what are the key risks associated with private investment in start up toll road projects in...
DESCRIPTION
MBA Dissertation, 2006TRANSCRIPT
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).
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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:
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 2 December 2006
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
private stakeholders in East Asian toll roads during a period of increasing price
inflation and interest rates”
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 3 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 4 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 5 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 6 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 7 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 8 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 9 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 10 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 11 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 12 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 13 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 14 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 15 December 2006
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)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 16 December 2006
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:
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 17 December 2006
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).
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 18 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 19 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 20 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 21 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 22 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 23 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 24 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 25 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 26 December 2006
(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
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 27 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 28 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 29 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 30 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 31 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 32 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 33 December 2006
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$4.633m per km of
Dual-2 was derived, to be used in Chapter 5’s simulation model.
There is a trade-off between construction and subsequent operations and maintenance
costs. The latter also affected by periodic major maintenance (e.g. immediately before
concession handback). Literature review found little agreement as to how to gauge such
costs, and whether they should be related to construction or traffic flow/ revenue. Table
2.5 shows some public domain values; some confidential sources suggested using 6% of
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 34 December 2006
initial construction cost. In practice, there is likely to be a fixed element which could be
taken as proportional to construction costs, plus a variable element proportional to
traffic/ revenue. For simulations in Chapter 5, it is recommended to adopt 2% of
construction cost (fixed) plus 3% of toll revenue (variable).
Table 2.4: Estimated Expressway Construction Costs
Expressway Length
Cost
(US$ 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)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 35 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 36 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 37 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 38 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 39 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 40 December 2006
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)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 41 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 42 December 2006
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$. Whilst time-series data on
other currencies were not available, Vietnamese Đong, Cambodian Riel, Myanmar
Kyats (free-market rate) and especially Lao Kip all depreciated substantially over this
period also. However, Figure 3.H16
shows that since January 2001 currencies have been
broadly stable; Myanmar Kyats (not shown) are the exception, continuing to devalue on
the free-market.
14 Raw data from: NESDB (2006) and www.fx.sauder.ubc.ca
15 Raw data from: www.bot.or.th and www.fx.sauder.ubc.ca
16 Raw data from: www.fx.sauder.ubc.ca
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 43 December 2006
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
19931994
19951996
19971998
19992000
20012002
20032004
20052006
Mil
lio
n B
ah
t (1
988 p
rices)
0%
10%
20%
30%
40%
50%
60%
Co
nstr
ucti
on
as %
of
GF
CF
Gross Fixed Capital Formation Private Construction
Government Construction Construction as % of GFCF
Figure 3.B: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter)
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
19931994
19951996
19971998
19992000
20012002
20032004
20052006
Mil
lio
n U
SD
(1988 p
rices)
0%
10%
20%
30%
40%
50%
60%
Co
nstr
ucti
on
as %
of
GF
CF
Gross Fixed Capital Formation Private Construction
Government Construction Construction as % of GFCF
Figure 3.C: Thai GFCF 1993-2006 (Rolling Annual Average by Quarter) in US$
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 44 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 45 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 46 December 2006
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).
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 47 December 2006
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
GDP Per Capita (USD p.a.)
Ln
(km
2 p
er
km
of
Paved
Ro
ad
)
Figure 3.J: Relationship between Wealth and Road Density
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 48 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 49 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 50 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 51 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 52 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 53 December 2006
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?
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 54 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 55 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 56 December 2006
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).
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 57 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 58 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 59 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 60 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 61 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 62 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 63 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 64 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 65 December 2006
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
19 Values transposed from raw data in Appendix 15.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 66 December 2006
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
20 Values transposed from raw data in Appendix 15.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 67 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 68 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 69 December 2006
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?
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 70 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 71 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 72 December 2006
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
Complete Sample Financial, Legal, Operator Transport Planner, Economist
Engineer, Architect Government, Aid Agency Academic
Other
No Market Nascent Developing Steady Over-Developed
Figure 4.N: Country Perceptions by Respondent Category
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 73 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 74 December 2006
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Fuel prices
General price inflation
Interest rates
Economic growth
Exchange rate
volatility
Tolling Acceptability
Complete Sample Financial, Legal, Operator Transport Planner, Economist
Engineer, Architect Government, Aid Agency Academic
Other
Significant
Decrease
Decrease to
an Extent No Change
Increase to
an Extent
Significant
Increase
Figure 4.P: Economic Expectations by Respondent Group
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 75 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 76 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 77 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 78 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 79 December 2006
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%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 80 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 81 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 82 December 2006
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,
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 83 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 84 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 85 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 86 December 2006
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 $308.9
Financial
Failure
# of Cases 55 114 1250
% of Cases 0.6% 1.1% 12.5% Note: * FIRR’s were not calculable once beneath 0%; hence 0% minimum value in all cases.
† Excludes 13 instances under Kondratieff case where no payback obtained.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 87 December 2006
0%
20%
40%
60%
80%
100%
5% 10% 15% 20% 25%
Cu
mu
lati
ve %
Conventional Respondents' Kondratieff
Figure 5.C: Cumulative Probability Distribution of FIRR (excluding FIRR<0%)
0%
20%
40%
60%
80%
100%
8 10 12 14 16 18 20 22
Cu
mu
lati
ve %
Conventional Respondents' Kondratieff
Figure 5.D: Cumulative Probability Distribution of Payback Period (years)
0%
20%
40%
60%
80%
100%
-500 -400 -300 -200 -100 0 100 200 300 400
Cu
mu
lati
ve %
Conventional Respondents' Kondratieff
Figure 5.E: Cumulative Probability Distribution of NPV at 16% ($m)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 88 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 89 December 2006
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,
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 90 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 91 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 92 December 2006
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
private stakeholders in East Asian toll roads during a period of increasing price
inflation and interest rates.
In addition to reviewing evaluation criteria (discussed in 6.2), the literature 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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 93 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 94 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 95 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 96 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 97 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 98 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 99 December 2006
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).
