 14.263

HHI (2003, pp

104-108)

Phase 1 West, Guangzhou,

China

14.7 km,

Dual 3 US$207m 14.082

HHI (2003, pp

114-118)

Hanoi – Lao Cai Expressway,

Vietnam

260km,

Dual 2 US$915m 3.519

Corbett, et al

(2006, p.VIII-4)

Nanning – Baise Expressway,

China

189km,

Dual 2 US$600m 3.175

Corbett, et al

(2006, p.VIII-5) Bien Hoa – Vung Tau

Expressway, Vietnam

90km,

Dual 2 US$680m 7.556

Dau Giay – Lien Khoung

Expressway, Vietnam

189km,

Dual 2 US$600m 3.175

Corbett, et al

(2006, p.VIII-9)

Hanoi – Haiphong

Expressway, Vietnam

100km,

Dual 2 US$410m 4.100

Corbett, et al

(2006, p.IX-4)

Da Nang – Quang Ngai

Expressway, Vietnam

140km,

Dual 2 US$700m 5.000

Saigon – Long Thanh – Dau

Day Expressway, Vietnam

55km,

Dual 2/ 3 US$350m 6.364

Hanoi Ring Road, Vietnam 65km US$600m 9.231

Total 1,140.7km US$5,604 4.913

Total (assuming Dual 2 throughout) 4.633

Table 2.5: Operations and Maintenance Costs

Highway

O&M as % of

Construction

(Mean)

O&M as % of

Conservative

Revenue (Mean) Source(s)

Hefei-Nanjing Expressway

(134km, Dual-2)

2.3% to 8.4%

(4.2%)

SWHK (1996a,

1996b)

Shanghai-Nanjing Expressway

(254km, Dual-2)

3.2% to 17.5%

(6.1%)

SWHK (1997a,

1997b)

Guangzhou E-S-W Ring 1.3% to 7.8%

(2.9%) PBA (2003)

Phase I West 2.1% to 5.1%

(3.1%) PBA (2003)




2.12 Other Considerations

Usually concessions are tendered. Ormerod (2005, pp.94-98) notes even with relatively

few pre-qualified firms (with technical, managerial, local and financial capability),

oligopolistic Nash equilibria are elusive. Possible strategic stakes, asymmetric

information, future expectations and track-records complicate bidding.

Given major highways’ perceived importance, local reputation/ guanxi11

may be

important. In Malaysia, Buchanan (1999) reports prospective concessionaires able to

identify then pursue projects uncontested. Gomez and Jomo (1999) observe well-

connected businesses getting lucrative contracts in exchange for undertaking less

lucrative 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 et

al (2005, p.92) note “pork barrelling” is prevalent in the Philippines, with an estimated

22.5% 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 of

over-investment. Whilst decentralisation is predicated on increasing responsiveness, a

lack of suitable local experience contributed to Mexico’s US$13bn 1989-94 toll road

programme amassing US$5.5bn in non-performing non-recourse loans (Irwin, 1999).

11 See Appendix 6 for detailed definition.




Whilst expecting the same “game rules” as in the West is unreasonable, corruption is a

concern. 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 unofficial

payments to each government tier involved. WB (2003b, 2004b) observes similar

problems in Indonesia, costing up to 30% of procurement budgets. Data in Appendix 9

show corruption is widely perceived as a problem, both within the region and by

Transparency International (2004); only Malaysia is (just) outside the “widespread

corruption” definition.

Brinkman (2003) and Kilsby (2004) identify other forecasting issues, such as models’

opaqueness, lack of resources to properly forecast, plus psychological and ethical

factors, overlapping to an extent with some of the “technical” issues above. This

includes modellers deluding themselves as to the infallibility and neutrality of their

forecasts, which are more often flawed and biased.




2.13 Summary of Key Issues

This Literature Review identified a number of evaluation metrics which may be used

(e.g. NPV, FIRR). Numerous project risks were also identified, whose importance may

vary between countries and projects (with some risks correlated), which may be

summarised 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 be

tested by questionnaire surveys (Chapter 4), in terms of how often such risks are

considered, whether they are deemed important and in the case of certain economic

parameters, 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 readily

testable by either questionnaire or risk simulation.

Addressing the hypothesis, excepting the use of interest rates in financial analysis, little

literature emphasised any importance of either price inflation or interest rates on

tollways. Are they unimportant? Or is this merely symptomatic of most literature being

based on Kondratieff downswing conditions? They are therefore included in the key

risks to be considered in both the questionnaire surveys and risk simulation.




Table 2.6: Summary of Key Risks and Issues

Risk Type

For Testing By

Questionnaire Risk Simulation

Macro-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 Strategy




3. Environmental Analysis

3.1 Introduction and PESTLE Analysis

Whilst Literature Review concentrated on generic project risks, environmental analysis

is 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 project

justifying its own framework. However, PESTLE analysis can identify external

dynamics affecting the market. Table 3.1 summarises key points (full analysis in

Appendix 10), showing a growing desire overall for inter-urban transport. The key

driving-force is economics; however, political/ legal constraints include corruption.

Table 3.1: Highlights of PESTLE Analysis

Element Description

Political Stability concerns in many countries, though not always deterring

infrastructure investment

Economic Economies generally growing relatively rapidly, although wealth

levels varied.

Social

Generally much/ growing inter-urban travel, in parallel with rapid

urbanisation. Demand suppressed in some cases by poor

infrastructure.

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.

Environmental Economic development predominates over environmental

considerations




3.2 Political, Legal and Stakeholder Issues

Cheong (1999) notes all countries experienced “maximum government” since 1945:

military (Indonesia, Myanmar, Thailand12

), emergency (Malaysia, Philippines) or

communist rule (Cambodia, China, Laos, Vietnam). Such potential for centralisation

remains either through current maximum government or switching from “rule of law"

to “rule by law.” Risks might be compounded by multi-tiered government with

overlapping authority and a lack of transparency. Stakeholder mapping can illustrate

opportunities and risks. Figure 3.A maps typical post-opening stakeholders13

.

Figure 3.A: Typical Concession Stakeholder Map

12 Thailand reverting to military rule during the preparation of this Dissertation. Although to date, little

resultant impact appears to have been made on economic sentiment.

13 Author’s own work.

Concession

Equity

Holders

Lenders

Government

(Concerned Dept.)

Users

Staff Suppliers

Rest of

Government

Competing

Projects

Society

Development

Agencies

(excepting

project donors/

lenders)

Supplying

Industries

(including

consultants,

contractors, etc)




Prior to award there are numerous potential concessionaires with equity/ debt financiers

(and in some cases possibly conflicting units within government allied with certain

bidders over others). During construction, suppliers would be more central.

In addition to competition with other routes/ concessions, there may be conflict between

different government departments having larger perceived stakes in other projects,

either through governmental equity involvement or guanxi. There is also a trade-off

between users and society; concession terms negotiated with government determine the

extent of user subsidisation/ penalisation. Likewise, there may be conflicts between

equity holders and government.

The above illustrates possible conflicting/ coinciding interests, which should be mapped

for each project individually. Equally, factors’ impacts may change: following the AFC,

connections with the Suharto family (previously critical to success in Indonesia) became

a business liability (Forsgren et al, 1999, p.152).

La Porta et al (1997) found shareholder protection, good accounting standards and rule

of law strongly negatively correlated with concentration of ownership, suggesting such

facets are important for good operation of capital markets to facilitate infrastructure

financing.




3.3 Economic Recovery

Section 2.8 discussed the causes and impacts to 2001 of the AFC in Thailand, showing

over-investment precipitated the currency crisis and investment-downturn. However,

there is evidence that economies (and construction as a proxy for infrastructure

investment 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, private

sector 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 as

indexed to 1995 in Baht and US$ respectively. However, Figure 3.F15

shows a recovery

in GFCF in recent years (indexed to 2000); GFCF appears relatively income elastic,

dipping lower than GDP in 2001, thereafter growing more rapidly. This suggests an

upturn in GFCF, likely to increase construction spending and possibly tollways.

Figure 3.G16

shows declines in a number of currencies following the AFC; only the

Chinese RMB was unscathed due to its pegging to the US$. Whilst time-series data on

other currencies were not available, Vietnamese Đong, Cambodian Riel, Myanmar

Kyats (free-market rate) and especially Lao Kip all depreciated substantially over this

period also. However, Figure 3.H16

shows that since January 2001 currencies have been

broadly stable; Myanmar Kyats (not shown) are the exception, continuing to devalue on

the free-market.

14 Raw data from: NESDB (2006) and www.fx.sauder.ubc.ca

15 Raw data from: www.bot.or.th and www.fx.sauder.ubc.ca

16 Raw data from: www.fx.sauder.ubc.ca


http://www.bot.or.th/






0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

19931994

19951996

19971998

19992000

20012002

20032004

20052006

Mil

lio

n B

ah

t (1

988 p

rices)

0%

10%

20%

30%

40%

50%

60%

Co

nstr

ucti

on

as %

of

GF

CF

Gross Fixed Capital Formation Private Construction

Government Construction Construction as % of GFCF

Figure 3.B: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter)

0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

19931994

19951996

19971998

19992000

20012002

20032004

20052006

Mil

lio

n U

SD

(1988 p

rices)

0%

10%

20%

30%

40%

50%

60%

Co

nstr

ucti

on

as %

of

GF

CF

Gross Fixed Capital Formation Private Construction

Government Construction Construction as % of GFCF

Figure 3.C: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) in US$




0

25

50

75

100

125

150

175

200

1994 1996 1998 2000 2002 2004

Ind

ex (

1995=

100)

GFCF GDP M2

Figure 3.D: Thai GFCF, GDP and M2 in Baht, Indexed to 1995

0

25

50

75

100

125

150

175

200

1994 1996 1998 2000 2002 2004

Ind

ex (

1995=

100)

GFCF GDP M2

Figure 3.E: Thai GFCF, GDP and M2 in US$, Indexed to 1995

50

75

100

125

150

2000 2001 2002 2003 2004

Ind

ex (

2000=

100)

GFCF GDP M2

Figure 3.F: Thai GFCF, GDP and M2 in US$, Indexed to 2000




0

20

40

60

80

100

120

Jan-9

4

Jan-9

5

Jan-9

6

Jan-9

7

Jan-9

8

Jan-9

9

Jan-0

0

Jan-0

1

Jan-0

2

Jan-0

3

Jan-0

4

Jan-0

5

Jan-0

6

Ind

exed

Valu

e v

s.

US

D (

Jan

96=

100)

Chinese RMB Indonesian Rupiah Malaysia Ringgit

Philippine Peso Thai Baht

Figure 3.G: Currency Performance since 1994

0

20

40

60

80

100

120

Jan-0

1

Jul-01

Jan-0

2

Jul-02

Jan-0

3

Jul-03

Jan-0

4

Jul-04

Jan-0

5

Jul-05

Jan-0

6

Ind

exed

Valu

e v

s.

US

D (

Jan

01=

100)

Chinese RMB Indonesian Rupiah Malaysia Ringgit

Philippine Peso Thai Baht

Figure 3.H: Currency Performance since 2001




3.4 Vehicle Ownership

Khan and Willumsen (1986) note correlation between car ownership and roadspace in

developing countries: statistically one proxying the other. ADB et al (2005, p.3) suggest

the following broad correlation between GDP and roadspace:

pitaofGDPperCaUSkminLandArea

PavedRoadsofkm$ln5.05.0

__

__ln

2

(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 and

2, comprising fits on the Study Area 9 countries, plus 5 others for benchmarking. These

suggest S-curve relationships for paved roads, railway and airports in terms of

kilometrages/ number of airports per km2 or per capita. Figures 3.I and 3.J show

equations fitted for roads per capita and per km2 respectively, with respect to GDP per

capita, suggesting substantial road build-out/ vehicle ownership growth are likely as

economies 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).




Population per km of Paved Road

MM

LA

KH

VN

ID

PH

CNTH

MX

MY

PO

KR

UK

US

3

4

5

6

7

8

9

10

0 10,000 20,000 30,000 40,000 50,000

GDP Per Capita (USD p.a.)

Ln

(Po

pu

lati

on

per

km

of

Paved

Ro

ad

)

Figure 3.I: Relationship between Wealth and Roads Per Capita

km2 per km of Paved Road

MM

LA

KH

VN

ID

PH

CN

TH

MX

MY

POKR

UK

US

-1

0

1

2

3

4

5

6

0 10,000 20,000 30,000 40,000 50,000


Ln

(km

2 p

er

km

of

Paved

Ro

ad

)

Figure 3.J: Relationship between Wealth and Road Density




3.5 Traffic Performance of Existing Toll Roads

For further analysis of income’s effect on traffic volumes, time series econometric

analyses 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 and

toll changes are not considered, the overall income elasticity of expressway traffic is

remarkably similar in all three instances; despite differences in vehicle ownership

sensitivities: although Guangdong Province is relatively more developed and thus might

be higher-up the S-curve (ownership growth smoothing off), this does not explain the

difference in vehicle ownership growth between Jiangsu and Zhejiang.

Table 3.2: Vehicle, Trip and Expressway Patronage Income Elasticities

Income Elasticity of:

Guangdong Province/

Guangzhou-Shenzhen

Superhighway

Jiangsu Province/

Shanghai-Nanjing

Expressway

Zhejiang Province/

Shanghai-Hangzhou-

Ningbo Expressway

Vehicle Ownership 1.02 1.41 1.78

Passenger-km 0.61 0.69 0.37

Passenger Trip Length 0.14 0.23 0.03

Freight MT-km 0.44 0.33 0.82

Freight Trip Length 0.68 0.30 0.37

Expressway 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 3.10

Expressway Revenue 1.38 n/a 2.11

Data from: 1995-2004 1997-2003 1998-2003




Whilst the largest vehicle categories’ expressway patronage is most responsive to GDP

growth, they constitute a small proportion of traffic as shown in Figures 3.K and 3.L for

Shanghai-Nanjing and Shanghai-Hangzhou-Ningbo Expressways respectively.

These analyses show the difference (and consequent risk) even between three leading

coastal 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 to

GDP, even in a country with relatively well developed highway networks relative to the

rest of the region (see Section 3.4) and thus may make an attractive investment.

Shenzhen Expressway (2006), Jiangsu Expressway (2006, p.143) and Zhejiang

Expressway (2006, p.7) levy quite similar tolls on interurban highways17

. Cars are tolled

at RMB0.40-0.60 per km (US$0.05-0.075 per km); and trucks (up 10 tonnes) at

RMB1.00-2.40 per km (US$0.12-0.30 per km), though on most routes at around

US$0.15 per km. Thus representative values of US$0.06 and US$0.15 per km for cars

and 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 vehicle

class and with toll rates not readily available.




0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

1997 1998 1999 2000 2001 2002 2003

Dis

tan

ce-W

eig

hte

d A

vera

ge V

eh

icle

s p

er

Day

Car Small Medium Large+Heavy

Figure 3.K: Traffic Growth on Shanghai-Nanjing Expressway

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

1998 1999 2000 2001 2002 2003

Dis

tan

ce-W

eig

hte

d A

vera

ge V

eh

icle

s p

er

Day

Revenue Small Medium Large Heavy

Figure 3.L: Traffic Growth on Shanghai-Hangzhou-Ningbo Expressway




3.6 Opportunities and Threats

Section 3.5 illustrated that different projects in the same country may perform

differently. Hence, it is not feasible to present Strengths and Weaknesses analysis for

the Study Area as a whole. Nevertheless a broad-brush Opportunities and Threats

analysis can summarise key potential macro-level downside and upside risks, based on

the clusters suggested in Section 3.4:

3.6.1 Cambodia, Laos and Myanmar

Opportunities include potential for substantial growth in car usage and expansion of

highway networks. Given their pressing development needs, favourable/ flexible

contract terms might be possible, possibly with partial funding from aid agencies

(subject to sanctions in case of Myanmar). Geographically these countries link the

stronger regional economies: Thai-Vietnamese land transport either via Laos or

Cambodia; Sino-Thai trade via Laos or Myanmar; Myanmar offering land-linkage

between East Asia and South Asia.

A key threat is that current poverty may lengthen ramp-up and limit toll affordability

and traffic levels. Significant sovereign and institutional risks persist, together with

corruption.

These countries are thus quite risky.

3.6.2 Philippines and Vietnam

High capacity trunk highway networks are largely undeveloped, meaning attractive

routes remain to be developed. Economic growth suggests that tolls might be

affordable. Vietnam has signalled intent to open-up to expanded FDI, whilst the

Philippines is the most culturally westernised country in the sample.




However, legal protection for investors remains weak and the extent of tolling

affordability is uncertain. Corruption remains a concern.

These countries are fairly risky.

3.6.3 China, Indonesia, Malaysia and Thailand

These countries have relatively strong economies with strong prospects, both in export-

oriented manufacturing and commodity markets. Furthermore, they have sizeable

domestic economies possibly providing some resilience to international economic

factors. They also have strong track records in attracting FDI into transport

infrastructure, including reasonable legal systems (as compared to other countries in the

region). Tolls are relatively easily afforded by many drivers.

However, these countries may risk over-investment in certain regions (as befell all bar

China in the AFC). There remains some legal/ institutional risk, as well as corruption or

a need to have business networks (e.g. guanxi). It might be argued that many of the

most attractive projects have already been built.




3.7 Postulated Position on K-Wave

Based on a peak in interest rates in most leading economies in the early 1980’s, along

with a peak in commodity prices (especially gold) and inflation rates, the last

downwave begun around 1980/1981. Equally, the bottoming out of many commodity

prices 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 interest

rates 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 the

USA are evidence of inflation; price indices are likely to accelerate. Prolonged low real

interest rates since the K-Wave bottom are likely to yield gold prices over-and-above

what 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 recently

postulated and examined the implications of significant fuel price increases; though

such work is not yet widespread.

Assuming the K-Wave exists, an upswing has likely begun. Given S-curves of vehicle

ownership and the above analyses on road build-out, this suggests an upturn in

investment prospects, also coinciding more-or-less with the Kuznets cycle ready to

rebound (based on roughly a half-cycle elapsed since AFC). Private sector project

financing developed since the 1980’s (during a Downswing); will the Upswing change

the rules of the game?




0

2

4

6

8

10

12

14

16

18

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

Tre

asu

ry B

ill

Inte

rest

Rate

s (

%)

0

100

200

300

400

500

600

700

800

Go

ld P

rice (

US

D p

er

oz)

10 Year T-Bill 3 Month T-Bill Gold(USD/oz) K-Wave

Figure 3.M: Interest Rates, Nominal Gold Price and Kondratieff Wave




4. Questionnaire Survey

4.1 Purpose

The Literature Review (Chapter 2) presented a number of project risks associated with

start-up toll road projects, as well as critiques of transport planners’ failures to take due

consideration of these. The Environmental Analysis (Chapter 3) intimated the potential

for toll roads in Study Area countries. In order to test both literature and environmental

analyses, a questionnaire survey was undertaken to test practitioners’ experience and

perceptions 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 literature

review and environmental analysis, expectations were measured to help define a

forecast scenario for risk simulation testing (see Chapter 5). Differences in attitudes

between different project stakeholders/ professional groups were also evaluated, both to

test how different stakeholders perceive risks and to identify gaps between transport

planners’ performance and others’ expectations of them.




4.2 Design Concept and Sample Selection

The questionnaire was designed to afford a relatively broad sample of opinion; various

advisory professions were sampled. Respondents were also asked to state the extent of

their working experience, the proportion of this spent in relevant fields and their

geographic experience. Questions covered economic, legal, engineering and

connectivity risks, as well as (for those with modelling experience) an investigation into

the reliability of transport demand forecasts, attempting to identify where practitioners

feel their art is weakest.

Given the relative obscurity of business cycle theory even amongst economists, only

one question relates directly to the use of business cycles, though others test

expectations regarding price inflation and other economic variables. In order to prevent

comparison 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 from

literature reviewed, using a number of internet-based newsgroups (“yahoogroups”), plus

review of professional databases (e.g. www.legal500.com for legal professionals). As

suggested by “sub-tribalism” (Morris, 1971), the best response rate was from those

known to the author and those in the same primary field (transport planning), so the

sample skewed towards transport planners/ economists. Such people also accounted for

most of the optional (text) responses on broader issues.

Given that some questions were designed specifically for transport planners this was not

a problem (those without such experience being screened out of such questions, though

not the rest of the survey, as described below).

http://www.legal500.com/




4.3 Questionnaire Design and Survey Execution

In order to expedite survey diffusion and result collation, the internet-based

www.surveymonkey.com was employed, allowing easy questionnaire dissemination

and 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, with

observations on the Pilot in Appendix 14 (also detailing actions taken to revise the

questionnaire to incorporate pilot feedback).

Approximately 40 respondents started the survey but dropped-out after just a few

questions. These responses were excluded from the analysis. In a number of cases,

respondents did not give answers to each question, but nonetheless gave answers to

many questions. Under such circumstances a “not sure” response was assumed for

omitted answers. And when evaluating answers, such “not sure” responses were

typically excluded, such that analysis would concentrate on stated opinions only. A total

of 162 responses were considered as valid for analysis (though due to “not sure” and

omitted 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 analyse

responses by headings under which questions were grouped.

http://www.surveymonkey.com/




4.4 The Survey Sample

The first question determined in which sector(s) respondents had experience, based on

14 categories, with multiple answers permitted. The 162 responses were aggregated into

six categories as shown in Table 4.1, based on which different stakeholders’ attitudes

could be examined (as shown later). The relative proportions are also shown in Figure

4.A (note: many respondents worked in multiple sectors).

Table 4.1: Aggregated Respondent Experience Categories

Group Components Number

Financial, Legal,

Operator

Expressway Developer/ Operator/ Equity

Investor

Lawyer/ Attorney/ Solicitor

Private Sector Lender

Investment Banker

Ratings Agency

Accountant/ Valuer

Insurer

29

Transport Planner/

Economist

Transport Planning Consultant

Economist 98

Engineer/ Architect Civil/ Structural/ Pavement/ Highway

Engineer/ Architect 37

Government/ Aid

Agency

Government

Aid Agency 43

Academic Academic 22

Other Other 24

The largest group was transport planners (for reasons explained in 4.2 above). The total

sample was relatively experienced (20.6 years mean working experience), as shown in

Figure 4.B. The sample’s working experience cross-tabulated with years of experience

in Table 4.2 shows the average respondent has spent over 10 years on transport

infrastructure projects and over 7 in developing countries. Although there were perhaps

not a great many respondents in the Financial/ Legal/ Operator category, respondents

did include a number of very senior figures within this category, including key decision-

makers/ backers of private infrastructure schemes.




0

20

40

60

80

100

Financial,

Legal,

Operator

Transport

Planner,

Economist

Engineer,

Architect

Government,

Aid Agency

Academic Others

Figure 4.A: Respondents by Experience Type

30+

26%

20 to 29

30%

10 to 19

31%

5 to 9

6%

1 to 4

7%

Figure 4.B: Respondents by Years of Experience




Table 4.2: Respondents’ Mean Years’ Experience in Various Fields

Project Type

Average

Years per

Respondent

Sample’s

Total Years’

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 1.12 173

Figure 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 Study

Area countries, showing substantial numbers with China experience (69 respondents),

through to few with Myanmar experience (5 respondents). As a significant proportion of

the sample have experience within East Asia and a familiarity with developing

economies, the sample appears suitable for analysis.

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

Other

Respondents with Experience in this Region

Figure 4.C: Respondents’ Global Experience




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

Vietnam

Respondents Experience by Country

Figure 4.D: Respondents with Experience in East Asia

Table 4.3: Respondents with Experience in Study Area

Tolled

Highways

Other

Transport

Projects

Other

Infrastructure

Projects

Non-

Infrastructure

Projects

Anything in

this

Country Cambodia 1 18 9 5 20

China 38 49 29 25 69 Indonesia 16 30 9 10 43

Laos 2 14 8 7 20 Malaysia 19 31 12 13 42 Myanmar 0 4 1 0 5

Philippines 19 30 13 13 44 Thailand 22 39 18 12 50 Vietnam 8 21 8 14 36




4.5 Tollway Appraisal

Figure 4.E shows respondents’ attitudes to macro-level risks (5 signifying critical and 1

unimportant18

; mean being the red bar and standard deviation the black line). Findings

are broadly consistent with literature, with sovereign and institutional risks (political

and legal) predominating, followed by economic factors which, as shown previously

drive much of a project’s likely success (e.g. Sections 2.10.1 and 3.4). Income

inequality and toll familiarity were not deemed important. Overall there was neutral

opinion towards corruption, currency risks, price inflation and interest rates. The latter

two possibly due to adaptive expectations from recent lows in both; business cycles

were also deemed unimportant. The literature review found very little written transport

literature 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 not

much difference between groups, though transport planners are less concerned about

corruption than other groups; possibly as neither parties to the concession proper nor to

construction, it affects them less. Figure 4.E showed a relatively large variance for

corruption; 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 opinion

between groups. Financial/ legal/ operators are relatively more concerned with

minimum income guarantees and toll affordability; and less concerned with

construction time, construction and running costs, relative to most other groups; i.e.

18 Values transposed from raw data in Appendix 15.




they are more sensitive to revenues relative to costs than others. Ramp-up is ranked

bottom by all, despite its potential impact on early project cashflow.

