UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Generalized economic model, risk analysis framework and decision support system for the analysis and.. 2000

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


ubc_2000-485943.pdf [ 87.19MB ]
JSON: 1.0064032.json
JSON-LD: 1.0064032+ld.json
RDF/XML (Pretty): 1.0064032.xml
RDF/JSON: 1.0064032+rdf.json
Turtle: 1.0064032+rdf-turtle.txt
N-Triples: 1.0064032+rdf-ntriples.txt

Full Text

GENERALIZED ECONOMIC MODEL, RISK ANALYSIS FRAMEWORK AND DECISION SUPPORT SYSTEM FOR THE ANALYSIS AND EVALUATION OF CAPITAL INVESTMENT PROJECTS by A H M E D M O H A M E D A B D E L - A Z I Z B.Sc, Zagazig University, Egypt, 1986 M.Sc, Zagazig University, Egypt, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES DEPARTMENT OF CIVIL ENGINEERING We accept this thesis as conforming tp-the required standard The University of British Columbia May 2000 © Ahmed M . Abdel-Aziz, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department D a t e flffil, P.ft DE-6 (2/88) ABSTRACT The goal of this thesis is to contribute to the knowledge base for making high quality decisions about capital investment projects. To achieve this goal, a generalized economic model, risk analysis framework, and decision support system for the evaluation and risk analysis of capital investment projects are developed. A detailed investigation of the characteristics of capital investment projects for which the economic model, risk analysis framework and decision support system are to be built is made. A framework called "requirement structure" is introduced in order to investigate the characteristics of a project during its life cycle. A number of infrastructure transportation projects developed under alternative procurement modes in the U.S., U.K., and Canada are used in this study. Capital investment projects are a feature of several industries, markets and business sectors. These markets and business sectors have different characteristics and require the use of a variety of methods in preparing estimates and forecasts. An objective of the thesis deals with modeling such diversity. To achieve this objective, a generalized economic model is developed with a multipurpose hierarchical network-based time function structure. One concept behind the generalized model is that of cash flow classification. A classification represents a domain, e.g. maintenance or finance, and possesses the properties and methods of that domain. This allows a cash flow of a classification type to inherit its domain's properties and methods. Another concept behind the generalized model is shape functions, which allows the variables of the generalized model to change over time according to a selected pattern. More importantly, shape functions serve in converting an estimate into an expenditure flow. The model structure is organized in four components reflecting four classification domains in capital investment projects namely, capital expenditure, revenue, operation and maintenance, and project financing. The basic elements in a component are called constructs. Each construct represents a cash flow that has the same classification type of the component and consequently inherits its properties and methods. With the generalized economic model, a project economic structure can be formulated with any required properties and methods. The generalized model embraces a broad range of periodic and cumulative cash flows and performance measures such as net present value, internal rate of return, total costs (e.g. total construction cost), life cycle cost, total revenues, debt service coverage ratio, loan life cover ratio, and benefit cost ratios. ii To model the uncertainties inherent in the estimates of variables and economic indicators of capital investment projects, a risk analysis framework is introduced. The framework uses an analytical two- and four-moment approach that directly derives the four moments of the performance measures in the generalized model regardless of how complicated their economic structure might be. The framework reduces the necessity of computing intermediate moments as in other moment approaches. A rigorous and expanded derivation for the four moments of a system function is introduced for the framework in order to enhance the accuracy over the standard moment approach. Considerable flexibility in terms of several types of methods, e.g. percentile values, moments, and full probability distribution is introduced for modeling the uncertainty of variables in the generalized model. This provides flexibility over the simulation risk analysis approach that works only with full probability distributions. Pearson and Schmeiser-Deutsch distribution families are used to qualify/fit a distribution model for a performance measure based to its moments. A practical implementation of the generalized model and the risk analysis framework through a decision support system, called Evaluator, is presented. The system has three components: data, model, and interface/dialogue components and makes use of existing software tools. Two examples are presented in order to validate the output of the system and to show application of the system to a transportation project. Decision makers in the public and private sector should find the system to be an effective tool to assist in making decisions regarding the procurement, investment, financing, and risk allocation of capital investment projects. iii Table of Contents Abstract i i Table Of Contents iv List Of Figures x List Of Tables xv P A R T I: A N A L Y S I S 1 Introduction 1 1.1 General 1 1.2 Scope Definition 2 1.3 Problem Definition 6 1.4 Research Objectives and Methodology 11 1.4.1 Capital Investment Projects 11 1.4.2 Generalized Economic Model 12 1.4.3 Risk Analysis Framework 14 1.4.4 Decision Support System 16 1.5 Research Outline 19 2 Characteristics of Capital Investment Projects 23 2.1 Introduction 23 2.2 Public-Private Partnerships (PPP) 25 2.2.1 PPP Arrangements 25 2.2.2 PPP Project Company 29 2.2.3 PPP Project Evaluation 32 2.2.4 PPP Project Risks 34 2.3 Government Requirement Structure 38 2.3.1 Requirement Structure Description 38 2.3.2 Projects and Acts Considered 40 2.3.3 Rights Dimension 43 Rights Dimension: Possession Attribute 43 iv Rights Dimension: Revenues Attribute 49 2.3.4 Obligations Dimension 55 Obligation Dimension: Development, Operation, and Environment Attributes 55 Obligation Dimension: Financing Attribute 63 2.3.5 Liabilities Dimension 70 Liabilities Dimension: General Liability Attribute 70 Liabilities Dimension: Risk Attribute 74 Liabilities Dimension: Taxes Attribute 80 2.4 Requirements Structure: Conclusions 83 3 Proposed Economic Model Characteristics 89 3.1 Introduction 89 3.2 Economic Model Underlying Concepts 89 3.3 Previous Models and Systems for Project Appraisals 94 3.4 Characteristics of the Proposed Model and Support System 101 PART II: DESIGN 4 Generalized Economic Model 105 4.1 Introduction 105 4.2 Generalized Economic Model 106 4.2.1 Model Structure and General Concepts 106 4.2.2 Components Common Properties I l l 4.2.3 Components Common Methods: Shape Functions 116 Rate Functions 117 Area Functions 117 4.3 Capital Expenditure Component 133 4.3.1 CE Component Structure and Properties 134 4.3.2 C E Component Methods and Cash Flow Formulation 141 C E Semi-detailed Methods 141 C E Detailed Methods 148 a) Material Estimating Methods 149 v b) Labor and Equipment Estimating Methods 150 c) Subcontracted/Indirect Estimating Methods 152 C E Crude Methods: Discrete Costs 153 C E Component Cash Flow 153 4.3.3 C E Component Aggregation Levels 158 4.3.4 C E Component Cumulative and Discounted Costs 161 4.4 Revenue Component 164 4.4.1 RV Component Structure and Properties 166 4.4.2 RV Component Methods and Cash Flow Formulation 169 RV Semi-detailed Methods 169 RV Detailed Methods 171 a) Service Charge Rate Variable 171 b) Total Volume of Demand 173 c) Project Demand Function 177 i. Simplified Trend Analysis 177 ii. Elasticity Based Methods 178 iii. Individual Choice Based Methods 182 RV Crude Methods: Discrete Revenues 188 4.4.3 RV Component Cumulative and Discounted Revenues 188 4.5 Operation and Maintenance Component 192 4.5.1 O M Component Structure and Properties 194 4.5.2 O M Component Methods 195 O M Semi-detailed Methods 196 O M Detailed Methods 198 O M Crude Methods: Discrete Costs 200 4.5.3 O M Component Cumulative and Discounted Costs 200 4.6 Finance Component 203 4.6.1 Background and Project Financing 203 4.6.2 FN Component Structure and Properties 209 4.6.3 FN Component Common Methods 212 Debt Drawing Methods 214 Debt Repayment Methods 217 a) Special Repayment Clauses 222 v i b) Repayment Methods Summary 225 Debt Interest Rate Methods 227 a) Fixed Interest Rate 227 b) Floating Interest Rate 232 Debt Fees 238 Debt Currency and Exchange Rates 242 4.6.4 FN Component Project Financing Instruments 245 General Term Loans 245 Syndicated Term Loans 250 General Bonds 255 Private Placement Bonds 258 Floating Rate Notes 262 4.6.5 FN Component Discounted Flows 265 4.7 Economic Model Cash Flows and Performance Measures 267 4.7.1 Constructs, Components and Project Cash Flows 267 4.7.2 Component Performance Measures 268 4.7.3 Project Performance Measures 270 4.8 Brief Summary 273 5 Risk Analysis Framework 275 5.1 General 275 5.2 Analytical Risk Analysis Framework 277 5.2.1 Probabilistic Characteristics of Model Variables 279 5.2.2 Four Moments of Performance Measures 298 Background 298 The Four Moments 299 5.2.3 Probability Model of Performance Measures 301 Pearson Distribution Family 303 Schmeiser-Deutsch Distribution Family 311 5.2.4 Comparison of the Approximated Moments to Exact Cases 314 Example 1: Summation of Gamma Variables 314 Example 2: Product of Log Normal Variables 317 5.3 Sensitivity Analysis of Performance Measures 320 vii PART III: IMPLEMENTATION 6 Decision Support System Design 325 6.1 Introduction and System Structure 325 6.2 System Database 326 6.3 System Model Base 327 6.4 System Interface 330 7 Decision Support System Validation and Applications 337 7.1 Introduction 337 7.2 Example 1: Validation of Model Performance Measures 338 7.2.1 Example Project Variables 340 7.2.2 Net Present Value 350 7.2.3 Internal Rate of Return on Equity and Total Capital 359 7.2.4 Total Cost, Total Revenues, and Construction Completion Time 359 7.2.5 Debt Service and Loan Life Cover Ratios 362 7.2.6 Sensitivity Analysis of Project Variables 362 7.3 Example 2: DSS Modeling and Analysis Features 372 7.3.1 Case 1: Preliminary Appraisal Study and Deterministic Analysis 372 7.3.2 Case 2: Semi-Detailed Appraisal Study and Probabilistic Analysis 380 a) Total Investment Cost 382 b) Traffic Demand and Project Revenues 382 c) Project Operation and Maintenance 385 d) Project Inflation Variables 385 e) Project Financing and Financing Measures 386 f) Uncertainty Modeling of Project Variables 390 g) Project Measures and Risk Analysis 390 7.3.3 Case 3: Detailed Modeling and Sensitivity Analysis 399 a) Detailed Cost Estimating 399 b) Detailed Traffic Demand 406 c) Cost Aggregation 417 d) Summary 419 v i i i 8 Conclusion and Recommendation 421 8.1 Conclusions 421 8.1.1 Capital Investment Proj ects 421 8.1.2 Generalized Economic Model 422 8.1.3 Risk Analysis Framework 423 8.1.4 Decision Support System 424 8.2 Recommendations for Future Work 426 8.2.1 Capital Investment Projects 426 8.2.2 Economic Models 426 8.2.3 Risk Analysis Framework 427 8.2.4 Decision Support Systems 428 Bibliography 430 Appendices 450 A The First Four Moments of a System Function 451 A . l General 451 A.2 The Expected Value 453 A,3 The Second Central Moment 454 A.4 The Third Central Moment 455 A. 5 The Fourth Central Moments 458 B Calculation Sheets for the Example Project 465 B. l Distribution Models And Variables In The Example Project 466 B.2 Net present Value And Sensitivity Analysis 469 B.3 Internal Rate Of Return On Equity 473 B.4 Internal Rate of Return On Total Capital 477 B.5 Total Capital Expenditure 481 B.6 Total Revenues 484 B.7 Construction Completion Time 486 B.8 Debt Service and Loan Life Cover Ratios 489 ix List of Figures Figure 1.1: Evaluator's analysis menu 17 Figure 1.2: A debt stream using Syndicated term loan financing method 17 Figure 1.3: Evaluator's probabilistic output for the total cost of a group of work packages 18 Figure 1.4: Thesis Guide 21 Figure 2.1: Project company structure and agreements 30 Figure 2.2: Requirement structure 39 Figure 3.1: Flow chart for road project appraisal, Reutlinger (1970) 95 Figure 4.1: Economic model - Interrelated components and basic constructs 109 Figure 4.2: A construct has properties and methods 109 Figure 4.5: Shape functions 132 Figure 4.6: Shape function and global reference of material inflation of a work package 132 Figure 4.7: Capital expenditure component structure 135 Figure 4.8: Time and network properties in the C E component 136 Figure 4.9: Capital expenditure common and specific variables 139 Figure 4.10: Inflation strategies for the CE, RV and O M components 140 Figure 4.11: Work package cost calculation using capital forecasting methods 147 Figure 4.12: Material cost cash flow function 150 Figure 4.13: Work package cost calculation using decomposed methods and quantity windowl55 Figure 4.14: Work package material and labor cost window methods 156 Figure 4.15: Work package indirect and discrete cost methods 157 Figure 4.16: Capital expenditure component areas and levels 159 Figure 4.17: A work package may belong to any or all levels in the C E component 160 Figure 4.18: Total capital expenditure cash flow 163 Figure 4.19: Structure of the revenue and operation components 167 Figure 4.20: Properties of a revenue stream 168 Figure 4.21: Aggregated revenues modeled as sinusoidal rate function, see Table 4.1 170 Figure 4.22: Toll bands for shadow toll structure 172 Figure 4.23: Unit rate revenues using trend analysis for demand forecasting 174 Figure 4.24: Forecasting demand using polynomial regression and demographic indicator 176 Figure 4.25: GDP (demographic indicator) modeled as a fourth degree polynomial 176 Figure 4.26: Forecasting project demand by elasticity-based methods 181 Figure 4.27: Uniform rate function for modeling base current demand 181 Figure 4.28: Project demand by stated preference technique 187 Figure 4.29: Total revenues cash flow 191 Figure 4.30: "Aggregated Methods" for operating cost 197 Figure 4.31: Normal rate function for marketing cost 197 Figure 4.32: Detailed methods in the O M component 199 Figure 4.33: Periodical and cumulative cash flow for the O M component streams 203 Figure 4.34: Financing component structure 210 Figure 4.35: Properties and methods of the FN classification 211 Figure 4.36: Description of debt elements, interest and fees are not shown 213 Figure 4.37: Debt drawdown as a percentage of capital expenditure 216 Figure 4.38: Debt drawdown at start of specified work packages 216 Figure 4.39: Amortized repayment methods with flexible maturity clause 226 Figure 4.40: Bond interest/coupon rates during repayment period 229 Figure 4.41: Interest payments on loan Tranches during grace period 231 Figure 4.42: London Inter-bank Offered Rates on six-month deposits 233 Figure 4.43: Commitment fee payments 241 Figure 4.44: General term loan with fixed interest rate 247 Figure 4.45: General term loan with floating rate of interest 249 Figure 4.46: Syndicated loans in the international market 251 Figure 4.47: Syndicated term loan 253 Figure 4.48: General bond issue 257 Figure 4.49: Private placement bond 261 Figure 4.50: Floating rate note 264 Figure 4.51: Tranches, interest payments and principal repayments of two debt streams 266 Figure 4.52: Project cash flows 269 Figure 5.1: Deviation between forecasted and actual revenues, (Dedeitch 1993) 276 Figure 5.2: Probabilistic risk analysis framework 278 Figure 5.3: Triangular distribution 289 Figure 5.4: Modeling uncertainty of time variables 292 Figure 5.5: Modeling the uncertainty of a work package cost 292 Figure 5.6: Uncertainty of total cost (Parameterl) modeled as Log Normal distribution 292 Figure 5.7: Modeling uncertainty of service rate in a revenue stream 294 xi Figure 5.8: Linear demand modeled as an uncertain 294 Figure 5.9: Elasticity of demand modeled using a Normal distribution 295 Figure 5.10: 6-month floating interest rate described as a sinusoidal function with uncertain parameter variables 296 Figure 5.11: 6-month floating interest rate described as a sinusoidal function. 296 Figure 5.12: Discrete costs and their application times modeled as risk variables in an O M stream 296 Figure 5.13: Discrete cost modeled as Gamma model 297 Figure 5.14: Pearson distribution types against the criterion k (Elderton and Johnson 1969) 304 Figure 5.15: Regions in (p 1, (32) plane for various distributions 304 Figure 5.16: Total capital expenditure fitted to a Pearson distribution 310 Figure 5.17: Total capital expenditure fitted by S-D distribution 313 Figure 5.18: Cumulative distribution of Gamma and Pearson Type III for T 316 Figure 5.19: Density function of Gamma and Pearson Type III for T 316 Figure 5.20: Cumulative distribution of Gamma and Pearson Type V for O 319 Figure 5.21. Sensitivity coefficients of capital expenditure variables in the NPV formulation 323 Figure 6.1: Components of the decision support system 327 Figure 6.2: Data flow diagram of the decision support system 331 Figure 6.3: Building a new proj ect in Evaluator 333 Figure 6.4: Selecting a project component to work with in the system 333 Figure 6.5: The analysis menu in the system 334 Figure 6.6: Too many forms opened at the same time triggers errors 335 Figure 7.1: Cash flow structure of example proj ect 339 Figure 7.2: Design work package 345 Figure 7.3: Construction work package 346 Figure 7.4: Revenue stream duration and calculation method 347 Figure 7.5: Total number of units for the project 348 Figure 7.6: Unit selling price defined as Triangular distribution 348 Figure 7.7: Drawdown, interest payments, and repayments of term loan for the project 349 Figure 7.8: Net present value results for the example project 356 Figure 7.9: Cumulative distribution of the net present value 357 Figure 7.10: Comparison between DSS and simulation for NPV distribution 358 Figure 7.11: Internal rate of return on equity 360 xii Figure 7.12: Internal rate of return on total capital 361 Figure 7.13: Total capital expenditure analysis 363 Figure 7.14: Total revenue analysis 364 Figure 7.15: Construction completion time analysis 365 Figure 7.16: Debt elements for the example project 366 Figure 7.17: Debt service cover ratio for the example project 367 Figure 7.18: Loan life cover ratio of the example project 368 Figure 7.19: Normalized sensitivity coefficients of the capital expenditure variables 369 Figure 7.20: Normalized sensitivity coefficients of revenue variables 370 Figure 7.21: Normalized sensitivity coefficients of revenue variables 371 Figure 7.22: Economic structure of the highway project, preliminary analysis 372 Figure 7.23: Design cost profile 373 Figure 7.24: Road construction cost profile, see Figure 7.26 373 Figure 7.25: Road structures cost profile 373 Figure 7.26: Modeling the cost and time of the road construction work package 375 Figure 7.27: Project cash flows for the base preliminary case of the highway project 376 Figure 7.28: Project cash flows after government contribution 377 Figure 7.29: Revenues cash flow before government contribution 378 Figure 7.30: Revenues cash flow after government contribution 379 Figure 7.31: Semi-detailed economic structure for the highway project 381 Figure 7.32: Actual A A D T on highway project example and forecasting models 384 Figure 7.33: Actual and forecasted A A D T for the life of the highway project 384 Figure 7.34: Toll inflation modeled as sinusoidal function during project life cycle 385 Figure 7.35: Bond cash flow payments for the highway project 387 Figure 7.36: Debt service cover ratio for the highway project 388 Figure 7.37: Loan life cover ratio for the highway project 389 Figure 7.38: Distribution models of some variables in the highway project example 392 Figure 7.39: Risk analysis of total cost of total design and construction cost 393 Figure 7.40: Probability distribution of construction completion time for the highway project 395 Figure 7.41: Risk analysis of current-dollar total revenues for the highway project 395 Figure 7.42: Risk analysis of net present value of the highway project 398 Figure 7.43: Work packages and network for bridge "Bl" of the highway project example 400 Figure 7.44: Detailed estimating of the "excavation, structural" work package 403 x i i i Figure 7.45: NPV normalized sensitivity coefficients for some capital expenditure variables 404 Figure 7.46: GDP models and RFP estimates 407 Figure 7.47: Modeling cars revenue stream using stated preference technique 408 Figure 7.48: Relationship between cars traffic, toll rate and time 412 Figure 7.49: NPV normalized sensitivity coefficients for some revenue variables 413 Figure 7.50: NPV sensitivity coefficients for some revenue variables 415 Figure 7.51: Pavement work package assigned to road construction C E area 1 417 Figure 7.52: Probabilistic characteristics of the total cost of the road construction work packages- C E area 1 418 xiv List of Tables Table 2.1: Rights Dimension: Possession-attribute characteristics 45 Table 2.2: Rights Dimension: Revenue-attribute characteristics 50 Table 2.3: Obligation-dimension for the selected PPP projects and related acts 57 Table 2.4: Liabilities-dimension for the selected projects and acts 71 Table 3.1: Road project appraisal 96 Table 4.1: Rate functions 118 Table 4.2: Area functions 126 Table 4.3: Properties of work package variables 137 Table 4.4: Disbursement profiles for project cost in current price terms 143 Table 4.5: Capital forecasting functions 144 Table 4.6: Components of total operating cost 193 Table 4.7: Sequencing of a BOT financial package 208 Table 4.8: Repayment methods 219 Table 4.9: London Inter-Bank Offered Rates on US dollar deposits 233 Table 4.10: Spreads (in basis points) on international bank loans 234 Table 4.11: Currency distribution of syndicated loans 251 Table 4.12: Construct and component cash flow modeling equations 267 Table 5.1: Four moments of probability distributions in the risk analysis framework 282 Table 5.2: Pearson distributions and calculations 307 Table 5.3: Characteristics of three Gamma variables 315 Table 5.4: Characteristics of the summation T of Gamma variables 315 Table 5.5: Percentile values of the exact Gamma and Pearson III of the summation variable 316 Table 5.6: Moment characteristics of the three cost variables 318 Table 5.7: Characteristics of the product of Log-Normal variables for the C2 318 Table 5.8: Percentile values of the exact Log Normal and Pearson Type V for C2 319 Table 6.1: Table Definitions for the capital expenditure database 328 Table 7.1: Explicit formulation of the example project (Mathcad sheet) 341 Table 7.2: Variables used in the example project (Mathcad sheet) 342 Table 7.3: Description of variables used in the example project (Mathcad sheet) 343 Table 7.4: Calculation sheet for the risk analysis of NPV 351 Table 7.5: Comparison between DSS and simulation results for NPV 355 XV Table 7.6 Moment characteristics of internal rate of return on equity and total capital 359 Table 7.7: Moment characteristics of total cost, revenues, and construction completion time 359 Table 7.8: Constant dollars costs and revenues for the preliminary analysis 373 Table 7.9: DSS results for the preliminary study 375 Table 7.10: Constant-dollars costs and revenues for semi-detailed analysis 380 Table 7.11: A A D T on the highway example project 382 Table 7.12: Modeling the uncertainty of project variables 391 Table 7.13: Cost items in a single bridge of the highway project and estimated duration 401 Table 7.14: Detailed labor and equipment cost for the structural excavation 401 Table 7.15: Work packages in the highway example project 404 Table 7.16: A A D T (vehicles-day) and GDP (in millions of dollars) 406 Table 7.17: Work packages in the three case studies of the highway example project 419 xvi Chapter 1 Introduction 1.1 General Economic modeling and risk analysis are two processes that constitute an important body of knowledge necessary for performing appraisal studies of capital investment projects. Decisions in the public and private sectors depend largely on the findings of such appraisal studies. A number of methods and techniques have been developed to address these processes. The purpose of this thesis is to make further contributions to the development of these methods and techniques, in order to improve the quality of appraisal studies and consequently the quality of decisions. Specifically, this thesis describes a generalized economic model, a risk analysis framework and a decision support system developed for the economic evaluation and risk analysis of capital investment projects at the appraisal stage of a project's life cycle. The next section overviews the scope of the thesis. This is followed by a section that describes some of the characteristics of capital investment projects and the problem issues raised in the design of economic models for such projects. Research objectives and methodology are then introduced and supplemented by some of the major thesis findings. Finally, a guide to the chapters of the thesis is introduced in the last section. 