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 100 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 101 December 2006
“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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 102 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 103 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 104 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 105 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 106 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 107 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 108 December 2006
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
Guangdong Provincial Bureau of Statistics (2000), Guangdong Statistical Yearbook
2000, China Statistics Press, Beijing, 1 August 2000
Guangdong Provincial Bureau of Statistics (2003), Guangdong Statistical Yearbook
2003, China Statistics Press, Beijing, 1 August 2003
Guangdong Provincial Bureau of Statistics (2005), Guangdong Statistical Yearbook
2005, China Statistics Press, Beijing, 1 August 2005
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 109 December 2006
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
Jiangsu Provincial Statistics Bureau (2002), Jiangsu Statistical Yearbook 2002, China
Statistics Press, Beijing, 1 July 2002
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 110 December 2006
Jiangsu Provincial Statistics Bureau (2003), Jiangsu Statistical Yearbook 2003, China
Statistics Press, Beijing, 1 July 2003
Jiangsu Provincial Statistics Bureau (2004), Jiangsu Statistical Yearbook 2004, China
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 111 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 112 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 113 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 114 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 115 December 2006
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
Conference, Loughborough, 14-18 September 1998
Willumsen, L. and Russell, C. (1998) “Reducing Revenue Risk”, European Transport
Conference, Loughborough, 14-18 September 1998
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 116 December 2006
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
Zhejiang Provincial Bureau of Statistics (2000), Zhejiang Statistical Yearbook 2000,
China Statistics Press, Beijing, 1 August 2000
Zhejiang Provincial Bureau of Statistics (2002), Zhejiang Statistical Yearbook 2002,
China Statistics Press, Beijing, 1 July 2002
Zhejiang Provincial Bureau of Statistics (2004), Zhejiang Statistical Yearbook 2004,
China Statistics Press, Beijing, 1 July 2004
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 117 December 2006
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)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 118 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 119 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 120 December 2006
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$) Real GDP
Growth
Rate
Proportion of GDP by Sector Gross Fixed
Investment
as % of GDP PPP
Official
Exchange Rate
Agri-
culture Industry Services
Cambodia 29.89 4.791 6.0% 35.0% 30.0% 35.0% 22.8%
China 8182 1790 9.3% 14.4% 53.1% 32.5% 43.6%
Indonesia 901.7 270 5.4% 14.7% 30.6% 54.6% 21.5%
Laos 11.92 2.541 7.2% 48.6% 25.9% 25.5% n/a
Malaysia 248.7 121.2 5.2% 7.2% 33.3% 59.5% 20.3%
Myanmar 76.36 8.042 1.5% 54.6% 13.0% 32.4% 11.5%
Philippines 451.3 90.3 4.6% 14.8% 31.7% 53.5% 16.3%
Thailand 545.8 177.2 4.4% 9.3% 45.1% 45.6% 31.7%
Vietnam 253.2 44.66 8.4% 20.9% 41.0% 38.1% 38.7%
South Korea 965.3 801.2 3.9% 3.7% 40.1% 56.3% 28.9%
Poland 489.8 242.7 3.5% 2.8% 31.7% 65.5% 18.5%
Mexico 1068 699.5 3.0% 4.0% 26.5% 69.5% 21.1%
UK 1869 2218 1.7% 1.1% 26.0% 72.9% 16.3%
USA 12410 12470 3.5% 1.0% 20.7% 78.3% 16.8%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 121 December 2006
Country
Public Debt as
% of GDP/
External Debt
(US$ 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.)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 122 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 123 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 124 December 2006
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)
Profit(Loss) Before Interest and Tax
Turnover
Expenses as Percentage of
Turnover (EPT)
Expenses
Turnover
Sales to Capital Employed (SCE)
. Turnover .
Total Assets less Current Liabilities
Sales to Fixed Assets (SFA)
. Turnover .
Fixed Assets
Sales to Working Capital
. Turnover .
Net Current Assets
LIQUIDITY/ WORKING CAPITAL MANAGEMENT RATIOS
Working Capital Requirement
(WCR)
Current Assets less Current Liabilities
Current Ratio
. Current Assets .
Current Liabilities
Asset Turnover
. Turnover .
Total Assets less Current Liabilities
Interest Cover/ Debt Service
Coverage Ratio
Profit(Loss) Before Interest and Tax
Interest Payable
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 125 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 126 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 127 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 128 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 129 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 130 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 131 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 132 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 133 December 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)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 134 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 135 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 136 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 137 December 2006
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
economic development.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 138 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 139 December 2006
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:
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 140 December 2006
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
GDP Per Capita (USD p.a.)
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
GDP Per Capita (USD p.a.)
Ln
(km
2 p
er
Air
po
rt)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 141 December 2006
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
GDP Per Capita (USD p.a.)
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
GDP Per Capita (USD p.a.)
Ln
(km
2 p
er
km
of
Railw
ay)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 142 December 2006
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
)
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
GDP Per Capita (USD p.a.)
Ln
(km
2 p
er
km
of
Paved
Ro
ad
)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 143 December 2006
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).
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 144 December 2006
Regression Results
(1A) PopAP = 107000 – 23.65 GDPpc R2 = 0.4%
(740700) (133.65)
(1B) PopAP =
668300
x 0.999970GDPpc
R2 = 0.8%
(0.7005) (0.000126)
(2A) Km2AP = 8425 – 0.2643 GDPpc R2 = 2.0%
(3860) (0.6964)
(2B) Km2AP =
7595
x 0.999925GDPpc
R2 = 7.2%
(0.5628) (0.000102)
(3A) PopRail = 37830 – 1.239 GDPpc R2 = 1.7%
(22240) (3.809)
(3B) PopRail =
32287
x 0.999945GDPpc
R2 = 5.8%
(0.5253) (0.000090)
(4A) Km2Rail = 247.2 – 0.008482 GDPpc R2 = 9.6%
(62.23) (0.01066)
(4B) Km2Rail =
232.6
x 0.999960GDPpc
R2 = 9.4%
(0.2984) (0.000051)
(5A) PopRoad = 7947 – 0.8623 GDPpc R2 = 32.9%
(2580) (0.4655)
(5B) PopRoad =
7835
x 0.999731GDPpc
R2 = 53.9%
(0.5215) (0.000094)
(6A) Km2Road = 104.5 – 0.01264GDPpc R2 = 34.4%
(36.6) (0.00660)
(6B) Km2Road =
89.04
x 0.999686GDPpc
R2 = 58.3%
(0.5554) (0.000100)
Standard errors associated with each parameter are shown in parenthesis beneath the
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 145 December 2006
Population per km of Paved Road
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
GDP Per Capita (USD p.a.)