0

1

2

3

4

5

Poli

tical S

yst

em

Leg

al S

yst

emE

con

om

ic G

row

th

Corr

up

tion

Rep

atr

iati

ng P

rofi

tsC

urre

ncy

Ris

ks

Pri

ce I

nfl

ati

on

Inte

rest

Rate

sB

usi

nes

s C

ycl

esT

oll

Fam

ilia

rity

Inco

me

(In

)Eq

uali

ty

Figure 4.E: Attitudes to Macro-Level Risks

Table 4.4: Rankings of Macro-Level Risks by Respondent Category

All

Financial,

Legal,

Operators

Transport

Planner,

Economist

Engineer,

Architect

Government,

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 4

Repatriating 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 10

Income (In)Equality 11 11 11 11 10 11 11




0

1

2

3

4

5

Leg

al/ co

ntr

actu

al fo

undati

ons

Const

ruct

ion c

ost

Com

pet

ing

route

s

Tol

l in

crea

se e

nfo

rcea

bilit

y

Soci

al/ e

conom

ic b

enef

its

Const

ruct

ion t

ime

Conce

ssio

n len

gth

Oper

atin

g &

main

tenan

ce c

ost

s

Tol

l af

ford

abilit

y (la

rge

vehic

les)

Connec

ting r

oute

s

Tol

l af

ford

abilit

y (oth

er v

ehic

les)

Guanxi

Min

imum

inco

me

guar

ante

esT

oll le

akag

e

Ram

p u

p

Figure 4.F: Attitudes to Project-Level Risks

Table 4.5: Rankings of Project-Level Risks by Respondent Category

All

Financial,

Legal,

Operators

Transport

Planner,

Economist

Engineer,

Architect

Government,

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 3 Toll affordability (large

vehicles) 9 7 9 12 12 12 11 Connecting routes 10 7 8 14 9 5 13

Toll 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 15




4.6 Transport Modelling Issues

Confined to those with modelling experience, Figures 4.G shows respondents’

experience of data availability, for model calibration and forecasting, with 5 signifying

always and 1 never19

. This shows little difference between calibration and forecast data

availability and reliability, that reliability is typically slightly worse than availability,

but that data are generally more available and reliable in developed countries, as would

be expected.

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

5.00

Developed

countries;

sufficient data

Developed

countries;

reliable data

Developing

countries;

sufficient data

Developing

countries;

reliable data

Calibration Forecast

Figure 4.G: Data Availability and Reliability





Figure 4.H presents attitudes to different transport model types (defined in Section 2.9),

with 5 representing Strongly Agree and 1 Strongly Disagree20

. Four stage and

assignment models are both seen as slightly reliable, with spreadsheets marginally less

so. On balance no model type is seen as too data hungry, simplistic or complicated (by

degree). Interestingly, four stage models are seen as least inappropriate for tollways

(contrasting with Willumsen and Russell, 1998), though they are perceived as “black

boxes” (echoing much literature). There is little difference in perceived suitability for

developing 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 Spreadsheet

Figure 4.H: Attitudes to Transport Model Types





4.7 Forecast Performance and Evaluation Criteria

Figure 4.I presents perceptions of forecast performance. It appears fairly rare for outturn

traffic to significantly exceed forecasts. All groups experienced significant

overforecasting (consistent with literature); whilst Financial/ Legal/ Operators have

strongest experience of this, Transport Planners/ Economists cite the next strongest

experience 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 and

debt perspectives. This is surprising; one pilot respondent noted (by follow-up email), it

would be “utterly wrong" if equity- and debt-side forecasts were the same, given the

different risk/ reward profiles of either side.

This raises concern as to practitioners’ and users’ understanding of forecasting. It may

reinforce Brinkman’s (2003) assertion of forecasters being self-deceived as to the

supposed inscrutable neutrality of their models; and to systematic forecast errors

observed by Bain and Polakovic (2005).

Figure 4.J shows how often respondents’ consider various forecast outputs and other

factors when appraising projects. Case study congestion is most often considered, then

base/ central case traffic and revenue, then conservative forecasts, followed by

congestion on competing then feeder routes. Conservative forecasts are used more often

than optimistic ones (perhaps allaying some of the concerns regarding differences

between equity- and debt-side forecasts).

Figure 4.K shows how often respondents’ consider various aspects of a project. NPV

and FIRR are most commonly used, suggesting the primacy of financial returns over




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 considered

least often. Some respondents also cited (by text entry) use of the Debt Service

Coverage 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

Other

Figure 4.I: Perceptions of Forecast Performance




0.0 1.0 2.0 3.0 4.0 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 Forecasts

Never Rarely Sometimes Usually Always

Figure 4.J: Which Forecast Outputs are Considered?

0.0

1.0

2.0

3.0

4.0

5.0

Net Present

Value (NPV)

Financial

Internal Rate

of Return

(FIRR)

Economic

Internal Rate

of Return

(EIRR)

Social Cost/

Benefit

Ratios

Risk

correlation

versus other

projects in

portfolio

Counterparty

risks

Sovereign/

Institutional

other

country/

legal risks

Always

Usually

Sometimes

Rarely

Never

Never

Figure 4.K: How Often Are Which Criteria Considered?




4.8 Countries’ Outlooks

Respondents’ expectations of the toll road market presented in Figure 4.L broadly

concur with the categorisation of countries presented in Section 3.4 (countries coloured

according to those categories), though Indonesia’s market is perceived as less developed

than that of the Philippines. Malaysia, China and Thailand have the most developed

markets. Cambodia, Laos and Myanmar are perceived as being almost nascent on

average (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 respondents

have worked in, showing a general positive bias by those with country experience.

Differences are typically small (Indonesia and Malaysia having the largest); although

Cambodia and Myanmar are rated as nascent and Indonesia overtakes the Philippines

being 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 and

particularly 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’etat

immediately prior to the survey). Yet they are more bullish on Indonesia and Malaysia.




1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

Malaysia China Thailand Philippines Indonesia Vietnam Cambodia Myanmar Laos

Developing

Nascent

No Market

Steady

Maturing

O ver-

Developed

Figure 4.L: Perceived Tollway Market Opportunities by Country

1.0 2.0 3.0 4.0 5.0 6.0

Cambodia

China

Indonesia

Laos

Malaysia

Myanmar

Philippines

Thailand

Vietnam

Full Sample Those With Country Experience

No Market Nascent Developing Steady Maturing Over-

Developed

Figure 4.M: Impact of Experience on Country Perceptions




1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Cambodia

China

Indonesia

Laos

Malaysia

Myanmar

Philippines

Thailand

Vietnam



Other

No Market Nascent Developing Steady Over-Developed

Figure 4.N: Country Perceptions by Respondent Category




4.9 Economic Outlook

Figure 4.O shows overall respondents anticipate substantially higher fuel prices in the

future, as well as increased tolling acceptability and general price inflation. Interest

rates, economic growth and exchange rate volatility are also expected to increase.

Figure 4.P illustrates that there are no major differences of perception between

respondent groups. Perceptions are largely consistent with the economic outlook posited

by the K-Wave as set out in Section 3.7, even if the perceived impacts of rising interest

rates and price inflation are not deemed significant (Section 4.5)

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

Significant

Increase

Increase to an

Extent

No Change

Decrease to an

Extent

Significant

Decrease

Figure 4.O: Economic Expectations




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



Other

Significant

Decrease

Decrease to

an Extent No Change

Increase to

an Extent

Significant

Increase

Figure 4.P: Economic Expectations by Respondent Group




4.10 Other Comments

A variety of text comments were also received. With regards transport modelling, the

key theme was of tailoring models to local conditions and of the potential validity of all

approaches, subject to circumstances (e.g. data availability, timescale, phase of project

life-cycle, client requirements, etc.)

In all 71 respondents requested information on findings (23 on the survey, 48 on

broader research), perhaps intimating that this research is of perceived importance.

4.11 Key Conclusions from the Questionnaire Survey

There was a significant response rate from transport planners and economists, with a

lower number of responses from other groups. Nevertheless, it was deemed that there

were sufficient data to analyse different stakeholder perceptions (using groupings in

Table 4.1). With mean working experience of 20.6 years, the sample has substantial

experience; the “average” respondent has just over one year’s experience in rural tolled

highways in developing countries. The sample is thus deemed sufficient for the

purposes of this Dissertation.

There is perceived primacy of legal and political factors on viability; though once

modelling commences, economic factors predominate. Business cycles, toll familiarity

and 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 are

perceived as fairly reliable, but also as opaque.




Whilst under-forecasting appears relatively rare, over-forecasting happens much more

often. There is some acknowledgement of transport planners adjusting forecasts to meet

clients’ expectations. There appears a fundamental misunderstanding of the purposes of

equity- and debt-side forecasts; this based on only weak acceptance of differences

between forecasts for either side (as the author suspected a priori).

NPV is the most often-used evaluation criterion, followed by FIRR, then economic

metrics. Counterparty risks and risk correlation versus other projects are used more

rarely.

Country categorisation in Section 3.4 is broadly supported, but with Indonesia seen as

less 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 and

Vietnam as nascent-to-developing; and Cambodia, Myanmar and Laos as sub-nascent.

However, those with Indonesia experience rate the country as developing-to-steady; and

those with local experience regard Cambodia and Myanmar as nascent.

There is a reasonable acceptance of symptoms of a K-Wave upswing, in terms of

increasing price inflation (especially fuel prices), interest rates and to a lesser extent,

economic growth. Tolling acceptability is predicted to increase. However, respondents

did not deem the impacts of rising interest rates and price inflation to be significant.




5. Risk Simulation Modelling

5.1 Introduction

Chapter 4 presented inter alia perceptions of different risks, as well as respondents’

expectations of economic parameters. For example, on average respondents believed

inflation and interest rates would increase, but would have little impact on project risk.

Monte Carlo simulation is used to quantitatively estimate the relative importance of

different risks, to test whether respondents might have underrated such risks. Although

each project has specific locational and institutional risks, such are excluded here

through use of a simplified, fictional case; the aim being to concentrate on the relative

importance 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.




5.2 The Case Study and Its Parameterisation

The case study is necessarily contrived to give a distribution of outcomes suitable for

analysing the relative importance of different risk elements; and in particular to test the

hypothesis of the impact of increasing price inflation and interest rates on project risk.

As far as practicable, parameters were taken from public sources (and where

appropriate, influenced by questionnaire responses); inevitably some use had to be

made of confidential sources. Finally, values were adjusted (primarily base travel

demand) to guarantee the distribution of financial outcomes outlined above, together

with a “base case” showing FIRR≈16% (see 5.4), corresponding to a typically required

investment threshold (see 2.4.1).

The case study network topology is shown in Figure 5.A, constituting six zones (for trip

origins and destinations) and eight links, including the fictional tolled highway. The

lengths and freeflow speeds of each link are shown in Table 5.1. Though the number of

lanes 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 on

each origin-destination movement); shaded cells correspond to movements that could

potentially use the tollway. Both the total number of trips and road capacities are

included amongst the simulation variables; all of which are shown in Appendix 16. 27

parameters were common for all three Cases (Conventional, Respondents’ and

Kondratieff). A further 6 parameters had values specified for each Case differently,

though for each Monte Carlo iteration, the values were inter-related.




A

G

H

B F

C D E

Start of

Highway

End of

Highway

Zone 1

Zone 2 Zone 3 Zone 4 Zone 5

Zone 6

Figure 5.A: Case Study Notional Network

Table 5.1: Basic Link Characteristics of Case Study Network

Road

Section

(“Link”)

Length

(km)

Road

Standard

Lanes per

Direction

Freeflow Speed (kph)

Small

Vehicles

Large

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 2 120 100

Table 5.2: Assumed Trip Distribution (% by O-D Pair)

To Zone

1 2 3 4 5 6 Total

Fro

m Z

on

e

1 3% 2% 4% 4% 5% 18%

2 3% 5% 3% 3% 3% 17%

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%




Sections 2.10 and 2.11 gave evidence of serially overoptimistic errors in tollway

appraisal and transport models, supported by questionnaire findings (Section 4.7) with

outturn parameters more often being low than high. Thus it was not felt reasonable to

adopt symmetric probability distributions for all parameters either side of a notional

“mean”, “median” or “base” value in many instances. Rather a modal average was

specified, together with a probability of below-modal value21

; a different standard

deviation either side of this modal value was also applied in many cases. Finally, to

preclude 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 the

forecast horizon (e.g. gradually declining economic growth rates); such detail was

deemed superfluous for this exercise. It might be argued that a key risk of the K-Wave

upswing pertains to those seeking downstream refinancing (typically ever pricier as

opposed to cheaper during a downswing); however, the impacts of inter alia different

interest rates are tested across the three scenarios and via simulation.

Thus a single economic growth rate, together with a single elasticity across time in each

case (though specified separately for small and large vehicles) was adopted. The use of

Monte Carlo techniques ought anyway to proxy for such uncertainty. Moreover, it

permits the analysis of economic growth, T

y , p

D or y

VOT per se, which would not be

readily feasible if such parameters changed across the forecast horizon. The relative

importance of unforeseen changes in these parameters can be gauged to an extent from

analysing changes in project value due to changes in these parameters. More detailed

21 And by implication probability of above-modal values.




analysis might be appropriate for a specific case study, wherein parameter sets tailored

to specific local conditions would be used in lieu of generic values and ranges, such as

those befitting and employed by this Dissertation. Such simplification also permits the

consideration of a wider range of forecast parameters. Furthermore, in the case of

interest rates, it is reasonable to assume prospective concessionaires would size and

acquire debt based on current interest rates (e.g. through issuance of bonds) and that

changes to interest rates will primarily affect bridging loans or overdrafts required

downstream (i.e. unbudgeted when deciding whether to proceed and on finance

structuring).

The Respondents’ and Kondratieff Cases adjusted Conventional Case values, with

Kondratieff Case and price inflation and interest rates greater than or equal to

Respondents’ Case and these at least as great as Conventional values. Vehicle Operating

Costs in Kondratieff and Respondents’ cases were the same, as respondents largely

predict significantly higher fuel prices, in line with Kondratieff-based forecasts. All

random parameters are specified in Appendix 16.

Fixed parameters are summarised in Appendix 17. The modelled concession length was

30 years (modelled as 120 quarters), including construction time.




5.3 Methodology

10,000 iterations of Monte Carlo simulation were employed, using the following

methodology:

5.3.1 Defining Random Parameters

As 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 assumed

constant across all years; though different between Conventional, Respondents’ and

Kondratieff cases and with different values for each iteration. Distributions used are

shown in Appendix 16.

5.3.2 Applying Parameters to Derive Variables for Each Quarter

Quarterly 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 are

given in Appendix 18. In the case of toll rates where increases were not uniform, but

rather at certain intervals, assumed toll rates were kept fixed, being updated to the

correct theoretical toll rate every x intervals (x= number of quarters between increases).

5.3.3 Traffic Assignment for Each Quarter

For each quarter, two-class assignment (small and large vehicles) was performed using

a 10-iteration incremental loading (to take account of congestion) and logit equations to

apportion loads on each iteration between expressway-using and non-expressway paths.

Speeds were initially set to freeflow values. Generalised costs were then determined

based on these speeds and on each iteration 1/10 of the matrices were assigned, being

split between expressway-using and non-expressway routes using a logit relationship,




given in Appendix 18. Link loadings were then increased as appropriate based on the

split of this traffic (10% of the matrix further split by the logit curve). Speeds were

revised based on the new loadings, using volume/capacity-to-speed relationships given

in Figure 5.B. The process was repeated until 10 iterations had been completed.

0

20

40

60

80

100

120

0 0.5 1 1.5 2 2.5

Volume/Capacity Ratio

Sp

eed

(k

ph

)

Tollway (Small Vehicles) Tollway (Large Vehicles)

Local Road (Small Vehicles) Local Road (Large Vehicles)

Figure 5.B: Volume/Capacity-to-Speed Relationships

5.3.4 Financial Analysis

Having obtained loadings of small and large vehicles on the expressway link for each

quarter, financial analysis followed.

For any quarters prior to completion of the expressway, flows were set to zero, and the

appropriate quarterly construction costs were accrued. For subsequent quarters

following opening (where revenues were generated), ramp-up was applied; which was

assumed to be linear from its first to final quarter. For later quarters, any flows over the

expressway’s capacity were capped-off, in line with industry standard practice and




initial revenues were calculated. Variable operations & maintenance costs were

subtracted from these on a percentage basis as were losses from toll leakage. This gave

a net revenue from which fixed operations & maintenance costs were subtracted.

Any “surplus” revenues were used first to pay-off extra debts incurred, first paying off

interest 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 any

quarter 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, comprising

FIRR, payback period and NPV at various interest rates. For purposes of comparison

between cases and iterations, FIRR was used. Also, any run where FIRR≤0% or there

was no payback within 120 quarters was deemed to constitute “financial failure” (i.e.

bankruptcy). Comparative probabilities of “failure” were also used to compare between

runs (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 each

parameter usually randomised. The objective being to ensure a realistic return on the

base scheme and to provide an ample spread of performance (i.e. a meaningful but not

overwhelming prevalence of “failure”); also to enable an initial comparison between the

three cases.

22 Initially, debt was sized at 10% more than the envisaged construction cost providing a small buffer

against interest rates.




The results are shown in Table 5.3. This suggests that despite delayed Payback (due to

higher price inflation and interest rates) the Kondratieff case may give superior returns

than the Conventional case, but that the Respondents’ case would yield the best returns;

perhaps indicating more optimism amongst practitioners (transport planners being the

largest respondent group, as per Table 4.1). However, the risky nature of forecasting

means 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 $27,524,725

5.5 Comparison of Simulation Results between Cases

Though the results in 5.4 suggest that the Kondratieff case might be more beneficial to

investors than the Conventional case (based on recent past experience), does this

translate into less risk? Equally, is the apparent optimism of the Respondents’ case

consistent over risk-testing also? Do the higher price inflation and interest rates inherent

in the Kondratieff case (and to a lesser extent in the Respondents’ case) increase

riskiness when tested using Monte Carlo risk simulation techniques?

Summary results from the 10,000 simulations for each case are shown in Table 5.4, with

cumulative probability distributions of FIRR, payback and NPV (at 16%) shown in

Figure 5.C, 5.D and 5.E respectively. Comparing Table 5.4 against Table 5.3, mean

FIRR’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 both

FIRR and payback, the general pattern of Respondents’ Case being the most optimistic




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 substantially

higher 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 large

negative NPV’s. In all cases, Kondratieff standard deviations are the greatest and

Conventional standard deviations the smallest. Furthermore, 12.5% of Kondratieff runs

resulted 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 that

notwithstanding its superior mean values, the Respondents’ case is riskier than the

Conventional case; however, the Kondratieff case is substantially riskier still. The next

section analyses the impacts of different risk elements, underlying these results.

Table 5.4: Summary Results from Simulation Runs

Metric Statistic

Case

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

Period

(years)†

Mean 10.83 10.85 11.60

Minimum 6.44 6.31 6.53

Maximum 29.98 29.23 29.66 Standard Deviation 2.32 2.41 2.73

NPV

at 16%

($ million)

Mean $28.1 $46.4 -$37.7

Minimum -$377.8 -$936.0 -$5,600.3

Maximum $296.6 $357.3 $465.0 Standard Deviation $80.3 $96.3 $308.9

Financial

Failure

# of Cases 55 114 1250

% 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.




0%

20%

40%

60%

80%

100%

5% 10% 15% 20% 25%

Cu

mu

lati

ve %

Conventional Respondents' Kondratieff

Figure 5.C: Cumulative Probability Distribution of FIRR (excluding FIRR<0%)

0%

20%

40%

60%

80%

100%

8 10 12 14 16 18 20 22

Cu

mu

lati

ve %


Figure 5.D: Cumulative Probability Distribution of Payback Period (years)

0%

20%

40%

60%

80%

100%

-500 -400 -300 -200 -100 0 100 200 300 400

Cu

mu

lati

ve %


Figure 5.E: Cumulative Probability Distribution of NPV at 16% ($m)




5.6 Analysis of Individual Risks

Appendix 19 shows the impact of different simulation variables on FIRR and the

probability of financial failure. As some variables have a large but poorly correlated

effect on FIRR, whilst others have a smaller but better correlated effect, Appendix 20

takes 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 coefficient23

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-relationship

The simulation variables were then grouped by category, so as to determine the relative

importance of such risk categories, so as to avoid distortions due to the number of

variables tested within each category. The categories were then ranked for each of the

three cases, as shown in Table 5.5. The Case Study appears to give very low importance

to the Value of Time, but this is likely a consequence of the nature of network

modelled. What is more critical in the context of this Dissertation is the change in

impacts and rankings between cases. Excepting Vehicle Operating Costs and Toll

Leakage (the latter in the Respondents’ case), all parameters increase their impact on

FIRR between Conventional and Respondents’ and Respondents’ and Kondratieff

cases, indicating increased forecast risk volatility overall.

23 i.e. coefficient rather than constant from linear regression in Appendix 19. Goodness-of-fit between

linear and log-linear equations was very similar, so for simplicity the linear regressions were used here.




Table 5.5: Rankings of Risk Categories’ Importance by Case

Risk Group Conventional Respondents' Kondratieff

Impact*R2 Ranking Impact*R

2 Ranking Impact*R

2 Ranking

Road Capacities 0.30% 11 0.53% 10 1.35% 9 Construction Cost &

Duration 5.68% 4 6.40% 4 8.75% 5

All O&M Costs 0.59% 9 1.02% 8 1.98% 8 Value of Time & Its

Income Elasticity 0.00% 13 0.03% 13 0.05% 13

Vehicle Operating Costs 0.30% 10 0.19% 11 0.27% 12 Demand (Initial &

Income Elasticity) 12.13% 1 12.69% 2 16.05% 2

Toll Revenue Leakage 0.95% 8 0.94% 9 1.20% 10 Ramp-Up: Amplitude &

Duration 2.14% 6 2.30% 7 3.66% 7

Logit Model Parameters 0.09% 12 0.12% 12 0.32% 11 Toll Escalation Rate and

Frequency 1.73% 7 2.56% 6 4.69% 6

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% 1

Interest rates increase markedly in importance in both Respondents’ and Kondratieff

cases; this may be slightly overstated as initial interest rates feed into interest rates on

any extra (subsequent) borrowings. However, in the Kondratieff case both sets of

interest rates would individually outrank all other categories; illustrating the exponential

increase in their impact as they rise, thus signifying that interest rates increase markedly

in importance in times of high (or increasing) interest rates. Demand ranks as most

important in the Conventional case and remains second only to interest rates in the other

cases, followed by GDP growth (which itself permeates many other parameters as

explained in Sections 2.10.1 and 3.4). Price inflation and construction costs/ duration

are 4th

and 5th

most important (precise ranking case-dependent). The importance of toll

escalation rates and frequency of increases also increases in the Respondents’ and

especially the Kondratieff case, as might be expected: with price inflation increasing,

the impact of delayed or incomplete adjustments increases. Indeed the impact of price

inflation does not appear simple. General price inflation accounts for most price impact,




but is positively correlated with outturn performance; likely because it decreases the real

value of initial debt and is abated to an extent by toll increases, so downside risks

associated with increased price inflation are statistically associated more strongly with

toll escalation-associated variables.

Such inter-relationships between simulation system variables often occur; isolation of

individual variables’ impacts is not always possible (Pindyck and Rubinfeld, 1981).

With regards the Hypothesis, this suggests the impacts of increasing price inflation and

interest rates on various project risk elements are not always wholly linear; rather, more

complex system-wide interactions are possible.




5.7 Discussion of Results

In terms of the Hypothesis, the importance and riskiness of interest rates increase

markedly as they increase. In the Case Study in both Respondents’ and Kondratieff

cases interest rates become the most important determinant of project FIRR. The extent

of 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 interest

rates increase the risk-free return on cash (via bank deposits) would also increase and

with it investors’ required project returns. Hence the overall trend remains reasonable.

Coupled with increasing price inflation, the impact of variance in almost all forecast

variables/ 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). This

suggests that greater caution should be exercised by stakeholders when evaluating

schemes under such circumstances; and more investigation of risk be undertaken.

Adopting K-Wave Theory, the increase in price inflation and moreover interest rates

should be weighed against possible increased economic growth (also potentially

affecting initial demand). In fact, given that the K-Wave appears to still be in the early

stages of upswing, the impacts of increasing interest rates may not be too substantial at

present, 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 be

advisable to refinance projects down-stream; rather, sufficient fixed-rate debt should be

acquired at project outset.




6. Discussion and Conclusions

6.1 Introduction

Having undertaken literature review, environmental analysis, questionnaire surveys and

risk simulation modelling, this section discusses and summarises findings with a view to

answering 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


inflation and interest rates.

In addition to reviewing evaluation criteria (discussed in 6.2), the literature review

identified 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 6.7

makes observations and recommendations for both transport planners and project

finance. Finally, formal evaluation of the hypothesis is performed (Section 6.8).

Inevitably there is a certain degree of overlap between sections.




6.2 Evaluation Criteria and Implications of the Time-Nature of Risk

Investors tend to focus on a scheme’s NPV and FIRR. The questionnaire survey

revealed that NPV is used slightly more often than FIRR. Economic criteria are used

fairly often. Ratings agencies and lenders also consider Debt Service Coverage Ratio.

The questionnaire survey found that sovereign and institutional risks, counterparty risks

and 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 full

due diligence proceeds (bringing more appraisers, e.g. transport planners, into the

process).

The capital-intensive nature of infrastructure projects and in particular their dependence

on heavy up-front investment means that many standard financial ratios, e.g. Return on

Capital Employed, Gross Profit Margin are unlikely to be that reliable in early years of

a project. Faber (2002, p.69) notes that returns are likely to be volatile in capital-hungry

projects, especially in emerging economies and “emerging companies” (as start-up

tollways 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’s

investigation of risk is directly relevant to many aspects of the tollway industry and may

also add value to other infrastructure investment sectors.




6.3 Macro-Level Risks and Opportunities

Use of private finance in infrastructure increased since the 1980’s in both developed and

developing economies; toll roads being one of the recipients of such investment,

especially in developing countries. Although activity slowed in the aftermath of the

Asian Financial Crisis (at least partially attributable to previous over-investment),

recently it picked-up again and has begun spreading into some of the poorest countries

in 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 2.8

discussed 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 economic

growth rates, coupled with increased incidence of social and political upheaval.