1 1.2 Scope Definition An investment for the development of a project can be defined as: "A long-term commitment of economic resources made with the objective of producing and obtaining net gains in the future. The main aspect of the commitment is the transformation of liquidity - the investor's own and borrowed funds- into productive assets, represented by fixed investment and net working capital, as well as the generation of liquidity again during the use of these assets" (Behrens and Hawranek 1991). Capital investment projects as defined in this thesis embrace this definition and include those projects that feature large capital needs for initial development, substantial recurring and replacement costs, an extended service life, and exposure to many uncertainties during all phases of the project life cycle. Infrastructure projects are typical examples of capital investment projects, and generally include several types of projects as follows (World Bank 1994, Dias 1994): 1. Public utilities: e.g. power, gas, water and sanitation systems, and telecommunications; 2. Public works: e.g. roads, bridges, airports, and urban transportation; and 3. Social works: e.g. education and health care facilities. The realization of such capital investment projects can be arranged through several delivery systems. Traditionally, public infrastructure has been delivered by the public sector through public funds. Recently, and especially in the past decade, several revenue-generating infrastructure projects have been delivered by the private sector through public-private partnerships (PPP) - a system that provides for more involvement of the private sector in the development, financing, marketing and operation of public infrastructure (Augenblick and Custer 1990, Price Waterhouse 1993, UNIDO 1996). Examples of these projects are the Fixed Link or Channel Tunnel (U.S. $ 16 billions) between U K and France; Northumberland Strait Crossing 2 Project - Confederation Bridge (Cdn $ 840 million, US$ 661 in 1992) between New Brunswick and Prince Edward Island, Canada; the Second Severn Crossing (£ 300 million, US$ 549 million in 1992) in UK; and, the State Route 91 Median Improvement project (US$ 126 million) in California, USA. The business environment of capital investment projects and the PPP delivery system will be emphasized in the thesis. The delivery of capital investment projects whether through the private sector, public sector or a public-private partnership, passes through several stages that start with the appraisal stage in which the viability of the project or the investment is assessed by the interested parties. These parties may include government or public sector agency, private sector developer (promoter), lenders, and investors (e.g. equity investors). Capital investment projects under the PPP delivery system typically feature the interaction of these four parties. Project appraisals represent a crucial requirement for these parties before making decisions on whether or not to undertake the project. A total project appraisal comprises several types of appraisals including (Frohlich et al. 1994, Behrens and Hawranek 1991): commercial, technical, environmental and economic. Generally, a technical appraisal involves an analysis to determine if the project design, development and operation will be technically sound, completed on time and within budget, and according to the required standards and specifications. Therefore, the technical appraisal analyzes, for example, the various technologies that will be employed, time and cost estimates and allocations, and resource requirements. A commercial appraisal seeks an analysis of the market for a proposed project in terms of: potential customers; prices, production and demand; competitiveness; and, industry regulations. The main output of the commercial analysis is the formulation of a demand function(s) essential for making demand and revenue forecasts for the proposed project. An environmental study is an analysis required to determine the impact of a project both during construction and after on the ecological (air, water, soil) and socio-economic environment. Thus, it deals with determining how environmentally sound the project is and how responsive the design, development, and operation of the project is to necessary environmental requirements. The economic1 appraisal, the fourth appraisal type, draws on and integrates the information from the other appraisals in order to perform an economic analysis and evaluation for use by the decision-makers. The general objective of the economic analysis for the private and public parties to a project is to assess the economic viability of the project under the prevailing and future conditions of the project. Specifically, the objectives of economic analysis are: to determine whether the project will generate an acceptable return; to determine the most attractive project alternative that will achieve the greatest economic return; to determine the most critical variables that will affect the investment; to determine the impact of different proportion of debt to equity in financing a project and whether the project can service its debt; and, to determine the flow of financial resources required during a project life cycle. The output of the analysis can further be used by the public sector to decide on which delivery system to use when procuring the project, to assess the bids submitted in a competitive tender, to check the strength of project cash flows and determine whether contributions will be required (UNIDO 1996). Economic analysis and evaluation of capital investment projects will be emphasized throughout the thesis. Essential to the economic analysis of a project are an economic model and an analysis of risks. Nearly all the frameworks that have been developed or proposed to date for risk analysis and management of projects have a risk analysis stage in which economic models and quantitative 1 The focus in this thesis is on the investment or financial analysis of a project, mainly from the perspective of the private sector, but also for use by the public sector in assessing projects that are required to be self-financing. For such an analysis, the focus is on actual or realizable cash flows, not imputed cash flows. It is recognized that other analyses such as benefit cost analyses would be conducted by government in order to account for imputed flows as well as other societal value systems. Having said the foregoing, the words economic, financial and investment analyses have tended to be used as synonyms throughout the thesis. 4 risk analysis techniques are employed (APM 1997; ICE et al. 1998; PMI 1996; Chapman and Ward 1997; Thompson and Perry 1992; Hertz and Thomas 1983; Perry and Hayes 1985; Tummala and Burchett 1999; Al-Bahar and Crandall 1990). A project economic model is an economic structure depicting all cash flows (costs and benefits) to be experienced during all the phases of a project's life cycle. Attached to the structure is a set of one or more performance measures (economic indicators) that use these cash flows to provide a value or values with which a conclusion or a statement on a project's viability can be made. Risk analysis is performed on the economic model to provide a picture of the behavior of the performance measures under conditions of uncertainty, with the objective being to provide for better conclusions on project viability and for the establishment of risk management strategies. To help achieve the objectives of economic analysis for a project, the formulated economic model and the chosen risk analysis technique are generally integrated together as an essential component of a decision support system (DSS). A DSS is an interactive computer-based system that is intended to help decision-makers identify and solve problems and make decisions in strategic and tactical situations based on mathematical models and data. The main functions of a DSS are to facilitate in the formulation of alternatives, the analysis and interpretation of their impacts and the selection of an appropriate option for implementation (Sage 1991). A decision support system has three primary components: a data management component; a model management component; and, a dialogue management component (Pearson and Shim 1995). The development of a generalized and versatile economic model, risk analysis framework and decision support system for the analysis and evaluation of capital investment projects represents the main theme of the thesis. 5 1.3 Problem Definition The thesis problem statement is introduced in this section by way of a project example that highlights the main issues and problem areas which need to be addressed when appraising capital investment projects. The example is for a toll highway to be developed, financed, and operated by the private sector in a public-private partnership with a government. A technical appraisal for the toll highway would involve, as explained earlier, an analysis of the design, construction, operation and maintenance (O&M) of the project. Part of the analysis would cover the construction phase in terms of the following: • Construction Methods: to decide on the construction methods to develop the project; • Planning: to identify all project segments (e.g. work packages) and the order in which they are to be accomplished, i.e. logic that links them; • Scheduling: to determine the duration of each project segment and its times (e.g. early and finish times) using the logic devised in the planning stage (e.g. critical path method); • Resource Estimating: to determine for the resources (material, labor and equipment) of each project segment their quantity, pricing (unit prices, wages), escalation rates and production rates; and, • Cost Estimating: to apply to each project segment the appropriate cost estimating method (e.g. preliminary and detailed) and calculate total cost of the segment and the project. Part of the technical appraisal would cover the operation and maintenance phases as follows: • O & M Methods: to decide on the methods used for operation and maintenance, e.g. tolling system such as conventional vs. intelligent transportation system ITS (e.g. automatic vehicle identification AVI); • Planning and Scheduling: to decide on the O & M tasks, their type, order, duration and timing; and, • Resource and Cost Estimating: to decide for each O & M task how cost will be estimated, and the relevant quantities, prices and escalation rates for the required resources. 6 A commercial appraisal for the toll highway would involve an analysis to determine likely revenue sources and forecasts for the project during its operational life. Relevant analyses would include: • an analysis of the classes of vehicles that will use the highway and traffic volumes; • an analysis of traffic growth for a planned time horizon; • an analysis of traffic volume to be captured by the toll highway using, for example, an urban transportation modeling system or transportation choice models (e.g. Stated Preference Technique); and, • an analysis of tolling rates and adjustment mechanism (e.g. due to inflation rate changes). Part of the appraisal work involves carrying out a pure financial analysis to determine whether the project would attract the necessary finance if placed in the market. This analysis include: • sources and methods of financing (debt and equity); • type of debt, likely amount, and method of drawing (money advances); • debt currency and applicable exchange rates; • type of interest (e.g. floating and fixed) and applicable rates; • methods of repayment and term of debt (maturity); and, • insurance and guarantees required by lenders. Findings from the technical, commercial, and financial appraisals provide essential input to the economic analysis of the toll highway. All the information from the various analyses are integrated in an economic model to determine the value of one or more project performance measures. These measures include, for example, total costs (e.g. total construction cost), total revenues and associated cash flow function(s), life cycle cost (LCC), net present value (NPV), internal rate of return (IRR), benefit-cost ratios (B/C) and debt-service-cover-ratio (DSCR). The analysis at this level is generally referred to a deterministic, fixed-value, analysis. 7 A sensitivity analysis and/or risk analysis is then performed on the measures to determine their behavior under conditions of uncertainty. How comprehensive is the sensitivity and/or risk analysis applied to the measure depends on how detailed the model structure in representing the project. For example, the economic model may accept total revenues as mere single sums for each year during the operation period of a project. The uncertainty around any of these single revenue sums might then be modeled, say, by a normal distribution. Alternatively, if the model is comprehensive in representing project revenues, then the uncertainty in each of the model parameters that model a single revenue sum, e.g. initial volume of demand, traffic growth rate, toll rate and inflation rate, might be modeled separately. Generally, project appraisal requires the investigation of several alternatives or scenarios under, for example, different design methods, different construction methods, different financing methods and different tolling methods. The economic structure for the scenarios may require the development of several economic models that reflect the methods required by the specified alternatives or scenarios. Further, sensitivity and risk analyses would be revised to deal with the parameters of each model. That would all be required in timely manner so as to inform decision makers of the appropriate alternatives or scenarios and their relevant results. All capital investment projects are required to pass through the foregoing appraisal processes. Since the economic analysis integrates the other analyses and since the appraisal process involves several issues as described above, the structure of the economic model and the risk analysis framework need to address and deal with the issues raised in the appraisal process. These issues can be described as challenges facing the designers of economic models and risk analysis frameworks, and can be summarized as follows: 8 1. Several Industries and Methods There are several project types, and depending on the type of project several industries, markets and/or sectors (e.g. transportation, maintenance, finance) need to be consulted for carrying out the various appraisals of a project. Since each industry, market or sector has its own business practices and calculation methods (e.g. transportation demand methods and finance mechanisms), the model should be equipped with such methods. 2. Multiple Phases and Sub-Phases When appraising a project, the analysis usually should be based on an economic model that mostly closely replicates the market place. A project usually includes several segments or phases such as design, construction, O & M , revenue, and finance. Some phases may be arranged in terms of work packages where each package has its own duration, logic and timing. Other phases, e.g. finance and revenue, have its own characteristics that the model needs to reflect in its structure. 3. Cash Flow Characteristics Cash flows can be described in three forms. The first is direct in terms of an assignment of discrete values at discrete points of time. The second is indirect where an estimate e.g. total cost of a work package, is distributed over the work package duration according to a specified pattern. The third is driven where an estimate computed according to a specified method is converted into cash flow distributed over time following the way the parameters of the estimating method actually change over time. The economic model should be capable of dealing with such cash flow methods. 4. Time Variables Most project variables change their values over time, such as resource prices; escalation, interest and exchange rates; and, project demand. Further, work package quantities may have consumption rates or patterns that can predict the amount of work done at a 9 specified period of time. The model needs to include several rate functions and patterns and needs to allow its variables to change over time following any of such patterns. 5. Several Measures There are several performance measures that can be used in the assessment of a project economic viability. Depending on the objectives of decision-makers, a project appraisal may be required to carry out evaluation and analysis on more than one the performance measures. The model structure should be equipped to deal with multiple measures. 6. Uncertainty Models and their parameters are mere abstracts of reality. Similarly, estimates of costs, revenues, and financial indicators represent the best estimates that can be made at the time of analysis. Uncertainty is inherent in all such estimates. Risk analysis frameworks need to model the uncertainty of the estimates and the variables making these estimates. 7. Several Alternatives Economic models should be able to formulate several project alternatives or scenarios in order to investigate the project under varying assumptions or conditions. The foregoing issues describe how detailed and general an economic model needs to be in order to realistically model capital investment projects. Further, there is a need to address the above issues when appraising capital investment projects in order to improve the quality of appraisal studies and consequently the decisions based on them. Current support systems with their underlying economic models and risk analysis frameworks lack the ability to address both the detailed and generality aspects of capital investment projects as described in the foregoing seven issues or problem areas. The current thesis tries to fill this gap, which is the main thesis contribution, by developing a generalized economic model, risk analysis framework and decision support system for the appraisal of capital investment projects. 10 1.4 Research Objectives and Methodology The main objectives of the thesis are: 1. To develop an understanding of the characteristics of capital investment projects; 2. To develop a generalized economic model that can reflect the detailed and generality aspects of capital investment projects; 3. To develop a risk analysis framework that deals with the detailed aspects of the generalized economic model; and 4. To develop a prototype decision support system implementing both the generalized economic model and the risk analysis framework. 1.4.1 Capital Investment Projects The study explores the characteristics of capital investment projects in its first objective and uses an analysis framework, called requirements structure, which classifies and links the requirements in a project to the essential elements of a project structure (e.g. construction and operation) the execution of which takes a project into full development until the end of its life cycle . The output of this analysis is a description of the requirements attached to the development (e.g. design, construction), finance, operation, revenues, and liabilities of a project. The accomplishment of such analysis required an extensive knowledge about how projects are executed. The study was fortunate that it had access to a set of five capital investment projects in terms of their tender and/or contractual documents including contractual agreements between government and developers, request for proposals (RFPs), request of expressions of interest (EOIs) and government acts legislated specifically for the projects. The set of projects studied represents large revenue-generating infrastructure transportation projects, developed under the 2 Part of the PPP characteristics can be reviewed in: Russell A. D., and Abdel-Aziz, A. M. (1997)."Public-Private Partnerships and Public Infrastructure." 1st International Conference on Construction Industry Development: Building the Future Together, Dec. 9-11, Singapore. 11 public-private partnership delivery system, as examples of capital investment projects. In addition to the contractual documents, the study was supplemented by an extensive literature search and insights from work done by the author for the use of public-private partnership in the procurement of educational infrastructure3. The fulfillment of the first objective in understanding capital investment projects represents a first stage in systems development, i.e. system analysis (Kendal and Kendal 1999). This stage helps to identify the characteristics of such projects and consequently helps in the delineation of the structure of the proposed economic model and decision support system. 1.4.2 Generalized Economic Model The generalized economic model developed in this study is a multipurpose hierarchical network- based time function structure. In its simplest form, the economic model provides the amount of money spent/acquired at any given point of time. In its complicated forms, the model's time functions can be arranged to serve a specific purpose; for example, to compute financial and economic performance indicators at any given point of time. One concept behind the generalized economic model is the formation of cash flow classifications. A classification represents a domain, such as maintenance, and possesses the properties and methods of that domain. This allows a cash flow of that classification type to inherit the domain's properties and methods. Properties of a classification cover, for example, time units/intervals, duration, and logic; and methods depending on the classification type include cost estimating methods, demand and revenue modeling methods, financing instruments 3 Abdel-Aziz, A . M . (1998). "Public-Private Partnerships In Infrastructure Development: The North East Burnaby High School Case Study." Rep. Public-Private Partnership Advisory Committee, Economic Partnership Branch, Ministry of Employment and Investment, Ministry of Finance, B.C., Canada. . 12 (e.g. syndicated term loans, private placement bonds). Typically, each method has a number of variables and one or more mathematical expressions. The "structure" of the generalized economic model consists of four components reflecting four classification domains in capital investment projects namely, capital expenditure, revenue, operation and maintenance, and project financing. The basic elements in a component are called constructs, which are called work packages for the capital expenditure component and streams for the other components. Each construct of a component represents a cash flow that has the same type of classification representing the component. With the generalized economic model's concept and structure, a component (e.g. revenue component) in the model is represented by a time function formulated by all of the cash flow time functions representing the constructs of that component. Arranging the components' time functions, then, serves to compute a specific performance measure as explained earlier. The generality aspect of the economic model is reflected by the richness of the properties and methods of each of the four classifications represented in the model structure. Along with the earlier survey on the characteristics of capital investment projects, an extensive literature search was made to acquire the properties and calculation methods of each of the four classification domains and to acquire the knowledge about the economic modeling of projects. Numerous properties and methods of the four classifications were then employed in the hierarchical time- function structure of the generalized economic model. th The concept of classifications has served in addressing the 7 problem issue in economic model design (to model several alternatives). Any alternative can be formulated since the properties and methods used in the scenario will be selected from those of the model classifications without the need to build new models. The challenge was in dealing with the different types of properties 13 and methods of each classification knowing that each new construct would have its own values of such properties and would have any of the available methods; the final model time function would have to combine all such differences before processing any performance measure. While several of the estimating methods and economic indicators used in building the generalized model are not new, the contribution stems from refining and assembling them in a consistent manner and making them available in a generalized modeling framework suitable for the economic and risk analysis of investment projects. 