Ln
(Po
pu
lati
on
per
km
of
Paved
Ro
ad
)
km2 per km of Paved Road
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
GDP Per Capita (USD p.a.)
Ln
(km
2 p
er
km
of
Paved
Ro
ad
)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 146 December 2006
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
economic development.
Only roads data were regressed, as follows:
Regression Results
(5A) PopRoad = 804.6 – 0.01809 GDPpc R2 = 40.3%
(330.5) (0.01268)
(5B) PopRoad =
915.6
x 0.999943GDPpc
R2 = 47.5%
(0.9017) (0.000035)
(6A) Km2Road = 10.41 – 0.0002622GDPpc R2 = 26.1%
(6.644) (0.0002548)
(6B) Km2Road =
4.765
x 0.999962GDPpc
R2 = 15.7%
(1.323) (0.000051)
Standard errors associated with each parameter are shown in parenthesis beneath the
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 147 December 2006
Population per km of Paved Road
MX
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
)
km2 per km of Paved Road
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
GDP Per Capita (USD p.a.)
Ln
(km
2 p
er
km
of
Paved
Ro
ad
)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 148 December 2006
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)
GDSY00: Guangdong Provincial Bureau of Statistics (2000)
GDSY03: Guangdong Provincial Bureau of Statistics (2003)
GDSY05: Guangdong Provincial Bureau of Statistics (2005)
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 149 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 150 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 151 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 152 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 153 December 2006
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)
JSSY02: Jiangsu Provincial Statistics Bureau (2002)
JSSY03: Jiangsu Provincial Statistics Bureau (2003)
JSSY04: Jiangsu Provincial Statistics Bureau (2004)
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 (%)
1997 6,680.34 12.0% -0.7% 11.9%
1998 7,199.95 11.0% -2.9% 10.9%
1999 7,697.82 10.1% -2.9% 10.1%
2000 8,582.73 10.6% 0.8% 10.4%
2001 9,511.91 10.2% 0.6% 10.2%
2002 10,631.75 11.6% 0.2% 11.6%
2003 12,460.83 13.6% 3.2% 13.6%
Source JSSY04, p61 JSSY04, p62 derived derived
Year Civil Vehicle Ownership Source
1997 519,930 JSSY99, p228
1998 561,129 JSSY99, p228
1999 639,152 JSSY02, p241
2000 745,106 JSSY02, p241
2001 871,191 JSSY02, p241
2002 1,044,960 JSSY03, p288
2003 1,317,673 JSSY04, p288
Year Highway Passenger
Trips (10,000 people)
Highway Passenger-
km (100 million)
Average Highway
Passenger Trip Length (km)
1997 88,826 504.05 56.7
1998 92,215 527.62 57.2
1999 95,564 554.04 58.0
2000 101,713 594.48 58.4
2001 105,105 682.25 64.9
2002 110,139 719.08 65.3
2003 118,046 774.11 65.6
Source JSSY04, p285 JSSY04, p285 derived
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 154 December 2006
Year Highway Freight
Transport (10,000 MT)
Highway MT-km
(100 million)
Average Freight Trip
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%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 155 December 2006
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%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 156 December 2006
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
Indices thus obtained from above data:
GDP
Civil Vehicle
Ownership Passenger-km
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
Income elasticities thus calculated (1997 to 2003):
Income elasticity of: Value
Civil Vehicle Ownership 1.41
Passenger-km 0.69
Passenger Trip Length 0.23
Freight MT-km 0.33
Freight Trip Length 0.30
Total Expressway Traffic 1.43
Expressway Cars 1.14
Expressway Small 1.54
Expressway Medium 1.35
Expressway Large+Heavy 2.46
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 157 December 2006
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)
GDP Civil Vehicle Ow nership
Passenger-km Passenger Trip Length
Freight MT-km Freight Trip Length
Jiangsu Expressw ay Traff ic
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 158 December 2006
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)
ZJSY00: Zhejiang Provincial Bureau of Statistics (2000)
ZJSY02: Zhejiang Provincial Bureau of Statistics (2002)
ZJSY04: Zhejiang Provincial Bureau of Statistics (2004)
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 (%)
1998 4,988 10.10% -2.3% 10.0%
1999 5,365 10.00% -2.2% 10.0%
2000 6,036 11.01% 1.4% 10.9%
2001 6,748 10.50% 1.2% 10.5%
2002 7,796 12.50% 2.7% 12.4%
2003 9,395 14.40% 5.3% 14.4%
Source ZJSY04, p24 ZJSY04, p26 derived derived
Year Civil Vehicle Ownership Source
1998 478,297 ZJSY99, p395
1999 575,882 ZJSY00, p379
2000 680,586 ZJSY02, p415
2001 855,642 ZJSY02, p415
2002 1,078,311 ZJSY04, p445
2003 1,358,209 ZJSY04, p445
Year Highway Passenger
Trips (10,000 people)
Highway Passenger-
km (100 million)
Average Highway
Passenger Trip Length (km)
1998 111,847 436.34 39.01
1999 111,771 433.50 38.78
2000 116,996 449.51 38.42
2001 126,008 479.53 38.06
2002 128,980 519.20 40.25
2003 133,968 531.63 39.68
Source ZJSY04, p447 ZJSY04, p448 derived
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 159 December 2006
Year Highway Freight
Transport (10,000 MT)
Highway MT-km
(100 million)
Average Freight Trip
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%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 160 December 2006
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%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 161 December 2006
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
Indices thus obtained from above data:
GDP
Civil Vehicle
Ownership Passenger-km
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
Income elasticities thus calculated (1997 to 2003):
Income elasticity of: Value
Civil Vehicle Ownership 1.78
Passenger-km 0.37
Passenger Trip Length 0.03
Freight MT-km 0.82
Freight Trip Length 0.37
Zhejiang Expressway Traffic 1.54
Zhejiang Expressway Revenue 2.11
Expressway Traffic Small 1.70
Expressway Traffic Medium 1.05
Expressway Traffic Large 1.34
Expressway Traffic Heavy 3.10
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 162 December 2006
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)
GDP Civil Vehicle Ow nership
Passenger-km Passenger Trip Length
Freight MT-km Freight Trip Length
Zhejiang Expressw ay Traff ic Zhejiang Expressw ay Revenue
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 163 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 164 December 2006
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%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 165 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 166 December 2006
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:
(rating scale) Mandatory
One answer per row, from:
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:
(rating scale) Mandatory
One answer per row, from:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 167 December 2006
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?