Questionnaire survey respondents ranked the political system, legal system, ease of

profit 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 with

potentially explosive mobility growth. Based on Khan and Willumsen (1986)’s

equivalencing of roadspace and vehicle ownership, regression analyses (in Section 3.4)

identified potential for rapid demand growth; with the greatest ultimate growth potential

being 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 and

Vietnam as nascent-to-developing. Cambodia, Myanmar and Laos were rated not-yet-

nascent, signalling that whilst there may be significant percentage growth in vehicle




ownership, tolling affordability and absolute market size may be insufficient to provide

a viable near-term tollroad market; though those with respective local experience rated

Cambodia and Myanmar as nascent and Indonesia as developing-to-steady (Figure

4.M). This suggests Malaysian investments should be seen as the least risky

(commanding a smaller CAPM risk premium), with those in Cambodia, Myanmar and

Laos as the riskiest (requiring a higher forecast return to proceed).

Whilst interest rates do not feature greatly in transport planning literature, they are very

important in project finance: with investment risk increasing substantially as interest

rates escalate; this corroborated by the risk simulation modelling (see 5.6 in particular).

Price inflation can affect both construction and operating/ maintenance costs and the

impacts 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 were

seen 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). Economic

growth feeds through many aspects of market and forecasting risks (both discussed

below). Survey respondents expected increasing price inflation and interest rates (and

especially fuel prices), yet their importance was not rated that highly.

To test these expectations, risk simulation modelling was undertaken based on three

economic scenarios. The first (“Conventional Case”) assumed similar trends to those

experienced in recent years; the second (“Respondents’ Case”) incorporated

respondents’ expectations of higher fuel prices and slightly higher interest rates and

economic growth; the third case (“Kondratieff Case”) was based on an upswing in the

K-Wave (Kondratieff, 1926), resulting in markedly higher general price inflation and




interest rates, as well as higher economic growth. Based on FIRR, the Respondents’

Case generally gave the most optimistic returns, with mean Kondratieff Case returns

also higher than those from the Conventional Case. However, 12.5% of Kondratieff

Case runs resulted in project failure (i.e. no payback and/or negative FIRR), versus

0.6% in the Conventional Case and 1.1% in the Respondents’ Case.

Given that economic growth correlates positively with FIRR, the impacts of increased

price inflation and interest rates, where these outstrip economic growth would appear to

have a significant negative impact on project performance. Furthermore, the apparent

volatility of the Kondratieff Case would tend to support Faber’s (2002) assertion

regarding the riskiness of the K-Wave Upswing (higher mean returns, but with an

inherent danger of short-term reversals which can lead to bankruptcy), based on his

analysis of the 19th

Century American railroad industry.

However, it also appears that the K-Wave upswing is being facilitated by the economic

emergence 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 Asian

Financial Crisis, a further driver of infrastructure growth in the region could be posited.

6.4 Market Risks

Rigby and Penrose (2001) define project-level risks as the most critical. Though this

Dissertation concentrates on demand-side risks, construction cost and delay are

important, affecting early-year performance (Willumsen and Russell, 1998). There is

evidence of serial-underestimation of these costs (Flyvberg and COWI, 2004), which

are very significant to project performance: respondents ranked construction cost as

second only to contractual foundations (Table 4.5), with construction time ranked sixth

out of fifteen project-risk criteria. The risk simulation model showed construction cost




and time as fourth or fifth most significant risk category out of thirteen (dependent on

forecast 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 sixth

or seventh most important, the impact of toll increases is predicted to increase by 48%

between the Conventional and Respondents’ Cases and by 170% between Conventional

and Kondratieff Cases (see Table 5.5)24

. This risk is linked to contract enforceability

(institutional risk). Minimum income guarantees ranked only thirteenth overall, but

financial/ legal/ operators ranked them sixth; arguably because they are not of primary

concern 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 risk

simulation did not consider competing routes (beyond an existing local road) as such

impacts are very location-specific. It can be a contractual/ institutional issue, pertaining

to 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 additional

routes). Conversely, an absence of good connecting routes can subtract from project

performance. Questionnaire respondents asserted that they usually consider congestion

levels on both their study routes and competing and feeder routes (Figure 4.J).

24 Increases quantified by the percentage difference in “Impact*R

2” between cases in Table 5.5.




Whilst survey respondents did not rank toll affordability highly amongst market risks, it

likely 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 Risks

Start-up tollways do not have business track records for analysis. Every aspect of their

performance has to be predicted (cost and revenue). However, transport planners

generally over-forecast demand and revenue (Bain and Polakovic, 2005); and the more

uncertain the environment the poorer the forecasting record (Bain and Wilkins, 2002).

Questionnaire respondents broadly concurred with these assertions (Figure 4.I). They

also deemed the availability and reliability of data poorer in developing countries. Such

environments are typically more economically volatile, further compounding risk.

Those with experience of using or developing transport models did not hold any model

form as significantly inherently better or worse than others (Figure 4.H). Whilst it was

acknowledged that clients sometimes pressure transport planners to manipulate

forecasts (Figure 4.I; corroborating Brinkman, 2003), there was an ambivalent attitude

to whether equity-side forecasts should be higher than debt-side forecasts, suggesting

many forecasters do not understand the requirements of different financial perspectives.

Economic growth underpins demand forecasting parameters. In addition to uncertainty

regarding economic forecasts, both price sensitivities (e.g. Value of Time (VOT)) and

income elasticities of traffic growth and of VOT are rarely known. Even where

historical data are available, given S-curve relationships (Sections 2.7, 2.9 and 3.4) and

the impact of locational specifics on any given project, uncertainty is inherent over-and-

above that concerning economic growth per se.




Whilst Section 3.4 postulated cross-sectional income-car ownership relationships, in

themselves they are not sufficient to derive risk-free time-series sensitivities; Section

3.5 highlighted significant time-series differences within a single country (China). Gunn

and Sheldon (2001) advocate income elasticity of VOT in the range 0.35-0.70; despite

the implications of adopting 0.35 versus 0.70, there is not strong consensus as to what

values 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 (Table

5.5), with the demand-level (itself driven by GDP) ranking first or second most

significant on financial outcome of the case study.

For new roads, induced traffic may result and this may be substantial, boosting local

economic growth (Corbett et al, 2006). However, forecasting induced traffic is beset

with 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, it

may 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 market

risk by all questionnaire respondent groups (Table 4.5).

Due to practical limitations of scope, this dissertation did not investigate the direct

impacts of interest rates on travel demand (compound errors through the economic

linkages determining disposable income for car purchase, discretionary travel, etc

prohibited such analyses). However, should interest rates rise substantially it is

reasonable to postulate reduced car purchases (often loan-financed) and travel (as an

increased portion of income is used to service existing debts).




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 when

using the same frameworks and assumptions, there can be substantial variance in

forecasts. Conversely, despite the centrality of economics to demand forecasting, there

is 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 to

an extent, followed by general price inflation, economic growth, interest rates and

exchange rate volatility. Thus it might be argued that this is evidence of acceptance of

principles 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 direction

of the K-Wave (before adaptive expectations have completely caught-up with the

qualitative change of the K-Wave). However, given that the K-Wave is far from

common acceptance this perhaps states the case too strongly. Nevertheless, it does

suggest that some economic change is anticipated; and there was relatively little

disagreement between different stakeholders (Figure 4.P).

Figure 4.E showed that whilst economic growth was seen as third-most-important

macro-level risk, price inflation and interest rates were ranked only seventh and eighth

respectively (out of 11), though they still rated as “important.” Although the risk

simulation case study is simplistic based on a shift from the “Conventional Case” to the




“Respondents’ Case” (based on survey responses), it suggests that interest rates will

become approximately 80% more important, with GDP growth and price inflation

becoming approximately 15% more significant (using “Impact*R2” in Table 5.5

26).

However, based on the Kondratieff Case, interest rates become substantially more

important still and price inflation more than doubles in importance, amid more

widespread forecast volatility.

6.7 What Lessons for Practitioners?

Although practitioners appear to accept the likelihood of increased price inflation and

interest rates, comparing the outcomes of risk simulation modelling between

Conventional, Respondents’ and Kondratieff Cases, it appears that optimism-bias

persists. This despite a track-record of demand overforecasting.

Uncertainty is inherent in forecasting, particularly for start-up facilities and in

developing economies. Given this, reliance on base/ central case forecasts can be

misleading. Whilst full-blown Monte Carlo testing of traditional assignment models (let

alone four-stage models) may not be practical, it is advisable that risk simulation testing

be undertaken on traffic and revenue forecasts; as well as to cost forecasts. This might

be achieved through use of simplified forecasting models in spreadsheets, with key

values and sensitivities derived from orthodox traffic assignment models. If interest

rates and price inflation escalate, the impacts of variance in these variables will

25 “Peak oil” theorists hold that global oil production has either already or will shortly start to decline as

reserves 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-bound

of 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.




permeate many aspects of forecasting models (further compounded should escalating

economic growth in line with the K-Wave also be assumed).

Inevitably this will require additional resources. However, the costs of project

evaluation pale in comparison with infrastructure costs. Those commissioning transport

planners should also resist pressuring their consultants from boosting forecasts; the

environment is likely to get riskier, so distorted forecasts will result in an increased rate

of project financial failure.

Furthermore, it appears that many fail to appreciate the difference between equity- and

debt-side perspectives. Simply put, equity-side perspectives seek the mean value of a

project; whilst debt-side perspectives concentrate on all downside-risk elements. With

increasing interest rates, the implications of mis-structuring finance will escalate;

downstream re-financing tending to get more expensive (versus experience in the

1980’s and 1990’s when interest rates generally decreased). This also suggests that

fixed-rate debt should be arranged wherever possible (e.g. bonds) and that prospective

operators ought to err on the side of taking-on extra up-front debt at lower rates, rather

than risking downstream re-financing at a premium to initial rates. (Though not taking

so much debt as to incur excessive debt servicing requirements.)

Finally, it is suggested that further research is undertaken into the economic linkages

underlying many transport models, with particular emphasis on developing countries.




6.8 Conclusions: Evaluation of Hypothesis

Developing countries face more uncertainty than developed countries. Their increased

economic growth potential offset against economic volatility, corruption, sovereign and

institutional issues and in particular contract enforceability. Yet East Asian economies

are increasingly important to global trade, as manufacturing centres and commodity

producers. Historical experience would support the view that such trends would drive

demand for transport infrastructure, including tollways. Given funding constraints,

private participation is likely to remain important. However, performance is likely to be

volatile; this based on historical experience (e.g. 19th

Century American railroads) and

the results of risk simulation modelling, with most forecast parameters exerting

increased 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 of

project finance risks for private stakeholders in East Asian toll roads during a period of

increasing price inflation and interest rates.” Practitioners generally held that both

price inflation and interest rates would increase to an extent, though were less certain as




to the significance of such increases. The risk simulations showed both were strongly

correlated 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 as

increasing the impact of delayed toll increases (and increases beneath price inflation

rates). Risk simulation showed that rising price inflation and especially interest rates are

likely 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 (rising

price inflation decreasing the real debt burden) and subsequent increases in interest rates

are less important (so long as re-financing is not necessary).

Furthermore, increasing price inflation and interest rates could be associated with

accelerating economic growth; though the K-Wave posits this, acceptance of the K-

Wave is not necessarily required to accept the linkage between economic growth, price

inflation and interest rates. And economic growth is strongly positively correlated with

project performance, permeating most aspects of demand forecasting and potentially

mitigating some of the impacts of rising interest rates.

Indeed if one accepts the K-Wave upswing scenario, then notwithstanding likely

periodic reversals, economic prospects for East Asia are likely good. Furthermore, there

is a window of opportunity to set-up projects to take advantage of this growth before

price inflation and interest rates escalate markedly.

In conclusion, rising price inflation and interest rates do appear likely to change the

nature and extent of project finance risks for private stakeholders in East Asian toll

roads.




REFERENCES: LITERATURE

The following are referenced either in the Main Text (above) or in the Appendices:

Anhui Expressway Ltd (1996) Prospectus for Placing and New Issue, Crosby Capital

Markets (Asia) Limited and CEF Capital Limited, Hong Kong, 31st October 1996

Asian Development Bank (ADB) (2003), Transit Fee and Tolling for Routes 3 and 9,

Lao PDR, TA-3348 Final Report, Manila, April 2003

Asian Development Bank (ADB), Japan Bank for International Cooperation and

Development (JBIC), and World Bank (WB), (2005) Connecting East Asia: A New

Framework for Infrastructure, Advance Edition, Manila, Washington, D.C. and Tokyo,

May 2005

Azfar, O., Gurgur, T., Kähkönen, S., Lanyi, A. and Meagher, P. (2000)

“Decentralization and Governance: An Empirical Investigation of Public Service

Delivery in the Philippines” IRIS Center, University of Maryland, College Park and The

World 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 August

2005

Bain, R. and Wilkins, M. (2002) “Credit Implications of Traffic Risk in Start-Up Toll

Facilities”, Infrastructure Finance, Standard & Poor’s, London, September 2002

Beaverstock, J.V. and Doel, M.A. (2001) “Unfolding the Spatial Architecture of the

East Asian Financial Crisis: The Organizational Response of Global Investment Banks”,

Geoforum, 32(1), pp.15-32




Brinkman, P.A. (2003) The Ethical Challenges and Professional Responses of Travel

Demand Forecasters, PhD Dissertation, University of California at Berkeley

Buchanan, C.P. (1999) “The Role of BOT in the Highway Sector – Experience from

Malaysia and Elsewhere”, PIARC 21st World Roads Congress, Kuala Lumpur, October

1999

Central Intelligence Agency (CIA) (2005/2006) The World Factbook,

www.odci.gov/cia/publications/factbook/index.html

Chan, C. (2003) “New World Divests in Mainland Toll Roads and Bridges”, South

China Morning Post, Hong Kong, 17 November 2003

Cheong, Y.M. (1999) “The Political Structures of Independent States”, in Tarling, N.

(Ed.) The Cambridge History of Southeast Asia, Volume Two Part Two (From World

War II to the Present), Cambridge University Press, Cambridge, pp 59-138

Christensen, P. and Mertner, J. (1994), Assessment of Modal Competitiveness and

Traffic Potential of a Rehabilitated Railway in Cambodia, Final Report, COWI Consult,

Copenhagen and Asian Development Bank, Manila

Corbett, V.F. and Di Bona, R.F. (2006) “Discussion of Future Development,

Management and Use of the GMS Transport Model”, Workshop on ADB TA

6195-REG: GMS Transport Sector Strategy Study, Vientiane, Laos, 21 March 2006

Corbett, 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 2006

http://www.odci.gov/cia/publications/factbook/index.html




Di Bona, R.F. (2002) Surviving Bahtulism: A Financial and Economic Appraisal of

Scott Wilson Kirkpatrick Thailand Ltd, Managing Financial Resources assignment,

Henley Management College.

Di Bona, R.F. (2005) “GMS Transport Model”, Workshop on Draft Final Report of

ADB TA 6195-REG: GMS Transport Sector Strategy Study, Ho Chi Minh City,

Vietnam, 8-9 December 2005

Dizon, N. (2002) “Laguna Folk Sue Skyway Managers for Breach of Contract”, Sun

Star, 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, Hong

Kong, 2002

Faber, M. (2003) “The Financial Implications of Reflation”, The Gloom, Boom & Doom

Report, Marc Faber Limited, Hong Kong, June 23, 2003

Faber, M. (2005) “Why are Investment Markets Stalling?”, Market Commentary,

http://www.gloomboomdoom.com/marketcoms/mcdownloads/060509.pdf, 2 September

2005

Faber, M. (2006) “Poor Mr. Bernanke”, Market Commentary, 9 May 2006,

http://www.gloomboomdoom.com/marketcoms/mcdownloads/060509.pdf

Flyvbjerg, B. and COWI (2004) Procedures for Dealing with Optimism Bias in

Transport Planning – Guidance Document, The British Department of Transport,

London, and COWI, Copenhagen, 10 June 2004

http://www.newsflash.org/2002/01/ht/ht002241.htm






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’s

Infrastructure Finance, October 1999, pp149-157

George, C., Trommer, S. and McDermott, M. (2004) “The Continuing Search for Bliss:

Flexible Toll Road Structures”, Project Finance Special Report, Fitch Ratings, New

York, 20 October 2004

Gini, C. (1912) “Variabilità e Mutabilità”, Studi Economico-Giuridici, Anno III, Parte

2, Facoltà di Giurisprudenza, Regia Università di Cagliari, 1912

Gomez, E.T. and Jomo K.S. (1999) Malaysia’s Political Economy: Politics, Patronage

and Profits, 2nd

Edition, Cambridge University Press.

Guangdong Provincial Bureau of Statistics (1998), Guangdong Statistical Yearbook

1998, China Statistics Press, Beijing, 1 August 1998







Guislain, P. and Kerf, M. (1995) “Concessions – The Way to Privatize Infrastructure

Sector Monopolies”, Public Policy for the Private Sector, Note No. 59, The World

Bank, Washington, D.C.




Gunn, H. and Sheldon, R. (2001) “The Value of Time”, The Transport Economist,

Volume 28 Number 1, Spring 2001

Hayek, F.A., (1933) Monetary Theory and the Trade Cycle, London, 1933

Higson, C. (1995) Business Finance, Second Edition, Butterworths, London.

Hopewell Highway Infrastructure Limited (HHI) (2003) Global Offering Prospectus,

Hong Kong, 28 July 2003

Hymans, S.H. (1970) “Consumer Durable Spending: Explanation and Prediction”,

Brookings Papers on Economic Activity, no. 2, pp. 173-199, 1970

International Project Finance Association (IPFA) (2006), “About Project Finance”,

http://www.ipfa.org/about_pf.shtml, sighted in June 2006

Irwin, C. (1999) “Decentralizing Transportation Infrastructure”, in Standard & Poor’s

Infrastructure Finance: Project Finance, Utilities and Concessions, October 1999,

pp.146-148

Jiangsu Expressway Co. Ltd. (1997), Prospectus for Placing and New Issue, HSBC

Investment Bank Asia, Hong Kong, 18 June 1997

Jiangsu Expressway Co. Ltd. (2006) 2005 Annual Report, Nanjing

Jiangsu Provincial Statistics Bureau (1999), Jiangsu Statistical Yearbook 1999, China

Statistics Press, Beijing, 1 July 1999



http://www.ipfa.org/about_pf.shtml







Statistics Press, Beijing, 1 June 2004

Juglar, C. (1863) Des Crises Commerciales et leur Retour Périodique en France, en

Angleterre, et aux Etats-Units, Imprimerie de Veuve Berger-Levrault, Strasbourg, 1863

Khan, A. and Willumsen, L.G. (1986) “Modelling Car Ownership and Use in

Developing Countries”, Traffic Engineering and Control, 27(11), pp554-60.

Kilsby, D. (2004) “Ethical Challenges and Professional Responses of Travel Demand

Forecasters – 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 Alternative

Transport Fuels, National Committee on Transport, Engineers Australia, downloaded

from www.kilsby.com.au (in October 2006)

Kilsby, D. (2006b) “Peak Oil – It’s Coming (Fast)”, Article in Cityscape, July and

August 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 1996

Kitchin, J. (1923) “Cycles and Trends in Economic Factors”, Review of Economic

Statistics, Volume 5, pp. 10-16

http://www.kilsby.com.au/





Klein, M., So, J. and Shin, B. (1996) “Transaction Costs in Private Infrastructure

Projects – Are They Too High”, Public Policy for the Private Sector, Note No. 95, The

World Bank, Washington, D.C., October 1996

Kondratieff, N.D. (1926) “Die Langen Wellen der Konjunktur”, Archiv für

Sozialwissenschaft und Sozialpolitik, 1926, Volume 56, No. 3, pp. 573-609; translated

by W. F. Stolper and reprinted as “Long Waves of Economic Life”, The Review of

Economics and Statistics, Volume 17, Number 6, November 1935, pp. 20-42

Krugman, P. (2000) The Return of Depression Economics, Paperback Edition, Penguin,

London

Kuznets, S. (1930) Secular Movements in Production and Prices: Their Nature and

their 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 1983

La Porta, R., López-de-Silanes, F., Shleifer, A., and Vishny, R. (1997) “Which

Countries Give Investors the Best Protection?”, Public Policy for the Private Sector,

Note No. 109, The World Bank, Washington, D.C., April 1997

Lawrence, A. (1999) “The Curse Bites: Skyscraper Index Strikes”, Property Report,

Dresdner Kleinwort Benson Research, 3 March 1999




Luu, H.S. (2006) “Financing Sources for Transport Development in Vietnam”, National

Graduate Institute for Policy Studies, Tokyo, 26th

August 2006,

www.grips.ac.jp/vietnam/VDFTokyo/Doc/25LHSon26Aug06Slides.pdf

Maddison, A. (1995) Monitoring the World Economy, 1820-1992, Washington, D.C.:

Organization for Economic Cooperation and Development

Manasan, R.G. (2004) “Infrastructure and Decentralization”, Background Paper for

Philippines AAA, World Bank, Washington D.C.

Markowitz, H.M. (1952), “Portfolio Selection”, The Journal of Finance, Volume VII,

No. 1, March 1952, pp77-91

Mendez, C. (2004) “DOTC to Block Tollway Rate Hike”, STAR, Manila, July 7, 2004,

http://www.newsflash.org/2004/02/hl/hl100616.htm

Montlake, S. (2005) “Cambodia’s Killing Fields Get Privatized”, Christian Science

Monitor, 3 May 2005, http://www.csmonitor.com/2005/0503/p06s01-woap.html

Morris, D. (1971) The Human Zoo, London: Corgi

National and Economic Social Development Board (NESDB) (2006) Gross Domestic

Product: 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 2004

Ormerod, 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.

http://www.grips.ac.jp/vietnam/VDFTokyo/Doc/25LHSon26Aug06Slides.pdf

http://www.newsflash.org/2004/02/hl/hl100616.htm

http://www.csmonitor.com/2005/0503/p06s01-woap.html





Orosz, C. (1998) “The Effects of Tolled Motorways in Hungary”, European Transport

Conference, Loughborough, 14-18 September 1998

Ortúzar, J.d.D. and Willumsen, L.G. (1994) Modelling Transport, 2nd

Edition, John

Wiley & Sons.

Parsons Brinckerhoff (Asia) Ltd (PBA) (2003) “Traffic Consultant’s Report”, in

Hopewell Highway Infrastructure Limited (2003) Global Offering Prospectus

Pigou, A.C. (1920) The Economics of Welfare, London.

Pindyck, R.S. and Rubinfeld, D.L. (1981) Econometric Models and Economic

Forecasts, 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 Preference

Surveys”, presented at 26th

Conference of Australian Institutes of Transport Research

(CAITR), Melbourne, December 2004

Rigby, P.N. (1999) “Project Finance: Technical Risk Criteria”, Standard & Poor’s

Infrastructure Finance, October 1999, pp 40-45

Rigby, P. and Penrose, J. (2001) “Project Finance Summary Debt Rating Criteria”, in

Standard & Poor’s Project & Infrastructure Finance Review: Criteria and Commentary,

October 2001, pp. 19-32

Rogers, E. (1962) Diffusion of Innovations, 1962

Ruster, J. (1996) “Mitigating Commercial Risks in Project Finance”, Public Policy for

the Private Sector, Note No. 69, The World Bank, Washington, D.C., February 1996




Schumpeter, J.A. (1939) Business Cycles: A Theoretical, Historical and Statistical

Analysis of the Capitalist Process, Philadelphia, 1939

Schumpeter, J.A. (1950) Capitalism, Socialism and Democracy, 3rd

Edition, New York:

Harper & Brothers, 1950

Scott 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, pp167-182

Scott 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, pp183-190

Scott Wilson Kirkpatrick (Hong Kong) Ltd (SWHK) (1996a) “Letter from the Traffic

Forecast Consultant”, in Anhui Expressway Ltd (1996) Prospectus for Placing and New

Issue, pp105-111

Scott Wilson Kirkpatrick (Hong Kong) Ltd (SWHK) (1996b) “Letter from the

Operation Review Consultant”, in Anhui Expressway Ltd (1996) Prospectus for

Placing and New Issue, pp112-115

Shenzhen Expressway (2006), “Toll Roads & Bridges >> Toll Standards”, webpage

found October 2006 at www.sz-expressway.com

Standing Advisory Committee on Trunk Roads Assessment (SACTRA) (1994), Trunk

Roads and the Generation of Traffic, Her Majesty’s Stationery Office, London

Streeter, W., Hermans, M., McCarthy, B., Stephenson, K., Monnier, L., Kolotas, P. and

Osako, M. (2004) “Toll Road Securitizations: Wherein the Future Flows”, International

http://www.sz-expressway.com/




Structured Finance/ Project Finance Special Report, Fitch Ratings, New York, 20

January 2004.

Streeter, W. and McManus, K. (1999) “Challenges of Start-Up Toll Roads”, Project

Finance Special Report, Fitch ICBA, Duff & Phelps, New York.

Thornton, M. (2003) “Skyscrapers and Business Cycles”, Working Paper, Von Mises

Institute, http://www.mises.org/workingpapers.asp, 29 May 2003

Transparency 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 Estimated

From Traffic Counts”, Transportation Research, 14B(3), pp. 281-293

Wardman, M. (1998) “Review of Service Quality Valuations”, European Transport


Willumsen, L. and Russell, C. (1998) “Reducing Revenue Risk”, European Transport


Wong, M. and Moy, P. (2004) “Higher Tolls Prove a Drag on Western Tunnel”, South

China Morning Post, Hong Kong, 2 November 2004

World Bank (WB) (2003a), Private Participation in Infrastructure: Trends in

Developing Countries 1990-2001, Washington D.C.

http://www.mises.org/workingpapers.asp

http://www.unescap.org/stat/data/statind/pdf/index.asp#VolXXXV




World Bank (WB), (2003b) “Combating Corruption in Indonesia”, World Bank Poverty

Reduction and Economic Management Unit Report 27246, Washington D.C., 2003

World Bank (WB) (2004a), Civil Society Perceptions. Background for the Infrastructure

Flagship http://www.worldbank.org/eapinfrastructure

World Bank (WB), (2004b) Averting an Infrastructure Crisis – Indonesia. A

Framework for Policy and Action, Washington D.C., 2004

World Bank (WB), (2005) China Quarterly Update, Washington D.C., February 2005

Yepes, 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 The

World Bank (WB), Washington, D.C.