1.4.3 Risk Analysis Framework Given the detailed aspects of the generalized economic model in representing any required project alternative or scenario and computing any of the mentioned performance measures, it is necessary to develop a framework for risk analysis that can deal with such details. In probabilistic estimating and subject to the available information about a variable, it is possible to deal with either a limited number of percentiles, the first two moments, the first four moments, or a full probability distribution for the variable. One approach or framework is to use Monte Carlo simulation, however, this will always require the use of full probability distributions to model the uncertainty of the variables. This might not be practical or feasible at all times particularly at the appraisal stage of a project. Further, the processing time for simulation of highly detailed economic structure of an alternative might be significant. Another approach is to use an analytical method that derives the four moments in all the hierarchical levels of an economic model until it reaches the top level that represent a performance measure (Ranasinghe 1990; Russell and Ranasinghe 1992). This approach, however, with the detailed aspects of the generalized model means that the four moments will need to be calculated several times within each construct (e.g. moments of labor cost, moments of material cost, ...etc), then for each construct of the total number of constructs that may be 14 added to a component, then to each of the four components, then finally to the performance measure. This approach, while possible, seems prohibitive in terms of calculation time needed for highly detailed project representation, which is usually the case in investment projects. The framework adopted in this study, instead, uses an analytical approach that derives the four moments directly for the performance measure function that represents the hierarchical cash flow structure of the generalized economic model. This serves to derive the uncertainty of any required function in the hierarchical structure, i.e. any project performance measure modeled through the economic model regardless of the form or the mathematical expression of the measure. The uncertainty of a performance measure is derived through the uncertainties of the variables used in the measure in two steps. The first derives the four moments of the performance measure and the second determines the distribution of the measure. The uncertainty of a variable is modeled in a probabilistic manner using 16 uncertainty modeling methods defined in three categories: (1) two and four moments; (2) three and five percentiles (Pearson and Tukey 1965, Keefer and Bodily 1983, Pfeifer et al. 1991); and, (3) full probability distributions (Bury 1999, Evans et al. 1993). This provides a general capability to model the uncertainty of a variable according to the data available for modeling the uncertainty. The four moments of any performance model are obtained by approximating the measure by a multivariate Taylor series expansion and taking expectations to derive the four moments of the measure using the moments of the variables (Kottas and Lau 1982; Siddall 1972; Russell and Ranasinghe 1992). Using this approach an expanded formulation is developed for deriving moments of performance measures in the risk analysis framework. Finally, a probability distribution for any performance measure is determined using its four moments and the characteristics of the four-parameters systems of frequency curves (Elderton and Johnson 1969; Hahn and Shapiro 1994; Johnson et al. 1994; Ord 1972; Schmeiser and Deutsch 1977). 15 1.4.4 Decision Support System4 The decision support system (DSS) represents the implementation part of the generalized economic model and risk analysis framework; the system has been named Evaluator. The system is designed in line with the three components of a DSS, i.e. data, model, and the interface components. The system's database uses the Jet database engine of Microsoft (Visual Basic 1998) and holds the data, methods, results, and graphs for any project defined in the system. The model base represents the generalized economic model and the risk analysis framework and is built using Mathcad programming (Mathcad 1998); graphs are processed using Excel (Excel 1996). The interface is designed using the development environment of Visual Basic (Visual Basic 1998). The interface is the managing module of the system and interacts with the other parts of the system using the Object Linking and Embedding (OLE) protocol in the Microsoft environment. The system works through windows and menus. With the system it is possible to experiment with a project during its appraisal by forming any required alternative or scenario, maintaining, reproducing, and processing information for the alternatives and obtaining results in several formats, e.g. tables and graphs, all through the interface only. Figure 1.1 shows the summarized analysis menu through which any desired processing could be chosen. Figure 1.2 shows part of the data for a debt stream that uses a syndicated term loan as a financing method in the financing classification (component). This figure shows the level of detail supported by the generalized economic model in representing capital investment projects. Figure 1.3 shows a typical output of Evaluator for the probabilistic analysis of a performance measure - in this case total cost of a group of work packages. 4 Abdel-Aziz, A. M , and Russell, A. D. (1999). "Decision Support System for Infrastructure Project Appraisal and Risk Analysis." The Canadian Society of Civil Engineers, 27th Annual CSCE Conference, June 2-5, Regina. 16 Die Eroject Analysis Widow Help Cash Row: Capital Expenditrue Cosh Flow: Revenues Cash Row; Operation & Maintenance Cash Flow: Finance Cash Flow: Project Loan Life Coverage Ratio Debt Service Coverage Ratio Sensitivity: Capital Expenditrue Variables Sensitivity; Revenue Variables Sensitivity: O&MVeriables Sensitivity: Financing Variables Probabilistic: £jet Present Value Probabilistic: Internal Rate of Return Probabilistic: Ufe Cycle Cost Probabilistic Aggregate B/C Probabilistic. Net B/C Probabilistic Capital Expenditure Level 1 Probabilistic Total Capital Expenditure Probabilistic Total Revenues Probabilistic: Total Operation and Maintenance Area 2 Area 3 Area 4 Area 5 Area 6 Area? Area B Area 9 Area 1D Figure 1.1: Evaluator's analysis menu ( I . FvnluHtnr [FX-001] - |Pioject Financing] li"> Debt Streams J s l K l , Debits [l Financing Method | Syndicated Term Loan j j General | 1 Month Interest | 3 Month Interest | 6 Month Interest | 12 Month Interest | Exchange Identification Stream ID [FNirJI Comments Syndicated term loan arranged by Bank of America. Floatig interest rate using LIBOR and split margin. Debt Time Zero "DTZ" (in project time unit) Rate | Physical Pay Exchange | Loan Drawdown j %s of Total Loan drawn at start of WPs M Total draw numbers and % of each in decimal 1.5 Term & Issue Debt Time Unit"DTUM | quarter (3-month) T) Debt Term, in DTU [To" Deb i Amount Draws D#1 % |D#2% |D#3% |D#1% 3 0.2 ;0.4 0.2 0 <l I For each Draw, erttBrwp/t = WPcode if draw is [3Q0E6 Debt/Project currency | Different Physical Payments | Yes [•] Dlwp/t |D2wp/t |D3wp/t |D4wp/t CEid2 CEid4 CEidIO <l 1 Interest period for each draw above. Interest Fixing Day | Start of Period T] IP after Grace P. | semi-annual (6-mo_»J Margin/Spread Type | Split. 5 Margins. WP-Uj«j Interest Ref. Time | Start of Project j*] D#1 IP |D#2 IP |D#3lP DM IP 3 3 l3 0 <l I _ WP Code |CEid2 M5, Margin Ml. Margin | M2. Margin 0 0.1 Dates are [periods after EF ofWP,in DTU. |0.05 M2 Last Date M3, Margin Joo? M 3 Last Date MA, Margin MA Last Date Use Interest Period Codes as follows: 2- (1-Month) 3- Quarter (3-Month) A • semi-annual 5 - annual Sinking Fund/Debt Repayment Grace Period (in DTU) ff? Repayment Method Note Repayment Periods rust be greoter than or equal to Interest Periods oftei Grace Period Amortized Principal. Balloon & Separate Interest Repayment Periods | semi-annual (6-month) _J Balloon %. dec. |o.2 Fees Manag. Fee. dec. at DTZ Expenses $ (on DTZ) [300000 • Agency Fee $. annually 50000 Commitment Fee. decimal m.n.2 Commitment Period j semi-annual (B-mo 0.045 Figure 1.2: A debt stream using Syndicated term loan financing method 17 Q. 0 . CO CO CM CSJ o 4 o + o + a a LU LU LU LU LU a in i— *r n r*. T— Ln r~- CD TP a r— TT o V ; CD CD CD o os_ CO CO iri ro'_ ca S5 > is m 0. =3 ci> >s 0 c < a La 3 CI) 0 01 2 3 < S 0 Id .2 5 CO >> c < in m '5 !E 1 I £ & 2 01 03 Q Q id > E > E = CO a a o a + + + LU LU LU CO c i ro C\J rs. : en CM LO cn Ps. un CO t—'• o > Q T3 cd T3 £ s 55 o o Q . X _2 Q . cd O "D a) LL "O > LU LU OJ o O 2 o "a Q) £ o n o a a CO o s o o o + + + + + LU LU LU LU LU CO a a o o CM a a o CO CSJ CD a a o r~- o a o CD CO CD o CO CD T3 a) Q Q 1.6 Research Outline The thesis is structured in a system development life cycle sequence as shown in Fig. 1.4. "Part I" is the analysis part that establishes the real characteristics and criteria of capital investment projects which should be reflected when building economic models and support systems. Therefore, chapter two provides a detailed analysis of a number of real infrastructure revenue- generating projects developed under the PPP delivery system. To explain the key features of capital projects, the requirement structure is introduced as an analysis framework. The structure has three dimensions, namely, rights, obligations and liabilities; each dimension consists of a number of attributes. Project possession (ownership) and revenues are the attributes covered for the rights dimension. Development, operation and financing are the attributes of the obligation dimension. General liability, risks and taxes are the attributes of the liability dimension. The analysis of such attributes explains much about the characteristics of capital projects. Chapter three is the point of departure for the work in this thesis; the chapter covers three main themes. The first establishes, based on the attributes of the requirement structure, that the properties and calculation methods used in the business environment of the eight attributes of the requirement structure should be the basis for the design of economic models. A review of previous work by other researchers on economic models and support systems is then introduced in the second part of the chapter. Finally, the chapter concludes with a list of some of the characteristics that should be recognized by the proposed model and support system. "Part II" is the design part, where the generalized economic model and the risk analysis framework are introduced. Chapter 4 presents a detailed description of the concept, structure and mathematical formulations of the economic model and its performance measures. The chapter starts with a general description of the classifications, methods, properties and shape functions 19 r PART I: System Analysis Chapter 1: Introduction Chapter 2: Characteristics of Capital Investment Projects Chapter 3: Proposed Economic Model Characteristics PART II: System Design Chapter 4: Generalized Economic Model Capital Expenditure Modeling Demand and Revenue Modeling Operation & Maintenance Modeling Finance Methods Modeling Performance Measures Chapter 5: Risk Analysis Framework r PART III: System Implementation Chapter 6: Decision Support System Design Chapter 7: System Application and Validation Chapter 8: Conclusions and Recommendation Figure 1.4: Thesis Guide 20 used by the model. Then, four subsections are introduced to provide a detailed description of the four components that comprise the model. A review of the methods used by the industry of each component is introduced followed by a development of the model formulations for that component. Formulations of the model performance measures are then introduced. Chapter 5 describes the risk analysis framework used by the decision support system. A detailed description of the analytical framework is introduced explaining its three main parts: modeling uncertainty of a variable, four moments of a performance measure, and distribution of the performance measure. A review of uncertainty modeling methods is introduced with an explanation of their implementation in the framework. The second part derives the first four moments of a performance measure using a multivariate Taylor series expansion. The third part explains how a probability distribution of a performance measure is determined using its four moments of the measure and the characteristics of the Pearson and Schmeiser-Deutsch distribution families. The framework integrates the three parts for risk analysis. "Part III" is the implementation part of the thesis where the design, testing and application of the decision support system are introduced followed by conclusions. Chapter 6 describes the design process of the system. It covers the three parts of the system; its database, model base and interface. Chapter 7 presents a simple example to verify the system output. This is followed by a detailed example of a highway project in order to demonstrate the generality and flexibility of the generalized economic model, risk analysis framework and the decision support system. Finally, conclusions and future work are described in Chapter 8. 21 22 Chapter 2 Characteristics of Capital Investment Projects 2.1 Introduction A cash flow model generally reflects the expenditures and revenues that are relevant to the party using the model. For example, if maintenance activities of a PPP project are the responsibility of government under a project agreement, then maintenance costs will not be seen on the cash flow model of the private sector developer. Instead, they will be included in the government cash flow model. Similarly, if land costs are subsidized by government, then these costs may not be included in a private developer's cash flow model as part of his ownership costs, unless the government requires otherwise. These requirements vary among project delivery systems and among projects. Thus, as a prerequisite to developing an economic model and support system for such projects to study their business environment in order to establish the various requirements which the model and the system will have to address. In delivering an infrastructure project, government may adopt a conventional delivery system or engage in a partnership with a private sector developer as an alternative delivery system (PPP). For the latter, several, if not all aspects of the project life cycle are covered in the project documents. These include contractual agreements, legislative acts/regulations, request for proposals and call for expressions of interest. These documents explain all of the requirements that have to be addressed in the developers' proposals. Consequently, before submitting proposals developers will carry out project appraisals as outlined in the previous chapter and 23 build models that express these requirements in cash flow elements if they can be quantified, or highlight them along with other non-financial aspects of the project (Lopes and Flavell 1998) for further assessment and possible negotiation if they are qualitative. Developers would do most of the same steps even for unsolicited proposals (Bederman and Trebilcock 1997). Based on its requirements government too would develop its own economic models to appraise the project and to evaluate the bids submitted. It may require the developer to submit a detailed cash flow as well ("Highway 104" 1995). Project lenders and investors would similarly appraise a project under the given requirements, particularly those that are related to revenues and guarantees. In conclusion, the study of project requirements as established by government or proposed by developers in an unsolicited bid constitute an essential step to identify the information necessary for project economic analysis and more importantly for the development of economic models. This chapter presents the results of a study of the characteristics of a number of PPP revenue generating infrastructure transportation projects as representatives of capital investment projects. PPP projects are emphasized since they typically embrace all phases in a project life cycle and experience inputs from nearly all project participants including government, developer, lenders and investors. The following section provides a brief background on PPP followed by a description of a structure proposed to organize project requirements. Then, a brief general description of the projects reviewed in the study is given followed by a detailed description of the attributes of the requirements structure. While transportation projects are emphasized, the requirements structure is broadly applicable to a diverse range of projects. Finally, the chapter ends with a set of conclusions relating to the use of the proposed requirements structure in the analysis of public- private partnerships and in building economic models. 24 2.2 Public-Private Partnerships (PPP) 2.2.1 PPP Arrangements The conventional delivery system for infrastructure involves government assuming full responsibility for financing, development (planning, acquisition, design, and construction), operation and maintenance of projects. The private sector is involved in this process through the provision of consulting, design, and/or construction services. For this delivery system government usually acts as a provider and deliverer of services through delivery methods designed to facilitate rigorous standards and control and to promote indicators of measurable performance. This approach is not intended to realize a financial reward or speculative gain for government, and the services are not withheld from those who cannot afford them (Flynn 1997; Baldry 1998). Financing such projects has traditionally been done using pay-as-you-go financing and debt financing (Robinson and Leithe 1990; Feldman et al. 1988). With pay-as-you-go financing, funds to pay for infrastructure costs are secured directly from government current revenues such as taxes, fees and user charges, interest earnings, and grants. Debt financing, on the other hand, requires government to tap credit markets to raise the necessary funds through issuance of debt, e.g. general obligation and revenue bonds and revolving loans. Despite the use of rigorous standards and control, Baldry (1998) explained that: "In practice, however, the history of performance of such projects [public sector projects] in general, and certain notable projects in particular, has indicated a less than satisfactory performance resulting in substantial cost and time overruns, inappropriate project outcomes and significant secondary effects in terms of disruption and frustration of operational and strategic activity." Baldry explained further that: "A service which is provided monopolistically by an arm of government, and which is free at the point of consumption, is divorced completely from the market economy, 25 resulting in inefficiencies of delivery with consequent cost and time penalties." A Public-private partnership (PPP) is an alternative project delivery system that provides for an increased involvement of private sector organizations in the delivery of functions that were previously the exclusive domain of government. This delivery system represents a viable alternative to the conventional approach as it provides a solution for the financial problem expressed by increased government debt and a capability to increase the efficiency of the delivery of public infrastructure in terms of time, money, quality and management (Kay 1993; Blaiklock 1992). For the developing countries, reducing budget deficits and government debt appears to be a strong motivation for using PPP arrangements (World Bank 1994; Augenblick and Custer 1990). For developed countries, e.g. the U.K., the objective of using PPP is directed more at harnessing the entrepreneurial, financial and management skills of the private sector in the provision of infrastructure ("New Roads" 1989; "Paying" 1993; Blaiklock 1992; Winfield 1996). A Public-private partnership (PPP) can be defined as a contractual arrangement between the public and private sectors for the development of public infrastructure where the two parties share resources, risks and rewards appropriately for the successful implementation of the project (Price Waterhouse 1993; CCPPP 1998; IBI 1995). Specifically, PPP is a partnership that defined the responsibilities for project design, construction, financing, operation and maintenance. The contractual arrangement usually takes the form of a concession or franchise; a concession "is the award of a right or license to build, own and operate a public infrastructure for a given period" (Blaiklock 1992). According to the allocation of responsibilities, a PPP arrangement can take several forms/modes, for example, Build-Operate-Transfer (BOT), Build-Own-Operate (BOO), Build-Own-Operate-Transfer (BOOT), Build-Transfer-Operate (BTO), Lease-Develop-Operate 26 (LDO), Buy-Build-Operate (BBO), and design-build-finance-operate (DBFO) (Price Waterhouse 1993; UNIDO 1996; Walker and Smith 1995; CCPPP 1996). The BOT arrangement is the most referred to approach; it is defined as (UNIDO 1996): "A contractual arrangement whereby a private sector entity undertakes the construction, including design and financing, of a given infrastructure facility and the operation and maintenance thereof. The private sector entity operates the facility over a fixed term during which it is allowed to charge facility users appropriate fees and other charges not exceeding those proposed in its bid and incorporated in the project agreement to enable the private sector entity to recover its investment and operation and maintenance expenses in the project, plus a reasonable return thereon. At the end of fixed term the private sector entity transfers the facility to the government agency or to a new private entity through public bidding." The BOOT arrangement differs mainly from the BOT arrangement in that the private sector entity (developer) owns the facility during the term of agreement. In BOO arrangements, the developer owns the facility in perpetuity. Financing for PPP projects is usually raised by the developer using both equity and debt markets. Equity comes in part through project developers and mainly through equity investors. Debt is usually is raised through "Project financing" instruments such as syndicated term loans and private placement bonds (see chapter four). The PPP approach has received much attention worldwide and several acts have been legislated promoting this alternative delivery system. Examples include the New Roads and Street Works Act ("New" 1991) and Private Finance Initiative in the U.K. (Moore 1994); the Virginia Public- Private Transportation Act of 1995 and its implementation guidelines ("Public" 1995), and the Minnesota Toll Road Enabling Legislation ("Toll" 1993) in the U.S.A. As of 1998, Public Works Financing reports in its database 2208 infrastructure concessions worth $1.1 trillion for various types of projects and services. "Of the 2,208 projects, 795 have been awarded since 1985 by governments in 64 countries for development and operation of $335 billion worth of power, road, rail, airport, water, institutional buildings, and other infrastructure facilities" (PWF 1998). 27 However, the PPP delivery system has not been successful in several projects because of political, social and financial reasons. For example, Texas High Speed Rail, a $5.6 billion 50- year concession project, was cancelled in 1994 because the developers were unable to raise financing for the environmental studies ($170 million) resulting in a loss of approximately $40 million by the developers ("Franchise" 1991). TH212 in Minnesota was cancelled in 1996 after one of the communities affected by the road project vetoed the project after the signing of the initial agreement ("TH 212" 1996). In Washington State, Substitute House Bill 1006 (SHB1006) was enacted in 1993 to allow for BOOT/BTO project procurement. A number of demonstration transportation projects were initiated ("Public" 1994). Due to a political change from Democratic to Republican coupled with a public outcry over the prospect of tolls, SHB1006 was amended in 1995 by SHB 1317 and followed by Substitute Senate Bill 6044 which dramatically affected the demonstration projects resulting in the cancellation of some. The literature regarding project and implementation aspects of PPP arrangements has been growing. Some best practice guidelines have been published which provides general description of the implementation process of PPP arrangements (World Bank 1990; UNIDO 1996; CCPPP 1996; Price Waterhouse 1993; ACPPP 1998; Merna and Smith 1996a). Other literature explains the analysis process regarding the evaluation and negotiation of proposals and appropriateness of a project for PPP procurement (Tiong and Alum 1997a, 1997c; Ngee et al. 1997; Ashley et al. 1998; Dias and Ioannou 1996). Still other literature covers general issues dealing with financing, risks and guarantees of BOT projects (Tiong 1990b; Tarn 1995; Levy 1996; Walker and Smith 1995; Shen 1996; Merna and Smith 1996b). General contractual and financial aspects of BOT projects have also been treated (McCarthy and Tiong 1991; Tiong and Alum 1997b; Haley 1992). Critical success factors for PPP projects have also been examined (Tiong et al. 1992; Tiong 1995a, 1995b, 1997; Tiong and Yeo 1993; Keong et al. 1997; Blaiklock 1992). 