(rating scale) Mandatory
One answer per row, from:
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?
(rating scale) Mandatory
One answer per row, from:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 168 December 2006
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:
(rating scale) Mandatory
One answer per row, from:
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:
(rating scale) Mandatory
One answer per row, from:
Such models are reliable
Such models are too data hungry to be relied upon
Such models are too simplistic 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
20 Regarding spreadsheet-based traffic/ revenue models,
please indicate how strongly you agree or disagree
with the following statements:
(rating scale) Mandatory
One answer per row, from:
Such models are reliable
Such models are too data hungry to be relied upon
Such models are too simplified 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 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:
(text response) Optional
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 169 December 2006
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:
(rating scale) Mandatory
One answer per row, from:
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?
(rating scale) Mandatory
One answer per row, from:
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)?
(rating scale) Mandatory
One answer per row, from:
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:
(text response) Optional
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 170 December 2006
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:
(rating scale) Mandatory
One answer per row, from:
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:
(rating scale) Mandatory
One answer per row, from:
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):
(text response) Optional
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.
(text response) Optional
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 171 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 172 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 173 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 174 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 175 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 176 December 2006
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
Rural or inter-urban
tolled highway projects in
developing economies
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
Rural or inter-urban
tolled highway projects 1.70 0 0 4 10 79 66
Rural or inter-urban
tolled highway projects in
developing economies
1.12 0 0 2 7 65 85
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 177 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 178 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 179 December 2006
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:
Excluding “Not Sure” and null responses:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 180 December 2006
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 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
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
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 181 December 2006
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?
Excluding “Not Sure” and null responses:
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
sufficient data available 1 9 36 52 2
In developing countries, data are
reliable 1 8 37 47 6
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 182 December 2006
Question 16: From your experience, do you feel there are sufficient data to
successfully prepare meaningful traffic and revenue forecasts?
Excluding “Not Sure” and null responses:
1 2 3 4 5
Always Usually Sometimes Rarely Never
In developed countries, there are
sufficient data available 8 69 28 4 0
In developed countries, data (e.g.
land use/economic forecasts) are
reliable 6 55 41 4 1
In developing countries, there are
sufficient data available 1 11 51 34 2
In developing countries, data (e.g.
land use/economic forecasts) are
reliable 1 9 42 43 4
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 183 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 184 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 185 December 2006
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:
Excluding “Not Sure” and null responses:
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
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 186 December 2006
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:
Excluding “Not Sure” and null responses:
1 2 3 4 5
Strongly
Agree Agree Neutral Disagree
Strongly
Disagree
Such models are reliable 2 39 47 9 2 Such models are too data hungry to
be relied upon 1 10 37 37 9
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 not suitable for
toll-road work 1 12 28 40 11
Economic uncertainties/ pace of
change makes them irrelevant in
developing economies 3 13 30 38 7
Such models are too much of a
black box for non-specialists to
properly critique model outputs 4 19 30 36 10
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
Responses Mean
Standard
Deviation Such models are reliable 99 2.70 0.74
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
them irrelevant in developing economies 91 3.36 0.93
Such models are too much of a black box for non-
specialists to properly critique model outputs 99 3.29 1.02
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 187 December 2006
Question 20: Regarding spreadsheet-based traffic/ revenue models, please indicate
how strongly you agree or disagree with the following statements:
Excluding “Not Sure” and null responses:
1 2 3 4 5
Strongly
Agree Agree Neutral Disagree
Strongly
Disagree
Such models are reliable 0 24 53 16 4 Such models are too data hungry to
be relied upon 1 2 27 48 16
Such models are too simplified to
be of worth 3 12 39 34 5
Such models are not suitable for
toll-road work 2 10 38 32 5
Economic uncertainties/ pace of
change makes them irrelevant in
developing economies 1 14 32 36 5
Such models are too simplistic to
provide meaningful outputs 5 10 40 39 3
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
Responses Mean
Standard
Deviation Such models are reliable 97 3.00 0.76
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
them irrelevant in developing economies 88 3.34 0.85
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..
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 188 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 189 December 2006
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:
Excluding “Not Sure” and null responses:
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
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 190 December 2006
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
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
TpEc = Transport Planners and Economists
E&A = Engineers and Architects
G&A = Government and Aid Agencies
Acad = Academics
Oth = Others
Question 23: How often do you explicitly consider the following in appraising
tolled highways?
Excluding “Not Sure” and null responses:
1 2 3 4 5
Always Usually Sometimes Rarely Never
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 191 December 2006
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
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)?