Yuen, J. (2005) China on the March – Again: A Business Culture Perspective, MBA

Dissertation, Henley Management College, Henley.

Zhejiang Expressway Co. Ltd. (2006) 2005 Annual Report, Hangzhou

Zhejiang Provincial Bureau of Statistics (1999), Zhejiang Statistical Yearbook 1999,

China Statistics Press, Beijing, 1 July 1999


China Statistics Press, Beijing, 1 August 2000





http://www.worldbank.org/eapinfrastructure




REFERENCES: INTERNET RESOURCES

The 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.org

Asian Development Bank (for reports and economic statistics): www.adb.org

Bank of Thailand (for economic statistics): www.bot.or.th

Central Intelligence Agency (CIA) World Factbook (for data on various countries):

www.odci.gov/cia/publications/factbook/index.html

Christian Science Monitor (for news article on Cambodia): www.csmonitor.com

Foreign and Commonwealth Office, UK (for data on various countries): www.fco.go.uk

GoogleEarthTM

(for map in Figure 1.A): earth.google.com

Henley Management College (for survey dissemination in addition to structural

guidance on the Dissertation): www.henleymc.ac.uk

Hopewell Highway Infrastructure Limited (for expressway traffic and revenue data):

www.hopewellhighway.com

International Project Finance Association (IPFA) (for background information on

history of project finance): www.ipfa.org

Jiangsu Expressway Co. Ltd. (for expressway traffic data): www.jsexpressway.com

(formerly www.jsexpressway.com.cn)

Kilsby Australia (for articles): www.kilsby.com.au

Legal500.com (for identifying suitable legal professionals for the survey):

www.legal500.com

Dr. Marc Faber/ Gloom Boom Doom Report website (for reports and market

commentaries): www.gloomboomdoom.com (note: in mid-2006 much of this content

became subscriber-only)

http://www.apec.org/

http://www.adb.org/

http://www.bot.or.th/


http://www.csmonitor.com/

http://www.fco.go.uk/

http://www.earth.google.com/

http://www.henleymc.ac.uk/

http://www.hopewellhighway.com/

http://www.ipfa.org/

http://www.jsexpressway.com/

http://www.jsexpressway.com.cn/


http://www.legal500.com/

http://www.gloomboomdoom.com/




National Economic and Social Development Board, Thailand (for economic statistics):

www.nesdb.go.th

National Graduate Institute for Policy Studies, Tokyo, Japan (for papers):

www.grips.ac.jp

Newsflash.org (for a number of news articles on the Philippines): www.newsflash.org

Pacific Exchange Rate Service (for historical exchange rate information):

www.fx.sauder.ubc.ca

Shenzhen Expressway Co. Ltd. (for toll rates): www.sz-expressway.com

Survey MonkeyTM

(used for conducting the questionnaire survey):

www.surveymonkey.com

The Urban Transport Institute (for articles): www.tuti.com.au

United Nations Economic and Social Commission for Asia and the Pacific (for

economic data and reports): www.unescap.org

Von Mises Institute (for working papers and reports): www.mises.org

World Bank (WB) (for economic data and reports): www.worldbank.org

Wren Investment Advisers (for historical gold prices and interest rates):

www.wrenresearch.com.au/downloads/index.htm

Yahoo! Newsgroups (for survey dissemination):

EMME/2 users’ group: http://groups.yahoo.com/group/emme2users

TransCAD users’ group: http://tech.groups.yahoo.com/group/transcad

Zhejiang Expressway Co. Ltd. (for expressway traffic and revenue data):

www.zjec.com.cn


http://www.grips.ac.jp/

http://www.newsflash.org/


http://www.sz-expressway.com/


http://www.tuti.com.au/

http://www.unescap.org/

http://www.mises.org/

http://www.worldbank.org/

http://www.wrenresearch.com.au/downloads/index.htm

http://groups.yahoo.com/group/emme2users

http://tech.groups.yahoo.com/group/transcad

http://www.zjec.com.cn/




APPENDICES

Appendix 1: Headline Demographic and Economic Statistics ..................................... 120

Appendix 2: Headline Transport Statistics ................................................................... 122

Appendix 3: Examples of Listed Provincial Chinese Expressway Operators .............. 123

Appendix 4: Typical Financial Ratios........................................................................... 124

Appendix 5: Kondratieff Waves since 1787 ................................................................. 125

Appendix 6: Definition of Guanxi ................................................................................ 126

Appendix 7: Typical Structure of Four-Stage Transport Models ................................. 127

Appendix 8: Traffic Risk Index .................................................................................... 128

Appendix 9: Measures of Corruption and its Impact .................................................... 130

Appendix 10: PESTLE Analysis ................................................................................... 131

Appendix 11: Correlation between Wealth and Transport Networks ........................... 137

Appendix 12: Expressway and Economic Index Calculations ..................................... 148

Appendix 13: Survey Questionnaire: Question Specification and Logical Flow ......... 163

Appendix 14: Amendments Made to Questionnaire Following Pilot Survey .............. 171

Appendix 15: Questionnaire Responses........................................................................ 175

Appendix 16: Risk Simulation Modelling: Simulation Parameters Employed ............ 199

Appendix 17: Risk Simulation Modelling: Fixed Parameters ...................................... 203

Appendix 18: Risk Simulation Modelling: Equations Employed ................................. 204

Appendix 19: Risk Simulation Modelling: Results by Parameter ................................ 206

Appendix 20: Risk Simulation Modelling: Comparison of Parameters’ Impacts ........ 245




Appendix 1: Headline Demographic and Economic Statistics

Data 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 values

given. In most cases, data are from 2005, although in some cases older data were quoted. The

data collection method is not known, but could be expected to vary from country-to-country

based 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 other

countries for comparison and benchmarking purposes.

Country

Land

Area

(km2)

Population

(Estimated

July 2006)

Population

(Annual

Change)

% Below

Poverty

Line

Median

Age

(years)

Life

Expectancy

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

Country

GDP (billion US$) Real GDP

Growth

Rate

Proportion of GDP by Sector Gross Fixed

Investment

as % of GDP PPP

Official

Exchange Rate

Agri-

culture Industry Services

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%





Country

Public Debt as

% of GDP/

External Debt

(US$ billion)27

Annual

Consumer

Price

Inflation

GDP per

capita

(PPP

method)

Labour

Force

(millions)

Unemploy-

ment (%)

Gini Index

on Family

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 Debt

in 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 Cambodia

was 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 and

underemployment in rural areas; an official Chinese journal estimated overall unemployment (including

rural 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 taken

from CIA Factbook.)




Appendix 2: Headline Transport Statistics

Airports

Railway

(km)

Roadway (km)

Waterway

(km) Notes Total

With Paved

Runway Total Paved

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 Factbook

Notes: (1) Estimate of length of Cambodia's 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 Thailand's waterways are restricted to vessels with draft up to 0.9m.

(5) 5,000km of Vietnam's 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 Korea's waterways mostly navigable only by small craft.

(7) Only 620km of UK's waterways used for commerce.

(8) Only 19,312km of USA's 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.




Appendix 3: Examples of Listed Provincial Chinese Expressway Operators

Company Year Bourse

Anhui Expressway 1996 Hong Kong

Guangdong Provincial Expressway 1996 Shenzhen

Jiangsu Expressway 1997

1999

Hong Kong

Shanghai

Shandong Infrastructure 2000 Shanghai

Sichuan Expressway 1997 Hong Kong

Zhejiang Expressway 1997

2000

Hong Kong

London




Appendix 4: Typical Financial Ratios

PROFITABILITY RATIOS

Return on Capital Employed

(ROCE)

Profit(Loss) Before Interest and Tax

Total Assets less Current Liabilities

Gross Profit Margin (GPM)

Gross Profit

Turnover

Profit on Sales (POS)


Turnover

Expenses as Percentage of

Turnover (EPT)

Expenses

Turnover

Sales to Capital Employed (SCE)

. Turnover .


Sales to Fixed Assets (SFA)

. Turnover .

Fixed Assets

Sales to Working Capital

. Turnover .

Net Current Assets

LIQUIDITY/ WORKING CAPITAL MANAGEMENT RATIOS

Working Capital Requirement

(WCR)

Current Assets less Current Liabilities

Current Ratio

. Current Assets .

Current Liabilities

Asset Turnover

. Turnover .


Interest Cover/ Debt Service

Coverage Ratio


Interest Payable




Appendix 5: Kondratieff Waves since 1787

Source: Faber, M. (2002, p.120)

P

anic

of

18

19

Pan

ic o

f 1

83

7P

anic

of

18

73

Cra

sh 1

92

9C

rash

19

73

Cra

sh 1

98

7

18

15

18

66

19

21

19

76

20

33

17

87

18

42

18

96

19

49

20

04

20

58

Dep

ress

ion

Har

d T

ime

Dep

ress

ion

Dep

ress

ion

Gre

at D

epre

ssio

n

17

87

- 1

84

21

84

2 -

18

96

18

96

- 1

94

91

94

9 -

20

04

20

04

- 2

05

8

Dis

pla

cem

ent

Can

als

Pro

gre

ss i

n:

Ele

ctro

nic

s

Ro

ads

Ele

ctri

city

Aer

osp

ace

Bri

dges

Co

nsu

mer

ism

Fif

th K

on

dra

tief

f

Go

ld D

isco

ver

ies

in C

alif

orn

ia a

nd

Au

stra

lia

Co

mm

un

icat

ion

, ch

emic

al a

nd

au

to

ind

ust

ry

Ser

vic

es i

ncl

ud

ing h

ealt

h c

are,

lei

sure

,

etc

Fir

st K

on

dra

tief

fS

eco

nd

Ko

nd

rati

eff

Th

ird

Ko

nd

rati

eff

Fo

urt

h K

on

dra

tief

f

Up

swin

g:

fro

m 1

99

5-2

00

4 t

o p

erio

d

20

25

-20

35

Do

wn

swin

g:

fro

m 2

02

5-2

03

5 t

o p

erio

d

20

55

-20

65

Op

enin

g o

f n

ew m

ark

ets,

Ch

ina,

Eas

tern

Eu

rop

e, R

uss

ia

Tel

eco

mm

un

icat

ion

s

Info

rmat

ion

tec

hn

olo

gy,

etc

Up

swin

g:

fro

m e

arly

18

90

s to

per

iod

19

14

-19

20

Do

wn

swin

g:

fro

m b

egin

nin

g o

f 1

91

4-

19

20

s

Up

swin

g:

fro

m 1

94

0s

to 1

97

0s

Do

wn

swin

g:

fro

m l

ate

19

70

s to

ear

ly

20

00

s

Up

swin

g:

fro

m l

ate

18

40

s to

ear

ly

18

70

s

Do

wn

swin

g:

fro

m e

arly

18

70

s to

ear

ly

18

90

s

Rai

lro

adis

atio

n o

f A

mer

ica

Up

swin

g:

fro

m l

ate

17

80

s to

per

iod

18

10

-18

17

Do

wn

swin

g:

fro

m p

erio

d 1

81

0-1

81

7 t

o

late

18

40

s

Ap

pli

cati

on

of

new

in

ven

tio

ns

to

man

ufa

ctu

rin

g (

Ind

ust

rial

Rev

olu

tio

n)

En

try o

f A

mer

ica

into

wo

rld

mar

ket

s




Appendix 6: Definition of Guanxi

Strictly speaking, guanxi applies to China and to relationships amongst Chinese, yet

there are parallels in other Asian cultures. The following definition of guanxi is

reproduced with permission from Yuen (2005, pp.75-76):

There is no direct English translation of the term guanxi, which would fully convey its

meaning of connections or relationships defined by reciprocity and mutual obligation

and underpinned by a sense of goodwill and personal affection. Guanxi is based on

mutual trust and shared experiences. Guanxi is a manifestation of China’s Confucian

heritage. Its origins can be traced back to ancient Chinese social customs, in which

reciprocity and mutual obligation were used to build and maintain interpersonal

relationships throughout society.

Guanxi exists in various forms. These differ depending on the closeness of the

relationship between the parties involved. Chinese see relationships as existing on one

of 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 is

used to express the reciprocation of outstanding favours, which accrue through guanxi.




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

(by Trip Purpose and/or Vehicle Type and/or Time of Day)

Second Stage: Trip Distribution

Linkage between Zones (e.g. by Purpose/ Vehicle Type)

Output: Trip Patterns Zone-to-Zone

(by Trip Purpose and/or Vehicle Type and/or Time of Day)

Third Stage: Mode Split

Proportions of Trips by Transport Mode

Output: Zone-to-Zone Trips by Mode

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, etc

Zo

ne-t

o-Z

on

e

co

sts

by

mo

de

Zo

ne-t

o-Z

on

e

co

sts

Co

sts

to/f

ro

m

ea

ch

zo

ne




Appendix 8: Traffic Risk Index

This is the Traffic Risk Index to be applied to traffic forecasts, postulated by Bain and

Wilkins (2002). Its purpose is to provide guidance as to the likely reliability of traffic

forecasts.

Project Attributes Low High

Tolling Regime Shadow tolls User-paid tolls

Tolling Culture Toll roads well established; data

on actual use available

No toll roads in country;

uncertainty over acceptance

Tariff Escalation Flexible rate setting/ escalation

formula; no government approval

required

All tariff hikes require regulatory

approval

Forecast Horizon Near-term forecasts required Long-term (30+ year) forecasts

required

Toll-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

(including tolling points and

intersections)

Confused/ unclear road

objectives (not where people

want to go)

Alignment: strong economics Alignment: strong politics

Clear understanding of future

highway network

Many options for 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

(e.g. traffic calming, truck bans)

Autonomous authorities can do

what they want

Surveys/ data

collection

Easy to collect (laws exist) Difficult/ dangerous to collect

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

used)

Develop zone framework from

scratch

Users: private Clear market segment(s) Unclear market segments

Few, key origins and destinations Multiple origins and destinations

Dominated by single journey

purpose (e.g. commute, airport)

Multiple journey purposes

High income, time-sensitive

market

Average/ low income market




Project Attributes Low High

Tolls in line with existing

facilities

Tolls higher than the norm –

extended ramp-up?

Simple toll structure Complex toll structure (local

discounts, frequent users,

variable pricing, etc)

Flat demand profile (time-of-day,

day-of-week, etc)

Highly seasonal and/ or “peaky”

demand profile

Users: commercial Fleet operator pays toll Owner-driver pays toll

Clear time and operating cost

savings

Unclear competitive advantage

Simple route choice decision-

making

Complicated route choice

decision-making

Strong compliance with weight

restrictions

Overloading of trucks is

commonplace

Micro-economics Strong, stable, diversified local

economy

Weak/ transitional local/ national

economy

Strict land-use planning regime Weak planning controls/

enforcement

Stable, predictable population

growth

Population forecast dependent on

many, exogenous factors

Traffic growth Driven by/ correlated with

existing, established and

predictable factors

Reliance upon future factors, new

developments, structural changes,

etc

High car ownership Low/ growing car ownership




Appendix 9: Measures of Corruption and its Impact

NGO perceptions of corruption and its impact on infrastructure development in China,

Indonesia, Japan, Philippines, Thailand and Vietnam are summarised below from WB

(2004a):

Agree /

Serious Obstacle

Disagree/

Not Serious Obstacle

Extent to which corruption is an obstacle 95% 5%

Extent to which potential for corruption

should be taken into account 91% 4%

Government does not do enough to prevent

corruption in infrastructure development 77% 23%

The 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”; this

would be seen as biased from a western perspective. It reinforces the need to “know the

system” 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 2.6




Appendix 10: PESTLE Analysis

POLITICAL Official Name Government

Cambodia Kingdom of Cambodia Democracy; Cambodian People’s Party in notional

coalition with FUNCINPEC

China People’s Republic of China One-Party State (Chinese Communist Party)

Indonesia Republic of Indonesia Republic with Parliamentary Elections (Golkar

currently ruling)

Laos Lao People’s Democratic

Republic One-Party State (Lao People’s Revolutionary Party)

Malaysia Malaysia

Rotating Monarchy with democracy: rule by Barisan

Nasional coalition; United Malays National

Organisation (UMNO) is main party therein

Myanmar Union of Myanmar Military Junta (State Peace and Development

Council)

Philippines Republic of the Philippines US-Style Presidential Republic (Lakas Party ruling)

Thailand Kingdom of Thailand Transitional military rule (pending restoration of

bicameral democracy), under a Monarchy

Vietnam Socialist Republic of

Vietnam One-Party State (Communist Party of Vietnam)

POLITICAL Stability31

Foreign Relations

Cambodia

Public order fragile; whilst

Prime Minister Hun Sen seen

as a “strong man”, much of

state apparatus relatively weak

Generally good and improving. PM close to

Vietnam. Major recipient of development aid.

China

Generally stable. But increased

labour and social unrest in

some areas.

Improving. Strong trade with most neighbours.

Whilst still receiving development aid, has

expanded its own aid donations in the region.

Indonesia Unrest in outlying areas;

ongoing terrorist threat.

Generally good, but seen by some as weak on

Muslim Militants.

Laos Sporadic rural banditry

Improving following chairing of ASEAN. Close

to Vietnam. Increasing cooperation with Thailand.

Major recipient of development aid.

Malaysia Generally stable Generally good.

Myanmar Unsettled; insurgencies in some

areas; bombings in capital

Economic sanctions by much of the West plus

political pressure within ASEAN constrain

economic development and receipt of aid.

Philippines

High crime level; threat of

bombings and kidnappings;

ongoing Presidential crisis

Government seen as bulwark against terrorism,

but reputation is hurt by ongoing allegations of

Presidential vote-rigging and cronyism.

Thailand Generally stable, except in

south (unrest and bombings)

Generally good. Seen as the political and trade

centre of Mainland S.E. Asia. Increasingly

involved in development aid to its neighbours.

Vietnam Generally stable

Improving. Vietnam still acts as an influence in

Laos and Cambodia. Still some tension with

China, but ties improving.

Summary

and

Comments

Stability concerns in many

countries, though not

necessarily deterring

infrastructure investment.

Most foreign relations improving, with possible

exception of Myanmar.

31 Source: www.fco.gov.uk, 23 August 2005, supplemented by some of the author’s own observations.

http://www.fco.gov.uk/




ECONOMIC

GDP (PPP method) Real GDP

Growth

Rate

Proportion of GDP by Sector Gross Fixed

Investment

as % of GDP

GDP

(US$ bn)

Per capita

(US$)

Agri-

culture Industry Services

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%

Summary

Excepting Myanmar, growing rapidly,

albeit typically from a relatively low

base.

Varied.

Typically

substantial

GFI as % of

GDP.

Source: CIA Factbook (http://www.odci.gov/cia/publications/factbook/index.html) in May 2006





SOCIAL Urbanisation

32

Propensity to Travel 1990 2004

Cambodia 11.6% 15.0% Inter-urban travel restricted by poor highway networks

and affordability.

China 26.4% 41.8% Increasing inter-urban travel. Some restriction in remote

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.

Myanmar 24.8% 30.0% Inter-urban travel often arduous and some areas

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.

Summary Increasing urbanisation.

Often quite dramatic.

Generally much/ growing inter-urban travel. Suppressed

in some cases by poor transport networks.

SOCIAL Attitudes to Foreign Private Sector Involvement in Infrastructure Provision

Cambodia

In general, keen to attract foreign investment, although local partners required in

many investments. However, controversy over privatisation of Choeung Ek

(Killing Fields) and associated toll-road33

.

China

Substantial involvement of private sector in toll road provision, especially by

overseas Chinese. Stock market listings and Bond Issues of State majority-owned

operators. Local connections often critical. Revenue guarantees largely abolished.

Indonesia Pre-Asian Financial Crisis there was substantial activity in toll-road financing.

Activity once again picking-up, but local connections often critical.

Laos Sector not yet developed. However, State Railways of Thailand extending their

network into Laos (Nong Khai – Friendship Bridge – Vientiane Municipality).

Malaysia Much private sector involvement. However, concessions go mainly to well-

connected locals who may then raise finance from overseas34

.

Myanmar

Keen to attract foreign investment largely curtailed by sanctions and shareholder

activism. Some roads have been financed by domestic BOT arrangements, but

concessions go to well-connected individuals, rather than FDI PPP.

Philippines

Whilst local partners are usually required, much infrastructure has been financed/

developed by international companies. However, there have been problems

enforcing toll increases, especially on foreign-invested projects35

.

Thailand Overseas investors long active in Thailand.

Vietnam Relatively few foreign private investments in Vietnamese toll-roads to date.

Vietnam is tipped by some to develop this sector quickly in coming years.

Summary

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.

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)




TECHNO-

LOGICAL Toll Collection Systems

Cambodia Manual collection systems

China Manual collection on minor routes. Increased usage of automatic systems on

major routes.

Indonesia Primarily manual

Laos Manual

Malaysia Computerised systems on major routes, but substantial manual collection.

Myanmar Manual

Philippines Primarily manual.

Thailand Primarily manual

Vietnam Manual

Summary Largely manual. Use of computerised and automated systems increasing,

typically on higher-volume routes in richer countries.

Development of Manufacturing and Primary Industries

Cambodia Agriculture predominates, with some basic export-oriented industries. However,

export price competitiveness restrained by efficiency of transport networks.

China

In coastal areas China is a world-leader in manufacturing. However, other parts

of China are yet to catch-up. Large producer and consumer of many

commodities.

Indonesia Whilst a largely agricultural society, there is also substantial and growing

manufacturing, mining and timber industries.

Laos Primarily agricultural. Mining has been hindered by being land-locked with

under-developed internal transport networks.

Malaysia Well developed manufacturing and primary industries.

Myanmar

Agriculture predominates. Sanctions and shareholder pressure limit development

of export-oriented manufacturing. Commodity exploitation increasing

(especially exports to China).

Philippines

Agriculture predominates through much of the country. However, some areas

(e.g. Subic, Metro Manila) have significant manufacturing industries. Roll-out

elsewhere often hindered by transport networks.

Thailand

Whilst the north (and especially north-east) are predominantly agricultural, there

is substantial manufacturing especially around Greater Bangkok/ Laem Chabang

areas.

Vietnam Agriculture predominates. However, manufacturing is rapidly increasing. Given

Vietnam’s geography distances to sea are typically short.

Summary Manufacturing is relocating globally to East Asia. Some countries are also very

important as commodity producers.




LEGAL Legal System36

Cambodia

primarily a civil law mixture of French-influenced codes from the United

Nations Transitional Authority in Cambodia (UNTAC) period, royal decrees,

and acts of the legislature, with influences of customary law and remnants of

communist legal theory; increasing influence of common law in recent years

China

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

January 1980; continuing efforts are being made to improve civil, administrative,

criminal, and commercial law

Indonesia

based on Roman-Dutch law, substantially modified by indigenous concepts and

by new criminal procedures and election codes; has not accepted compulsory ICJ

jurisdiction

Laos based on traditional customs, French legal norms and procedures, and socialist

practice

Malaysia

based on English common law; judicial review of legislative acts in the Supreme

Court at request of supreme head of the federation; has not accepted compulsory

ICJ jurisdiction

Myanmar has not accepted compulsory ICJ jurisdiction

Philippines based on Spanish and Anglo-American law; accepts compulsory ICJ jurisdiction,

with reservations

Thailand based on civil law system, with influences of common law; has not accepted

compulsory ICJ jurisdiction

Vietnam based on communist legal theory and French civil law system

Summary varied

LEGAL Level of Corruption/ Transparency Indices37

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 2.6

Summary

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

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 5

indicate widespread corruption; and, scores below 3 indicate rampant corruption. Comments under

“Summary and Comments” are the author’s own.




ENVIRON-

MENTAL Perceived Importance of Environmental Considerations

Cambodia

Economic development first priority. However, environmental considerations

increasingly considered with potential for eco-tourism. Moves towards forestry

management, but sometimes hampered by poor local control.

China

Increased attention being paid to environment. However, many Chinese cities

remain amongst the worst polluted in the world. Environmental efforts sometimes

undermined by local failings in rule of law.

Indonesia In some areas environmental considerations increasingly important. However,

highly variable across this very large archipelago.

Laos

Economic development first priority, with increasing attention to environmental

considerations, aided by development agencies’ involvement. However, e.g. Nam

Theun II dam and hydroelectric project remains contentious.

Malaysia Whilst economic development remains a priority, environmental protection is

increasingly important, both in major cities and in indigenous areas.

Myanmar Economic development first priority. Environmental considerations very much

secondary to development of individual projects.

Philippines

Some aspects of environment increasingly important, especially regarding tourism

projects. However, environmental efforts often undermined by local failings in rule

of law or ability to enforce.

Thailand Environment increasingly important, especially in tourism areas. Environmental

efforts sometimes undermined by local failings.

Vietnam Environment increasingly important, however standards not always applied equally

across the country.

Summary

Economic development predominates over environmental considerations, but the

environment is of increasing importance, possibly correlating with extent of

economic development.

Geographic Considerations

Cambodia Currently the most convenient land route from Bangkok to southern Vietnam, even

with road rehabilitation yet to be completed (still ongoing). Varied topography.

China Very large and diverse geography. Southern and eastern coastal regions more

developed. Western hinterlands more mountainous.

Indonesia A massive archipelago with differing customs and levels of economic development.

However, as part of the “Ring of Fire” much of the country is mountainous.