28 2.2.2 PPP Project Company Early in the implementation process of a PPP project, a consortium of private sector companies is formed either to prepare an unsolicited proposal for a project of interest or to review an issued RFP and if warranted carry out appraisal study on the project and submit a proposal. The consortium usually enters into a preliminary consortium agreement in order to submit a solicited/unsolicited proposal to government. The consortium agreement represents the initial step before the establishment of a PPP Project Company. A PPP project company, sometimes referred to as developer, promoter, concessionaire or owning company, is the company which, in a typical BOT project will ultimately be responsible for project development, design, construction, finance, operation and maintenance of the project. A project company is usually a special purpose-company formed as a partnership or joint venture (Clough and Sears 1994; Beidleman et al. 1990). The company usually includes partners such as large engineering and construction firms, equipment suppliers, operation and maintenance companies, and equity investors (e.g. investment banks). Figure 2.1 shows typical participants to a PPP. A project company assembled for the final realization of a PPP project will have to enter into several contractual agreements as shown in Figure 2.1. These agreements include, for example, a development agreement (sometimes called concession agreement, omnibus agreement), a construction agreement, an operation and maintenance contract, financing contracts (loan agreement), insurance contracts, and supply and off-take contracts (McCarthy and Tiong 1991; Walker and Smith 1995; Pyle 1997; UNIDO 1996; Merna 1996a; Payne 1996). Several of these agreements are mandatory on the project company as being required as a satisfaction to the government or as a condition to obtain finance from project lenders. 29 Figure 2.1: Project company structure and agreements, adapted from McCarthy and Tiong 1991, UNIDO 1996, Haley 1992, and Merna and Smith 1996a. Several authors have examined various facets of the skill sets and factors that lead to the successful promotion and winning of PPP projects by a project company. For example, McCarthy and Tiong (1991) explained that a project company would have to play several roles during the term of a project including roles as consultant, sponsor, contractor and equity holders. "Strength of consortium" was one of six critical success factors in winning a BOT contract as explained by Tiong et al. (1992) and Tiong (1996). Dias and Ioannou (1996) introduced "Desirability Model, DM", a multiattribute evaluation model that assesses the capability of a private sector company to become a promoter for a project as well as the attractiveness of a 30 project to be promoted by a given company. DM included nine attributes relating to company competence and which were grouped under three categories: internal organization characteristics; production capability; and, financial resources and constraints. The diversity of participants to a PPP shown in Figure 2.1 helps to explain the range of skill sets and roles required by a project company to carry out and manage a PPP project. These roles and responsibilities can be usefully grouped under four categories of functions: (1) project company, (2) project, (3) investment, and (4) government. 1. Concession company related functions • Assess private company technical, financial and legal resources • Perform a project needs-assessment • Select qualified partners and form the concession company • Draft contractual agreements with various parties (e.g. contractors and suppliers) • Prepare technical, financial, and operating proposals • Carry out project administration • Perform company and shareholder administration 2. Project related functions regarding design, construction, and operation • Carry out project studies (e.g. technical and environmental) • Develop project conceptual design and related construction methods • Conduct value engineering, construction method studies and constructability reviews • Prepare operation and maintenance plans for facility management • Develop construction management framework (e.g. planning, scheduling and control) • Perform quality management 3. Investment related functions regarding cost, economic analysis and financing • Estimate capital, O&M, administration, and management costs • Conduct commercial studies (e.g. supply, demand, competition, and tolls) • Seek and evaluate various sources for project financing • Administer debt and prepare security package • Conduct inflation analysis • Perform economic evaluation (e.g. NPV, IRR, benefit-cost analysis, cash flows) • Perform risk analyses project viability 4. Government related functions • Negotiate government support to obtain all project approvals and legislation • Negotiate concession terms and conditions with government • Negotiate various government support and guarantees • Seek support to obtain financing at favorable terms • Seek guarantees on project revenues, minimum demand, and no-second facilities • Seek subsidization for market imperfections • Manage stakeholder involvement process 31 2.2.3 PPP Project Evaluation Project evaluation is an essential stage for all participants involved in a PPP project. Evaluation involves, as explained in the first chapter, several types of appraisals including technical, environmental, commercial and financial appraisals. For the provision of an infrastructure project, government will check, in a preliminary feasibility/appraisal study, the viability of the project under several delivery systems, e.g. conventional and PPP. Later, if a procurement decision is made to adopt a PPP approach, government during the selection process of a private developer will evaluate the submitted proposals against stated criteria with the overall objective being to achieve a technically sound, cost effective, and financially attractive solution. Evaluation criteria normally reflect the technical, environmental, commercial and financial aspects of the project, and generally include price and non-price criteria. Generally, decisions in the selection process may be based on the value of particular criterion (e.g. net present value, initial user charges) or on a weighting/scoring system of several criteria (Merna 1996; Tiong and Alum 1997a). For example, Tiong and Alum (1997a) explained a differentiation between the evaluation criteria as (1) MUST criteria for which a developer must comply if it is to continue in the process and (2) WANT criteria for which a weight or number of points is attached in a scoring system. Several studies analyzed the relative importance of the technical solution (including environmental issues) and the financial package (including commercial aspects) of a proposal when evaluating the submitted bids in competitive tendering (Tiong 1995b;Tiong and Yeo 1993). The ability to provide an attractive financial package was judged to be critical under the conditions that the project is technically certain, the level of tolls to be charged are the government's main concern, competition is keen, and project financing is uncertain (Tiong 1995a; Tiong and Alum 1997b; Merna 1996b). 32 The economic/financial criteria that are generally of concern to a government when appraising a project or evaluating proposals in the selection process are included in the list below (Tiong and Alum 1997a, 1997c; Moles and Williams 1995). The comprehensive evaluation of projects involves the treatment of risk and uncertainty, and hence the evaluation of these criteria generally involves sensitivity and risk analysis. The criteria, grouped under four categories, are: 1. Cost aspects • Acquisition costs • Development costs (e.g. design and construction) • Operation and maintenance costs • Life cycle cost 2. Financial aspects • Equity amount and debt/equity ratio • Sources of debt • Interest rates • Debt drawdown and repayment schedules • Currency of debt and repayments • Financial charges (e.g. management and syndication fees) • Financial commitments and security package 3. Commercial aspects • Initial service charges (e.g. tolls and tariff) • Adjustment mechanism of service charges (e.g. due to inflation or demand changes) • Length of project/concession period • Demand forecasts • Projected revenues • Lease payments 4. Investment aspects • Net present value • Rates of return on equity and total capital • Benefit/cost ratios • Payback period • Insurance policies • Project cash flows 33 Project lenders and investors would look at several of the foregoing criteria particularly those that cover the cost, commercial and investment aspects. The commercial aspects receive the greatest attention since project financing is generally raised based on the merits of future project cash flows, the robustness of the demand forecasts (Nevitt and Fabozzi 1995), and debt service cover ratios. Lenders would be much more interested in performing sensitivity and risk analyses on the project (Woodward 1995) since they are the main source of finance for capital projects and their funds would be exposed if the project experiences completion risk, cost overrun risk, and lower than expected demand. Developers, before submitting a solicited/unsolicited proposal, would also evaluate all of the foregoing criteria together with sensitivity and risk analyses in order to check the viability of the project. Typically developers would have to prepare for the prospective lenders and equity investors a project information memorandum detailing the findings from a project appraisal/feasibility study along with results of project sensitivity and risk analyses in order to explain the merits of the project and receive the required credits and funds (Nevitt and Fabozzi 1995; Rhodes 1993; McDonald 1982). 2.2.4 PPP Project Risks Risk analysis, identification and management is of major concern to government, developers, lenders and investors when appraising a PPP project. PPP project risks receive special attention since these projects involve exposing large capital investments before any revenues/return are obtained in the extended life cycle of the project. Among a large list of PPP project risks, the major risks that receive considerable attention by all project participants particularly the lenders and investors, an increase in capital costs, delay in construction completion, partially completed 34 project/construction and less than expected demand or revenues represents. To hedge against these risks, several security packages, insurance packages and guarantees are usually required. A considerable body of literature have been devoted to the analysis and identification of PPP project risks in general and those identified for particular projects (Tiong 1990a, 1995c; "Paying" 1993; Wang et al. 1999; Moles and Williams 1995; Woodward 1995; UNIDO 1996; Merna and Smith 1996a; Merna and Adams 1996; Stein and Pote 1997; Yeung 1997; Beidleman et al. 1990; Hurst 1996). In general, the literature tends to classify PPP risks into two broad categories: 1. General (or Country or Global) risks: i.e. those risks that are related to a country's political, economic and legal environment. This include, for example, risk of expropriation and nationalization or cancellation of the concession, change in laws and regulation (e.g. discriminatory taxation regimes), restriction on repatriation of revenues or profits, currency inconvertibility risk, fluctuation of foreign exchange rates, devaluation risk, and inflation risk. These risks are generally outside the management and control of a private sector developer and the estimation of their impact is usually problematic. These risks can affect the whole project as an investment and lead to a total loss, reduce demand for the output of the project, or lead to the deterioration of project cash flows and consequently the viability of the project. Lenders will not provide any credits to a project if they are not sufficiently comfortable with the political stability of the country of the project. Considerable negotiation is the norm in PPP projects to address several of the above risks, particularly if government does not provide for their allocation or management (e.g. by exchange rate guarantee) when it issues an RFP. Along with negotiation, some political insurance vehicles can be used to deal with such risks. Insurance products can be 35 obtained, usually at high premiums, to reduce the effect of political and exchange rate risks. Examples include those provided by the Overseas Private Insurance Corporation (OPIC), the Multilateral Insurance Guarantee Agency (MIGA), International Finance Corporation (IFC) and the World Bank (UNIDO 1996; Hurst 1996). Also, capital market instruments such as swaps, options and futures can be obtained to hedge against the movements in the currency and interest rates thus providing some risk relief to project developers (Coopers & Lybrand 1987). 2. Specific (or elemental) project risks: i.e. risks that are related to project construction, operation, finance, and revenues. The major risks in this category include construction cost overruns, completion delay risk, interest rate risk, demand/revenue risk (volume and/or price), supply risk (volume and/or price), and force majeure risks. Construction cost overruns and completion delay risks pose critical risks since they affect project return - the whole investment could be lost, particularly if the project is not completed. Demand risks affect project cash flow, which can directly affect project return and the ability to repay project debt on time. Interest rate fluctuation can pose a risk of raising project total cost, which can lead to a decrease in return if service charges (e.g. toll rates) or project/concession periods are not adjusted accordingly. Several of the risks under this risk category are within the control of, and usually allocated to, the private sector developer who may generally distribute such risks through secondary contracts with its participants. Insurance and bond packages are usually used for the comfort and security of government and lenders; RFPs may stipulate specific 36 coverage during construction and operation. Negotiation is usually involved for force majeure risks particularly for any event that is not covered by insurance. The effect or impact of risks in the specific risk category can generally be analyzed and quantified through a formal risk management process (APM 1997; ICE et al. 1998; PMI 1996). The success of a PPP project depends to large extent on the allocation and management of the general and specific risks; and the rational in the allocation usually calls for a particular risk to be allocated to the party most able to control or influence it. 37 2.3 Government Requirement Structure 2.3.1 Requirement Structure Description Government requirements under public-private partnership arrangements cover all of the contractual, technical and financial aspects of a project. Each requirement may be described by some specific aspects and consequences that need to be considered carefully under PPP arrangements. An example of a government requirement is project ownership; government may provide for specific forms such as public ownership, private ownership, or both where transfer of ownership may occur during the term of development. The domain of ownership can be the whole project or it may be defined for individual parts of the project, e.g. real property (land), facility (improvements), movable and immovable properties, and intellectual property rights. The domain of ownership may have consequences for both governments and developers in terms of tax treatments or on the availability of rights or licenses to use a technology after project transfer to government. During the appraisal stage, where possible, government requirements would be converted into financial terms in a cash flow model and the requirements or their effects on a project scenario analyzed. Therefore, the use of PPP calls for governments to address the range of conditions that they may stipulate for each requirement and the consequences of each. Based upon a detailed study of several PPP projects and acts (explained below), a useful structure for describing the key features of a PPP project during its life cycle has three major dimensions: rights; obligations; and, liabilities. These dimensions along with explanatory attributes are shown in figure 2.2. The rights dimension describes the various rights given by government to the private entity in return for carrying out a specified set of obligations. Possession of the facility and access to revenues constitute the primary attributes of the rights dimension. Obligations represent the promises that the developer and the government agree to be 38 bound to under the agreement. Obligations can be described by three attributes: development obligations (e.g. planning, design, construction, environmental); operating obligations (e.g. operation and maintenance); and, financing obligations. The liabilities dimension covers the most controversial issues in PPP negotiations and includes three attributes: general liability (e.g. tort or third party liability and facility damage); risk liabilities; and, tax liabilities. Project Requirements Rights Possession Revenues 1 Obligations Development Operation Financing Figure 2.2: Requirement structure / \ Liabilities Liability Risks Taxes A starting point is that all requirements and associated attributes belong to, or are the responsibility of, government, as is ownership of a facility. Under PPP, selected or all attributes of a requirement can be temporarily or permanently assigned to another party. Therefor, various allocations of the attributes of the requirement structure can be assembled which in turn leads to the spectrum of procurement modes commonly associated with PPP (e.g. BOT, BTO). These allocations provide for various agreement titles such as development, franchise and omnibus. The following subsection explains the projects and acts used for investigating government requirements. Following this description, a subsection is devoted to each of the attributes identified under the three requirement dimensions. Description of an attribute starts with a general summary of findings followed by a detailed description of main terms and conditions. 39 2.3.2 Projects and Acts Considered The Channel Fixed Link, UK/France (BOT) The Fixed Link is a twin bored tunnel rail link with associated service tunnel under the English Channel between England and France. The approximately 50 km link was developed at a cost of £ 9 billions. The invitation to promoters (equivalent to a RFP) was issued in 1985 ("Invitation" 1985) with no prior call for expressions of interest. Agreement with the successful developer was reached in 1986 ("Concession" 1986), and the project was legislated in U.K. by the Channel Tunnel Act in 1987 ("Channel" 1987). The project was inaugurated in 1994. The promoter, Eurotunnel, consisted of a consortium of British and French engineering and construction companies and banks: Channel Tunnel Group and France-Manche. Second Severn Bridge, UK (BOOT, DBFO) The Second Severn Bridge is a 920-m. cable-stayed bridge and two 2000 m. approach viaducts over the Severn Estuary between England and Wales with a total cost of £ 300 million. Following a Notice and Invitation for Prospective Tenderers ("Second" 1988) (equivalent to a call for expressions of interest) and Tender Invitation ("Second" 1989), the project was arranged as a DBFO (Design-Build-Finance-Operate); however, it was known also as a BOOT project. The project promoter, Severn River Crossing, SRC, a joint venture between UK's John Laing and France's GTM Entrepose, expanded after awarding the project in April 1990 to include Bank of America and Barclays' De Zoete and the concession agreement was signed in. 1990. Project approval by Parliament came in November 1991 in the form of the Severn Bridges Act in 1992 ("Severn" 1992). Construction started in 1992 and ended in 1996. Along with the provision of a new crossing, the government required the promoter to take over the responsibility for the existing crossing over the Severn Estuary and inherit its debt ("Severn" 1988, 1989). 40 Highway 104 Western Alignment, Nova Scotia, Canada (BOT) The Western Alignment is a 45-km four-lane highway which forms part of Highway 104 (Trans Canada Highway) in Nova Scotia. The total capital cost of the project is Cdn $ 113 million. The Request for Proposals, issued in 1995, was followed by six addenda ("Highway" 1995). The legislation required for the project forms the Western Alignment Act ("An Act" 1995), W-A Act. This Act provided for the creation of the Western Alignment Corporation as a single-purpose corporate vehicle, not a public authority or crown corporation. This corporation was created to assist the developer, Atlantic Highways Corp, a subsidiary of Canadian Highways International Corp., in contracting with the Province for the realization of the project. Northumberland Strait Crossing Project NSCP, New Brunswick/PEI, Canada (BOOT) The NSCP bridge crosses the Northumberland Strait between New Brunswick and Prince Edward Island, Canada. The estimated cost of the 13.5-km bridge was about Cdn $ 840 million although the actual cost was in excess of this. After receiving unsolicited proposals for the project, the government issued a CFEI in 1987 ("Northumberland" 1987) followed by a call for proposals and six addenda in 1988 ("Northumberland" 1988). The project was legislated by the Northumberland Strait Crossing Act ("An Act" 1993a) and financial closing with the developer, Strait Crossing Inc., was made in 1993 after a number of environmental assessments and challenges in the courts. A number of 39 separate agreements and 400 documents were executed including a development agreement, a construction contract, a project security agreement, a project trust agreement, an operation agreement, and a regional agreement (FHWA 1996). State Route 91 Median Improvement, California, US (BTO) The State Route 91 (SR91) median improvement is one of four demonstration projects in California authorized by Assembly Bill 680 ("Assembly" 1989). These projects were proposed 41 by the private sector after issuance of Guidelines for Conceptual Project Proposals ("Guidelines" 1990) by the California Department of Transportation, Caltrans. The SR 91 development franchise agreement signed in 1991 was granted to the developer, California Private Transportation Corporation (CPTC) with final approval of the agreement being contingent on meeting environmental requirements. The agreement was amended and restated in 1993 ("Amended" 1993). The project represents a 10-mile (16-km) all-electronic tolled new four express lanes within the center median of the State Route 91. Construction started in 1993 and the project opened in December 1995. Estimated cost of the project was U.S. $126 million. US Acts After California's initiatives as explained by Assembly Bill 680, several States enacted similar legislation for PPP projects. In Minnesota, Toll Road Enabling Legislation 1993 ("An Act" 1993b), TREL, was enacted to provide for the development of BOOT/BTO projects through the TRANSMART program ("Request" 1995). Highway TH 212, proposed following the initiation of TRANSMART, was first signed by the government ("TH 212" 1996). However, during a 30- day voting period required by the TREL Act for community approval, one of four cities on the proposed highway voted against it. In Virginia, the Public-Private Transportation Act ("Public" 1995) was enacted to provide further refinements for the implementation of PPP projects following the earlier Virginia Highway Corporation Act of 1988 ("Virginia" 1988) and the Qualifying Transportation Act of 1994 ("Qualifying" 1994). While the above projects receive detailed description in the following subsections, only the major requirements included in the US acts, i.e. those of Minnesota and Virginia, are considered in the description. 