Excluding “Not Sure” and null responses:
1 2 3 4 5
Always Usually Sometimes Rarely Never
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
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 192 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 193 December 2006
Question 26: How would you rate the potential for inter-urban tolled highways
over the next 10 years in each of the following countries:
Excluding “Not Sure” and null responses:
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
Using the values 1 to 6 (as per column headings above), the mean and standard
deviation of responses as follows:
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
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
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
TpEc = Transport Planners and Economists
E&A = Engineers and Architects
G&A = Government and Aid Agencies
Acad = Academics
Oth = Others
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 194 December 2006
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
Using the values 1 to 6 (as per column headings above), the mean and standard
deviation of responses as follows:
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 195 December 2006
Question 27: Comparing the next 10 years (2006-2016) with the last 5 years (2001-
2006), how do you feel each the following will change:
Excluding “Not Sure” and null responses:
1 2 3 4 5
Significant
Increase
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
Using the values 1 to 5 (as per column headings above), the mean and standard
deviation of responses as follows:
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
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
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
TpEc = Transport Planners and Economists
E&A = Engineers and Architects
G&A = Government and Aid Agencies
Acad = Academics
Oth = Others
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 196 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 197 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 198 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 199 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 200 December 2006
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)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 201 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 202 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 203 December 2006
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.
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 204 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 205 December 2006
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
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 206 December 2006
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
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 207 December 2006
Variable: Capacity per Local Lane (pcu's)
Minimum: 8,000
Maximum: 12,000
Mean: 9,994
Monte Carlo Settings:
Modal Value 10,000
% < Modal 50%
SD (<Modal) 1,000
SD (>Modal) 1,000
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
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
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
8,0
00
8,5
00
9,0
00
9,5
00
10,
000
10,
500
11,
000
11,
500
12,
000
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 208 December 2006
Variable: Total Construction Cost (Base Year million$)
Minimum: 166.79 million
Maximum: 240.92 million
Mean: 199.77 million
Monte Carlo Settings:
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
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
30%
167.
5
177.
5
187.
5
197.
5
207.
5
217.
5
227.
5
237.
5
Conventional Respondents Kondatrieff
Distribution of Variable
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%
Cumulative (Left Axis) Probability Density (Right Axis)
Mean FIRR by Variable Value and Forecast Case
10%
12%
14%
16%
18%
20%
22%
167.
5
177.
5
187.
5
197.
5
207.
5
217.
5
227.
5
237.
5
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 209 December 2006
Variable: Construction Duration (Quarters)
Minimum: 8
Maximum: 14
Mean: 11.08
Monte Carlo Settings:
Modal Value 10
% < Modal 20%
SD (<Modal) 1
SD (>Modal) 2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
8 9 10 11 12 13 14
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
8 9 10 11 12 13 14
0%
5%
10%
15%
20%
25%
30%
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%
8 9 10 11 12 13 14
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 210 December 2006
Variable: Fixed Annual Operations and Maintenance Costs (as % of Construction Cost)
Minimum: 0.1%
Maximum: 4.0%
Mean: 2.2%
Monte Carlo Settings:
Modal Value 2% *
% < Modal 50%
SD (<Modal) 0.5%
SD (>Modal) 1.0%
Note: * as % of initial construction cost
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
0.1% 1.1% 2.1% 3.1% 4.1%
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.1% 1.1% 2.1% 3.1% 4.1%
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 211 December 2006
Variable: Variable Operations and Maintenance Costs (as % of Revenues)
Minimum: 1.0%
Maximum: 5.0%
Mean: 3.0%
Monte Carlo Settings:
Modal Value 3% *
% < Modal 50%
SD (<Modal) 1%
SD (>Modal) 1%
Note: * as % of revenues
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
1.00% 1.80% 2.60% 3.40% 4.20% 5.00%
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
1.00% 1.80% 2.60% 3.40% 4.20% 5.00%
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 212 December 2006
Variable: Base Value of Time ($/hr) for Small Vehicles
Minimum: 2.00 $/hour
Maximum: 6.00 $/hour
Mean: 4.00 $/hour
Monte Carlo Settings:
Modal Value 4
% < Modal 50%
SD (<Modal) 1
SD (>Modal) 1
Regression Analysis: FIRR on Variable
Conventional Case
FIRR = 0.17092 + 0.000061 * Variable
R2= 0.001
FIRR = 0.17089 * 1.000384 ^ Variable
R2= 0.001
Respondents' Case
FIRR = 0.17837 + -0.00005 * Variable
R2= 0.001
FIRR = 0.17835 * 0.999741 ^ Variable
R2= 0.001
Kondratieff Case
FIRR = 0.15427 + 0.000071 * Variable
R2= 0.0004
FIRR = 0.15419 * 1.000505 ^ Variable
R2= 0.0005
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
2.0 2.8 3.6 4.4 5.2 6.0
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
2.0 2.8 3.6 4.4 5.2 6.0
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
Cumulative (Left Axis) Probability Density (Right Axis)
Mean FIRR by Variable Value and Forecast Case
10%
11%
12%
13%
14%
15%
16%
17%
18%
19%
20%
2.0 2.8 3.6 4.4 5.2 6.0