Laos

Dominated by Mekong River and in north and east by mountains. Offers shortest

crow-fly route between Thailand and Vietnam, but mountainous. Sparsely

populated.

Malaysia

Peninsular Malaysia is most economically advanced part of Study Area, with well

developed north-south highways, though east coast and east-west routes less

developed. East Malaysia (Sarawak and Sabah, plus Labuan) relatively less

developed; topology dominated by rivers and mountains.

Myanmar

Very large and diverse topography. Underdeveloped transport networks. 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.

Philippines

Varied archipelago. High-capacity trunk road network relatively limited outside

Metro Manila and immediate environs. There are long-term plans (with Japanese

funding) to link main islands with a series of bridges.

Thailand Whilst topography is varied, trunk highway network is relatively well developed.

Tollways concentrated around Greater Bangkok.

Vietnam

A “long” country and relatively “thin” excepting Red River area in the north.

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.




Appendix 11: Correlation between Wealth and Transport Networks

The econometrics presented in this Appendix, together with graphical representations

thereof are the author’s own work, using data presented in Appendices 1 and 2. Three

sets 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 countries

An amalgamation of the last two sets of equations is presented in Section 3.4.

1. Models on Full Sample

Countries in Study Area

KH Cambodia

CN China

ID Indonesia

LA Laos

MY Malaysia

MM Myanmar

PH Philippines

TH Thailand

VN Vietnam

Other Countries (1)

KR South Korea

MX Mexico

PO Poland

UK United Kingdom

US United States of America

Note: (1) Other countries presented for benchmarking and indication of trends given further





Structure of Equations Fitted

A series of equations were fitted. In each case GDP per capita was taken as the

explanatory variable. Two basic structural forms were fitted, as follows:

(A) GDPpcDependent

(B) GDPpcDependent

The following dependent variables were used:

(1) PopAP Population per Airport rportsNumberofAi

ationTotalPopul

(2) Km2AP Land Area per Airport rportsNumberofAi

kmreaTotalLandA )2(

(3) PopRail Population per km of Railway )(kmRailways

ationTotalPopul

(4) Km2Rail Land Area per km of Railway )(

)2(

kmRailways

kmreaTotalLandA

(5) PopRoad Population per km of Paved Road )(kmPavedRoad

ationTotalPopul

(6) Km2Road Land Area per km of Paved Road )(

)2(

kmPavedRoad

kmreaTotalLandA




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 the

parameter in question. As can be seen, in each instance equation form (B) gave a better

goodness-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:




Population per Airport

MM

LA

KH

VN

IDPH

CN

TH

MX

MY

PO

KR

UK

US

9

10

11

12

13

14

15

0 10,000 20,000 30,000 40,000 50,000


Ln

(Po

pu

lati

on

per

Air

po

rt)

km2 per Airport

MM

LA

KH

VN

ID

PH

CN

TH

MX

MYPO

KR

UKUS

5

6

7

8

9

10

0 10,000 20,000 30,000 40,000 50,000


Ln

(km

2 p

er

Air

po

rt)




Population per km of Railway

MM

KH

VNID

PH

CNTH

MX

MY

PO

KR

UK

US

6

7

8

9

10

11

12

0 10,000 20,000 30,000 40,000 50,000


Ln

(Po

pu

lati

on

per

km

of

Railw

ay)

Km2 per km of Railway

MM

KH

VN

IDPH

CNTHMX

MY

PO

KR

UK

US

2

3

4

5

6

0 10,000 20,000 30,000 40,000 50,000


Ln

(km

2 p

er

km

of

Railw

ay)





MM

LA

KH

VN

ID

PH

CNTH

MX

MY

PO

KR

UK

US

3

4

5

6

7

8

9

10

0 10,000 20,000 30,000 40,000 50,000


Ln

(Po

pu

lati

on

per

km

of

Paved

Ro

ad

)


MM

LA

KH

VN

ID

PH

CN

TH

MX

MY

POKR

UK

US

-1

0

1

2

3

4

5

6

0 10,000 20,000 30,000 40,000 50,000


Ln

(km

2 p

er

km

of

Paved

Ro

ad

)




2. Models on the 9 Study Area Countries

Countries in Study Area

KH Cambodia

CN China

ID Indonesia

LA Laos

MY Malaysia

MM Myanmar

PH Philippines

TH Thailand

VN Vietnam

The rationale of these analyses was to determine whether a different set of functions

exists within developing East Asian economies. The same equations were fitted as in

(1).




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)


parameter in question. As can be seen, results for airports and railway are not

statistically meaningful, with R2 in all instances below 10%. Only the equations

regarding paved road give results which could be deemed meaningful; and as with

Appendix 5, (B) form equations give better fits in terms of R2 than (A) equations. (5B)

and (6B) are plotted below.





MM

LA

KH

ID

PH

CN

TH

VN

MY6

7

8

9

10

0 2,000 4,000 6,000 8,000 10,000 12,000


Ln

(Po

pu

lati

on

per

km

of

Paved

Ro

ad

)


MM

LA

KH

ID

PH

CN

TH

VN

MY

1

2

3

4

5

6

0 2,000 4,000 6,000 8,000 10,000 12,000


Ln

(km

2 p

er

km

of

Paved

Ro

ad

)




3. Models on 5 Benchmarking Countries

Other Countries (1)

KR South Korea

MX Mexico

PO Poland

UK United Kingdom

US United States of America

Note: (1) Other countries presented for benchmarking and indication of trends given further


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)


parameter in question. As can be seen, whilst a relationship appears to hold for roads

per capita, geographic road density gives unsatisfactory results. This sample is not

meant to be necessarily significant, merely to give some guidance as to an S-curve for

road provision with respect to economic development, as presented in Section 3.4.





MX

PO

KR

UK

US

3

4

5

6

7

8

9

10

0 10,000 20,000 30,000 40,000 50,000


Ln

(Po

pu

lati

on

per

km

of

Paved

Ro

ad

)


MX

PO

KR

UK

US

-1

-0.5

0

0.5

1

1.5

2

2.5

3

0 10,000 20,000 30,000 40,000 50,000


Ln

(km

2 p

er

km

of

Paved

Ro

ad

)




Appendix 12: Expressway and Economic Index Calculations

Appendix 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

Expressway

1. Data on Guangdong Province and Guangzhou-Shenzhen Superhighway

Socio-economic data from Guangdong Statistical Yearbooks relating to Guangdong Province as

a whole:

GDSY98: Guangdong Provincial Bureau of Statistics (1998)




Year GDP current

price

(100m RMB)

GDP growth rate year-

on-year

(comparable price)

Implied price

inflation year-on-

year (%)

Real GDP

growth year-on-

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, p387




Year Highway Passenger

Trips (10,000 people)

Highway Passenger-

km (100 million)

Average Highway

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

Transport (10,000 MT)

Highway MT-km

(100 million)

Average Freight Trip

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 derived

Traffic and revenue data for Guangzhou-Shenzhen Superhighway from Hopewell Highway

Infrastructure (www.hopewellhighway.com):

Year and

Month

Average

Daily

Traffic

(vehicles)

Average Daily

Revenue

(thousand

RMB)

Year and

Month

Average

Daily

Traffic

(vehicles)

Average Daily

Revenue

(thousand

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 3554

http://www.hopewellhighway.com/




Year and

Month

Average

Daily

Traffic

(vehicles)

Average Daily

Revenue

(thousand

RMB)

Year and

Month

Average

Daily

Traffic

(vehicles)

Average Daily

Revenue

(thousand

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 7965




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 7,426,615

Indices thus obtained from above data:

GDP

Civil Vehicle

Ownership Passenger-km

Passenger Trip

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 MT-km

Freight Trip

Length

Superhighway

Traffic

Superhighway

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 149.3 186.5 424.4 418.8

Income 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 1.38




Graph of Guangdong Province/ Guangzhou-Shenzhen Superhighway data (indexed to

1995):

0.0

50.0

100.0

150.0

200.0

250.0

300.0

350.0

400.0

450.0

19951996

19971998

19992000

20012002

20032004

Year

Ind

ex (

1995=

100)

GDP Civil Vehicle Ow nership

Passenger-km Passenger Trip Length

Freight MT-km Freight Trip Length

Superhighw ay Traff ic Superhighw ay Revenue




2. Jiangsu Province and Shanghai-Nanjing Expressway (Jiangsu Section)

Socio-economic data from Jiangsu Statistical Yearbooks relating to Jiangsu Province as a

whole:

JSSY99: Jiangsu Provincial Statistics Bureau (1999)




Year GDP current

price

(100m RMB)


on-year

(comparable price)

Implied price

inflation year-on-

year (%)

Real GDP

growth year-on-

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


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



Highway Passenger-

km (100 million)

Average Highway


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 derived






Highway MT-km

(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 derived

Traffic data for Jiangsu Section of Shanghai-Nanjing Expressway from Jiangsu Expressway Co.

Ltd. (www.jsexpressway.com.cn):

Year

and

Month

Average Daily

Traffic

(vehicles)

Composition (%)

Car Small Medium Large Heavy

1997_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%

http://www.jsexpressway.com.cn/




Year

and

Month

Average Daily

Traffic

(vehicles)

Composition (%)

Car Small Medium Large Heavy

1999_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%




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 3,952


GDP

Civil Vehicle


Passenger Trip

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 MT-km

Freight Trip

Length

Expressway

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 122.7 120.3 256.1




Passenger-km 0.69


Freight MT-km 0.33


Total Expressway Traffic 1.43

Expressway Cars 1.14

Expressway Small 1.54

Expressway Medium 1.35

Expressway Large+Heavy 2.46




Graph of Jiangsu Province/ Shanghai-Nanjing Expessway data (indexed to 1997):

0.0

50.0

100.0

150.0

200.0

250.0

300.0

19971998

19992000

20012002

2003

Year

Ind

ex (

1997=

100)




Jiangsu Expressw ay Traff ic




3. Zhejiang Province and Shanghai-Hangzhou-Ningbo Expressway

Socio-economic data from Zhejiang Statistical Yearbooks relating to Zhejiang Province as a

whole:

ZJSY99: Zhejiang Provincial Bureau of Statistics (1999)




Year GDP current

price

(100m RMB)


on-year

(comparable price)

Implied price

inflation year-on-

year (%)

Real GDP

growth year-on-

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


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



Highway Passenger-

km (100 million)

Average Highway


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 derived






Highway MT-km

(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 derived

Traffic and revenue data for Shanghai-Hangzhou-Ningbo Expressway from Zhejiang

Expressway Co. Ltd. (www.zjec.com.cn):

Year

and

Month

Average

Daily

Revenue

(thousand

RMB)

Average

Daily Traffic

(vehicles)

Composition (%)

0-2T

(Small)

2-5T

(Medium)

5-10T

(Large)

10-20T

(Heavy)

>20T

(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%

http://www.zjec.com.cn/




Year

and

Month

Average

Daily

Revenue

(thousand

RMB)

Average

Daily Traffic

(vehicles)

Composition (%)

0-2T

(Small)

2-5T

(Medium)

5-10T

(Large)

10-20T

(Heavy)

>20T

(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%




Annual Average Daily Traffic and Revenue thus derived:

Year

Revenue

(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 2,646 944


GDP

Civil Vehicle


Passenger Trip

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 MT-km

Freight Trip

Length

Expressway

Traffic

Expressway

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 100.0 100.0 100.0 100.0




Passenger-km 0.37


Freight MT-km 0.82


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 3.10




Graph of Zhejiang Province/ Shanghai-Hangzhou-Ningbo Expressway data (indexed to

1998):

0.0

50.0

100.0

150.0

200.0

250.0

300.0

350.0

400.0

19981999

20002001

20022003

Year

Ind

ex (

1998=

100)




Zhejiang Expressw ay Traff ic Zhejiang Expressw ay Revenue




Appendix 13: Survey Questionnaire: Question Specification and Logical Flow

www.SurveyMonkey.com was used to prepare the questionnaire and undertake the

survey. The survey is reproduced page-by-page as follows:

Page 1 Introduction

For ALL Respondents First of all, thank you for taking part in this survey. In most cases I hope this should take no

more than 10-15 minutes of your time.

You are free to complete this survey on an anonymous basis. However, if you would be

willing to let me know a little bit more about you, you may like to complete some (or all) of

the questions on this first page. But if you would rather remain anonymous, feel free to skip

these questions...

Should you have any problems completing this survey, or wish to make a comment where a

box for optional comments is not provided, please do email me at: [email protected]

Finally, 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) Optional

2 Your organisation: (text response) Optional

3 Your position/ role: (text response) Optional

4 Your email: (text response) Optional

5 Telephone: (text response) Optional

6 Would you be happy for me to contact you

directly for further discussions?

YES/ NO Optional





Page 2 Your Sector of Expertise

For ALL Respondents

7 Please select which sectors you have worked in: (multiple choice from

menu)

Mandatory

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

experience do you have?

(text response) Mandatory

9 What percentage of this time has been spent on:

(please answer for each row; as some categories

overlap the total time across all rows may

exceed 100%)

(rating scale) Mandatory

One answer per row, from:

Transport infrastructure projects

All infrastructure projects (transport & non-transport)

Projects in developing economies

Tolled highway projects (urban and/or rural, anywhere in world)

Rural or inter-urban tolled highway projects

Rural or inter-urban tolled highway projects in developing

economies

0%

1%-10%

11%-25%

26%-50%

51%-75%

76%-95%

96%-100%




Page 3 Your International Experience

For ALL Respondents

10 In which parts of the world have you worked

on projects?

(select all that apply) Optional

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

Cambodia

Mainland China (i.e. excluding Hong

Kong, Macau, Taiwan)

Hong Kong

Indonesia

Japan

North Korea

South Korea

Laos

Macau

Malaysia

Mongolia

Myanmar (Burma)

Philippines

Singapore

Taiwan

Thailand

Timor-Leste (East Timor)

Vietnam

Tolled Highways

Other Transport Projects

Other Infrastructure Projects

Non-Infrastructure Projects




Page 4 Socio-Economic Risks

For ALL Respondents

12 Based on your overall experience in infrastructure

projects, with an emphasis on transport projects and

particularly tolled highways (if applicable), please rate

the importance of the following risks:



The prevailing political system and its stability

The legal system

Currency (exchange) risks

Ease of repatriating profits

Interest rates

Price inflation

Income (in)equality

Economic growth

Business cycles (as distinct from recent growth)

Drivers’ familiarity with highway tolls

Corruption

Critical

Strong

impact

Important

Limited

impact

Not usually

considered

Not sure

13 Similarly, please rate the importance of the following

to a project’s likely performance/ riskiness:



The project’s social/ economic benefits

Guanxi/ the importance of business connections

The project’s overall legal/ contractual foundations

The length of the operating concession

Construction time/ risk of delayed opening

Construction cost/ risk of cost over-run

Reliability of operating and maintenance cost estimates

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)

Critical

Strong impact

Important

Limited impact

Not usually

considered

Not sure

14 Have you experience of using, undertaking or

reviewing traffic and/or revenue forecasts undertaken by

transport consultants/ economists?

(choice) Mandatory

(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 above




Page 5 Data Availability for Modelling

Only for those respondents replying (a), (b) or (c) to Question 14.

15 From your experience, do you feel there are sufficient

data to successfully calibrate/ validate models?



In developed countries, there are sufficient data available

In developed countries, data are reliable

In developing countries, there are sufficient data available

In developing countries, data are reliable

Always

Usually

Sometimes

Rarely

Never

Not sure

16 From your experience, do you feel there are sufficient

data to successfully prepare meaningful traffic and

revenue forecasts?



In developed countries, there are sufficient data available

In developed countries, data (e.g. land use/economic forecasts)

are reliable

In developing countries, there are sufficient data available

In developing countries, data (e.g. land use/economic

forecasts) are reliable

Always

Usually

Sometimes

Rarely

Never

Not sure

17 If you would like to be more specific regarding

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:

(text response) Optional




Page 6 Transport Modelling Issues

Only for those respondents replying (a), (b) or (c) to Question 14.

18 Regarding the applicability of full four-stage models

(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:



Such models are reliable

Such models are too data hungry to be relied upon

Such models are too complicated to be of worth

Such models are not suitable for toll-road work

Economic uncertainties/ pace of change makes them irrelevant

in developing economies

Such models are too much of a black box for non-specialists to

properly critique model outputs

Strongly

agree

Agree

Neutral

opinion

Disagree

Strongly

disagree

Not sure

19 Regarding network assignment models (e.g. based on

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:





Such models are too simplistic to be relied upon

Such models are too complicated to be of worth




Such models are too much of a black box for non-specialists to

properly critique model outputs

Strongly

agree

Agree

Neutral

opinion

Disagree

Strongly

disagree

Not sure

20 Regarding spreadsheet-based traffic/ revenue models,

please indicate how strongly you agree or disagree

with the following statements:





Such models are too simplified to be of worth




Such models are too simplistic to provide meaningful outputs

Strongly

agree

Agree

Neutral

opinion

Disagree

Strongly

disagree

Not sure

21 If you have any specific comments on issues with

developing/ calibrated/ forecasting with traffic and

revenue models (of any type), please give your

comments here:





Page 7 General Reliability of Traffic Forecasts

Only for those respondents replying (a), (b), (c) or (d) to Question 14.

22 From your experience, please state how often you feel

each of the following statements is true:



How often do projects significantly exceed forecast traffic/

revenue levels?

How often do projects fall well short of forecast traffic/

revenue levels?

Do you believe that transport planners are pressured by clients

to adjust forecasts to meet their expectations?

Do you believe that transport consultants’ forecasts are higher

if they are engaged by equity- rather than debt-side clients?

Very often

Quite often

Sometimes

Rarely

Never

Not sure

Page 8 Project Evaluation Criteria

For ALL Respondents

23 How often do you explicitly consider the following in

appraising tolled highways?



Traffic forecasts: Base and/or Central Case

Traffic forecasts: Optimistic and/or High Case

Traffic forecasts: Conservative and/or Low Case

Revenue forecasts: Base and/or Central Case

Revenue forecasts: Optimistic and/or High Case

Revenue forecasts: Conservative and/or Low Case

Congestion on competing/ alternative routes

Congestion on link-roads/ feeder routes

Capacity of the highway being considered

Always

Usually

Sometimes

Rarely

Never

Not sure

24 How often do you explicitly consider the following

criteria in appraising infrastructure projects (with an

emphasis on tolled highways if applicable)?



Net Present Value (NPV)

Financial Internal Rate of Return (FIRR)

Economic Internal Rate of Return (EIRR; including social

impacts)

Social Cost/ Benefit Ratios

Risk correlation versus other projects in company’s/ client’s

portfolio

Counterparty risks: can partners contribute equity/ debt

Sovereign/ Institutional other country/ legal risks

Always

Usually

Sometimes

Rarely

Never

Not sure

25 If you use any financial ratios when appraising

projects, please state which ratios you normally use:





Page 9 Future Prospects

For ALL Respondents

26 How would you rate the potential for inter-urban

tolled highways over the next 10 years in each of the

following countries:



Cambodia

Mainland China (i.e. excluding Hong Kong,

Macau, Taiwan)

Indonesia

Laos

Malaysia

Myanmar (Burma)

Philippines

Thailand

Vietnam

Sector overdeveloped (few

prospects)

Maturing market (decline)

Already strong and likely to

remain so (steady)

Fast developing (growing)

Only just starting (nascent)

Undeveloped and negligible

prospects (no market)

Not sure

27 Comparing the next 10 years (2006-2016) with the

last 5 years (2001-2006), how do you feel each the

following will change:



Fuel prices

General price inflation

Interest rates

Economic growth

Exchange rate volatility

Acceptability of road tolls and toll increases

Will be significantly greater

Will increase to an extent

No significant change

Will decrease to an extent

Will significantly decrease

Not sure

28 If you believe that there will be any other significant

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 Respondents

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.


30 If you would like information about the survey

responses, once collated, or about my broader

research, please indicate below:

(choose from

menu)

Optional

No, thank you.

Yes, regarding survey results.

Yes, regarding your research.




Appendix 14: Amendments Made to Questionnaire Following Pilot Survey

A number of useful comments were made during the piloting of the questionnaire.

Some comments could be readily incorporated through amending the wording of a

question. In other instances a multi-choice rather than single-choice response was

implemented. In one case a question was split into two separate questions. The Question

number references given refer to the question numbers in the Final Survey (as shown in

Appendix 13).

Question 7: changed from single-choice to multi-choice, following feedback from those

who have developed their career through consultancy and academia and/or the public

sector and/or aid agencies.

Question 9: following a comment received, the question was clarified through the

addition of the text: “(please answer for each row; as some categories overlap the total

time across all rows may exceed 100%)”

Question 11: “Timor-Leste” changed to “Timor-Leste (East Timor)” to provide greater

clarity; this a result of the Author’s own review of questions.

Questions 18 and 19: these were originally a single question, referring to network

assignment models and software. An initial comment was received via the pilot,

pointing out that approaches ought to be largely independent of software platforms. This

comment initially suggested use of the term “Four Stage Model” in lieu of software

platforms. However, subsequent consideration and discussion with the originator of the

specific comment (by email) and with another respondent (by telephone) resulted in the

initial question being subdivided into two, as now shown. Section 2.9 cited practical

difficulties of using Four Stage models; hence the subdivision was into a question




regarding Four Stage models, and a question regarding network assignment models

outwith Four Stage models.

Question 22: one respondent pointed out in their comments that it would be “utterly

wrong" if equity- and debt-side forecasts were the same, given the different risk/ reward

profiles of the different perspectives. It was clarified that the purpose of this question

was 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 Base

usually lower than Central. However, from experience the two terms are often used

interchangeably, 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 and

Low/ Conservative.

Question 24: one respondent claimed this question was ambiguous, with attention to

financial returns, social returns or portfolio-based risk-spreading being determined by

whom one is working for. This ambiguity was at least semi-intentional; the aim being to

see how often any group considers which set of objectives. Depending on eventual

survey returns, the intention being to see if one group are inherently more interested in

one set of factors than another. (A priori, government and aid agencies ought to be more

interested 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 questions

before answering. However, within the context of the software used (which includes

some logical branching) this is not feasible. A possible work-around would be to

complete the survey and then return to the beginning to revise answers. But it was felt




that the survey introduction would get overly complicated to include this possibility in

the preamble.

General Comments #2: following subsequent discussions with respondents, aside from

specific question-related comments (given above) no major areas appeared to have been

missed out.

General Comments #3: the time take to complete appeared to have been around 15

minutes by experienced transport planners/ modellers (the group likely to have the

longest response time as they need to answer all questions). This was felt to be a little

bit 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 the

likely response time once printing off and faxing back, etc were included). Also, there

were no obvious candidate questions to be omitted. Thus, whilst a 10-minute response

time might be preferable, it might not be attainable by those answering questions on

modelling. However, for those without hands-on modelling experience, a 10-minute

response ought to be feasible, hence the preamble was revised from “no more than 15

minutes of your time” to “no more than 10-15 minutes of your time.”

General Comments #4: one respondent suggested that a distinction between transport

infrastructure projects and infrastructure in general was not clear; that they cross-relate

to a great extent. Indeed, this is one of the rationales of the survey and respondent

targeting, that there are often substantial similarities. However, where appropriate the

intention remains for the respondent to concentrate on transport projects, should they

have such experience (and broader infrastructure experience where they do not).

Conversely, one respondent (non-transport planner) who successfully completed the

survey commented that he was unable to comment further as he was not an expert in the




transport field, notwithstanding his response to question 28 “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.”

General Comments #5: some respondents reported problems on pages with questions

containing mandatory answers. As such, logical control on giving mandatory answers

was over-ridden (enabling mandatory answers to be skipped in theory). The exception

was Question 14, where an answer was required to determine which, if any of questions

15 to 22 the respondent would be presented with.




Appendix 15: Questionnaire Responses

Questions 1 to 6

Relate to respondent identity; no “results” presented due to confidentiality

considerations.

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 of

different 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 10.9




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

projects 10 22 29 19 26 38 18

All infrastructure projects

(transport & non-

transport)

5 14 18 21 30 45 29

Projects in developing

economies 42 28 24 21 20 17 10

Tolled highway projects

(urban and/or rural,

anywhere in world)

49 54 30 18 6 4 1

Rural or inter-urban

tolled highway projects 69 51 24 11 2 4 1


tolled highway projects in


88 50 13 7 1 2 1

Assuming mid-range values (e.g. 5.5% for 1-10%) the following means and standard

deviations were calculated:

Mean Standard Deviation

Transport infrastructure projects 49.4% 35.8% All infrastructure projects (transport & non-

transport) 60.3% 34.9%

Projects in developing economies 31.3% 35.6% Tolled highway projects (urban and/or rural,

anywhere in world) 14.3% 22.5%

Rural or inter-urban tolled highway projects 10.3% 22.6% Rural or inter-urban tolled highway projects in

developing economies 6.7% 21.1%

Combining the percentages of time spent on each kind of work with number of years of

working experience, the following estimates were obtained of years per kind of work:

Mean 30+ 20-29 10-19 5-9 1-4 0

Transport infrastructure

projects 10.66 6 21 38 29 58 7

All infrastructure projects

(transport & non-

transport)

13.13 8 28 48 30 43 2

Projects in developing

economies 7.26 5 13 21 24 57 39

Tolled highway projects

(urban and/or rural,

anywhere in world)

2.57 0 0 9 20 84 46


tolled highway projects 1.70 0 0 4 10 79 66


tolled highway projects in


1.12 0 0 2 7 65 85




Question 10: In which parts of the world have you worked on projects? (150

responses)

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?

Tolled

Highways

Other

Transport

Projects

Other

Infrastructure

Projects

Non-

Infrastructure

Projects

Anything

in this

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 36

Note: * The last column can be smaller than the total of the previous four, as a

respondent may have worked on a number of different kinds of project within one

country/ territory.

Note: Countries being explicitly considered under this Dissertation are shown in bold.