42 2.3.3 Rights Dimension Rights Dimension: Possession Attribute The investigation of government requirements for this attribute has emphasized the types of properties and related government requirements for the possession and transfer of property. Several types of properties have been mentioned in the selected projects and acts. These include: 1. Land or real property needed for the project; 2. Improvements or the facility the developer agreed to construct on the land (e.g. highway, bridge, structure, movable and immovable properties, plant, equipment); 3. Airspace premises (e.g. over and under the right-of-way); and, 4. Intangible properties needed for the development, operation, and ownership (e.g. intellectual property rights, patent rights, project documents, reports, drawings, plans and specifications). Generally, not all of these properties have been explicitly identified and defined in the RFPs or agreements. Except for the U.S. experience, governments seemingly dislike to explicitly state that the developer will be the owner of the project. All-encompassing statements which treat the transfer of all property at the expiration of the agreement, such as with the Second Severn Bridge (BOOT), are typically featured in the agreements. When lease agreements are made for land or right-of-way, the reversion of the improvement (facility) may be written explicitly such as with the NSCP (work is deemed to be a fixture to the land) or implied to occur with the reversion of the land at the end of the lease such as with the Channel Tunnel. The two BOOT projects, Severn Bridge and the NSCP, leased the land to the developer at a minimal rent. For the Channel and Western Alignment BOT projects, the first provided land at cost and the second was free. Intangible properties such as intellectual property rights were a subject of transfer for the Channel Tunnel. However, for the Western Alignment, the RFP stated that it was to be under government possession at all times. 43 Table 2.1 provides a summary of the relevant characteristics of the possession attribute. Projects were generally required to be transferred or revert at no charge to the government at the end of the agreement. While this transfer requirement might be common in PPP projects, exceptions can be found. For example, the Texas High Speed Rail project had a requirement that at termination the government had the option to purchase the facility at its fair market value ("Franchise" 1991). Channel Fixed Link The Channel tender invitation explained that the chosen promoters would benefit from a concession to construct and operate the Link for a period of time, and the rights of the promoters would expire when the concession was terminated. The governments required the Link to be kept in the public domain ("Concession" 1986). The term Fixed Link was an all-encompassing term, defined to include a twin bored tunnel rail link with associated service tunnel, together with the terminal areas and dedicated facilities for control of, access to, an egress from, the tunnels. The term also included plant, machinery, movable and immovable equipment and railway shuttle rolling stock. Land, referred to as Operational Land and Construction Site Land, for the project was provided by the governments after compulsory acquisition and/or agreement and was leased to the developer. The agreement required the promoters to pay in respect of such lands the cost of acquisitions for land acquired after the agreement, market value for land acquired before the agreement, and the cost of vesting the foreshore and bed of the sea in the British Minister. Upon expiration or termination of the agreement, the Fixed Link will be handed over to the two governments. Immovable property will revert to government and land leases will end. 44 c o *S a U o g +: & 2 l« - i _, IH CU cu • ' a. o Cd ' u CU . C o s cr -. C3 B o OH O u cu CU o a. c o p. cd cd 3 > « 2 s 6 £ £ 1 - 1 6J3 a _cd _a cd s ° o ° ^ Id .2 i - — cu X> OH 3 O a > • — CD M 2 CUH C/l * _ H cd a — cd a cd —1 cu a a s H v a a cd a U -a OB S -S 2 cd O o " o a'a cu o S9 cd cn 1M cu § « u cu cn 5 a 2 -a + J p. <u cu I H CL, CU OH _o > CD T3 Is IT „ <D <~ 3 O <_> cn U cd £3 « c cd CD a o fi OH CU ^ o 3 OH 2 1 a a 1 3 £ o S 5 s E ° -2 S o l ^ S fi £ o . S cd cn a O CU o, cu OH a £ 1 3 °^ •73 e g a T. cd .a cu cu .p a x —' ia ^ t / i o )-H Crt O cd u" a 2 O — o a <2 a S O Cd ° a o fi cu OH o cd 3 T 3 1) o S ° cd o " cu a S cd T3 cu o '-5 cn « ^ T 3 a o cd oo C3\ pQ Pi < «5 T3 Id ^ •a '5 T3 H T3 cu u cu T3 . O _> cu ajj "o •° "§ a o o-.fi +j cu ° rS cu "c? OH . • CU O o o s ^ 5 > CU cu O . T 3 -2 >, cu X I > * J (D cn Q ° CD > O . fi B •— o ,cd 3 cd cu OH o cr -3 o a cd cd 2 a 2 ' « C3 H _ SJC o . m S3 H^ 2 u •a " 2 l ^ a 3 S cu cu cu cd • f i - d i i ^ > — . U S " o g cu (i ^ s 2 H H ^ ^ I O u y — cn ^ g cu « > Q . 2S cu a o cu OH O > o S o 3 CQ < o g o I D O Cd H>-* O cn CU 8 a o o t-l &, +-> o CO O O cd % o <u JO ? fi _o 'co co CU CO CO O fi 'co fi CU •8 o • i -H i — H OH C ^ +-» O fi -fi 2 CM CU In respect of intellectual property rights, the government required the developer to grant a non- exclusive royalty free license to use or sub-license any intellectual property which will be vested in the government for purposes of construction and operation of the Fixed Link after the expiration of the agreement. Second Severn Bridge The tender invitation ("Second" 1989) stated that both crossings would be highways for which the Secretary of State is the highway authority. The Severn Bridges Act ("Severn" 1992), S-B Act, granted the British Secretary of State the right to construct the new bridge and to delegate all relevant functions and power to levy tolls to a private promoter according to a concession agreement. The S-B Act authorized the acquisition of lands and to grant a lease or other interest in or right over any land according to a concession agreement. Such lease was to be provided at a peppercorn (insignificant) rent as mentioned by the Tender Invitation. The tender invitation provided for the transfer by the promoter of both crossings to the government at the end of the concession period. The S-B Act provided for the transfer of all property, rights, and liabilities of the concessionaire with no explicit definition of property. Highway 104 Western Alignment The PvFP explained that the project as developed will be part of the public highway system and the ownership of the project facility at all times is vested in the Province of Nova Scotia. All the needed lands were acquired by the government, at its cost, and made available to the project. The RFP provided for the project facility to include the road and all improvements, buildings, erections and structures, and all chattels, machinery, equipment, materials, tools, forming part thereof or used in the construction or operation. The first addendum provided for construction 46 equipment not to be part of the facility during project operation upon request from one of the RFP respondents. Along with public ownership of the facility, the RFP provided for the exclusive use and possession of the government of all project materials and information and their related patents, copyrights and other industrial and intellectual property rights including trade secrets. With such prior possession of the project by government, no transfer clause was included in the RFP. Northumberland NSCP The federal government explained as one of its objectives that the project be financed, designed, constructed, operated and maintained by the developer under a long-term subsidy agreement. The 1988 NSCP Proposal call explained that the development agreement would include a ground lease and a schedule of requirements, terms and conditions. The project facility was described in the NSCP Proposal Call to include collectively the lands, the work complete in all respects, with all operation and maintenance systems in place, and any other improvements or structures located on the lands. The work means all improvements and all appurtenances, which the developer agreed to construct on the lands. For project possession, the proposal call states that "the work shall be fixtures to the lands and shall become the absolute property of the Landlord [Minister of Public Works] without compensation upon the expiration or termination of this lease". In the first addendum, the government explained further its intention by stating that the contract with the developer would be to build, to own, and to operate the facility for 35 years, after which it would be transferred at a nominal amount to the government. The second and third addenda explained that the nominal amount was meant to effect that the facility will revert to the government after the 35 year ownership period, and that alternative private sector ownership could be considered and be the 47 subject of negotiation after selection. 'Own' was defined by the third addendum to mean: "to own the leasehold interest in the facility". In the sixth addendum the government stated that for purposes of financing and taxation, the project was a private sector venture. State Route 91 AB680 ("Assembly" 1989) authorized Caltrans to enter into agreements for the construction by and lease to private entities of transportation projects. The Bill mentioned three types of lease: lease of rights-of-way (real property); lease of airspace over or under state highways; and, lease of the facility (private transportation project). Lease terms would be up to 35 years during which private entities would charge fees for the use of the facilities. Facilities would be state-owned at all times and revert to the state after expiration of the lease term at no charge. In the SR 91 agreement ("Amended" 1993) Caltrans agreed to lease all of its rights, title to and interest in the real property, together with all improvements including the facility for a "construction lease" term and a 35-year "operating lease" term. Caltrans made available its power of eminent domain to be used in right-of-way acquisitions if requested by the developer (CPTC). Acquisitions would be made at all times at CPTC's cost. The SR 91 agreement provided for Caltrans to issue a notice of acceptance after the facility achieved substantial completion. On the acceptance date, the construction lease term would expire, CPTC would transfer title to Caltrans and the operating lease term would start. Upon expiration of the operating lease term, CPTC would be required to surrender the real property and the facility. Several grants and rights were identified in the SR agreement. A 1.5-mile Absolute Protection Zone was defined to protect CPTC's franchise rights and economic viability. Under this provision, Caltrans would not finance, grant or convey any franchise to any party other than 48 CPTC for the development or operation of a public transportation project within the protection zone, unless the proposed facility did not represent economic competition to the project. CPTC was given the right of first offer and first refusal with respect to the development and operation of any commercial airspace improvement, over, under, on or within the State Transportation Facility, State Route 91 right-of-way, in Orange County, California. Such airspace rights could run up to 99 years. Further, CPTC was granted an option for the development of three phased extensions to the current facility to be exercised during the term of the agreement. Rights Dimension: Revenues Attribute The revenue attribute is the second right assigned by governments to developers. The investigation emphasized the requirements in connection with toll and revenue arrangements along with other rights that may be given to developers; a summary is given in Table 2.2. Typically, BTO and BOOT modes as implemented in US projects and acts, provide for freedom in toll setting and application of congestion pricing (except for the Virginia act, Table 2) while setting up caps on the rates of return. Governments, in general, tend to control the term of agreement through statements directed at early termination if debt or revenues are satisfied (Western Alignment, Severn Bridge) and with an indirect statement if rates of return are met (SR91 and US acts). Generally, governments seem to have two objectives: (1) control the amount of revenue generated by the project and (2) control the amount of revenue the developer is entitled to earn. Both works to achieve the equity1 principle used in evaluating finance methods for public projects (Blackburn and Dowall 1991, Robinson and Leithe 1990). 1 Equity is determined in two ways. The first holds that those who benefit from a service should pay for it. A user's ability to pay for a service is the second equity principle (Robinson and Leithe 1990). 49 u eg u cs 18 0) 3 s ID OS A 0 DH O ft. g I* o u IU "o" I  £ o £ o ts ^ s • 3 2 <2 § T3 60 C - o 3 o o <U • J3 Z a. ". ° vi 3 v© o I • a <2 3 i - vi o c3 £ u " (J V) <u • s i s § H 60 o H s o S ^ — o •a .a D t+-i CJ (J a. cs 3, >- aj <2 £ 8 g>C0 I s 8-8 ~— CB v ; ' S ^ ro 3 s § T Y T-I = <U « ca a> • S iu 5 x> 2 «° Q) £ X> T3 3 CS CS S2 E . tu oo <L> i - T; o j ; « H u ts s > (» T 3 <u c o 2 u tu 'C e/3 03 CS a • 2 . 2 cs S <U CS I- kH £ «< o (A O M X> H 7 3 •B - o V H 3 ,0 CS ^ s _u § x 'C cs cs U 3 <N o «/} <K CS 2 u 3 <u :rr> ^•s m i 2 H TS, XJ r° 'U &, _o 13 > 2-5 CD L H • a c" — . o 3 S c S 2 * « o o M l il T3 CS T 3 £3 CS u o V) S E E m o c CS u 3 C - o (U v i 3 cs = C o c/l T3 (U <U >< — 3 X) CB v i (U ? 3 ° (U cs A-1 X! - o o -3 v i a, u= CS _ Q u Id IU T3 <U J3 C v i cs 3 J3 cs o to E o . 2 § ' 5 3 J) T 3 • IU v) -ic CS 1) >. . ^ S3 3 C 3 <u E - S (N 3 -a U 73 V 3 XI 3 00 3 o U so 3 '5b 3 CS T3 60 3 § 2 CD C > O •a 5 D 3 •â  cs cs £ SO s = ^ s 'U ^ k. cs • a cs H « 0 O ON « ha oj pa « R -< u o CO 3 t; CS o. 6 E • ° 2 T 3 c*_, 1) o 5 <u cs u 5J J3 o CS IU 3 3 > (U — V J V I o X W ̂ e - s 3 ^ ( Z 3 2 2 ^ s- E 2 • 3 cj S > CO <D E • « C 3 S 2 l-i CS •< oo 0\ < -5 OH ~ CM > O 60 3 o o >• CS T3 X) 00 CO 3 <T t- <~ a. ° o 2 u 5 > X ) o cs *- 3 V ) O O v i 3 co 3 < £ ° s H O CO CQ IT) J 2 £ 1 o C3 CJ <D '5* )-i ex <u -*-» o CD "53 c/> c2 V 2 O CD O ca ^3 o 3 1 (D C3 CD > CD c o •»—I C3 CD s Q +-» (N <N CD H Channel Fixed Link Governments through the Channel Invitation and Concession Agreement offered political guarantees not to intervene in the conduct and operation of the Link and not to terminate the promoters' right to construct or operate a Link provided that the concession terms are adhered to. The developer was given commercial freedom in setting tunnel tariffs. The Agreement stated that "The Concessionaires will be free to determine their tariffs and commercial policy and the type of service to be offered". Earlier in the Channel Invitation ("Invitation" 1985), the government explained that the duration of the concession would consider the type of project selected and would be sufficient to allow repayment of debt and permit a reasonable return on equity. The concession period was initially set by the agreement as 55 years. Due to delays and cost overruns the concession was extended 10 years by the governments involved (Huot 1995). Second Severn Bridge The S-B Act provided the authority to levy tolls on both bridges to be exercised by the concessionaire. The government in the Tender Invitation required promoters to state the initial tolls required by each class of traffic proposed. Further, the invitation required the basis for any subsequent adjustment of tolls due to inflation, the index of inflation to be used and its weighting in the adjustment formula, components of cost to which it would be applied, minimum toll increases, and the time period between adjustments. The Tender Invitation allowed for differential tolls by day or date provided road safety was not impacted. The government allowed for additional proposals for adjusting the toll level (and/or concession period) in order to take account of actual traffic flows diverging from the bid assumptions. For this case, detailed information was required regarding the mechanism for adjusting toll levels and/or the concession period, the traffic demand assumptions, and the upper limit for the 51 concession period. As enacted later in the U.K., specifications for maximum tolls to be levied on new roads have been described in the New Roads Act ("New" 1991). This act provides for specifying maximum tolls if the road consists of a major crossing for which there is no reasonably convenient alternative. Toll periods in this act may end on a specific date, or be determined by the achievement of specific financial objectives, or passage of specified number of vehicles, or the earlier or later of specified dates. The S-B Act provides the concessionaire with the power to levy tolls for a maximum of 30 years. The S-B Act provided, however, for early termination of this right if the revenue requirement had been met, i.e. the toll income received is equal to or greater than the amount the concessionaire is entitled to receive by the concession agreement. Highway 104 Western Alignment The W-A Act granted the Corporation the right to collect tolls and this became the responsibility of the selected respondent. The RFP provided for initial tolls and any proposed mechanism for increasing the initial tolls during the concession period to be established during negotiation of the Omnibus Agreement. However, the RFP stated that tolls would be sufficient to (a) pay the debt incurred to build the facility, (b) establish an operating and maintenance reserve, and (c) provide for required repair and rehabilitation work. The government included with the RFP a study of the current and future traffic volume and revenue forecasts for a range of toll road options. Beck, president of Canadian Highways, explained that tolls were initially set at Cdn $3 per car, Cdn $2 per axle for trucks and Cdn $4 for recreational vehicles (Beck 1997). Further, he explained that if debt service coverage were not met, tolls were to be adjusted automatically and that tools would be adjusted for inflation. 52 Government required that the Omnibus Agreement term be limited to the length of time toll revenues were needed to repay all the money borrowed or made available to pay for construction as well as to pay for any reserve requirements as mentioned previously. The government reinforced through the first addendum that the selected respondent must earn its return from the construction contract, operating contract, and a return on any debt, which the respondent chooses to hold. In the RFP addendum the, the government stated further that: "DOTC will give no assurance or guarantee that a fair market rate of compensation will be achieved by the Selected Respondent within the Concession Period to be fixed in the Omnibus Agreement." Northumberland NSCP The NSCP Act allowed the government to make regulations prescribing tolls for the use of the crossing. Toll collection was the responsibility of the developer and tolls were to be adjusted annually for 75% of the consumer price index (CPI). As explained in the proposal call and its first and fifth addenda, a toll revenue floor was established to be the greater of either $8 million in 1988 dollars or the actual toll revenues experienced by the ferry service in the full year preceding the date of substantial completion of the facility. It was explained also that toll rates may be increased by more than the permitted 75% of the CPI should toll revenues be lower than the established floor and additionally if tax changes or insurance premiums result in cost increases. Short falls in toll revenues were to be recouped in the succeeding year. With no explicit cap on toll revenues, the government required a separate account for toll revenues where the distribution of revenues would follow certain priorities. The toll distribution priorities included (1) payment of insurance premiums on a $150 million accident policy, (2) payment of interest and capital for the financing secured against toll revenues, (3) payments into a facility repair and maintenance fund, and (4) payment of the balance to the developer. 53 State Route 91 Caltrans entitled CPTC to establish, levy and collect tolls, fees and charges for the use of the facility. Toll adjustments and arrangements were at the discretion of CPTC without prior approval or evaluation of Caltrans. Further, CPTC was authorized by the SR 91 agreement to implement a congestion pricing arrangement to respond to dynamic traffic flows and to maintain the highest levels of service. According to the daily demand patterns toll rates move from $0.5, $1.0, $1.5 and $2.0 in four time zones with the rate being $0.25 for off-peak hours. Rates for Monday to Thursday differ from those for Friday and the weekend (PWF 1995) while High- occupancy vehicles (HOV) pay no tolls. However, the SR 91 agreement provides for tolling HOVs after two years of operation if the debt coverage ratio is not met. While Caltrans established no cap or control on toll rates, it established a 17% base return rate (BRR) for use in discounting calculations; this rate is to be adjusted annually, and upward only, according to the average yield on 5-year U.S. Treasury Bonds. CPTC is entitled to a reasonable return on investment (ROI), comprised of a base ROI and an incentive ROI for any fiscal year. CPTC is entitled to retain the available cash in any fiscal year as a base ROI whenever the base NPV calculated using BRR is less than zero. The incentive ROI is implemented to encourage CPTC to modify and improve the facility to maximize the number of vehicle occupants travelling during peak demand periods on the combined facility, SR 91. An incentive return rate gives 20 basis points (0.2%) increase on the base return rate for each 1% increase in the annual peak hour vehicle occupant volume; however, incremental increases may not exceed six hundred basis points (6.0%) for any fiscal year. If the base NPV is equal to or greater than zero, CPTC will share available cash for the fiscal year with Caltrans only if the total NPV calculated at the incentive return rate is less than zero; otherwise excess revenues will be directed to the State Highway Fund. 54 2.3.4 Obligations Dimension Obligations Dimension: Development and Operation Attributes This section deals with government requirements under the first three obligation attributes. For all projects, emphasis was placed on the spectrum of functions that governments require developers to be responsible for (planning, design, construction, environment, operation and maintenance); and the power governments have in project review, inspection and approvals. Generally, the projects studied showed that all project functions are the responsibility of the developer unless government suggested or required certain functions to be its responsibility such as maintenance, traffic management, and police services. For example maintenance was highly encouraged to be provided by the government for the SR 91(BTO), Western Alignment (BOT), and the Minnesota TREL Act (BOOT/BTO). Bylaws required by developers (e.g. for traffic management) were generally subject to government approvals and could not compromise safety. Under traditional procurement arrangements, governments have an active involvement in all project functions. For PPP, governments seek to maintain a role in those functions for which they have a responsibility for the public at large. The investigation showed that governments would provide for (1) the appointment of a representative or agent and consultants or independent engineers, (2) the default and substituted entity clauses in project agreements to handle cases of defaults by the proponent, and (3) the monitoring functions during development and operation. Supervision and approval duties may undergo more scrutiny in public-private partnerships. Supervision provides for checking compliance with standards and specifications, and takes place while work progresses. Approval provides for accepting the work after it has been reviewed or checked. Approval may hinder the progress of work if it takes time to be done. Government will 55 generally carry out both processes and promoters will seek strategies to speed them up, for example by having an independent engineer perform such functions, as in the NSCP case. Generally, the approval process has been substituted or replaced by one or more processes dealing with inspection/monitoring, quality control and quality assurance QC/QC, with the possible role of an independent engineer who may provide (1) approval of design and construction as in the NSCP project, (2) quality control services during construction as in the Western Alignment project, and (3) review of performance during the design and construction as in the Channel Tunnel and Severn Bridge (BOOT) project. The SR91 project provided for government approval of design and inspection of construction and operation. Generally, however, governments provide for final inspection of completed work before they accept the work or authorize operation such as in the Channel Tunnel, Western Alignment, and SR91 projects. Table 2.3 summarizes the major characteristics of the obligation dimension. The projects examined demonstrate that governments may direct or authorize changes to the work at their discretion as in the Western Alignment project or based on pre-agreed reasons such as in the Channel Tunnel and Northumberland Crossing projects where reasons included safety, defense, security, the environment, errors and omissions, or non-conformity. Time and cost adjustments as a consequence of changes may be subject to negotiations as with the Severn Bridge project, or added to the capital or operational costs as with the SR91 project. 