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 213 December 2006
Variable: Base Value of Time ($/hr) for Large Vehicles
Minimum: 1.00 $/hour
Maximum: 5.00 $/hour
Mean: 3.01 $/hour
Monte Carlo Settings:
Modal Value 3
% < Modal 50%
SD (<Modal) 1
SD (>Modal) 1
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
1.0 1.8 2.6 3.4 4.2 5.0
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
1.0 1.8 2.6 3.4 4.2 5.0
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 214 December 2006
Variable: Income Elasticity of Value of Time (Small Vehicles)
Minimum: 0.20
Maximum: 0.80
Mean: 0.50
Monte Carlo Settings:
Modal Value 0.5
% < Modal 50%
SD (<Modal) 0.15
SD (>Modal) 0.15
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
0.20 0.40 0.60 0.80
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
0.20 0.40 0.60 0.80
0%
2%
4%
6%
8%
10%
12%
14%
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%
0.20 0.40 0.60 0.80
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 215 December 2006
Variable: Income Elasticity of Value of Time (Large Vehicles)
Minimum: 0.20
Maximum: 0.80
Mean: 0.50
Monte Carlo Settings:
Modal Value 0.5
% < Modal 50%
SD (<Modal) 0.15
SD (>Modal) 0.15
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
0.20 0.40 0.60 0.80
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
0.20 0.40 0.60 0.80
0%
2%
4%
6%
8%
10%
12%
14%
16%
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%
0.20 0.40 0.60 0.80
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 216 December 2006
Variable: Base Expressway Vehicle Operating Costs ($/km) for Small Vehicles
Minimum: 0.03
Maximum: 0.09
Mean: 0.06
Monte Carlo Settings:
Modal Value 0.06
% < Modal 50%
SD (<Modal) 0.015
SD (>Modal) 0.015
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
0.030 0.050 0.070 0.090
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
0.030 0.050 0.070 0.090
0%
2%
4%
6%
8%
10%
12%
14%
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%
0.030 0.050 0.070 0.090
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 217 December 2006
Variable: Base Expressway Vehicle Operating Costs ($/km) for Large Vehicles
Minimum: 0.05
Maximum: 0.15
Mean: 0.10
Monte Carlo Settings:
Modal Value 0.1
% < Modal 50%
SD (<Modal) 0.025
SD (>Modal) 0.025
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
0.050 0.070 0.090 0.110 0.130 0.150
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.050 0.070 0.090 0.110 0.130 0.150
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 218 December 2006
Variable: Local Roads VOC Multiplier for Small Vehicles (Expressway VOC * Factor)
Minimum: 1.00
Maximum: 2.00
Mean: 1.50
Monte Carlo Settings:
Modal Value 1.5
% < Modal 50%
SD (<Modal) 0.25
SD (>Modal) 0.25
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
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
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
1.0
00
1.2
00
1.4
00
1.6
00
1.8
00
2.0
00
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 219 December 2006
Variable: Local Roads VOC Multiplier for Large Vehicles (Expressway VOC * Factor)
Minimum: 1.50
Maximum: 2.50
Mean: 2.00
Monte Carlo Settings:
Modal Value 2
% < Modal 50%
SD (<Modal) 0.25
SD (>Modal) 0.25
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
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
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
1.5
00
1.7
00
1.9
00
2.1
00
2.3
00
2.5
00
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 220 December 2006
Variable: Base Traffic Factor (Small Vehicles) to Expand/Contract Initial Demand
Minimum: 0.70
Maximum: 1.30
Mean: 1.00
Monte Carlo Settings:
Modal Value 1
% < Modal 50%
SD (<Modal) 0.15
SD (>Modal) 0.15
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
0.7
00
0.9
00
1.1
00
1.3
00
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.7
00
0.9
00
1.1
00
1.3
00
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 221 December 2006
Variable: Base Traffic Factor (Large Vehicles) to Expand/Contract Initial Demand
Minimum: 0.70
Maximum: 1.30
Mean: 1.00
Monte Carlo Settings:
Modal Value 1
% < Modal 50%
SD (<Modal) 0.15
SD (>Modal) 0.15
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
0.7
00
0.9
00
1.1
00
1.3
00
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.7
00
0.9
00
1.1
00
1.3
00
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 222 December 2006
Variable: Income Elasticity of Traffic (Small Vehicles)
Minimum: 0.85
Maximum: 1.65
Mean: 1.25
Monte Carlo Settings:
Modal Value 1.25
% < Modal 50%
SD (<Modal) 0.2
SD (>Modal) 0.2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
0.9
00
1.1
00
1.3
00
1.5
00
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.9
00
1.1
00
1.3
00
1.5
00
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 223 December 2006
Variable: Income Elasticity of Traffic (Large Vehicles)
Minimum: 0.70
Maximum: 1.50
Mean: 1.10
Monte Carlo Settings:
Modal Value 1.1
% < Modal 50%
SD (<Modal) 0.2
SD (>Modal) 0.2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
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
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.7
00
0.9
00
1.1
00
1.3
00
1.5
00
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 224 December 2006
Variable: Toll Revenue Leakage (%)
Minimum: 5%
Maximum: 20%
Mean: 11%
Monte Carlo Settings:
Modal Value 10%
% < Modal 50%
SD (<Modal) 2.5%
SD (>Modal) 5%
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
5% 9% 13%
17%
21%
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
5% 9% 13%
17%
21%
0%
2%
4%
6%
8%
10%
12%
14%
16%
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%
5% 9% 13%
17%
21%
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 225 December 2006
Variable: Initial Amplitude of Ramp-Up (%)
Minimum: 20%
Maximum: 80%
Mean: 44%
Monte Carlo Settings:
Modal Value 40%
% < Modal 50%
SD (<Modal) 10%
SD (>Modal) 20%
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
20.0
%
40.0
%
60.0
%
80.0
%
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
20% 40% 60% 80%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
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%
20.0
%
40.0
%
60.0
%
80.0
%
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 226 December 2006
Variable: Ramp-Up Duration (Quarters)
Minimum: 4
Maximum: 20
Mean: 9.79
Monte Carlo Settings:
Modal Value 8
% < Modal 40%
SD (<Modal) 2
SD (>Modal) 5
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
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