Question 12: Based on your overall experience in infrastructure projects, with an

emphasis 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

Critical Strong

Impact Important Limited

Impact Not Usually

Considered The prevailing political system and

its stability 49 56 31 7 4

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

recent growth) 3 19 48 46 15

Drivers’ familiarity with highway

tolls 2 22 44 47 22

Corruption 23 31 42 21 17

Using the values 1 to 5 (as per column headings above), the mean and standard

deviation of responses as follows:

Responses Mean

Standard

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 1.25




The mean scores from above were also compared with mean scores based on the 6

aggregated 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 2.80 Key: FLO = Financial, Legal, Operator

TpEc = Transport Planners and Economists

E&A = Engineers and Architects

G&A = Government and Aid Agencies

Acad = Academics

Oth = Others

Question 13: Similarly, please rate the importance of the following to a project’s

likely performance/ riskiness:


1 2 3 4 5

Critical Strong

Impact Important Limited

Impact Not Usually

Considered The project’s social/ economic

benefits 20 49 59 17 2

Guanxi/ the importance of business

connections 9 40 59 25 3

The project’s overall legal/

contractual foundations 33 45 58 8 1

The length of the operating

concession 15 46 62 14 2

Construction time risk 15 56 52 19 1 Construction cost/ risk of over-run 22 58 53 9 2

Reliability of operating and

maintenance cost estimates 13 43 65 22 2

The enforceability of toll/ tariff

increases 27 45 45 12 8

Minimum income guarantees and

their enforceability 14 35 57 19 9

The threat of competing routes/

alternatives to the project 26 47 51 14 5

Standard of connecting routes 7 44 73 15 3 Toll affordability for large vehicles 12 37 68 12 6 Toll affordability for other vehicles 13 40 57 19 8

Toll leakage/ evasion 13 28 58 30 7 Ramp up length 3 27 53 39 12






Responses Mean

Standard

Deviation The project’s social/ economic benefits 147 2.54 0.91

Guanxi/ the importance of business connections 136 2.80 0.89 The 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 0.94




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

foundations 2.30 1.90 2.35 2.18 2.34 2.44 2.57

Operating 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 2.43

Operating and maintenance

costs 2.70 2.52 2.79 2.76 2.64 2.76 2.48

Toll/ tariff increase

enforceability 2.48 1.83 2.44 2.45 2.57 2.19 2.70

Minimum income guarantee

enforceability 2.81 2.31 2.81 2.83 2.94 2.88 2.67

Threat of competing routes 2.48 2.14 2.43 2.36 2.68 2.11 2.70 Standard of connecting

routes 2.74 2.41 2.70 2.88 2.89 2.44 3.00

Toll affordability for large

vehicles 2.73 2.41 2.71 2.83 2.94 2.71 2.85

Toll affordability for other

vehicles 2.77 2.21 2.79 2.80 2.94 2.47 3.05

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 3.44




Question 14: Have you experience of using, undertaking or reviewing traffic

and/or revenue forecasts undertaken by transport consultants/ economists? (156

responses)

Yes, I have prepared forecasts myself 61 39%

I have supervised forecasts made by my staff, but have

not prepared them 12 8%

Yes, I have reviewed or used forecasts undertaken by

others, but have not prepared them 44 28%

I have worked alongside transport consultants, but

have not used their forecasts 16 10%

No, none of the above 23 15%

Question 15: From your experience, do you feel there are sufficient data to

successfully calibrate/ validate models?


1 2 3 4 5

Always Usually Sometimes Rarely Never

In developed countries, there are

sufficient data available 9 65 29 6 1

In developed countries, data are

reliable 2 53 48 6 0

In developing countries, there are


In developing countries, data are




Responses Mean

Standard

Deviation In developed countries, there are sufficient data

available 110 2.32 0.74

In developed countries, data are reliable 109 2.53 0.63 In developing countries, there are sufficient data

available 100 3.45 0.73

In developing countries, data are reliable 99 3.49 0.77




Question 16: From your experience, do you feel there are sufficient data to

successfully prepare meaningful traffic and revenue forecasts?


1 2 3 4 5


In developed countries, there are


In developed countries, data (e.g.

land use/economic forecasts) are


In developing countries, there are


In developing countries, data (e.g.

land use/economic forecasts) are




Responses Mean

Standard

Deviation In developed countries, there are sufficient data

available 109 2.26 0.64

In developed countries, data (e.g. land

use/economic forecasts) are reliable 107 2.43 0.70

In developing countries, there are sufficient data

available 99 3.25 0.72

In developing countries, data (e.g. land

use/economic forecasts) are reliable 99 3.40 0.75




Question 17: If you would like to be more specific regarding 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: (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 People's 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 IFI's 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 more




useful. Traffic counts are quick and cheap to conduct, shouldn't 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. It's 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.




Question 18: Regarding the applicability of full four-stage models (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:


1 2 3 4 5

Strongly

Agree Agree Neutral Disagree

Strongly

Disagree

Such models are reliable 2 35 56 7 2 Such models are too data hungry to

be relied upon 0 29 35 30 5

Such models are too complicated to

be of worth 4 12 26 47 11

Such models are not suitable for

toll-road work 0 11 26 42 14

Economic uncertainties/ pace of

change makes them irrelevant in

developing economies 2 20 31 35 6

Such models are too much of a

black box for non-specialists to

properly critique model outputs 10 37 22 27 6



Responses Mean

Standard

Deviation Such models are reliable 102 2.73 0.70

Such 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

them irrelevant in developing economies 94 3.24 0.93

Such models are too much of a black box for non-

specialists to properly critique model outputs 102 2.82 1.11




Question 19: Regarding network assignment models (e.g. based on 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:


1 2 3 4 5

Strongly


Strongly

Disagree



Such models are too simplistic to be

relied upon 0 21 33 37 5

Such models are too complicated to

be of worth 0 4 28 49 16






Such models are too much of a

black box for non-specialists to

properly critique model outputs 4 19 30 36 10



Responses Mean

Standard


Such 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


Such models are too much of a black box for non-

specialists to properly critique model outputs 99 3.29 1.02




Question 20: Regarding spreadsheet-based traffic/ revenue models, please indicate

how strongly you agree or disagree with the following statements:


1 2 3 4 5

Strongly


Strongly

Disagree



Such models are too simplified to

be of worth 3 12 39 34 5






Such models are too simplistic to

provide meaningful outputs 5 10 40 39 3



Responses Mean

Standard


Such 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


Such models are too simplistic to provide

meaningful outputs 97 3.26 0.88

Question 21: If you have any specific comments on issues with developing/

calibrated/ forecasting with traffic and revenue models (of any type), please give

your 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 It's 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..




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.




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 the

following statements is true:


1 2 3 4 5

Very

Often

Quite

Often Sometimes Rarely Never

How often do projects significantly

exceed forecast traffic/ revenue

levels?

5 18 43 48 0

How often do projects fall well

short of forecast traffic/ revenue

levels?

10 51 42 12 0

Do you believe that transport

planners are pressured by clients to

adjust forecasts to meet their

expectations?

18 43 45 12 3

Do you believe that transport

consultants’ forecasts are higher if

they are engaged by equity- rather

than debt-side clients?

10 22 44 14 3



Responses Mean

Standard

Deviation How often do projects significantly exceed forecast

traffic/ revenue levels? 114 3.18 0.85

How often do projects fall well short of forecast

traffic/ revenue levels? 115 2.49 0.80

Do you believe that transport planners are

pressured by clients to adjust forecasts to meet

their expectations?

121 2.50 0.95

Do you believe that transport consultants’ forecasts

are higher if they are engaged by equity- rather

than debt-side clients? 93 2.76 0.94







Significantly exceed

forecast traffic/ revenue 3.18 3.42 3.25 3.17 3.00 3.13 2.70

Fall well short of forecast

traffic/ revenue 2.49 2.23 2.45 2.67 2.69 2.40 2.40

Are transport planners

pressured by clients to

adjust forecasts?

2.50 2.32 2.41 2.56 2.38 2.40 2.55

Are forecasts are higher

for equity- rather than

debt-side clients? 2.76 2.86 2.64 3.14 2.86 2.69 2.63

Key: FLO = Financial, Legal, Operator




Acad = Academics

Oth = Others

Question 23: How often do you explicitly consider the following in appraising

tolled highways?


1 2 3 4 5


Traffic forecasts: Base and/or

Central Case 53 31 15 1 5

Traffic forecasts: Optimistic and/or

High Case 22 37 24 16 6

Traffic forecasts: Conservative

and/or Low Case 35 44 15 6 5

Revenue forecasts: Base and/or

Central Case 46 31 16 3 6

Revenue forecasts: Optimistic

and/or High Case 21 34 22 16 8

Revenue forecasts: Conservative

and/or Low Case 32 41 16 7 6

Congestion on competing/

alternative routes 37 38 16 6 7

Congestion on link-roads/ feeder

routes 37 37 20 4 8

Capacity of the highway being

considered 57 33 10 1 5






Responses Mean

Standard

Deviation Traffic forecasts: Base and/or Central Case 105 1.80 1.04

Traffic forecasts: Optimistic and/or High Case 105 2.50 1.15 Traffic 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 1.21

Revenue 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 1.01

Question 24: How often do you explicitly consider the following criteria in

appraising infrastructure projects (with an emphasis on tolled highways if

applicable)?


1 2 3 4 5


Net Present Value (NPV) 52 40 19 2 4 Financial Internal Rate of Return

(FIRR) 39 43 19 6 7

Economic Internal Rate of Return

(EIRR; including social impacts) 28 35 27 14 10

Social Cost/ Benefit Ratios 25 31 34 20 9 Risk correlation versus other

projects in company’s/ client’s

portfolio 19 12 31 27 20

Counterparty risks: can partners

contribute equity/ debt 17 23 19 23 21

Sovereign/ Institutional other

country/ legal risks 26 22 20 20 20



Responses Mean

Standard

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;

including social impacts) 114 2.50 1.23

Social Cost/ Benefit Ratios 119 2.64 1.20 Risk correlation versus other projects in

company’s/ client’s portfolio 109 3.16 1.33

Counterparty risks: can partners contribute

equity/ debt 103 3.08 1.38

Sovereign/ Institutional other country/ legal risks 108 2.87 1.44




Question 25: If you use any financial ratios when appraising projects, please state

which 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 doens't 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 thresholds




Question 26: How would you rate the potential for inter-urban tolled highways

over the next 10 years in each of the following countries:


1 2 3 4 5 6

Over-

developed Maturing Steady Developing Nascent No

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 5



Responses Mean

Standard

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 0.96




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 4.54

Key: FLO = Financial, Legal, Operator




Acad = Academics

Oth = Others




Repeating the analysis for Question 26, but this time including responses ONLY from

those with experience in the country in question (from answer to Question 11) and once

again excluding any “Not Sure” responses:

1 2 3 4 5 6

Over-

developed Maturing Steady Developing Nascent No

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 2



Responses Mean

Standard

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 0.72

Comparing the mean of all respondents who expressed an opinion with the sub-set of

those with experience in the country:

All Respondents

(A) Those with Country

Experience (B) Difference

(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 0.02




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:


1 2 3 4 5

Significant

Increase

Increase to

an Extent No Change

Decrease to

an Extent

Significant

Decrease

Fuel prices 49 70 2 2 0 General price

inflation 4 74 42 1 0

Interest rates 7 41 63 5 0 Economic growth 5 60 41 17 0

Exchange rate

volatility 6 43 52 8 1

Acceptability of

road tolls and toll

increases 22 63 28 2 2



Responses Mean

Standard

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 0.79




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 2.68 Acceptability of road tolls

and toll increases 2.14 2.42 2.14 2.29 2.36 2.36 2.11 Key: FLO = Financial, Legal, Operator




Acad = Academics

Oth = Others




Question 28: If you believe that there will be any other significant 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): (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 IFI's 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.




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 There's 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 don't 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.




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! I'm sorry that I don't have

more time to take more of an interest in your research. In any case, I'm 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 don't 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, once

collated, or about my broader research, please indicate below: (80 responses)

No, thank you 9

Yes, regarding survey results 23

Yes, regarding your research 48




Appendix 16: Risk Simulation Modelling: Simulation Parameters Employed

Ca

seV

ari

ab

le

Mo

da

l

Va

lue

Ch

an

ce o

f

Sm

all

er

Va

lue

Sta

nd

ard

Dev

iati

on

(Lo

w V

alu

es)

Min

imu

m

Va

lue

Sta

nd

ard

Dev

iati

on

(Hig

h V

alu

es)

Ma

xim

um

Va

lue

So

urc

e/ J

ust

ific

ati

on

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Ex

pre

ssw

ay C

apac

ity/

Lan

e (p

cu’s

per

day

)2

0,0

00

50

%2

,00

02

0,0

00

2,0

00

28

,00

0F

rom

oth

er s

tud

ies

(co

rro

bo

rate

d

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Lo

cal

Ro

ad C

apac

ity/

Lan

e (p

cu’s

per

day

)1

0,0

00

50

%1

,00

08

,00

01

,00

01

2,0

00

Fro

m o

ther

stu

die

s (c

orr

ob

ora

ted

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Co

nst

ruct

ion

Co

st (

$)*

$1

85

.32

m2

5%

5%

90

%1

5%

30

%

$4

,63

3,0

00

per

km

* 4

0 k

m;

cost

over

run

mo

re l

ikel

y t

han

un

der

run

(se

e 2

.11

)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Co

nst

ruct

ion

Du

rati

on

(in

Qu

arte

rs)

10

20

%1

82

14

tim

e o

ver

run

mo

re l

ikel

y t

han

un

der

run

(se

e 2

.11

); s

pec

ifie

d i

n

wh

ole

qu

arte

rs (

i.e.

in

teger

s)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Op

erat

ion

s &

Mai

nte

nan

ce F

ixed

Co

sts

(% o

f C

on

stru

ctio

n C

ost

)

2%

50

%0

.50

%0

.10

%1

%4

%F

rom

oth

er s

tud

ies

(co

rro

bo

rate

d

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Op

erat

ion

s &

Mai

nte

nan

ce V

aria

ble

Co

sts

(% o

f R

even

ues

)

3%

50

%1

%1

%1

%5

%F

rom

oth

er s

tud

ies

(co

rro

bo

rate

d

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Bas

e V

alu

e o

f T

ime

for

Sm

all

Veh

icle

s ($

/ho

ur)

$4

5

0%

$1

$

2

$1

$

6

Fro

m o

ther

stu

die

s (c

orr

ob

ora

ted

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Bas

e V

alu

e o

f T

ime

for

Lar

ge

Veh

icle

s ($

/ho

ur)

$3

5

0%

$1

$

1

$1

$

5

Fro

m o

ther

stu

die

s (c

orr

ob

ora

ted

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Inco

me

Ela

stic

ity o

f V

alu

e

of

Tim

e (S

mal

l V

ehic

les)

0.5

50

%0

.15

0.2

0.1

50

.8se

e 2

.10

.1

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Inco

me

Ela

stic

ity o

f V

alu

e

of

Tim

e (L

arge

Veh

icle

s)0

.55

0%

0.1

50

.20

.15

0.8

see

2.1

0.1




Ca

seV

ari

ab

le

Mo

da

l

Va

lue

Ch

an

ce o

f

Sm

all

er

Va

lue

Sta

nd

ard

Dev

iati

on

(Lo

w V

alu

es)

Min

imu

m

Va

lue

Sta

nd

ard

Dev

iati

on

(Hig

h V

alu

es)

Ma

xim

um

Va

lue

So

urc

e/ J

ust

ific

ati

on

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Bas

e S

mal

l V

ehic

le

Op

erat

ing C

ost

s o

n

Ex

pre

ssw

ays

($/k

m)

$0

.06

5

0%

$0

.01

5

$0

.03

$

0.0

15

$

0.0

9

Fro

m o

ther

stu

die

s (c

orr

ob

ora

ted

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Bas

e L

arge

Veh

icle

Op

erat

ing C

ost

s o

n

Ex

pre

ssw

ays

($/k

m)

$0

.10

5

0%

$0

.02

5

$0

.05

$

0.0

25

$

0.1

5

Fro

m o

ther

stu

die

s (c

orr

ob

ora

ted

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Sm

all

Veh

icle

Lo

cal

Ro

ad

Op

erat

ing C

ost

(re

lati

ve

to

exp

ress

way

)

1.5

50

%0

.25

1.0

0.1

52

.0F

rom

oth

er s

tud

ies

(co

rro

bo

rate

d

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Lar

ge

Veh

icle

Lo

cal

Ro

ad

Op

erat

ing C

ost

(re

lati

ve

to

exp

ress

way

)

2.0

50

%0

.25

1.5

0.2

52

.5F

rom

oth

er s

tud

ies

(co

rro

bo

rate

d

by o

ther

tra

nsp

ort

pla

nn

ers)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Fac

tor

for

Bas

e S

mal

l

Veh

icle

Dem

and

Mat

rix

10

0%

50

%1

5%

70

%1

5%

13

0%

arb

itra

ry t

o r

efle

ct p

oss

ible

err

or

ran

ge

in i

nit

ial

surv

eys

on

a "

real

wo

rld

" p

roje

ct

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Fac

tor

for

Bas

e L

arge

Veh

icle

Dem

and

Mat

rix

10

0%

50

%1

5%

70

%1

5%

13

0%

arb

itra

ry t

o r

efle

ct p

oss

ible

err

or

ran

ge

in i

nit

ial

surv

eys

on

a "

real

wo

rld

" p

roje

ct

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Inco

me

Ela

stic

ity o

f S

mal

l

Veh

icle

Tra

ffic

1.2

55

0%

0.2

00

.85

0.2

01

.65

see

3.5

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Inco

me

Ela

stic

ity o

f L

arge

Veh

icle

Tra

ffic

1.1

05

0%

0.2

00

.70

0.2

01

.50

see

3.5

(sm

alle

r val

ue

to a

llo

w f

or

larg

er t

ruck

s an

d c

oac

hes

an

d

incr

ease

d e

ffic

ien

cy)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

To

ll R

even

ue

Lea

kag

e (%

)1

0%

50

%2

.50

%5

%5

%2

0%

Fro

m o

ther

stu

die

s (c

orr

ob

ora

ted

by o

ther

tra

nsp

ort

pla

nn

ers)




Ca

seV

ari

ab

le

Mo

da

l

Va

lue

Ch

an

ce o

f

Sm

all

er

Va

lue

Sta

nd

ard

Dev

iati

on

(Lo

w V

alu

es)

Min

imu

m

Va

lue

Sta

nd

ard

Dev

iati

on

(Hig

h V

alu

es)

Ma

xim

um

Va

lue

So

urc

e/ J

ust

ific

ati

on

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Init

ial

Am

pli

tud

e o

f R

amp

-

Up

(%

tra

ffic

dec

reas

e)4

0%

50

%1

0%

20

%2

0%

80

%se

e 2

.10

.4

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Ram

p-U

p D

ura

tio

n

(in

Qu

arte

rs)

84

0%

24

52

0se

e 2

.10

.4

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Sm

all

Veh

icle

s' T

oll

ing

Pen

alty

(m

inu

tes)

10

50

%5

05

20

see

2.9

(an

d c

orr

ob

ora

ted

by o

ther

tran

spo

rt p

lan

ner

s)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Lar

ge

Veh

icle

s' T

oll

ing

Pen

alty

(m

inu

tes)

15

50

%5

55

25

see

2.9

(an

d c

orr

ob

ora

ted

by o

ther

tran

spo

rt p

lan

ner

s)

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Sm

all

Veh

icle

Ro

ute

ing

Sen

siti

vit

y "

Lam

bd

a" f

or

Lo

git

Su

b-M

od

el

0.0

55

0%

0.0

12

50

.25

0.0

12

50

.75

arb

itra

ry v

alu

e

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Lar

ge

Veh

icle

Ro

ute

ing

Sen

siti

vit

y "

Lam

bd

a" f

or

Lo

git

Su

b-M

od

el

0.0

55

0%

0.0

12

50

.25

0.0

12

50

.75

arb

itra

ry v

alu

e

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

To

ll E

scal

atio

n R

ate

(% o

f

Ret

ail

Pri

ce I

nd

ex

Infl

atio

n)

90

%5

0%

15

%6

0%

5%

10

0%

see

2.1

0.3

Co

nven

tio

nal

Res

po

nd

ents

'

Ko

nd

rati

eff

Qu

arte

rs B

etw

een

To

ll

Incr

ease

s1

24

0%

28

42

0se

e 2

.10

.3




Ca

seV

ari

ab

le

Mo

da

l

Va

lue

Ch

an

ce o

f

Sm

all

er

Va

lue

Sta

nd

ard

Dev

iati

on

(Lo

w V

alu

es)

Min

imu

m

Va

lue

Sta

nd

ard

Dev

iati

on

(Hig

h V

alu

es)

Ma

xim

um

Va

lue

So

urc

e/ J

ust

ific

ati

on

Co

nven

tio

nal

6%

50

%2

%2

%2

%1

0%

bas

e fo

r o

ther

cas

es a

lso

Res

po

nd

ents

'+

1%

50

%0

.5%

+0

%0

.5%

+2

%in

ad

dit

ion

to

Co

nven

tio

nal

val

ue

Ko

nd

rati

eff

+1

%5

0%

0.5

%+

0%

0.5

%+

2.5

%in

ad

dit

ion

to

Res

po

nd

ents

val

ue

Co

nven

tio

nal

2.5

%5

0%

1%

0.5

%1

%4

.5%

bas

e fo

r o

ther

cas

es a

lso

Res

po

nd

ents

'+

2.5

%5

0%

1.5

%+

0%

1.5

%+

5.5

%in

ad

dit

ion

to

Co

nven

tio

nal

val

ue

Ko

nd

rati

eff

+0

%n

/an

/an

/an

/an

/asa

me

as R

esp

on

den

ts v

alu

e

Co

nven

tio

nal

2.5

%5

0%

1%

0.5

%1

%4

.5%

bas

e fo

r o

ther

cas

es a

lso

Res

po

nd

ents

'+

0.7

5%

50

%0

.25

%+

0.2

5%

0.2

5%

+1

.25

%in

ad

dit

ion

to

Co

nven

tio

nal

val

ue

Ko

nd

rati

eff

+1

.0%

50

%0

.50

%+

0%

0.5

0%

+2

.5%

in a

dd

itio

n t

o R

esp

on

den

ts v

alu

e

Co

nven

tio

nal

2.5

%5

0%

1%

0.5

%1

%4

.5%

bas

e fo

r o

ther

cas

es a

lso

Res

po

nd

ents

'+

0.7

5%

50

%0

.25

%+

0.2

5%

0.2

5%

+1

.25

%in

ad

dit

ion

to

Co

nven

tio

nal

val

ue

Ko

nd

rati

eff

+1

.0%

50

%0

.50

%+

0%

0.5

0%

+2

.5%

in a

dd

itio

n t

o R

esp

on

den

ts v

alu

e

Co

nven

tio

nal

5%

50

%1

%3

%1

%7

%b

ase

for

oth

er c

ases

als

o

Res

po

nd

ents

'+

1%

50

%0

.50

%+

0%

0.5

0%

+2

%in

ad

dit

ion

to

Co

nven

tio

nal

val

ue

Ko

nd

rati

eff

+2

%5

0%

1%

+0

%1

%+

4%

in a

dd

itio

n t

o R

esp

on

den

ts v

alu

e

Co

nven

tio

nal

+2

%5

0%

1%

+0

%1

%+

4%

in a

dd

itio

n t

o i

nit

ial

deb

t ra

te

Res

po

nd

ents

'+

2%

50

%1

%+

0%

1%

+4

%in

ad

dit

ion

to

in

itia

l d

ebt

rate

Ko

nd

rati

eff

+2

%5

0%

1%

+0

%1

%+

4%

in a

dd

itio

n t

o i

nit

ial

deb

t ra

te

GD

P G

row

th (

% p

.a.)

Veh

icle

Op

erat

ing C

ost

Pri

ce I

nfl

atio

n (

% p

.a.)

Co

nst

ruct

ion

, O

per

atio

ns

& M

ain

ten

ance

Co

st

Infl

atio

n (

% p

.a.)

Gen

eral

Pri

ce I

nfl

atio

n (

%

p.a

.)

Inte

rest

Rat

es f

or

Init

ial

Deb

t (%

p.a

.)b

ased

on

co

nst

ruct

ion

co

sts

Inte

rest

Rat

es f

or

Ex

tra

Deb

t (%

p.a

.)fo

r su

bse

qu

ent

cash

sh

ort

fall

s




Appendix 17: Risk Simulation Modelling: Fixed Parameters

Concession Length

30 years (120 quarters), including construction time (commencing in quarter 1)

Network Road Lengths

As described in Section 5.2

Pcu Factors

(to equivalence different vehicle types to a common unit for congestion analysis)

Small Vehicles = 1

Large Vehicles = 2

Speeds by Road Capacity

As shown in Figure 5.B.