56 u CUD e _o %— u cu c o "O a CS H—1 c cu E c o 73 > cu Q I g j i~ SJ 5 «5 U B S CU X! 2 SJ «" in" O- cu o B. cn CU OS s CU B O c_ cn cu Qi q- § © o I a- s 'I a. 2 !/) C+_( o o £ Z ca a c n CU " X . 3 a C jj •H-g S?l •s -3 s a J a o CN cu cu o . o 60 . 3 cd x xs <u a - a = o * B 60-g . 3 -a S - i = • s 72 . 3 60 . 2 g S <-> a - S o c2 B" a o cu c £ g §•1 1 1 •8 E — T3 cd > o *-. o J o . cd O 4-» o cu x> 3 a 60 < o 5P § cd cu • a ; a o o cn cu cn cu X ! ~ a cd o i-a. las cd cd g S% ' a S . g a < • S O S , — > C7 o 60 a cu pa T j 6 cn X ) o IN 3 « •ST " T3 w 73 ^ a 3 • 5 S f l C O § . 3 . 3 a o x B cu cd a c | o E Mpq . 2 Z eg U o J Cd g B ^ C cu O cn a . _• cn ^ E o 'cu a cd T3 g js a - 3 „ - — X ! CO cm • T3 B. a g « a c § . 2 Q. cu s -C on CJX> C 2 8 . 2 a' t3 . 2 cu O a - 3 <= eg B U a o 2 a E g §•1 cu . a > cd cu c T3 c — -a < cd cu O 4> S g . cd cu c a B cd . a X! cd ° E r cd • a 73 .2? M ^ 2 a a § J3 & & S a a.« < I a | cn cu S u 2 c cd o a a <u o a S? a cu I cu | M I . I I S ^ < H H cn . a cu o CJ) o a „ S a • g .SP cn O cu T3 cn co P e X3 O, cd cd o . cu cn cu cu a t- •B hn 01) o a fe 6D cd .a-so 3 3 o ° a - a cd cu C cn • a cd H , X I a cu s s CU CU , 60 C . cd Qj cu « .2 cn X . O o T3 a cd cu" CJ g a a. a 60 S 2 ?3 cu cd cu c n Q. H cu a -5 a : cu ° cf I • a .3 1 3 5 g E P . B cj ^ g S 5 - 5 id < 2 a «« a o & a 1 0 cd O X V cu -> « cu J "O cd = ~> < CQ •o a CO o — 00 ON OS M cu cd cn E a a cd o cd —' C cu cu -zri a - S 2 a . Q. CU cu •£ ° o ^ •a . 3 .2 I s S . 3 e ^—* cd cj o cu OH cn T3 e cd e o o g O . B 3 eg cd j _ S c 2 e cu C 60 • - cd cu cu — w cu 5 £ . O. u & ^ cd Q T3 cn 3 = cd o cu S '> o . i J e . a a td . 2 B CJ a - 3 o cS B" U a o E g & s — E cu .3 -> cd • § E — T3 Z? a < cd CM > .1 £ Id . 2 C J o a - 3 o H 2 B" U a cj 2 a E g 7j . 3 > cd cu e T3 a — - o ^? a < cd o 2 H O OQ CQ i n CS o cn —1 2 o Q. T3 cn a a cd o cu o o e <D a S cu . 3 3 is 3 J5 cd — E ao cu <4 T3 pi '> O cn CU CJ a CD a o c o LH o . If CD kH ' 3 cr cu _cd ICQ o CCJ T3 13 -a o CU ' 5 " OH &H &H PL, T 3 (U H-» o CU CU X3 •4—• <H3 c o • H CO fi CU o • i-H bp O <N (U Channel Fixed Link The Agreement explained the developer's obligations to develop the Link in terms of design, construction, operation, and maintenance. The C-T Act provided for the concessionaires to make bylaws regulating the operation and use of the tunnel system, which were the subject of approval by the two governments. Promoters were required to carry out an environmental impact assessment in both the UK and France. Promoters were required to be aware of the procedures of the International Maritime Organization (IMO) before starting the development such that no permanent structure (e.g. ventilation shafts, artificial islands) would hamper the freedom and safety of navigation. Other requirements included provisions for facilities and installations for policing the tunnel and for frontier controls (customs, immigration, and animal health checks) which the Concessionaires would pay for but which would be organized and performed by the two governments. For the supervision of construction and operation, the governments authorized an intergovernmental commission and safety authority for the performance of these functions and required the concessionaire to comply with its directions. However, no strict approval process was mentioned in the agreement. The governments provided for their inspection of the completed work before they would authorize operation. An independent project manager, "Maitre d' Oeuvre", was appointed to review the performance during design and construction. The agreement explained that the concessionaire could proceed with the works relating to the 'Avant Projet' (project outline drawings and documentation list) unless the governments raised an objection to such Avant Projet. Huot (1995) noted that government imposition of the latest innovations, safety and other regulations, after the start of construction led to severe design changes and increased costs. 5 8 Second Severn Bridge Through the Tender Invitation and the S-B Act, government required the promoter to take responsibility for the design, construction, maintenance and operation of the second crossing as well as for the maintenance and operation of the existing crossing. Other specific requirements included quality assurance, navigational requirements and environmental aspects. The S-B Act gave a power of temporary prohibition or restriction of traffic to be exercisable by the concessionaire. The New Roads Act ("New" 1991) provided for similar power and for all highway functions to be exercisable by the concessionaire except for the power to make schemes, regulations, or give directions under the Road Traffic Regulations Act of 1984, U.K. The government explained in the tender invitation for this project that special procedures would replace the normal technical approval arrangements. The invitation explained the appointment of a consulting engineer to work as a government agent to monitor design and construction, audit the promoter's quality assurance system, and possibly the maintenance and operation of the works. Further requirements explained that the promoter was required to employ the services of a designer under a formal contractual relationship such that the contract would ensure the designer was sufficiently independent from the promoter. That was required to enable the designer to comply with government requirements, check the promoter's proposed construction methods, materials, and each element of work. Along with that, the detailed design was required to be checked by an independent checker. The government agent was to receive certificates of satisfactory completion from the designer and the checker. Design changes were allowed such that if the government issued a change, then the implications for the promoter's program and financial adjustment would be subject to negotiation. If the 59 promoter issued a change, then it would be subject to the agent's approval with no financial adjustment to the promoter, who also would bear the consequences of any delay. Highway 104 Western Alignment The respondent obligations as set in the RFP included design, construction, operation and maintenance, repair and rehabilitation. The government required an environmental management plan for the facility. The government set a 20-month objective for completion and a guaranteed maximum price for design and construction. Also, it required a marketing plan to maximize the use of the facility. Approvals and permits were the respondent's responsibility. The government was prepared to provide maintenance, repair and rehabilitation services. Beck (1997), president of Canadian Highway International Corporation, the developer, explained that an annual maintenance agreement was signed with the government for regular maintenance services. The RFP explained that the government at any time might direct or authorize changes in the work to be performed. The government through the RFP reserved its rights to undertake its own quality assurance activities. However, it was stressed that quality control and quality assurance (QA/QC) for the development, design, and construction were the responsibility of the respondents. The first addenda explained that members of the developer could perform QC, however QA had to be performed by an independent material testing firm and laboratory. The respondent was to be responsible for any corrective action due to non-compliant test results. For the operation and maintenance of the road, the government required the preparation of a road maintenance management plan explaining performance specifications, maintenance functions, and how the respondent would perform such functions. The government would evaluate periodically the performance of the respondent according to this plan. 60 North umberland NSCP The developer's obligations included all the development and operation functions including design, construction, operation and maintenance. A service life of 100 years was a design requirement. Extensive environmental reviews and assessments of the biophysical (e.g. air, marine and terrestrial life) and socioeconomic (e.g. labor issues, regional benefits and affected businesses) consequences of the crossing were required for the project and the developer was required to comply with all the requirements. A fixed crossing was considered to pose a threat of delaying the clearance of ice from the Strait. It was thought that such an "ice-out" could delay the start of the fishing season and could reduce the local temperature which in turn could delay the spring planting of crops (FHWA 1996). All designs were assessed against a 2-day delay in ice-out in any year over a period of 100 years. Developers were required to comply with this maximum ice-out delay among other requirements, which were addressed by the developer in its commitment to develop a plan for the management of all environmental aspects of the project in the Final Registration for the project ("Northumberland" 1993). Among the other obligations, the developer was required to maximize the economic and industrial benefits to the Atlantic region regarding businesses, employment, purchasing (material, equipment, supplies, and services), and technology amongst others. The regional benefit agreement signed for the project included several covenants on the developer such that 70% of all materials, 96% of labor, at least Cdn $20 million of engineering work after closing, and 75% of all marine workers had to be procured from the Atlantic Provinces region (FHWA 1996). Monitoring performance during design, construction, commissioning, operation and maintenance, was one of the government's roles for the project. The government at no cost to 61 itself could request changes to the work if the reason of change was due to errors, omissions, or non-compliance on the part of the developer. Time for additional work resulting from changes authorized by the government would be negotiated. An independent engineer was appointed for the review of design, construction, operation and maintenance procedures. Work approvals were the subject of negotiation. The government wished to retain the right to approve construction work and progress payments. A compromise was reached where the independent engineer would approve construction work and monitor the cost to complete of major work items (Pirie 1996). State Route 91 Caltrans through the SR 91 agreement required CPTC to design, develop, acquire, construct, install and operate the project transportation facility. Along with allocating such responsibilities to CPTC, Caltrans offered to assist CPTC in preparing and presenting documents required to obtain any permits and approvals needed for the project. For the operation of the project CPTC was responsible for performing the administrative, toll collection and traffic management activities. The AB680 and the SR 91 agreement encouraged CPTC to pursue possible contracts with Caltrans to perform traffic management activities and maintenance, and with the California Highway Patrol for police services. Environmental studies for the facility were CPTC's responsibility. CPTC was required to prepare all documents for environmental clearance and analysis in order to obtain all of the necessary permits and approvals. Final approval of the project and commencement of construction were contingent on meeting requirements of the California Environmental Quality Act. The SR91 agreement explained that Caltrans had the right to review and approve the design prior to commencement of construction. The approval process was limited to validating that the design 62 was in accordance with the Caltrans design standards cited in the agreement, and provided for Caltrans objections or approvals within twenty-one days. Construction of the facility was required to be in accordance with standards and specifications described in the agreement. Caltrans provided for overseeing CPTC compliance with such standards. Obligations Dimension: Financing Attribute For all of the projects studied, emphasis was placed on the promoter's financing responsibilities, the security used in raising finance, and the form of government support to the project, if any. Generally, for the projects and acts examined, government provided for all financing risks to be carried by the developer. Further, no financial guarantees were provided. However, support was provided in terms of (1) a direct subsidy as in the NSCP project, (2) operation of existing facilities as in the Severn Bridge, and (3) establishing a policy in favor of the facility such as in the Western Alignment. For purposes of calculating capital, operating and maintenance costs, governments generally require developers to maintain reserve funds such as a working capital reserve fund, maintenance and capital improvement reserve fund, and a debt service fund. To enable lenders to provide finance or credit support governments generally, allow the developer to use an umbrella of security instruments that cover the developer's interests in and rights under development, lease, and any project related agreements; tolls, income and project revenues; and, all developer's shares. However, as explained below for BOT and BOOT projects and as shown in Table 2.3, governments restrict the use of the project land and facility (i.e. improvements) as security. This restriction is imposed as a government requirement even for a project for which the developer has private possession/ownership, i.e. BOOT, such as the Northumberland NSCP project, and the Severn Bridge project. 63 However, there are cases where such restrictions may be relaxed until the occurrence of a stated condition or phase such as in BTO procurements. For example, the SR 125 franchise agreement ("Development" 1991a) explained that the financing assignment used as debt security might cover the developer's interests in all or any portion of "(i) the franchise documents, (ii) the project [toll highway, real property on which such toll highway will be located, personal property and intangible property], (iii) project revenues and/or (iv) any other property or rights (including operating rights) of developer." It was explained that such a financing assignment should not be made in a manner that precludes passing of the project title to Caltrans on the title transfer date before the start of operation. A similar assignment was made for the Mid State Tollway project ("Development" 1991b), however, it covered only the real property of the project. Both the SR 125 and Mid State Tollway projects were under the AB680 bill ("Assembly" 1989). The Virginia Act ("Public" 1995) provides the power necessary to the project developer/operator such that it could acquire, construct, improve or operate the facility. The act stated that the operator may "... secure any financing with a pledge of, security interest in, or lien on, any or all of its property, including all of its property interest in the qualifying transportation facility". Similar provisions for the use of facility as security were included in the cancelled Texas High Speed Rail franchise agreement ("Franchise" 1991). Channel Fixed Link Governments through the terms of the Channel Invitation ("Invitation" 1985) and the Concession Agreement (1986) ruled out all support from public funds or government guarantees and required financing to meet all construction and likely cost overruns and delays. In its White Paper ("The Channel" 1986), the U.K. government explained that for the evaluation of proposals, solid financing commitments coupled with the ability to attract financing were the final test for 64 the evaluation, which was best met by Eurotunnel's proposal. Financing was entirely the responsibility of the promoters and was to be raised based on the rights conferred in the agreement to the promoter. The amount of equity capital was left to the determination of the promoters, however, it was expected to be substantial. The Channel Invitation explained that full information on the promoter's anticipated capital structure, proposed time for calls on the various markets, and expected amounts to be raised on each of these markets were required by government. Further, as evidence of the robustness and viability of proposals, a detailed financial plan and a cash flow forecast along with related assumptions were required from the promoters. Detailed annual financial forecasts up to ten years after repayment of debt were also required including assessment of costs, traffic, measures of profitability, and related assumptions. Promoters were required to show the sensitivity of the project's economics to variations in traffic flow, cost overrun, delays in completion, and changes in interest and exchange rates. Similar analyses were performed by the promoters in the project information memorandum presented to the lenders in preparation to raise credits (Roger 1990). Holiday et al. (1991) reported through Eurotunnel reports that initial construction costs for the project was £ 2.3 billion in 1985. In 1990, total project cost was £ 7.608 billion, of which £ 4.208 billion for construction and £ 3.4 billion for corporate, inflation, and financing costs. Project cost at completion in 1994 was £ 9 billion, as reported in the Eurotunnel web site . Finance was raised through several debt and equity tranches. Equity was in four tranches totaling £ 1.589 billion. Debt finance, £ 7.123 billion, was arranged though a large syndicate of lenders from around the world. Total debt and equity finance, therefore, was £ 8.712 billion raised between 1987 and 1990 and was advanced in several currencies (Holiday 1990). 2 http://www.eurotunnel.co.uk. and http://www.channeltunneI.co.uk 65 Second Severn Bridge The Tender Invitation (1989) required the promoter to finance both the existing and new crossings and inherit an estimated debt of £ 122 million for the existing bridge. The government required that "proposals involve no material risk on financial grounds regarding the completion of the second crossing to time and specification, the acquisition of the concession and existing crossing, and the operation and maintenance of both crossings." Financing of the Second Severn Crossing involved a fully underwritten facility of £ 340 million arranged by the Bank of America in 1992. For the lenders the commercial viability of the project was very promising because of the higher traffic levels and congestion on an existing Severn crossing (5 km from the Second Severn Crossing) and because the project company would operate both bridges, creating a monopolistic structure that attracted lenders. Moreover, as mentioned in the tender documents for this project, adjustments for project tolls were allowed during the concession period based on changes in inflation rates. Highway 104 Western Alignment The government required the project to be entirely self-financing apart from Cdn $ 29 million under the SHIP Agreement (Canada-Nova Scotia Strategic Highways Improvement Program), which was raised to Cdn $ 55 million by the second addendum. The corporation was to borrow money without recourse to the government. The government explained that "it will not guarantee any debt incurred by the selected respondent or corporation". The government established a policy whereby all heavy trucks, except for local traffic, would use the Western Alignment. The W-A act (1995) explained that the corporation could borrow money based on its own credits, and could secure its borrowings against any or all of its assets and undertakings and the 66 revenue arising from the collection of tolls. The act explained that "no debt of the corporation constitutes any lien or other charge on the Western Alignment". Beck (1997) explained that Cdn $62 million toll revenue bonds were used to finance the project. A detailed project cash flow model and pro-forma financial statements were required reflecting forecasts and estimates for each year of the concession period. For evaluation purposes, government required the preparation of two sets of financial statements with accompanying cash flow models for two sets of toll revenue forecasts provided with the RFP. Assumptions for both sets included a 2.35% inflation rate, a 20-month completion period, a 35-year concession period, Cdn $ 650,000 annual maintenance cost, and an 8.25% yield on 30-year Canada Bonds. North umberland NSCP The government explained in the proposal call and the NSCP Act that its annual subsidy to the existing ferry service would be Cdn $42 million (1992 prices). This subsidy was provided to reduce the government's cost to maintain its obligation for continuous communication with PEL The annual subsidy was to continue for 35 years commencing with the operation of crossing and indexed 100% to the consumer price index. The government in the sixth addendum insisted that investors should be aware that the project was a private sector venture and the subsidy should be considered as income to support toll revenues. The goal of the federal government was to have its participation "off book". However, the Auditor General of Canada subsequently ruled that the NSCP project financing had to be considered a debt obligation 'on balance sheet' of the federal government (TFPPP 1996). This subsidy was used by the developer to raise about Cdn $660 million. Pirie (1996), Vice President of Strait Crossing Inc., explained that based on the subsidy, real-rate bonds paying a yield of 67 4.5% plus the annual inflation rate were issued and were taken up mainly by pension funds. Later the developer negotiated a reinvestment strategy for the bond proceeds to maximize the use of the loan considering the project's anticipated drawdown schedule. Equity for the project, as explained by the fifth addenda was required to be the lesser of 10% of total project cost (including direct and indirect cost, interest during construction, contingencies, start-up costs and working capital) or Cdn $ 75 million. Instead of requiring the deposit of equity up-front in a trust account, the government allowed the developer to pay in equity pro-rata (supported by a letter of credit) with debt proceeds during the course of the project. The developer was also required to designate a 'Prime Cost Sum' of Cdn $ 30 million (reduced to Cdn $10 million) for disbursement at the government's discretion for fisheries compensation. Along with the above requirements, the government required that the agreement between the developer and his financier include provisions to reflect that progress payments be disbursed only after progress certificates were signed and issued by the developer and an independent consultant. Pirie (1996) explained that during negotiations the involvement of the independent consultant was kept to major work items only. The government emphasized in the RFP and first addenda that neither the crossing nor the lands could be mortgaged or pledged as collateral by the developer in any way and were incapable of seizure by the developer's creditors. By the third addenda the government explained that it would permit some form of mortgage or pledge to the extent necessary to permit the placement of the required mezzanine financing (subordinated loan). However, by the fifth addenda the government emphasized its earlier restriction and added that "lenders will have available an assignment of cash flow security through the trust accounts and certain insurance proceeds". 