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 227 December 2006
Variable: Small Vehicles' Tolling Penalty (Minutes)
Minimum: 0
Maximum: 20
Mean: 10
Monte Carlo Settings:
Modal Value 10
% < Modal 50%
SD (<Modal) 5
SD (>Modal) 5
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
0 4 8 12 16 20
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
0 4 8 12 16 20
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
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%
0 4 8 12 16 20
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 228 December 2006
Variable: Large Vehicles' Tolling Penalty (Minutes)
Minimum: 5
Maximum: 25
Mean: 15
Monte Carlo Settings:
Modal Value 15
% < Modal 50%
SD (<Modal) 5
SD (>Modal) 5
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
5 9 13 17 21 25
Conventional Respondents Kondatrieff
Distribution of Variable
0%
20%
40%
60%
80%
100%
5 9 13 17 21 25
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
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%
5 9 13 17 21 25
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 229 December 2006
Variable: Routeing Sensitivity ("Lambda") for Small Vehicles
Minimum: 0.025
Maximum: 0.075
Mean: 0.050
Monte Carlo Settings:
Modal Value 0.05
% < Modal 50%
SD (<Modal) 0.0125
SD (>Modal) 0.0125
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
0.0
25
0.0
45
0.0
65
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.0
25
0.0
45
0.0
65
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 230 December 2006
Variable: Routeing Sensitivity ("Lambda") for Large Vehicles
Minimum: 0.025
Maximum: 0.075
Mean: 0.050
Monte Carlo Settings:
Modal Value 0.05
% < Modal 50%
SD (<Modal) 0.0125
SD (>Modal) 0.0125
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
0.0
25
0.0
45
0.0
65
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
0.0
25
0.0
45
0.0
65
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 231 December 2006
Variable: Toll Escalation Rate (% of RPI Inflation)
Minimum: 60%
Maximum: 100%
Mean: 86%
Monte Carlo Settings:
Modal Value 90%
% < Modal 50%
SD (<Modal) 15%
SD (>Modal) 5%
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
62.5% 72.5% 82.5% 92.5%
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
62.5% 72.5% 82.5% 92.5%
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 232 December 2006
Variable: Quarters between Toll Increases
Minimum: 8
Maximum: 20
Mean: 13.29
Monte Carlo Settings:
Modal Value 12
% < Modal 40%
SD (<Modal) 2
SD (>Modal) 4
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
8 9 10 11 12 13 14 15 16 17 18 19 20
Conventional Respondents Kondatrieff
Distribution of Variable
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%
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%
8 9 10 11 12 13 14 15 16 17 18 19 20
Conventional Respondents Kondatrieff
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 233 December 2006
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%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Respondents'
0%
20%
40%
60%
80%
100%
2% 4% 6% 8% 10% 12%
0%
2%
4%
6%
8%
10%
12%
Cumulative (Left Axis) Probability Density (Right Axis)
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%
Cumulative (Left Axis) Probability Density (Right Axis)
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%
Cumulative (Left Axis) Probability Density (Right Axis)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 234 December 2006
Variable: GDP Growth (% p.a.) Page 2 of 2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
10%
20%
30%
40%
50%
60%
2% 4% 6% 8% 10% 12% 14%
Conventional Respondents Kondatrieff All
Mean FIRR by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
2% 4% 6% 8% 10% 12% 14%
Conventional Respondents Kondatrieff All
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 235 December 2006
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
Distribution of Variable: Conventional
0%
20%
40%
60%
80%
100%
0% 1% 2% 3% 4%
0%
5%
10%
15%
20%
25%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Respondents'
0%
20%
40%
60%
80%
100%
0% 2% 4% 6% 8% 10%
0%
2%
4%
6%
8%
10%
12%
14%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Kondratieff
0%
20%
40%
60%
80%
100%
1% 3% 5% 7% 9% 11% 13%
0%
2%
4%
6%
8%
10%
12%
14%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: All Cases
0%
20%
40%
60%
80%
100%
0% 2% 4% 6% 8% 10% 12%
0%
2%
4%
6%
8%
10%
12%
Cumulative (Left Axis) Probability Density (Right Axis)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 236 December 2006
Variable: Price Inflation for Vehicle Operating Costs (% p.a.) Page 2 of 2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
0% 2% 4% 6% 8% 10% 12%
Conventional Respondents Kondatrieff All
Mean FIRR by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
0% 2% 4% 6% 8% 10% 12%
Conventional Respondents Kondatrieff All
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 237 December 2006
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 Value 0.75% added to Conventional
% < Modal 50%
SD (<Modal) 0.25%
SD (>Modal) 0.25%
Kondratieff
Minimum: 1%
Maximum: 8%
Mean: 4%
Modal Value 1% added to Respondents'
% < 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
Distribution of Variable: Conventional
0%
20%
40%
60%
80%
100%
0% 1% 2% 3% 4%
0%
5%
10%
15%
20%
25%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Respondents'
0%
20%
40%
60%
80%
100%
0% 1% 2% 3% 4% 5%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Kondratieff
0%
20%
40%
60%
80%
100%
1% 2% 3% 4% 5% 6% 7% 8%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: All Cases
0%
20%
40%
60%
80%
100%
0% 1% 2% 3% 4% 5% 6% 7%
0%
2%
4%
6%
8%
10%
12%
14%
16%
Cumulative (Left Axis) Probability Density (Right Axis)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 238 December 2006
Variable: Price Inflation for Construction and (Fixed) Operations & Maintenance Costs (% p.a.) Page 2 of 2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
0% 1% 2% 3% 4% 5% 6% 7%
Conventional Respondents Kondatrieff All
Mean FIRR by Variable Value and Forecast Case
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
0% 1% 2% 3% 4% 5% 6% 7%
Conventional Respondents Kondatrieff All
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 239 December 2006
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 Value 0.75% added to Conventional
% < Modal 50%
SD (<Modal) 0.25%
SD (>Modal) 0.25%
Kondratieff
Minimum: 1%
Maximum: 8%
Mean: 4%