Appendix 18: Risk Simulation Modelling: Equations Employed

Growing Prices in Line with Appropriate Inflation Rate

Applies To: Using:

Construction Cost per Quarter (charged

during construction period only)

Construction, Operations & Maintenance

Cost Inflation

Operations & Maintenance Fixed Costs

(charged following start of operations)

Construction, Operations & Maintenance

Cost Inflation

Vehicle Operating Costs for Expressways

and Local Roads, for Small and Large

Vehicles (used in path-building on

assignment)

Vehicle Operating Cost Price Inflation

Toll Rates (toll rates updated every X

quarters following start of operations,

where X is the number of quarters between

toll increases)

General Price Inflation * Toll Escalation

Factor

4

1

1 1 ateInflationRPRICEPRICE qq

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 Traffic

4

1

1 1 ElasticityateGDPGrowthRTRIPSTRIPS qq

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

4

1

4

1

1 11 ElasticityateGDPGrowthRInflationVOTVOT qq




Generalised Costs of Routes Using and Not Using Expressway

Travel time for each link:

)(

)(60

kphSpeed

kmLinkLengthimeMinutesOfT

Monetary cost for each link:

sInDollarsratingCostVehicleOpekmLinkLengtharsCostInDoll )(

Generalised Time for each link:

)/($60

houreValueOfTim

arsCostInDollimeMinutesOfTdTimeGeneralise

Total Generalised Time for non-expressway route;

nksnumberofliforl

ldTimeGeneralisewayTimeNonExpress...1

Total Generalised Time for expressway route;

nksnumberofliforl

ldTimeGeneralisehoureValueOfTim

ollDollarsOfTTimeExpressway

...1)/($

60

Share of expressway trips (logit relationship):

altyTollingPenTimeExpresswaywayTimeNonExpressLambdaeShareExpressway

1

1




Appendix 19: Risk Simulation Modelling: Results by Parameter

Variable: Capacity per Expressway Lane (pcu's)

Minimum: 20,000

Maximum: 28,000

Mean: 23,995

Monte Carlo Settings:

Modal Value 24,000

% < Modal 50%

SD (<Modal) 2,000

SD (>Modal) 2,000

Regression Analysis: FIRR on Variable

Conventional Case

FIRR = 0.14863 + 0.000001 * Variable

R2= 0.400

FIRR = 0.14985 * 1.000005 ^ Variable

R2= 0.398

Respondents' Case

FIRR = 0.14586 + 0.000001 * Variable

R2= 0.505

FIRR = 0.14832 * 1.000007 ^ Variable

R2= 0.500

Kondratieff Case

FIRR = 0.09223 + 0.000003 * Variable

R2= 0.656

FIRR = 0.10269 * 1.000017 ^ Variable

R2= 0.654

Chance of Failure by Variable Value and Forecast Case

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20,

000

21,

000

22,

000

23,

000

24,

000

25,

000

26,

000

27,

000

28,

000

Conventional Respondents Kondatrieff

Distribution of Variable

0%

20%

40%

60%

80%

100%

20,

000

21,

000

22,

000

23,

000

24,

000

25,

000

26,

000

27,

000

28,

000

0%

1%

2%

3%

4%

5%

6%

Cumulative (Left Axis) Probability Density (Right Axis)

Mean FIRR by Variable Value and Forecast Case

10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

20,

000

21,

000

22,

000

23,

000

24,

000

25,

000

26,

000

27,

000

28,

000





Variable: Capacity per Local Lane (pcu's)

Minimum: 8,000

Maximum: 12,000

Mean: 9,994


Modal Value 10,000

% < Modal 50%

SD (<Modal) 1,000

SD (>Modal) 1,000


Conventional Case

FIRR = 0.17226 + -1.24E-07 * Variable

R2= 0.004

FIRR = 0.17223 * 0.999999 ^ Variable

R2= 0.004

Respondents' Case

FIRR = 0.17887 + -1.04E-07 * Variable

R2= 0.002

FIRR = 0.17885 * 0.999999 ^ Variable

R2= 0.002

Kondratieff Case

FIRR = 0.16067 + -6.81E-07 * Variable

R2= 0.032

FIRR = 0.16086 * 0.999996 ^ Variable

R2= 0.033


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

8,0

00

8,5

00

9,0

00

9,5

00

10,

000

10,

500

11,

000

11,

500

12,

000



0%

20%

40%

60%

80%

100%

8,0

00

8,5

00

9,0

00

9,5

00

10,

000

10,

500

11,

000

11,

500

12,

000

0%

1%

2%

3%

4%

5%

6%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

8,0

00

8,5

00

9,0

00

9,5

00

10,

000

10,

500

11,

000

11,

500

12,

000





Variable: Total Construction Cost (Base Year million$)

Minimum: 166.79 million

Maximum: 240.92 million

Mean: 199.77 million


Modal Value 185.3 USDm*

% < Modal 25%

SD (<Modal) 5% USDm**

SD (>Modal) 15% USDm**

Note: * USD4.633 per km

Note: ** SD as % of Modal Value


Conventional Case

FIRR = 0.2952068 + -0.00062 * Variable

R2= 0.960

FIRR = 0.3574268 * 0.996323 ^ Variable

R2= 0.955

Respondents' Case

FIRR = 0.3122438 + -0.00067 * Variable

R2= 0.961

FIRR = 0.3837237 * 0.996162 ^ Variable

R2= 0.955

Kondratieff Case

FIRR = 0.3220731 + -0.00083 * Variable

R2= 0.941

FIRR = 0.4713168 * 0.994423 ^ Variable

R2= 0.929


0%

5%

10%

15%

20%

25%

30%

167.

5

177.

5

187.

5

197.

5

207.

5

217.

5

227.

5

237.

5



0%

20%

40%

60%

80%

100%

167.

5

177.

5

187.

5

197.

5

207.

5

217.

5

227.

5

237.

5

0%

1%

2%

3%

4%

5%

6%



10%

12%

14%

16%

18%

20%

22%

167.

5

177.

5

187.

5

197.

5

207.

5

217.

5

227.

5

237.

5





Variable: Construction Duration (Quarters)

Minimum: 8

Maximum: 14

Mean: 11.08


Modal Value 10

% < Modal 20%

SD (<Modal) 1

SD (>Modal) 2


Conventional Case

FIRR = 0.19502 + -0.00217 * Variable

R2= 0.989

FIRR = 0.19672 * 0.987384 ^ Variable

R2= 0.989

Respondents' Case

FIRR = 0.20836 + -0.00276 * Variable

R2= 0.988

FIRR = 0.21105 * 0.984577 ^ Variable

R2= 0.987

Kondratieff Case

FIRR = 0.21041 + -0.00504 * Variable

R2= 0.968

FIRR = 0.2212 * 0.967949 ^ Variable

R2= 0.969


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

8 9 10 11 12 13 14



0%

20%

40%

60%

80%

100%

8 9 10 11 12 13 14

0%

5%

10%

15%

20%

25%

30%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

8 9 10 11 12 13 14





Variable: Fixed Annual Operations and Maintenance Costs (as % of Construction Cost)

Minimum: 0.1%

Maximum: 4.0%

Mean: 2.2%


Modal Value 2% *

% < Modal 50%

SD (<Modal) 0.5%

SD (>Modal) 1.0%

Note: * as % of initial construction cost


Conventional Case

FIRR = 0.18088 + -0.46747 * Variable

R2= 0.170

FIRR = 0.17939 * 0.088485 ^ Variable

R2= 0.128

Respondents' Case

FIRR = 0.19135 + -0.58727 * Variable

R2= 0.329

FIRR = 0.19064 * 0.044776 ^ Variable

R2= 0.316

Kondratieff Case

FIRR = 0.17735 + -0.96474 * Variable

R2= 0.504

FIRR = 0.17678 * 0.002841 ^ Variable

R2= 0.536


0%

5%

10%

15%

20%

25%

0.1% 1.1% 2.1% 3.1% 4.1%



0%

20%

40%

60%

80%

100%

0.1% 1.1% 2.1% 3.1% 4.1%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.1% 1.1% 2.1% 3.1% 4.1%





Variable: Variable Operations and Maintenance Costs (as % of Revenues)

Minimum: 1.0%

Maximum: 5.0%

Mean: 3.0%


Modal Value 3% *

% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

Note: * as % of revenues


Conventional Case

FIRR = 0.17554 + -0.15022 * Variable

R2= 0.468

FIRR = 0.17557 * 0.415625 ^ Variable

R2= 0.468

Respondents' Case

FIRR = 0.1825 + -0.15471 * Variable

R2= 0.434

FIRR = 0.18253 * 0.419220 ^ Variable

R2= 0.434

Kondratieff Case

FIRR = 0.15889 + -0.15932 * Variable

R2= 0.159

FIRR = 0.1588 * 0.361829 ^ Variable

R2= 0.154


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

1.00% 1.80% 2.60% 3.40% 4.20% 5.00%



0%

20%

40%

60%

80%

100%

1.00% 1.80% 2.60% 3.40% 4.20% 5.00%

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

1.00% 1.80% 2.60% 3.40% 4.20% 5.00%





Variable: Base Value of Time ($/hr) for Small Vehicles

Minimum: 2.00 $/hour

Maximum: 6.00 $/hour

Mean: 4.00 $/hour


Modal Value 4

% < Modal 50%

SD (<Modal) 1

SD (>Modal) 1


Conventional Case

FIRR = 0.17092 + 0.000061 * Variable

R2= 0.001

FIRR = 0.17089 * 1.000384 ^ Variable

R2= 0.001

Respondents' Case

FIRR = 0.17837 + -0.00005 * Variable

R2= 0.001

FIRR = 0.17835 * 0.999741 ^ Variable

R2= 0.001

Kondratieff Case

FIRR = 0.15427 + 0.000071 * Variable

R2= 0.0004

FIRR = 0.15419 * 1.000505 ^ Variable

R2= 0.0005


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

2.0 2.8 3.6 4.4 5.2 6.0



0%

20%

40%

60%

80%

100%

2.0 2.8 3.6 4.4 5.2 6.0

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

2.0 2.8 3.6 4.4 5.2 6.0





Variable: Base Value of Time ($/hr) for Large Vehicles

Minimum: 1.00 $/hour

Maximum: 5.00 $/hour

Mean: 3.01 $/hour


Modal Value 3

% < Modal 50%

SD (<Modal) 1

SD (>Modal) 1


Conventional Case

FIRR = 0.17087 + 0.00004 * Variable

R2= 0.001

FIRR = 0.17086 * 1.000212 ^ Variable

R2= 0.001

Respondents' Case

FIRR = 0.17881 + -0.00031 * Variable

R2= 0.032

FIRR = 0.1788 * 0.998246 ^ Variable

R2= 0.032

Kondratieff Case

FIRR = 0.15591 + -0.00055 * Variable

R2= 0.032

FIRR = 0.15588 * 0.996423 ^ Variable

R2= 0.032


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

1.0 1.8 2.6 3.4 4.2 5.0



0%

20%

40%

60%

80%

100%

1.0 1.8 2.6 3.4 4.2 5.0

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

1.0 1.8 2.6 3.4 4.2 5.0





Variable: Income Elasticity of Value of Time (Small Vehicles)

Minimum: 0.20

Maximum: 0.80

Mean: 0.50


Modal Value 0.5

% < Modal 50%

SD (<Modal) 0.15

SD (>Modal) 0.15


Conventional Case

FIRR = 0.1719 + -0.00074 * Variable

R2= 0.007

FIRR = 0.17187 * 0.995842 ^ Variable

R2= 0.007

Respondents' Case

FIRR = 0.17828 + 0.000159 * Variable

R2= 0.0003

FIRR = 0.17825 * 1.001120 ^ Variable

R2= 0.0004

Kondratieff Case

FIRR = 0.15602 + -0.00228 * Variable

R2= 0.027

FIRR = 0.15593 * 0.986271 ^ Variable

R2= 0.024


0%

2%

4%

6%

8%

10%

12%

14%

16%

0.20 0.40 0.60 0.80



0%

20%

40%

60%

80%

100%

0.20 0.40 0.60 0.80

0%

2%

4%

6%

8%

10%

12%

14%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.20 0.40 0.60 0.80





Variable: Income Elasticity of Value of Time (Large Vehicles)

Minimum: 0.20

Maximum: 0.80

Mean: 0.50


Modal Value 0.5

% < Modal 50%

SD (<Modal) 0.15

SD (>Modal) 0.15


Conventional Case

FIRR = 0.17136 + -0.00137 * Variable

R2= 0.031

FIRR = 0.17135 * 0.991996 ^ Variable

R2= 0.032

Respondents' Case

FIRR = 0.17885 + -0.00275 * Variable

R2= 0.136

FIRR = 0.17885 * 0.984583 ^ Variable

R2= 0.136

Kondratieff Case

FIRR = 0.15044 + 0.00561 * Variable

R2= 0.118

FIRR = 0.15042 * 1.037364 ^ Variable

R2= 0.120


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

0.20 0.40 0.60 0.80



0%

20%

40%

60%

80%

100%

0.20 0.40 0.60 0.80

0%

2%

4%

6%

8%

10%

12%

14%

16%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.20 0.40 0.60 0.80





Variable: Base Expressway Vehicle Operating Costs ($/km) for Small Vehicles

Minimum: 0.03

Maximum: 0.09

Mean: 0.06


Modal Value 0.06

% < Modal 50%

SD (<Modal) 0.015

SD (>Modal) 0.015


Conventional Case

FIRR = 0.16901 + 0.033129 * Variable

R2= 0.130

FIRR = 0.169 * 1.214432 ^ Variable

R2= 0.131

Respondents' Case

FIRR = 0.17596 + 0.030190 * Variable

R2= 0.102

FIRR = 0.17595 * 1.185655 ^ Variable

R2= 0.102

Kondratieff Case

FIRR = 0.1515 + 0.041592 * Variable

R2= 0.067

FIRR = 0.15152 * 1.305841 ^ Variable

R2= 0.066


0%

2%

4%

6%

8%

10%

12%

14%

16%

0.030 0.050 0.070 0.090



0%

20%

40%

60%

80%

100%

0.030 0.050 0.070 0.090

0%

2%

4%

6%

8%

10%

12%

14%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.030 0.050 0.070 0.090





Variable: Base Expressway Vehicle Operating Costs ($/km) for Large Vehicles

Minimum: 0.05

Maximum: 0.15

Mean: 0.10


Modal Value 0.1

% < Modal 50%

SD (<Modal) 0.025

SD (>Modal) 0.025


Conventional Case

FIRR = 0.16677 + 0.037643 * Variable

R2= 0.359

FIRR = 0.1668 * 1.247046 ^ Variable

R2= 0.361

Respondents' Case

FIRR = 0.17422 + 0.031820 * Variable

R2= 0.251

FIRR = 0.17424 * 1.196431 ^ Variable

R2= 0.251

Kondratieff Case

FIRR = 0.15235 + 0.01331 * Variable

R2= 0.013

FIRR = 0.15238 * 1.085659 ^ Variable

R2= 0.012


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

0.050 0.070 0.090 0.110 0.130 0.150



0%

20%

40%

60%

80%

100%

0.050 0.070 0.090 0.110 0.130 0.150

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.050 0.070 0.090 0.110 0.130 0.150





Variable: Local Roads VOC Multiplier for Small Vehicles (Expressway VOC * Factor)

Minimum: 1.00

Maximum: 2.00

Mean: 1.50


Modal Value 1.5

% < Modal 50%

SD (<Modal) 0.25

SD (>Modal) 0.25


Conventional Case

FIRR = 0.16777 + 0.002163 * Variable

R2= 0.141

FIRR = 0.1678 * 1.012712 ^ Variable

R2= 0.141

Respondents' Case

FIRR = 0.17628 + 0.001067 * Variable

R2= 0.038

FIRR = 0.1763 * 1.005942 ^ Variable

R2= 0.037

Kondratieff Case

FIRR = 0.15063 + 0.002226 * Variable

R2= 0.040

FIRR = 0.15068 * 1.014337 ^ Variable

R2= 0.039


0%

2%

4%

6%

8%

10%

12%

14%

16%

1.0

00

1.2

00

1.4

00

1.6

00

1.8

00

2.0

00



0%

20%

40%

60%

80%

100%

1.0

00

1.2

00

1.4

00

1.6

00

1.8

00

2.0

00

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

1.0

00

1.2

00

1.4

00

1.6

00

1.8

00

2.0

00





Variable: Local Roads VOC Multiplier for Large Vehicles (Expressway VOC * Factor)

Minimum: 1.50

Maximum: 2.50

Mean: 2.00


Modal Value 2

% < Modal 50%

SD (<Modal) 0.25

SD (>Modal) 0.25


Conventional Case

FIRR = 0.16216 + 0.004343 * Variable

R2= 0.258

FIRR = 0.16234 * 1.025811 ^ Variable

R2= 0.258

Respondents' Case

FIRR = 0.16966 + 0.003963 * Variable

R2= 0.229

FIRR = 0.1698 * 1.022631 ^ Variable

R2= 0.228

Kondratieff Case

FIRR = 0.13897 + 0.00737 * Variable

R2= 0.335

FIRR = 0.13942 * 1.049863 ^ Variable

R2= 0.335


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

1.5

00

1.7

00

1.9

00

2.1

00

2.3

00

2.5

00



0%

20%

40%

60%

80%

100%

1.5

00

1.7

00

1.9

00

2.1

00

2.3

00

2.5

00

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

1.5

00

1.7

00

1.9

00

2.1

00

2.3

00

2.5

00





Variable: Base Traffic Factor (Small Vehicles) to Expand/Contract Initial Demand

Minimum: 0.70

Maximum: 1.30

Mean: 1.00


Modal Value 1

% < Modal 50%

SD (<Modal) 0.15

SD (>Modal) 0.15


Conventional Case

FIRR = 0.10093 + 0.068177 * Variable

R2= 0.994

FIRR = 0.11246 * 1.499343 ^ Variable

R2= 0.991

Respondents' Case

FIRR = 0.10692 + 0.068921 * Variable

R2= 0.996

FIRR = 0.11823 * 1.483236 ^ Variable

R2= 0.993

Kondratieff Case

FIRR = 0.06407 + 0.087354 * Variable

R2= 0.989

FIRR = 0.08385 * 1.795023 ^ Variable

R2= 0.980


0%

5%

10%

15%

20%

25%

0.7

00

0.9

00

1.1

00

1.3

00



0%

20%

40%

60%

80%

100%

0.7

00

0.9

00

1.1

00

1.3

00

0%

2%

4%

6%

8%

10%

12%

14%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.7

00

0.9

00

1.1

00

1.3

00





Variable: Base Traffic Factor (Large Vehicles) to Expand/Contract Initial Demand

Minimum: 0.70

Maximum: 1.30

Mean: 1.00


Modal Value 1

% < Modal 50%

SD (<Modal) 0.15

SD (>Modal) 0.15


Conventional Case

FIRR = 0.10347 + 0.066009 * Variable

R2= 0.975

FIRR = 0.1144 * 1.477431 ^ Variable

R2= 0.973

Respondents' Case

FIRR = 0.10919 + 0.067103 * Variable

R2= 0.970

FIRR = 0.1201 * 1.464043 ^ Variable

R2= 0.968

Kondratieff Case

FIRR = 0.07515 + 0.07755 * Variable

R2= 0.968

FIRR = 0.09152 * 1.660623 ^ Variable

R2= 0.970


0%

5%

10%

15%

20%

25%

0.7

00

0.9

00

1.1

00

1.3

00



0%

20%

40%

60%

80%

100%

0.7

00

0.9

00

1.1

00

1.3

00

0%

2%

4%

6%

8%

10%

12%

14%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.7

00

0.9

00

1.1

00

1.3

00





Variable: Income Elasticity of Traffic (Small Vehicles)

Minimum: 0.85

Maximum: 1.65

Mean: 1.25


Modal Value 1.25

% < Modal 50%

SD (<Modal) 0.2

SD (>Modal) 0.2


Conventional Case

FIRR = 0.13436 + 0.028688 * Variable

R2= 0.950

FIRR = 0.13782 * 1.183292 ^ Variable

R2= 0.949

Respondents' Case

FIRR = 0.1383 + 0.030938 * Variable

R2= 0.960

FIRR = 0.14214 * 1.190834 ^ Variable

R2= 0.957

Kondratieff Case

FIRR = 0.09821 + 0.043622 * Variable

R2= 0.921

FIRR = 0.10639 * 1.332811 ^ Variable

R2= 0.910


0%

5%

10%

15%

20%

25%

0.9

00

1.1

00

1.3

00

1.5

00



0%

20%

40%

60%

80%

100%

0.9

00

1.1

00

1.3

00

1.5

00

0%

2%

4%

6%

8%

10%

12%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.9

00

1.1

00

1.3

00

1.5

00





Variable: Income Elasticity of Traffic (Large Vehicles)

Minimum: 0.70

Maximum: 1.50

Mean: 1.10


Modal Value 1.1

% < Modal 50%

SD (<Modal) 0.2

SD (>Modal) 0.2


Conventional Case

FIRR = 0.14173 + 0.026119 * Variable

R2= 0.967

FIRR = 0.14394 * 1.165400 ^ Variable

R2= 0.969

Respondents' Case

FIRR = 0.14477 + 0.029486 * Variable

R2= 0.969

FIRR = 0.14747 * 1.180804 ^ Variable

R2= 0.971

Kondratieff Case

FIRR = 0.10672 + 0.04219 * Variable

R2= 0.933

FIRR = 0.11279 * 1.317465 ^ Variable

R2= 0.933


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

0.7

00

0.9

00

1.1

00

1.3

00

1.5

00



0%

20%

40%

60%

80%

100%

0.7

00

0.9

00

1.1

00

1.3

00

1.5

00

0%

2%

4%

6%

8%

10%

12%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.7

00

0.9

00

1.1

00

1.3

00

1.5

00





Variable: Toll Revenue Leakage (%)

Minimum: 5%

Maximum: 20%

Mean: 11%


Modal Value 10%

% < Modal 50%

SD (<Modal) 2.5%

SD (>Modal) 5%


Conventional Case

FIRR = 0.18011 + -0.07629 * Variable

R2= 0.833

FIRR = 0.18033 * 0.639348 ^ Variable

R2= 0.833

Respondents' Case

FIRR = 0.18704 + -0.07777 * Variable

R2= 0.805

FIRR = 0.18728 * 0.644238 ^ Variable

R2= 0.801

Kondratieff Case

FIRR = 0.16656 + -0.10153 * Variable

R2= 0.788

FIRR = 0.16707 * 0.514851 ^ Variable

R2= 0.787


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

5% 9% 13%

17%

21%



0%

20%

40%

60%

80%

100%

5% 9% 13%

17%

21%

0%

2%

4%

6%

8%

10%

12%

14%

16%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

5% 9% 13%

17%

21%





Variable: Initial Amplitude of Ramp-Up (%)

Minimum: 20%

Maximum: 80%

Mean: 44%


Modal Value 40%

% < Modal 50%

SD (<Modal) 10%

SD (>Modal) 20%


Conventional Case

FIRR = 0.17962 + -0.01845 * Variable

R2= 0.861

FIRR = 0.17983 * 0.897345 ^ Variable

R2= 0.862

Respondents' Case

FIRR = 0.18678 + -0.01924 * Variable

R2= 0.866

FIRR = 0.18701 * 0.896982 ^ Variable

R2= 0.867

Kondratieff Case

FIRR = 0.16915 + -0.03187 * Variable

R2= 0.826

FIRR = 0.16997 * 0.810841 ^ Variable

R2= 0.818


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

20.0

%

40.0

%

60.0

%

80.0

%



0%

20%

40%

60%

80%

100%

20% 40% 60% 80%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

20.0

%

40.0

%

60.0

%

80.0

%





Variable: Ramp-Up Duration (Quarters)

Minimum: 4

Maximum: 20

Mean: 9.79


Modal Value 8

% < Modal 40%

SD (<Modal) 2

SD (>Modal) 5


Conventional Case

FIRR = 0.18002 + -0.00094 * Variable

R2= 0.792

FIRR = 0.18042 * 0.994412 ^ Variable

R2= 0.785

Respondents' Case

FIRR = 0.18777 + -0.00103 * Variable

R2= 0.788

FIRR = 0.18822 * 0.994098 ^ Variable

R2= 0.780

Kondratieff Case

FIRR = 0.1693 + -0.00156 * Variable

R2= 0.834

FIRR = 0.17053 * 0.989561 ^ Variable

R2= 0.822


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20



0%

20%

40%

60%

80%

100%

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

0%

2%

4%

6%

8%

10%

12%

14%

16%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20





Variable: Small Vehicles' Tolling Penalty (Minutes)

Minimum: 0

Maximum: 20

Mean: 10


Modal Value 10

% < Modal 50%

SD (<Modal) 5

SD (>Modal) 5


Conventional Case

FIRR = 0.17277 + -0.00015 * Variable

R2= 0.240

FIRR = 0.17275 * 0.999112 ^ Variable

R2= 0.238

Respondents' Case

FIRR = 0.17982 + -0.00017 * Variable

R2= 0.291

FIRR = 0.1798 * 0.999047 ^ Variable

R2= 0.290

Kondratieff Case

FIRR = 0.15837 + -0.00035 * Variable

R2= 0.417

FIRR = 0.15833 * 0.997740 ^ Variable

R2= 0.422


0%

2%

4%

6%

8%

10%

12%

14%

16%

0 4 8 12 16 20



0%

20%

40%

60%

80%

100%

0 4 8 12 16 20

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0 4 8 12 16 20





Variable: Large Vehicles' Tolling Penalty (Minutes)