68 State Route 91 Financing was the responsibility of CPTC. Caltrans explained in the SR 91 agreement that it had no responsibility to meet any debt incurred by CPTC for the development and operation of the facility. Caltrans explained in the proposal guidelines ("Guidelines" 1990) that the development had to be performed and completed at no cost to the State. All services provided by Caltrans were to be reimbursed by the developers. This included reimbursement for optional services requested by the developer (e.g. traffic projection, maintenance, police services, etc.), and reimbursement for non-optional services performed to protect the State's interest (e.g. costs associated with proposal selection, review right-of-way acquisition, design and construction oversight and technical activities, etc.). CPTC was required to maintain a number of reserve funds for working capital, major maintenance, capital improvements, and debt service. Financing and debt security instruments referred to in the SR 91 agreement as leasehold mortgages were made based on CPTC's interest in the agreement, the lease, the project facility, and the tolls and profits of CPTC. Rights of leasehold mortgagees were subject to the provisions of the SR 91 agreement. The agreement stated that no CPTC default would be grounds for termination by Caltrans of the agreement or the lease until all remedies raised by Caltrans in a default notice in a cure period were met by CPTC or its leasehold mortgagees. Equity paid by the CPTC was $19 million. Taxable finance raised by CPTC included $35 million 17-year institutional debt, $65 million 14.5-year variable rate term loans, and a $7 million subordinated loan from the Orange County Transportation Authority (PWF 1995). 69 2.3.5 Liabilities Dimension Liabilities Dimension: General Liability Attribute The general liability attribute is the first attribute in the liability dimension. Table 2.4 summarizes the general characteristics of the general liability, risk, and tax attributes of this dimension for the projects examined. Governments generally require developers to maintain liability insurance policies sufficient to: (1) insure coverage of tort liability (claims arising on account of personal injury or death or damage to real or personal property) to third parties, users, and employees; (2) protect against physical loss or damage to the facility in order to ensure continued use of the facility; and, (3) provide protection against business interruption (loss of income or earnings due to an insured peril such as delay in start-up/completion). Other policies may be required particularly if government provides support or will carry risk if the project is not completed such as in the Northumberland Strait Crossing project. Exceptions, however, can be made to relieve a developer from part of the liability coverage as in the SR91 project. Channel Fixed Link The two governments required the promoters to be liable for damage caused to users of the Link and third parties. Two insurance programs were required, one during construction and one at start of operation which had to be renewable on a one, two, or three year basis. The risks to be insured included (1) physical damage to the Fixed Link, (2) tort liability to third parties, and (3) delay in start up and interruption of operations resulting from facility physical loss or damage. Such requirements proved to be invaluable when in November 1996 a fire erupted in the freight shuttle train and caused serious damage to the concrete lining. As a consequence the tunnel was closed for 16 days with revenue losses per day in the order of £ 1 million. The damage cost was approximately £ 230 million and insurance coverage repaid about 98% of the cost (Bennette 1997). 70 o cd o CP a T3 CD O CD "CD co CD X3 CI _o 'co CD CO CD -8 CN CD I Second Severn Bridge During construction the government required full contractor's all risk, third party, and employer's liability coverage. During operation, the invitation required insuring both crossings against all loss or damage. No explicit coverage was mentioned for liabilities of third parties during operation. However, the tender invitation required the promoter to indemnify the government against any liabilities to third parties arising out of development and operation of the two crossings. Highway 104 Western Alignment The RFP required the corporation to indemnify and hold harmless the government against any and all claims, damages, losses, liabilities, costs and expenses arising out of the performance or non-performance by the corporation in relation to the design, construction, maintenance and operation of the facility. The RFP required the respondent to maintain throughout the concession period a liability insurance coverage acceptable to the government. The fourth addendum described the insurance and bonding requirements during construction and operation to include (1) coverage for all risks of property damage to the facility and (2) coverage to protect against all claims of liability arising out of property damage, bodily injury including death and personal injury. North umberland NSCP To protect itself from being required to pay the subsidy payments and operate the ferry service or complete or repair the work (i.e. double payments), the government took certain precautions. A very expensive insurance coverage pre and after completion was required in the sixth addendum to (1) preserve the work (property) against all risks of physical damage, (2) to pay damages arising from claims from third parties for injury, death or loss of property, and (3) to reimburse 72 the government the cost of the subsidy or of providing ferry service if the completion date was not met. Pirie (1996) explained that Strait Crossing Inc. managed during negotiations to change the insurance limits so that they were based on the maximum foreseeable loss rather than the full replacement cost/value required by the government. State Route 91 Caltrans through the AB680 and the SR91 agreements provided for CPTC, the developer, to be protected and indemnified by the Tort Claim Act. Reasons and explanations for this protection included (1) Caltrans authority and obligation to supervise and provide specifications and operational requirements for the design, construction, and maintenance of the project, (2) Caltrans to hold title to the real property and facility, and (3) the designation of the facility to be deemed part of the state highway system. This enabled savings to CPTC which otherwise would have been reflected in the toll rates. Before the transfer of title to Caltrans, the agreement provided for CPTC to bear the risk of injury, loss or damage to the facility. Third party claims, except those that arise out of CPTC fault, were carried by Caltrans if these included claim that arise out of fault of Caltrans, any non- negligent actions taken or omitted by CPTC in compliance with any Caltrans permits or regulations, or design and construction which conforms to the standards in the agreement. The same also applied after the acceptance date. However, Caltrans also assumed further the tort claims arising out of any act or omission in connection with traffic management and maintenance activities for which it is responsible. CPTC was required to maintain throughout construction and operation, bodily injury and property damage liability coverage of at least $50 million general aggregate per year. 73 Liabilities Dimension: Risk Attribute Governments in general seek procurement by PPP in order to transfer more risks to the private sector than can be done using conventional procurement arrangements ("Paying" 1993). Three categories of requirements can be distinguished from the RFPs and agreements for the projects studied: (1) risks related to the developer's obligations including financing, (2) risks related to the developer's rights particularly those related to revenues, and (3) force majeure risks. An overview of the risk attribute for the selected projects is outlined in Table 2.4. For the first category of risks, governments generally allocate all development and operation risks to the developer in clear wording in the RFPs and agreements. Emphasis was placed on explaining that government monitoring, inspection, and quality assurance processes did not relieve the developer from his responsibility for the work. This is different than government directed changes to the work for which time and cost consequences may be negotiated as explained earlier. Further, for the allocation of risk, governments usually require completion guarantees, performance bonds, and labor and material payment bonds with amounts that vary according to each project's circumstances. As PPPs are vehicles used to derive private finance, financial risks are the responsibility of the developers. The second category under the risk-attribute deals with revenue risks. The general requirement is for developers to carry all such risks with no guarantees. Governments, may provide (1) adjustments for facility rate and/or term of agreement to account for some risks such as inflation, and actual traffic growth rates as in the Severn Bridge, and (2) policies to protect the developer's revenues from competing facilities through a 'no second facility' guarantee as in the Channel Tunnel, or 'absolute protection zone' for the SR91 project. 74 The third risk category deals with force majeure risks. The definition of force majeure varied among the projects studied. It is helpful to categorize force majeure risks in general to include (1) war actions - including war, invasion, acts of foreign enemy, and nuclear events (2) civil actions - including riots, insurrection, act of terrorism, sabotage, and strikes, (3) government actions - including expropriation, changes in law, interference by civil or military authorities, and (4) natural catastrophes - including floods, earthquakes, unforeseeable geological conditions, chemical contamination, and epidemics. When force majeure risks are realized, governments in general provide developers with a time extension for the performance of their obligations. Cost consequences, however, vary among projects and may more usefully be considered along with the insurance coverage for facility physical damage and loss generally required from developers. Governments in general provide no financial compensation for force majeure risks except for war actions, as defined above, for which the government provides compensation or retains the risk and carries the cost of repairs, such as in the Severn Bridge (BOOT), Channel Tunnel (BOT), and Northumberland Crossing (war and extreme catastrophes) (BOOT). In SR91 (BTO), for all force majeure risks government will restore land and reinforcements to restore the weight-bearing capacity of the real property. Channel Fixed Link Both the Channel Invitation and Concession Agreement emphasized that the Link would be constructed and operated at the promoters' own risk without recourse to the governments. For force majeure risks or exceptional circumstances, the agreement explained that the time allowed for the performance of obligation would be extended accordingly. However, no compensation would be made to the concessionaires due to interruption of construction or operation based on such risks. However, if interruption occurred based on national defense, 75 the concessionaires would be compensated. If these conditions/risks lead to the termination of the concession, "no compensation will be made to the concessionaires but the Principals may pay to the Concessionaires such amount which takes account of the net financial benefits, if any, to the Principals resulting therefrom". As mentioned earlier, insurance coverage was required for physical loss or damage to the facility arising from civil actions and natural catastrophes. For financial and revenue risk, along with the requirement for no-recourse to government funds, governments gave concessionaires the freedom to determine their tariffs and commercial policy. Further, the government undertook not to facilitate the construction of another fixed link whose operation would commence before the end of 2020. Second Severn Bridge Through the Tender Invitation, the government required all design and construction risks to be allocated to the developer. The government explained that its agent appointed to monitor design, construction, operation and maintenance would not relieve the promoter of any of his responsibilities. The government transferred all geo-technical risks arising from physical conditions and artificial obstructions to the developer. A substantial on-default performance bond and/or parent company guarantee was required from the developer. Basically, the developer was responsible for the care of works including cost of repairs from any causes except for force majeure risks for which compensation and time extension would be allowed. These force majeure risks did not include natural catastrophes. Insurance for physical loss or damage of the crossing was required as mentioned earlier. 76 Revenue risks related to changes from initial traffic volume and traffic growth forecasts were transferred to the developer. Further, the government stated that it would not be liable for any loss of revenues arising from any closure of approach roads to the crossings. The government provided traffic records on the existing crossing and projections for future levels of traffic. However, it assumed no liability from the use of such projections ("Second" 1988, 1989). Given the provisions dealing with toll adjustment and the variable concession period, which account for the actual traffic flows, the cost of such risks to the developer was reduced. Highway 104 Western Alignment The government stated in the RFP that the developer must assume all project risks and for that it is entitled to earn a fair market rate of return commensurate with the risks assumed. A guaranteed maximum price of design and construction was required along with an objective of 20 months to open the road. Performance and labor and material bonds in the amount of 50% of the maximum price were required for the construction phase. A performance bond was also required for the operation phase in an amount of 50% of the annualized contract for operation and maintenance. For force majeure risks, the RFP explained that an extension of time for completion of the road would only be allowed for the affected activities on the critical path of the project. No time extension for force majeure would be made unless it was filed within seven days of its first occurrence. An "All Risk" property insurance policy was required from the developer. For financing and revenue risks, the government stated that it would not guarantee any debt incurred by the respondent nor the corporation, forecasted traffic levels, and any factors that might impact revenues or costs. However, it covenanted all heavy trucks to use the highway. 77 North umberland NSCP The government explained in the RFP that the developer must bear all project risks during both the construction and operation periods with the exception of legal challenges and regulatory impediments risk (delays and cost increases directly attributable to government actions). The government explained that its inspection and independent check of the work did not relieve the developer of his responsibility for the work. The government through the NSCP proposal call ("Northumberland" 1988) and its addenda required the developer to provide a security package that assured the completion of the facility, assured the specified level of operating performance, assured the specified condition at the time of turnover, and assured the interim funding of the ferry service. Pirie (1996) explained that the security package against completion risk and cost overrun included along with parent company guarantees, a cover of Cdn $ 200 million performance bond and a Cdn $20 million labor and material bond along with a Cdn $ 73 million letter of credit for cost overrun risk. Further, the developer agreed to pay the operating cost of the ferry in case of completion delay. Through the terms of the NSCP proposal call and the first addendum, the government relieved the developer from its responsibilities for the normal operation of the facility and completion of the facility in four force majeure cases: (1) acts of the Queen's enemies, (2) government retroactive legislation, (3) earthquakes in excess of design criteria, and (4) a catastrophic event. A catastrophic event was defined as an event which damages the facility and renders it inoperable. Under such circumstances and where the government was bound by its constitutional obligation for "continuous communication", the government required the developer to provide as part of the security package reimbursement of an amount equivalent to the subsidy paid during the period of time the government assumed responsibility and operated the crossing service. 78 Pirie (1996) explained that during negotiations with the government, force majeure risks were replaced by what was defined as 'Project Risk Event'. Project risk events are retained by the government under the occurrence of acts of war, acts of government, extreme weather conditions, earthquakes beyond certain standards, and nuclear event; while the normal force majeure risks were carried by the developer. 'Project delay event' was another negotiated concept that described events beyond the developer's reasonable control such as contaminated material, third party strike or walkout. The realization of such delay events could, subject to negotiations, provide the developer with time extension and toll adjustments. State Route 91 In addition to liability requirements for facility damage and tort, Caltrans required CPTC in the SR 91 agreement to furnish payment and performance bonds or completion guarantees. Yet, with no specific amounts requested, Caltrans required such bonds to be acceptable to CPTC's lenders. For events of force majeure, CPTC's time to perform its obligations would be extended by an equal amount. Where the force majeure event damaged or destroyed all or any part of the real property, Caltrans would be obligated to restore the land, grading and reinforcements necessary to restore the weight-bearing capacity of the real property immediately prior to such event. However, the agreement explained that failing to restore the land should not be considered default if Caltrans had also declined to restore the land on the state transportation facility (SR 91 is a median improvement to State Route 91 that includes an adjacent SR 91 free highway). Strong protections were provided by the SR 91 agreement for CPTC against Caltrans' default, event of loss, and change in law. In the agreement Caltrans stated that it would not grant, nor convey to any other party other than CPTC, and would not finance with public funds, the development of a transportation facility that might present economic competition to the project 79 within the absolute protection zone. Failure of the application or performance of representations, warrants, and obligations would constitute a default by Caltrans, while failure to comply with covenants or requisition of title or requisition of use would constitute an event of loss. Both entitled CPTC to remedies, compensation, and/or termination of the agreement and the lease. The agreement explained that under a "change in law" that adversely impairs CPTC's exercise of its property, franchise and other contract rights, CPTC could elect to close the project and seek payment by Caltrans of all unrecovered costs at the date of calculation (capital and operating costs, interest on debt, distribution to equity investors minus total revenues at that date). Caltrans stated in the agreement that it would protect and defend CPTC against any challenges to the validity or enforceability of the acts and challenges to the enforceability of the agreement. Liabilities Dimension: Taxes Attribute The treatment of taxes is the third attribute under the liabilities dimension. Generally, governments require developers to be familiar with all tax rulings (e.g. corporate, income, and property taxes) that might apply to their proposed business structure. Further, governments make no representations or warrants to the tax consequences or accuracy of the developers' proposed business structure. Summarized in Table 2.4 are the tax attributes of the projects examined. Governments, according to the circumstances for each project, may provide for certain vehicles to support project development. These vehicles may include exemptions for certain types of taxes such as the exemption of property tax in the Northumberland NSCP project (BOOT), capital allowance such as in the Severn Bridge (BOOT), or creation of a corporate body with special characteristics such as in the Western Alignment (BOT). 80 Channel Fixed Link The Channel Invitation ("Invitation" 1985) explained that the principle of territoriality of taxation will be applied where each country will apply its normal laws to the construction, maintenance and operation of that part of the project falling within its jurisdiction. The requirement for the levying of taxes was set also in the Concession Agreement (1986) requiring that "all duties and taxes levied or to be levied, including taxes on immovable property, will be liabilities of the Concessionaires and will be applied according to the provisions of national law". Second Severn Bridge The promoter as mentioned in the tender invitation was to be treated as "trading" for corporation tax purposes and would be able to claim capital allowance for construction expenditure. The Tender invitation explained that value added taxes were payable on construction and exempted for project tolls. The invitation mentioned that local authority rates would be excluded. Highway 104 Western Alignment The RFP explained that the government made no representation or warrants concerning the tax or legislative consequences of any structure used by the respondent. Further, it explained that the respondents must satisfy themselves about the consequences of the provisions of Canadian and Provincial tax laws. The government did not entertain special tax concessions. The Western Alignment Corporation, not a public authority or crown corporation, was created by the W-A Act to assist the RFP respondent in the realization of the project (development and finance). The W-A Act stated that neither the corporation nor its property was liable to taxation including income tax under any enactment. The government required the RFP respondents to satisfy themselves as to the tax status of the corporation. 81 In the first addenda, August 1995, to the RFP the government was asked about sales taxes (GST) and it emphasized that "each respondent is responsible for obtaining its own advice as to all tax matters" and added "if necessary the corporation will be declared an agent of the Crown in relation to its toll collection activities". North umberland NSCP The government explained that the development was designed as a private sector venture and the developer's corporate structure was required to comply fully with both the letter and the spirit of the Income Tax Act of Canada in order to be accepted. The developer was required to satisfy itself and make appropriate allowances in regard to all taxes of every nature and kind that may be imposed on the facility, improvements, equipment, or any property brought on lands. The government explained that special tax concessions would not be entertained. A potential increase in sales tax liability was considered a business risk which must be assumed by the developer. The Provinces of New Brunswick and Prince Edward Island were considering exempting the crossing facility from municipal and provincial property taxes. State Route 91 The SR91 agreement explained that all taxes imposed on the real property and the project were the sole responsibility of CPTC as part of its capital and operating costs; despite the fact that the real property and the project were to be considered property of Caltrans at all times. The agreement, however, provided for franchise fees (base, variable, and excess) to be reduced by the amount of taxes after title transfer to Caltrans. CPTC was concerned about depreciating the project, after title transfer to Caltrans, and was advised it could depreciate the improvements. 