Modal Value 1% added to Respondents'
% < 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
Distribution of Variable: Conventional
0%
20%
40%
60%
80%
100%
0% 1% 2% 3% 4%
0%
5%
10%
15%
20%
25%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Respondents'
0%
20%
40%
60%
80%
100%
0% 1% 2% 3% 4% 5%
0%
5%
10%
15%
20%
25%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Kondratieff
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%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: All Cases
0%
20%
40%
60%
80%
100%
0% 1% 2% 3% 4% 5% 6% 7%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
Cumulative (Left Axis) Probability Density (Right Axis)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 240 December 2006
Variable: General Price Inflation (% p.a.) Page 2 of 2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
10%
20%
30%
40%
50%
60%
0% 2% 4% 6%
Conventional Respondents Kondatrieff All
Mean FIRR by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
0% 1% 2% 3% 4% 5% 6% 7%
Conventional Respondents Kondatrieff All
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 241 December 2006
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 Value 2% added to Respondents'
% < 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
Distribution of Variable: Conventional
0%
20%
40%
60%
80%
100%
3% 4% 5% 6% 7%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Respondents'
0%
20%
40%
60%
80%
100%
3% 4% 5% 6% 7% 8% 9%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Kondratieff
0%
20%
40%
60%
80%
100%
3% 5% 7% 9% 11%
0%
2%
4%
6%
8%
10%
12%
14%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: All Cases
0%
20%
40%
60%
80%
100%
3% 5% 7% 9% 11% 13%
0%
2%
4%
6%
8%
10%
12%
14%
Cumulative (Left Axis) Probability Density (Right Axis)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 242 December 2006
Variable: Interest Rates for Initial Debt (% p.a.) Page 2 of 2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
3% 5% 7% 9% 11% 13%
Conventional Respondents Kondatrieff All
Mean FIRR by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
30%
3% 5% 7% 9% 11% 13%
Conventional Respondents Kondatrieff All
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 243 December 2006
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 Value 2% added to Initial Interest Rate
% < Modal 50%
SD (<Modal) 1%
SD (>Modal) 1%
Kondratieff
Minimum: 4%
Maximum: 16%
Mean: 10%
Modal Value 2% added to Initial Interest Rate
% < 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
Distribution of Variable: Conventional
0%
20%
40%
60%
80%
100%
3% 5% 7% 9% 11%
0%
5%
10%
15%
20%
25%
30%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Respondents'
0%
20%
40%
60%
80%
100%
3% 5% 7% 9% 11% 13%
0%
5%
10%
15%
20%
25%
30%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: Kondratieff
0%
20%
40%
60%
80%
100%
4% 6% 8% 10% 12% 14% 16%
0%
5%
10%
15%
20%
25%
Cumulative (Left Axis) Probability Density (Right Axis)
Distribution of Variable: All Cases
0%
20%
40%
60%
80%
100%
3% 5% 7% 9% 11% 13% 15% 17%
0%
5%
10%
15%
20%
25%
Cumulative (Left Axis) Probability Density (Right Axis)
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 244 December 2006
Variable: Interest Rates for Extra Debt (% p.a.) Page 2 of 2
Regression Analysis: FIRR on Variable
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
Chance of Failure by Variable Value and Forecast Case
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
3% 7% 11% 15%
Conventional Respondents Kondatrieff All
Mean FIRR by Variable Value and Forecast Case
0%
5%
10%
15%
20%
25%
30%
3% 7% 11% 15%
Conventional Respondents Kondatrieff All
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 245 December 2006
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 $m)
Range 74 74 74
Impact 4.58% 4.96% 6.19%
R2 96% 96% 94%
Impact*R2 4.39% 4.76% 5.82%
+/- Impact -ve -ve -ve
Construction Duration
(Quarters)
Range 6 6 6
Impact 1.30% 1.66% 3.03%
R2 99% 99% 97%
Impact*R2 1.29% 1.64% 2.93%
+/- Impact -ve -ve -ve
Sum of Impact * R2 5.68% 6.40% 8.75%
All O&M Costs
Fixed O&M Costs
(as % of Construction
Costs)
Range 4% 4% 4%
Impact 1.81% 2.27% 3.73%
R2 17% 33% 50%
Impact*R2 0.31% 0.75% 1.88%
+/- Impact -ve -ve -ve
Variable O&M Costs
(as % of Revenue)
Range 4% 4% 4%
Impact 0.60% 0.62% 0.64%
R2 47% 43% 16%
Impact*R2 0.28% 0.27% 0.10%
+/- Impact -ve -ve -ve
Sum of Impact * R2 0.59% 1.02% 1.98%
Value of Time & Its Income Elasticity
Small Vehicle VOT
($/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
($/hr)
Range 4 4 4
Impact 0.01% 0.13% 0.22%
R2 0% 3% 3%
Impact*R2 0.00% 0.00% 0.01%
+/- Impact +ve -ve -ve
Income Elasticity of
VOT (Small Vehicles)
Range 1 1 1
Impact 0.04% 0.01% 0.14%
R2 1% 0% 3%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 246 December 2006
Parameter Conventional Respondents' Kondratieff
Impact*R2 0.00% 0.00% 0.00%
+/- Impact -ve +ve -ve
Income Elasticity of
VOT (Large Vehicles)
Range 1 1 1
Impact 0.08% 0.17% 0.34%
R2 3% 14% 12%
Impact*R2 0.00% 0.02% 0.04%
+/- Impact -ve -ve +ve
Sum of Impact * R2 0.00% 0.03% 0.05%
Vehicle Operating Costs
Small Vehicle
Expressway VOC
($/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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
Sum of Impact * R2 12.13% 12.69% 16.05%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 247 December 2006
Parameter Conventional Respondents' Kondratieff
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%
+/- Impact -ve -ve -ve
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%
+/- Impact -ve -ve -ve
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%
+/- Impact -ve -ve -ve
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%
+/- Impact -ve -ve -ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact -ve -ve -ve
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 248 December 2006
Parameter Conventional Respondents' Kondratieff
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%
+/- Impact +ve +ve +ve
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%
+/- Impact -ve -ve -ve
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%
+/- Impact +ve +ve +ve
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%
+/- Impact -ve -ve -ve
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%
+/- Impact -ve -ve -ve
Sum of Impact * R2 8.78% 15.73% 47.38%
Dissertation Richard F. DI BONA
Henley Management College (1005661)
DissFinal Page 249 December 2006
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