Minimum: 5

Maximum: 25

Mean: 15


Modal Value 15

% < Modal 50%

SD (<Modal) 5

SD (>Modal) 5


Conventional Case

FIRR = 0.17198 + -0.00006 * Variable

R2= 0.042

FIRR = 0.17195 * 0.999631 ^ Variable

R2= 0.040

Respondents' Case

FIRR = 0.17833 + -0.00003 * Variable

R2= 0.007

FIRR = 0.1783 * 0.999852 ^ Variable

R2= 0.007

Kondratieff Case

FIRR = 0.15341 + 0.00003 * Variable

R2= 0.005

FIRR = 0.1534 * 1.000199 ^ Variable

R2= 0.004


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

5 9 13 17 21 25



0%

20%

40%

60%

80%

100%

5 9 13 17 21 25

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

5 9 13 17 21 25





Variable: Routeing Sensitivity ("Lambda") for Small Vehicles

Minimum: 0.025

Maximum: 0.075

Mean: 0.050


Modal Value 0.05

% < Modal 50%

SD (<Modal) 0.0125

SD (>Modal) 0.0125


Conventional Case

FIRR = 0.16919 + 0.02656 * Variable

R2= 0.040

FIRR = 0.16917 * 1.170635 ^ Variable

R2= 0.040

Respondents' Case

FIRR = 0.17557 + 0.03624 * Variable

R2= 0.084

FIRR = 0.17553 * 1.231156 ^ Variable

R2= 0.086

Kondratieff Case

FIRR = 0.15357 + 0.00169 * Variable

R2= 0.0001

FIRR = 0.15348 * 1.018592 ^ Variable

R2= 0.0003


0%

2%

4%

6%

8%

10%

12%

14%

16%

0.0

25

0.0

45

0.0

65



0%

20%

40%

60%

80%

100%

0.0

25

0.0

45

0.0

65

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.0

25

0.0

45

0.0

65





Variable: Routeing Sensitivity ("Lambda") for Large Vehicles

Minimum: 0.025

Maximum: 0.075

Mean: 0.050


Modal Value 0.05

% < Modal 50%

SD (<Modal) 0.0125

SD (>Modal) 0.0125


Conventional Case

FIRR = 0.17011 + 0.01230 * Variable

R2= 0.030

FIRR = 0.1701 * 1.075460 ^ Variable

R2= 0.030

Respondents' Case

FIRR = 0.1773 + 0.00671 * Variable

R2= 0.010

FIRR = 0.17729 * 1.038873 ^ Variable

R2= 0.010

Kondratieff Case

FIRR = 0.15144 + 0.04419 * Variable

R2= 0.112

FIRR = 0.15142 * 1.336206 ^ Variable

R2= 0.114


0%

2%

4%

6%

8%

10%

12%

14%

16%

0.0

25

0.0

45

0.0

65



0%

20%

40%

60%

80%

100%

0.0

25

0.0

45

0.0

65

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

0.0

25

0.0

45

0.0

65





Variable: Toll Escalation Rate (% of RPI Inflation)

Minimum: 60%

Maximum: 100%

Mean: 86%


Modal Value 90%

% < Modal 50%

SD (<Modal) 15%

SD (>Modal) 5%


Conventional Case

FIRR = 0.13572 + 0.04035 * Variable

R2= 0.861

FIRR = 0.13862 * 1.270939 ^ Variable

R2= 0.856

Respondents' Case

FIRR = 0.13126 + 0.05336 * Variable

R2= 0.917

FIRR = 0.13604 * 1.358413 ^ Variable

R2= 0.912

Kondratieff Case

FIRR = 0.07763 + 0.08760 * Variable

R2= 0.964

FIRR = 0.09193 * 1.803628 ^ Variable

R2= 0.960


0%

5%

10%

15%

20%

25%

62.5% 72.5% 82.5% 92.5%



0%

20%

40%

60%

80%

100%

62.5% 72.5% 82.5% 92.5%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

62.5% 72.5% 82.5% 92.5%





Variable: Quarters between Toll Increases

Minimum: 8

Maximum: 20

Mean: 13.29


Modal Value 12

% < Modal 40%

SD (<Modal) 2

SD (>Modal) 4


Conventional Case

FIRR = 0.17765 + -0.00051 * Variable

R2= 0.565

FIRR = 0.17781 * 0.997028 ^ Variable

R2= 0.561

Respondents' Case

FIRR = 0.18744 + -0.00073 * Variable

R2= 0.690

FIRR = 0.18773 * 0.995904 ^ Variable

R2= 0.687

Kondratieff Case

FIRR = 0.17121 + -0.00129 * Variable

R2= 0.850

FIRR = 0.17223 * 0.991630 ^ Variable

R2= 0.851


0%

2%

4%

6%

8%

10%

12%

14%

16%

8 9 10 11 12 13 14 15 16 17 18 19 20



0%

20%

40%

60%

80%

100%

8 9 10 11 12 13 14 15 16 17 18 19 20

0%

2%

4%

6%

8%

10%

12%

14%

16%



10%

11%

12%

13%

14%

15%

16%

17%

18%

19%

20%

8 9 10 11 12 13 14 15 16 17 18 19 20





Variable: GDP Growth (% p.a.) Page 1 of 2

Conventional

Minimum: 2%

Maximum: 10%

Mean: 6%

Modal Value 6%

% < Modal 50%

SD (<Modal) 2%

SD (>Modal) 2%

Respondents'

Minimum: 2%

Maximum: 12%

Mean: 7%

Modal Value 1% added to Conventional

% < Modal 50%

SD (<Modal) 0.5%

SD (>Modal) 0.5%

Kondratieff

Minimum: 2%

Maximum: 14%

Mean: 8%

Modal Value 1% added to Respondents'

% < Modal 50%

SD (<Modal) 0.5%

SD (>Modal) 0.5%

All Cases

Minimum: 2%

Maximum: 14%

Mean: 7%

Modal Value n/a

% < Modal n/a

SD (<Modal) n/a

SD (>Modal) n/a

Distribution of Variable: Conventional

0%

20%

40%

60%

80%

100%

2% 4% 6% 8% 10%

0%

2%

4%

6%

8%

10%

12%


Distribution of Variable: Respondents'

0%

20%

40%

60%

80%

100%

2% 4% 6% 8% 10% 12%

0%

2%

4%

6%

8%

10%

12%


Distribution of Variable: Kondratieff

0%

20%

40%

60%

80%

100%

2% 4% 6% 8% 10% 12% 14%

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%


Distribution of Variable: All Cases

0%

20%

40%

60%

80%

100%

2% 4% 6% 8% 10% 12% 14%

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%





Variable: GDP Growth (% p.a.) Page 2 of 2


Conventional Case

FIRR = 0.07953 + 1.408371 * Variable

R2= 0.962

FIRR = 0.0923 * 8818 ^ Variable

R2= 0.908

Respondents' Case

FIRR = 0.07433 + 1.349827 * Variable

R2= 0.935

FIRR = 0.08764 * 6825 ^ Variable

R2= 0.860

Kondratieff Case

FIRR = 0.03262 + 1.384828 * Variable

R2= 0.938

FIRR = 0.05548 * 55033 ^ Variable

R2= 0.855

All Cases

FIRR = 0.09232 + 0.94170 * Variable

R2= 0.905

FIRR = 0.09977 * 472 ^ Variable

R2= 0.854


0%

10%

20%

30%

40%

50%

60%

2% 4% 6% 8% 10% 12% 14%

Conventional Respondents Kondatrieff All


0%

5%

10%

15%

20%

25%

2% 4% 6% 8% 10% 12% 14%





Variable: Price Inflation for Vehicle Operating Costs (% p.a.) Page 1 of 2

Conventional

Minimum: 0%

Maximum: 5%

Mean: 2%

Modal Value 2.5%

% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

Respondents'

Minimum: 0%

Maximum: 11%

Mean: 5%

Modal Value 2.5% added to Conventional

% < Modal 50%

SD (<Modal) 1.5%

SD (>Modal) 1.5%

Kondratieff

Minimum: 0%

Maximum: 11%

Mean: 5%

Modal Value 0% same as Respondents'

% < Modal n/a

SD (<Modal) n/a

SD (>Modal) n/a

All Cases

Minimum: 0%

Maximum: 11%

Mean: 4%

Modal Value n/a

% < Modal n/a

SD (<Modal) n/a

SD (>Modal) n/a


0%

20%

40%

60%

80%

100%

0% 1% 2% 3% 4%

0%

5%

10%

15%

20%

25%



0%

20%

40%

60%

80%

100%

0% 2% 4% 6% 8% 10%

0%

2%

4%

6%

8%

10%

12%

14%



0%

20%

40%

60%

80%

100%

1% 3% 5% 7% 9% 11% 13%

0%

2%

4%

6%

8%

10%

12%

14%



0%

20%

40%

60%

80%

100%

0% 2% 4% 6% 8% 10% 12%

0%

2%

4%

6%

8%

10%

12%





Variable: Price Inflation for Vehicle Operating Costs (% p.a.) Page 2 of 2


Conventional Case

FIRR = 0.17127 + -0.00581 * Variable

R2= 0.003

FIRR = 0.17126 * 0.967 ^ Variable

R2= 0.003

Respondents' Case

FIRR = 0.1764 + 0.041954 * Variable

R2= 0.090

FIRR = 0.17628 * 1.274 ^ Variable

R2= 0.097

Kondratieff Case

FIRR = 0.15218 + 0.036530 * Variable

R2= 0.031

FIRR = 0.15202 * 1.271 ^ Variable

R2= 0.031

All Cases

FIRR = 0.16869 + -0.00970 * Variable

R2= 0.009

FIRR = 0.16869 * 0.941 ^ Variable

R2= 0.010


0%

5%

10%

15%

20%

25%

0% 2% 4% 6% 8% 10% 12%



0%

5%

10%

15%

20%

25%

0% 2% 4% 6% 8% 10% 12%





Variable: Price Inflation for Construction and (Fixed) Operations & Maintenance Costs (% p.a.) Page 1 of 2

Conventional

Minimum: 0%

Maximum: 5%

Mean: 3%

Modal Value 2.5%

% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

Respondents'

Minimum: 1%

Maximum: 6%

Mean: 3%


% < Modal 50%

SD (<Modal) 0.25%

SD (>Modal) 0.25%

Kondratieff

Minimum: 1%

Maximum: 8%

Mean: 4%


% < Modal 50%

SD (<Modal) 0.5%

SD (>Modal) 0.5%

All Cases

Minimum: 0%

Maximum: 8%

Mean: 3%

Modal Value n/a

% < Modal n/a

SD (<Modal) n/a

SD (>Modal) n/a


0%

20%

40%

60%

80%

100%

0% 1% 2% 3% 4%

0%

5%

10%

15%

20%

25%



0%

20%

40%

60%

80%

100%

0% 1% 2% 3% 4% 5%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%



0%

20%

40%

60%

80%

100%

1% 2% 3% 4% 5% 6% 7% 8%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%



0%

20%

40%

60%

80%

100%

0% 1% 2% 3% 4% 5% 6% 7%

0%

2%

4%

6%

8%

10%

12%

14%

16%





Variable: Price Inflation for Construction and (Fixed) Operations & Maintenance Costs (% p.a.) Page 2 of 2


Conventional Case

FIRR = 0.17777 + -0.25581 * Variable

R2= 0.923

FIRR = 0.17789 * 0.2225 ^ Variable

R2= 0.920

Respondents' Case

FIRR = 0.18585 + -0.22984 * Variable

R2= 0.780

FIRR = 0.186 * 0.2737 ^ Variable

R2= 0.775

Kondratieff Case

FIRR = 0.15982 + -0.06411 * Variable

R2= 0.024

FIRR = 0.15966 * 0.6604 ^ Variable

R2= 0.025

All Cases

FIRR = 0.17853 + -0.31342 * Variable

R2= 0.503

FIRR = 0.17888 * 0.1471 ^ Variable

R2= 0.488


0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

0% 1% 2% 3% 4% 5% 6% 7%



0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

0% 1% 2% 3% 4% 5% 6% 7%





Variable: General Price Inflation (% p.a.) Page 1 of 2

Conventional

Minimum: 0%

Maximum: 5%

Mean: 3%

Modal Value 2.5%

% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

Respondents'

Minimum: 1%

Maximum: 6%

Mean: 3%


% < Modal 50%

SD (<Modal) 0.25%

SD (>Modal) 0.25%

Kondratieff

Minimum: 1%

Maximum: 8%

Mean: 4%


% < Modal 50%

SD (<Modal) 0.5%

SD (>Modal) 0.5%

All Cases

Minimum: 0%

Maximum: 8%

Mean: 3%

Modal Value n/a

% < Modal n/a

SD (<Modal) n/a

SD (>Modal) n/a


0%

20%

40%

60%

80%

100%

0% 1% 2% 3% 4%

0%

5%

10%

15%

20%

25%



0%

20%

40%

60%

80%

100%

0% 1% 2% 3% 4% 5%

0%

5%

10%

15%

20%

25%



0%

20%

40%

60%

80%

100%

1% 2% 3% 4% 5% 6% 7% 8%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%



0%

20%

40%

60%

80%

100%

0% 1% 2% 3% 4% 5% 6% 7%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%





Variable: General Price Inflation (% p.a.) Page 2 of 2


Conventional Case

FIRR = 0.14125 + 1.072175 * Variable

R2= 0.973

FIRR = 0.1429 * 554 ^ Variable

R2= 0.967

Respondents' Case

FIRR = 0.13892 + 1.069124 * Variable

R2= 0.931

FIRR = 0.14147 * 468 ^ Variable

R2= 0.929

Kondratieff Case

FIRR = 0.06937 + 1.802280 * Variable

R2= 0.945

FIRR = 0.07889 * 642111 ^ Variable

R2= 0.835

All Cases

FIRR = 0.14241 + 0.71620 * Variable

R2= 0.914

FIRR = 0.14446 * 61 ^ Variable

R2= 0.931


0%

10%

20%

30%

40%

50%

60%

0% 2% 4% 6%



0%

5%

10%

15%

20%

25%

0% 1% 2% 3% 4% 5% 6% 7%





Variable: Interest Rates for Initial Debt (% p.a.) Page 1 of 2

Conventional

Minimum: 3%

Maximum: 7%

Mean: 5%

Modal Value 5%

% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

Respondents'

Minimum: 3%

Maximum: 9%

Mean: 6%

Modal Value 1% added to Conventional

% < Modal 50%

SD (<Modal) 0.5%

SD (>Modal) 0.5%

Kondratieff

Minimum: 3%

Maximum: 13%

Mean: 8%


% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

All Cases

Minimum: 3%

Maximum: 13%

Mean: 6%

Modal Value n/a

% < Modal n/a

SD (<Modal) n/a

SD (>Modal) n/a


0%

20%

40%

60%

80%

100%

3% 4% 5% 6% 7%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%



0%

20%

40%

60%

80%

100%

3% 4% 5% 6% 7% 8% 9%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%



0%

20%

40%

60%

80%

100%

3% 5% 7% 9% 11%

0%

2%

4%

6%

8%

10%

12%

14%



0%

20%

40%

60%

80%

100%

3% 5% 7% 9% 11% 13%

0%

2%

4%

6%

8%

10%

12%

14%





Variable: Interest Rates for Initial Debt (% p.a.) Page 2 of 2


Conventional Case

FIRR = 0.23277 + -1.19631 * Variable

R2= 0.970

FIRR = 0.24606 * 8.E-04 ^ Variable

R2= 0.959

Respondents' Case

FIRR = 0.27426 + -1.68129 * Variable

R2= 0.848

FIRR = 0.33297 * 1.E-05 ^ Variable

R2= 0.752

Kondratieff Case

FIRR = 0.36567 + -2.78234 * Variable

R2= 0.936

FIRR = 1.16299 * 1.E-13 ^ Variable

R2= 0.681

All Cases

FIRR = 0.29774 + -2.14979 * Variable

R2= 0.853

FIRR = 0.70788 * 1.74E-11 ^ Variable

R2= 0.603


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

3% 5% 7% 9% 11% 13%



0%

5%

10%

15%

20%

25%

30%

3% 5% 7% 9% 11% 13%





Variable: Interest Rates for Extra Debt (% p.a.) Page 1 of 2

Conventional

Minimum: 3%

Maximum: 11%

Mean: 7%

Modal Value 2% added to Initial Interest Rate

% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

Respondents'

Minimum: 3%

Maximum: 13%

Mean: 8%


% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

Kondratieff

Minimum: 4%

Maximum: 16%

Mean: 10%


% < Modal 50%

SD (<Modal) 1%

SD (>Modal) 1%

All Cases

Minimum: 3%

Maximum: 16%

Mean: 8%

Modal Value n/a

% < Modal n/a

SD (<Modal) n/a

SD (>Modal) n/a


0%

20%

40%

60%

80%

100%

3% 5% 7% 9% 11%

0%

5%

10%

15%

20%

25%

30%



0%

20%

40%

60%

80%

100%

3% 5% 7% 9% 11% 13%

0%

5%

10%

15%

20%

25%

30%



0%

20%

40%

60%

80%

100%

4% 6% 8% 10% 12% 14% 16%

0%

5%

10%

15%

20%

25%



0%

20%

40%

60%

80%

100%

3% 5% 7% 9% 11% 13% 15% 17%

0%

5%

10%

15%

20%

25%





Variable: Interest Rates for Extra Debt (% p.a.) Page 2 of 2


Conventional Case

FIRR = 0.22287 + -0.65067 * Variable

R2= 0.796

FIRR = 0.22774 * 2.53E-02 ^ Variable

R2= 0.799

Respondents' Case

FIRR = 0.26441 + -1.16853 * Variable

R2= 0.631

FIRR = 0.3541 * 6.37E-05 ^ Variable

R2= 0.492

Kondratieff Case

FIRR = 0.33942 + -1.94655 * Variable

R2= 0.945

FIRR = 0.5787 * 2.75E-07 ^ Variable

R2= 0.767

All Cases

FIRR = 0.28635 + -1.56141 * Variable

R2= 0.882

FIRR = 0.39919 * 5.59E-06 ^ Variable

R2= 0.686


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

3% 7% 11% 15%



0%

5%

10%

15%

20%

25%

30%

3% 7% 11% 15%





Appendix 20: Risk Simulation Modelling: Comparison of Parameters’ Impacts

Parameter Conventional Respondents' Kondratieff

Road Capacities

Capacity per

Expressway Lane (pcus)

Range 8,000 8,000 8,000

Impact 0.74% 1.06% 2.05%

R2 40% 51% 66%

Impact*R2 0.30% 0.53% 1.34%

+/- Impact +ve +ve +ve

Capacity per Local Road

Lane (pcus)

Range 4,000 4,000 4,000

Impact 0.05% 0.04% 0.27%

R2 0% 0% 3%

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

Construction Cost

(Base Year $m)

Range 74 74 74

Impact 4.58% 4.96% 6.19%

R2 96% 96% 94%

Impact*R2 4.39% 4.76% 5.82%


Construction Duration

(Quarters)

Range 6 6 6

Impact 1.30% 1.66% 3.03%

R2 99% 99% 97%

Impact*R2 1.29% 1.64% 2.93%


Sum of Impact * R2 5.68% 6.40% 8.75%

All O&M Costs

Fixed O&M Costs

(as % of Construction

Costs)

Range 4% 4% 4%

Impact 1.81% 2.27% 3.73%

R2 17% 33% 50%

Impact*R2 0.31% 0.75% 1.88%


Variable O&M Costs

(as % of Revenue)

Range 4% 4% 4%

Impact 0.60% 0.62% 0.64%

R2 47% 43% 16%

Impact*R2 0.28% 0.27% 0.10%


Sum of Impact * R2 0.59% 1.02% 1.98%

Value of Time & Its Income Elasticity

Small Vehicle VOT

($/hr)

Range 4 4 4

Impact 0.02% 0.02% 0.03%

R2 0% 0% 0%

Impact*R2 0.00% 0.00% 0.00%

+/- Impact +ve -ve +ve

Large Vehicle VOT

($/hr)

Range 4 4 4

Impact 0.01% 0.13% 0.22%

R2 0% 3% 3%

Impact*R2 0.00% 0.00% 0.01%

+/- Impact +ve -ve -ve

Income Elasticity of

VOT (Small Vehicles)

Range 1 1 1

Impact 0.04% 0.01% 0.14%

R2 1% 0% 3%





Impact*R2 0.00% 0.00% 0.00%

+/- Impact -ve +ve -ve

Income Elasticity of

VOT (Large Vehicles)

Range 1 1 1

Impact 0.08% 0.17% 0.34%

R2 3% 14% 12%

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

Small Vehicle

Expressway VOC

($/km)

Range 0.06 0.06 0.06

Impact 0.20% 0.18% 0.25%

R2 13% 10% 7%

Impact*R2 0.03% 0.02% 0.02%


Large Vehicle

Expressway VOC

($/km)

Range 0.10 0.10 0.10

Impact 0.38% 0.32% 0.13%

R2 36% 25% 1%

Impact*R2 0.14% 0.08% 0.00%


VOC Multiplier (Small

Vehicles on Local

Roads)

Range 1.0 1.0 1.0

Impact 0.22% 0.11% 0.22%

R2 14% 4% 4%

Impact*R2 0.03% 0.00% 0.01%


VOC Multiplier (Large

Vehicles on Local

Roads)

Range 1.0 1.0 1.0

Impact 0.43% 0.40% 0.74%

R2 26% 23% 34%

Impact*R2 0.11% 0.09% 0.25%


Sum of Impact * R2 0.30% 0.19% 0.27%

Demand (Initial & Income Elasticity)

Small Vehicle Demand

Range 60% 60% 60%

Impact 4.09% 4.14% 5.24%

R2 99% 100% 99%

Impact*R2 4.07% 4.12% 5.18%


Large Vehicle Demand

Range 60% 60% 60%

Impact 3.96% 4.03% 4.65%

R2 98% 97% 97%

Impact*R2 3.86% 3.91% 4.50%


Traffic Income Elasticity

(Small Vehicles)

Range 0.8 0.8 0.8

Impact 2.30% 2.48% 3.49%

R2 95% 96% 92%

Impact*R2 2.18% 2.38% 3.22%


Traffic Income Elasticity

(Large Vehicles)

Range 0.8 0.8 0.8

Impact 2.09% 2.36% 3.38%

R2 97% 97% 93%

Impact*R2 2.02% 2.28% 3.15%


Sum of Impact * R2 12.13% 12.69% 16.05%





Toll Revenue Leakage

Toll Revenue Leakage

(%)

Range 15% 15% 15%

Impact 1.14% 1.17% 1.52%

R2 83% 80% 79%

Impact*R2 0.95% 0.94% 1.20%


Sum of Impact * R2 0.95% 0.94% 1.20%

Ramp-Up: Amplitude & Duration

Initial Amplitude of

Ramp-Up (%)

Range 60% 60% 60%

Impact 1.11% 1.15% 1.91%

R2 86% 87% 83%

Impact*R2 0.95% 1.00% 1.58%


Ramp-Up Duration

(Quarters)

Range 16 16 16

Impact 1.50% 1.65% 2.49%

R2 79% 79% 83%

Impact*R2 1.19% 1.30% 2.08%


Sum of Impact * R2 2.14% 2.30% 3.66%

Logit Model Parameters

Small Vehicles Toll

Penalty (minutes)

Range 20 20 20

Impact 0.31% 0.34% 0.71%

R2 24% 29% 42%

Impact*R2 0.07% 0.10% 0.30%


Large Vehicles Toll

Penalty (minutes)

Range 20 20 20

Impact 0.13% 0.06% 0.06%

R2 4% 1% 0%

Impact*R2 0.01% 0.00% 0.00%

+/- Impact -ve -ve +ve

Small Vehicle Toll

Sensitivity ("Lambda")

Range 0.05 0.05 0.05

Impact 0.13% 0.18% 0.01%

R2 4% 8% 0%

Impact*R2 0.01% 0.02% 0.00%


Large Vehicle Toll

Sensitivity ("Lambda")

Range 0.05 0.05 0.05

Impact 0.06% 0.03% 0.22%

R2 3% 1% 11%

Impact*R2 0.00% 0.00% 0.02%


Sum of Impact * R2 0.09% 0.12% 0.32%

Toll Escalation Rate and Frequency

Toll Escalation Rate

(% of RPI)

Range 40% 40% 40%

Impact 1.61% 2.13% 3.50%

R2 86% 92% 96%

Impact*R2 1.39% 1.96% 3.38%


Quarters between Toll

Increases

Range 12 12 12

Impact 0.61% 0.87% 1.54%

R2 57% 69% 85%

Impact*R2 0.34% 0.60% 1.31%






Sum of Impact * R2 1.73% 2.56% 4.69%

GDP Growth

GDP Growth (% p.a.)

Range 8% 10% 12%

Impact 11.27% 13.50% 16.01%

R2 96% 94% 94%

Impact*R2 10.84% 12.62% 15.01%


Sum of Impact * R2 10.84% 12.62% 15.01%

Price Inflation

Vehicle Operating Cost

Inflation (% p.a.)

Range 4% 11% 11%

Impact 0.02% 0.45% 0.39%

R2 0% 9% 3%

Impact*R2 0.00% 0.04% 0.01%

+/- Impact -ve +ve +ve

Construction,

Operations, Maintenance

Cost Inflation (% p.a.)

Range 4% 5% 7%

Impact 1.02% 1.15% 0.46%

R2 92% 78% 2%

Impact*R2 0.94% 0.90% 0.01%


General Price Inflation

(% p.a.)

Range 4% 5% 7%

Impact 4.29% 5.35% 12.33%

R2 97% 93% 95%

Impact*R2 4.17% 4.98% 11.65%


Sum of Impact * R2 5.12% 5.91% 11.67%

Interest Rates

Initial Interest Rate

(% p.a.)

Range 4% 6% 10%

Impact 4.79% 10.09% 26.46%

R2 97% 85% 94%

Impact*R2 4.64% 8.56% 24.76%


Interest Rate for Extra

Debt (% p.a.)

Range 8% 10% 12%

Impact 5.21% 11.37% 23.94%

R2 80% 63% 94%

Impact*R2 4.14% 7.17% 22.62%


Sum of Impact * R2 8.78% 15.73% 47.38%




Rankings of Risk Categories by Case

Risk Group Conventional Respondents' Kondratieff

Road Capacities 11 10 9

Construction Cost & Duration 4 4 5

All O&M Costs 9 8 8

Value of Time & Its Income Elasticity 13 13 13

Vehicle Operating Costs 10 11 12

Demand (Initial & Income Elasticity) 1 2 2

Toll Revenue Leakage 8 9 10

Ramp-Up: Amplitude & Duration 6 7 7

Logit Model Parameters 12 12 11

Toll Escalation Rate and Frequency 7 6 6

GDP Growth 2 3 3

Price Inflation 5 5 4

Interest Rates 3 1 1

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

Technology

east asian toll roads

construction risks

dissertation richard

asian financial crisis

east asian countries

di bonaid

forecasts of toll road

interurban toll roads