82 2.4 Requirements Structure: Conclusions The above investigation explained key parts of the scope of government requirements for each of the eight attributes of the requirement structure as summarized in Tables 2.1 to 2.4. The output or benefits of the requirement structure can be summarized in three points. • The requirement structure is a framework through which the characteristics and requirements of a project during its life cycle can be usefully studied and analyzed in a comprehensive manner. The requirement structure was useful in explaining how governments implemented BOT, BOOT, DBFO and BTO procurement modes for a number of PPP projects. Alternatively, the structure can be used to identify arrangements under several PPP modes and assist in formulating other modes. The structure shows that for the traditional public procurement all eight attributers are under the responsibility of the government, while under build- own-operate or full privatization all eight attributes are the responsibility of the private developer. Between these two extremes are a number of responsibility allocations which lead to other PPP procurement modes. The structure can be used by government in negotiation with developers such that the eight attributes can be assigned to achieve a balance between the rights, obligations, and liabilities of the private developer. • Categorizing the requirements through the eight attributes should help in distinguishing the features of each attribute such that their key characteristics and variables can be treated quantitatively and/or qualitatively in economic models and/or analytical 83 frameworks and the uncertainty or risk surrounding them can be analyzed. That represents the main theme of the following chapter and the rest of this thesis. A framework based on the requirement structure assists in making decisions about the details of the eight attributes and on which allocation of the attributes will result in the best value for the public at large. For example, it can be used to help gauge what the benefits would be (e.g. reduction in facility rates as a requirement in the revenue attribute) if ownership of the facility (possession attribute) were in the private sector hands during the concession period? Clarity with which the terms and conditions in tender documents, or specifically the eight attributes of the requirements structure, need to be emphasized for PPP projects. The range of terms and conditions of each attribute as explained by the requirement structure, the large number of supporting documents to RFPs, the long negotiation process between government and developers, and the involvement of several stakeholders in PPP decisions suggest that the rights, obligations and liabilities in PPP projects need to receive more analysis in general and particularly when appraising and preparing for PPP projects. The clear identification and articulation of government requirements is needed in order for developers to carry out properly their economic appraisals of projects and to respond with proposals that fit the requirements, reduce the amount of time spent in negotiations, and reduce the amount of RFP supplemental materials. Specific questions that need to be addressed under the various dimensions are identified below. 84 Rights Dimension A clear statement of possession requirements in general and with each property type is needed. Items that should be addressed include: 1. Types of project properties (land, improvements, airspace, intangible property); 2. Type of possession permitted for each type (e.g. public, private, lease); 3. Properties, if any, that can be taken as a security instrument; 4. Clear title statement during the different phases and terms of agreement; 5. Who will carry the responsibility for the acquisition of land and right-of-way and its related costs (e.g. government, developer, or both); and, 6. Properties which are the subject matter of reversion, transfer, or dedication at the expiration of the agreement or at default. While many of the revenue terms are kept for the negotiation phase, explicit statements regarding project revenues are needed for the following: 1. Term of agreement: type (e.g. fixed, variable); 2. Term of agreement: measure, (e.g. NPV, IRR pre and after tax, specified amount of revenues, specified number of vehicles); 3. Types of revenues permitted to the developer (tolls, charges); 4. Treatment of collateral revenues (e.g. revenues from airspace improvements) 5. Toll types allowed (e.g. direct, shadow, congestion, or at developer's discretion); 6. Toll setting authority (e.g. developer, government, or both); 7. Toll adjustment mechanism (e.g. formula for inflation, traffic demand, debt ratios); 8. Toll caps (e.g. maximum toll rates allowed); 9. Base returns allowed: measure and value (e.g. NPV, IRR, specified revenues); 10. Incentive returns allowed and related performance measure (e.g. achieving specified use of the facility, vehicle occupants, number of cars); and, 11. Excess revenues, their measures and their distribution (e.g. shared, or allocated). 85 Obligations Dimension The obligation dimension represents the purpose of the PPP venture and the core of the requirement structure. Explicit requirements have to be set for two issues. The first is the developer's extent of obligations and responsibilities. The second relates to the extent and terms of the government's power in performing inspection/supervision, approval, and the right to request changes. Public-private partnership acts, RFPs, and agreements have to consider details for the obligation requirements, some of which include the following: 1. Description of project functions for which the developer is responsible (e.g. planning, permits, acquisitions, design, construction, operation, maintenance, environmental assessments and compliance); 2. Project functions the government prefers, or is required, to perform (e.g. traffic management, maintenance, police services); 3. Statement of the applicable standards and specifications 4. Extent of government monitoring, inspection, and approval processes, and right to make changes; 5. Statement of quality control and quality assurance systems and the responsibility for performing such activities (e.g. developer, independent consultant, government); and, 6. Processes for addressing time and cost effects resulting from changes made by government (e.g. allocated to capital/operating costs, to be negotiated). While general statements are provided by government regarding project financing, it is important that CFEIs and RFPs treat the following: 1. Financial risks, if any, that may be absorbed by the government, (e.g. interest rate); 2. Type of financial support or guarantees that might be provided; and, 3. Type of security instruments permitted (e.g. project revenues and rights). 86 Liabilities Dimension Explicit statements are needed by government to explain its requirements regarding project general liabilities, risks and taxes. They should cover the following: 1. Types of liability coverage (e.g. facility damage, tort and business interruption); 2. Responsible party for each liability during project development and operation; 3. Amounts of each insurance coverage required during construction and operation 4. Types and amounts of project bonds needed for construction and operation; 5. Extent and conditions, if any, of government liability (e.g. due to developers compliance with government specifications and standards); 6. Statement regarding the allocation of risks in relation to the developer's obligations; 7. Explicit definition of force majeure risks; 8. Time and cost consequences of force majeure risks; and, 9. Statements about tax policies, exemptions or allowances for the project. 87 88 Chapter 3 Proposed Economic Model Characteristics 3.1 Introduction This chapter sets out in the following section some of the concepts that should be used in developing generalized economic models that fully reflect the attributes of the requirements structure described in Chapter 2. In light of these concepts, a review is made of the major economic models and support systems that have been developed to date by other authors and organizations for the appraisal of projects. The characteristics of this thesis' proposed model and support system are then introduced at the conclusion of the chapter. 3.2 Economic Model Underlying Concepts An economic model can be described as a generalized model if it attains the following: (1) its structure directly or indirectly supports, and differentiates between, the several attributes of a project requirements structure; (2) the structure can support for each attribute a generalized representation covering the calculation methods across several industrial sectors (e.g. power and transportation); (3) its structure is able to include any number of operations (e.g. work tasks, activities, or phases) in a project life cycle; and (4) its structure is flexible enough to formulate several performance measures (i.e. not restricted to one economic indicator). These four characteristics are described below. 89 The requirements structure was used in the previous chapter to explain the various characteristics and aspects of capital investment projects, as exemplified by PPP projects. Capital investment projects are realized, as explained, by satisfying those obligations represented by the requirements in the development (e.g. design, and construction), operation and financing attributes, and by acquiring those rights represented by the possession and revenue attributes. The successful realization of projects, however, also involves addressing and analyzing project liabilities as represented by the general liabilities, risks and tax attributes. When building an economic model for project appraisal, the design of the model should convert or interpret all the requirements of the foregoing attributes into cash flow elements. The degree to which a model represents all of the attributes is in turn a measure of how comprehensive is the model's capacity to represent the real economic life of a project. In other words, the rights, obligations and liabilities attributes of a project should be integral to the design of the economic model, independent of the level of detail sought in representing the project life cycle for a specified project. Therefore, the most important objective in phasing a project, e.g. in two phases as engineering and operational (Willmer 1991) or four phases as conceptual, design, construction and operation (Meyer and Cressman 1984), should be to distinguish between these phases in terms of the calculation methods and work tasks for each phase. So the challenge becomes how to represent or interpret the requirements of the different project attributes in the form of cash flow elements when designing economic models. It is asserted herein that the model design should include two essential structural elements for each attribute: operations (e.g. number of work tasks, activities, or cost/revenue items) and methods (e.g. estimating and forecasting methods). Once these operations and methods are defined for each attribute, a project life cycle may be modeled appropriately since a project phase that may be 90 represented by an attribute can be distinguished from other phases in a project life cycle. The two structural elements are discussed further below. The first structural element represents the set of operations, their duration and logic necessary to represent requirements of each attribute. Operations for the development obligation-attribute represent, for example, the tasks used in planning, preparation, design, and construction of a project. Tasks for the operation obligation-attribute cover the administrative, operation, maintenance and utility-services activities in a project. The same applies to the financing obligation-attribute where financing operations can be represented by several bonds and loans used by a project. Tasks for the revenue rights-attribute cover all potential revenue streams in a project. The same goes for the possession rights-attribute where the expenditures attached to the various properties in a project may represent ownership tasks. Similarly, the liabilities attributes can be represented in a model as expenditure tasks required for the fulfillment of the project. The second structural element represents the methods used in the estimation or calculation of the requirements of each of the eight attributes. A generalized model design should support a smorgasbord of various methods used in estimation. Each of the obligation attributes, for example, may have particular methods of calculation that can be found in the business environment of the attribute such as construction cost estimating methods, maintenance cost estimating methods, and project financing methods. Similarly, for the rights attributes, the spectrum of available demand analysis methods and service charge (e.g. toll) methods all need be recognized in the design of the model. An economic model can be structured to represent the expenditures and revenues in a project life cycle in one or a combination of three arrangements. The first arrangement comes through direct 91 assignment of cash outflows and discrete inflows at specific points of time or project periods without reference to how these net sums were obtained (this is referred to later as crude or aggregated estimating methods). This structural arrangement offers very little opportunities for risk analysis on the economic viability of a project. The second arrangement calls for the identification of packages for different phases of the project life to which expenditures/revenues can be attached and distributed over the duration of the relevant package (this is referred to later as semi-detailed estimating methods). The third arrangement provides for the inclusion of the methods used to derive a calculation/estimate along with their variables in the model design or structure (this is referred to later as detailed estimating methods). This arrangement provides the greatest opportunity for analyzing the project through risk analysis. While various business sectors (e.g. power, water, or transportation) may share some methods in common, often there are differences between sectors. Therefore, an economic model should allow for a "specialized" representation of a particular attribute so that the characteristics of the operations and methods of this attribute in a particular business sector can be treated. For example, a model that includes in its structure most of the transportation demand/revenue methods can be categorized to have a specialized revenue attribute that reflects the transportation sector. Alternatively, an economic model can have a "generalized" representation of a particular attribute if its structure reflects the characteristics of the operations and methods of this attribute across several business sectors, thus providing an open architecture where several methods can be added to the model. Further examples of the specialization/generalization of attribute methods include the following. The representation of the finance attribute of a project in an economic model may reflect only the bond instrument of the financial sector. However, when it reflects all the other instruments, e.g. 92 syndicated loans, term loans, and private placement bonds, it can be said to have a generalized representation of the finance attribute. Similarly, cost-estimating methods from the conceptual through detailed (Clough and Sears 1991) for the development attribute of a project can be utilized across several sectors or project types, and therefore they constitute generalized representations. The difficulty in developing generalized representation of an attribute lies in the necessity of acquiring the methods used by the several sectors or project types for the calculation of a particular attribute. One approach adopted herein to assist in this task is enriching the specialized representation of an attribute by crude and semi-detailed methods such that it becomes a generalized representation. The final point to address in model design is the flexibility of the model structure. With a flexible model structure, several economic performance measures can be modeled based on, or using, the same structure of the attribute design. For example, a model can be structured to calculate the total construction cost from the sum of the estimate of construction work packages. These same work packages should be available for the calculation of, or use by, any other performance measure such as net present value, internal rate of return, benefit-cost ratios, and loan-life-cover- ratio. With flexibility, (1) several performance measures can be built in a model without any repetition of calculations and without redundant model structure, and, (2) only one-time input to the model can be used to provide all the estimates of the performance measures in the model. 93 3.3 Previous Models and Systems for Project Appraisals To date, several economic models and computer-based support systems have been developed to help decision-makers at the appraisal stage of capital investment projects. This section reviews the most significant project appraisal models and systems in light of the concepts presented in the previous section. Since the 1970s the World Bank has been relying on economic models and risk analysis techniques for use in project evaluation. Pouliquen (1970) and Reutlinger (1970) explained the World Bank's formulation and use of economic models, sensitivity analysis and risk analysis using Monte Carlos simulation for the appraisal of projects. The models were built by the World Bank, as a lender, for the appraisal of projects as a basis to assist in making lending decisions. The models explained by Pouliquen and Reutlinger for the appraisal of road projects in Africa were constructed using a number of concise mathematical statements/ equations to determine an economic indicator such as net present value or internal rate of return. Figure 3.1 and Table 3.1 shows a flow chart and mathematical formulation for a road project appraisal as referred to by the Reutlinger (1970). In terms of operations and calculation methods, the model differentiated between construction, revenue and maintenance operations. An aggregated cost sum was used for construction, an exponential method for revenue/savings calculations and a linear method for maintenance cost calculations. However, the model made no use of a network structure to logically link operations and was not generalized enough to handle other road projects. New models would have to be constructed to include other methods for construction, revenue and maintenance, to handle other project operations or attributes (e.g. financing), and to calculate other measures such as completion time, cost-benefit ratios, and debt-service-cover ratio. 94 DETERMINE PROJECT COST (Equations 1 and 2) FORECAST DAILY TRAFFIC (Equations 3 and 7) i I 1 COMPUTE COMPUTE COMPUTE BENEFITS BENEFITS BENEFITS RESULTING FROM RESULTING FROM RESULTING FROM REDUCTION IN REDUCTION IN SAVING IN ROAD VEHICLE LENGTH OF THE MAINTENANCE OPERATING COST ROAD COST (Equations 8 and 17) (Equations 18 and 21) (Equations 23 and 26) 1 ADD A L L BENEFITS (Equations 27 and 29) J COMPUTE RATE OF RETURN (Equation 30) Figure 3.1: Flow chart for road project appraisal, Reutlinger (1970) 95 Table 3.1: Road Project Appraisal, Reutlinger (1970) (Pavement Base) + (Sub-Base/Shoulders) + (Earth Works) + (Borrow Materials) + (Others) (Project Cost) / (Construction Time) (n) if t < (Construction Time) C if t > (Construction Time) (1 + Traffic Growth Cars)' x (Initial Traffic Cars) (1 + Traffic Growth Trucks)' X (Initial Traffic Trucks) (1 + Traffic Growth Buses)' X (Initial Traffic Buses) (1 + Traffic Growth Trailers)' X (Initial Traffic Trailers) (1 + Traffic Growth Special)' X (Initial Traffic Special) Cost p.v.m. Old Road Car - Cost p.v.m. New Road Car Cost p.v.m. Old Road Truck - Cost p.v.m. New Road Truck Cost p.v.m. Old Road Bus - Cost p.v.m. New Road Bus Cost p.v.m. Old Road Trailer - Cost p.v.m. New Road Trailer Cost p.v.m. Old Road Special - Cost p.v.m. New Road Special (Cost Saving Car) x (Traffic Cars), x (Miles) x 365 (Cost Saving Truck) x (Traffic Trucks), x (Miles) x 365 (Cost Saving Bus) X (Traffic Buses), X (Miles) X 365 (Cost Saving Trailer) x (Traffic Trailers), x (Miles) x 365 (Cost Saving Special) x (Traffic Special), x (Miles) x 365 (Traffic Cars), x (Cost Travel Old Road Car) x (Reduction Miles) x 365 (Traffic Trucks), x (Cost Travel Old Road Truck) x (Reduction Miles) X 365 (Traffic Buses), X (Cost Travel Old Road Bus) X (Reduction Miles) X 365 (Traffic Trailers), x (Cost Travel Old Road Trailers) x (Reduction Miles) x 365 (Traffic Special), x (Cost Travel Old Road Special) x (Reduction Miles) x 365 (Traffic Cars), x 2(Traffic Trucks), X 2(Traffic Buses), X 3(Traffic Trailers), x (Traffic Special), 24. (Maintenance Cost Old Road), = (a + b) x (Traffic Units),, where a and b are constants 25. (Maintenance Cost New Road), = (c + d) X (Traffic Units),, where c and d are constants 26. (Maintenance Cost Saving), = (Maintenance Cost Old Road), - (Maintenance Cost New Road), x (Miles), 27. (Total Operating Cost Saving), = (Operating Cost Saving Car), + (Operating Cost Saving Truck), + (Operating Cost Saving Bus), + (Operating Cost Saving Trailer), + (Operating Cost Saving Special), 28. (Total Mileage Cost Saving), = (Mileage Cost Saving Car), + (Mileage Cost Saving Truck), + (Mileage Cost Saving Bus), + (Mileage Cost Saving Trailer), + (Mileage Cost Saving Special), 28. (Benefits), = when t > (Construction Time) then = (Total Operating Cost Saving), = (Total Mileage Cost Saving), + (Maintenance Cost Saving), when t < (Construction Time) then = C 29. Calculate r such that! (1 +r)"' (Cost), =1(1 + r)"' (Benefits), , r= 1, n Notes: Traffic refers to Average Daily Traffic; Cost p.v.m. is cost per vehicle mile; any variable followed by subscript t indicates amount per year. User specified data is in boldface. 1. (Project Cost) 2. (Cost), 3. (Traffic Cars), = 4. (Traffic Trucks), = 5. (Traffic Buses), = 6. (Traffic Trailers), = 7. (Traffic Special), = 8. (Cost Saving p.v.m. Car) = 9. (Cost Saving p.v.m. Truck) = 10. (Cost Saving p.v.m. Bus) = 11. (Cost Saving p.v.m. Trailer) = 12. (Cost Saving p.v.m. Special) = 13. (Operating Cost Saving Car), = 14. (Operating Cost Saving Truck), = 15. (Operating Cost Saving Bus), = 16. (Operating Cost Saving Trailers), = 17. (Operating Cost Saving Special), = 18. (Mileage Cost Saving Car), 19. (Mileage Cost Saving Truck), = 20. (Mileage Cost Saving Bus), = 21. (Mileage Cost Saving Trailer), = 22. (Mileage Cost Saving Special), = 23. (Traffic Units), 96 Clark and Chapman (1987) explained British Petroleum's (BP International) development of decision support systems for risk analysis of the time and cost of the construction phase of several projects including North Sea oil production platforms, oil processing facilities in refineries, and projects in chemical, mineral, and communication industries. The structure of the cost analysis model included decomposing total cost into cost line items through which risks could be identified and propagated to determine the total cost distribution. Time analysis incorporated the use of activities in a CPM network where risks can be identified for each activity. The technique used for risk analysis was the Controlled Interval and Memory (CIM) method (Chapman and Cooper 1983a; Cooper et al. 1985; Cooper and Chapman 1987; Chapman and Ward 1997). The structure of the cost model did not support the use of different construction cost estimating methods and did not treat time dependent costs. Thompson and Willmer (1987), Willmer (1991), Thompson and Perry (1992) and Thompson (1993) explained some features of CASPAR (Computer Aided Simulation for Project Appraisal and Review). CASPAR was used to analyze part of the costs of the Channel Tunnel and was applied to the modeling and appraisal of a number of build-operate-transfer (BOT) projects. The model structure underling CASPAR was designed to simulate the interaction between time and money during a project life cycle which was considered to consist of two phases - engineering and operation. The structure of the model is network-based made of inter-related activities to which costs and revenues can be attached as a lump sum and which can be made discrete at specific periods of time or spread uniformly over the duration of the activity, or as unit cost/revenue per unit of time or unit of quantity. Each of the two life cycle phases in the model can have up to seven cost centers (e.g. production, administration, marketing) to which the activity costs can be assigned for the purpose of structural analysis. Different inflation factors can be applied to each cost center. 97 CASPAR was implemented in two separate programs, one for time analysis, and the other for cost analysis. CASPAR has a special way of dealing with sensitivity and risk analyses. It defines a number of risk variables (a maximum of twenty) which are defined in terms of various elements (maximum twenty one in the cost program and fourteen in the time program) of the original deterministic data. A design delay risk variable, for example, may contain all those activities involving design where their duration and/or costs can be added to the risk variable. As stated by Willmer (1991): "In the simplest case, the duration of one activity may define a risk variable. In more complicated cases activity duration, costs, and resource quantities may all be combined in one variable". Each risk variable can be defined on a percentage range change. In sensitivity analysis, all data elements within a risk variable will be altered to the same extent when the risk variable changes in its range. In probability analysis and using the cost program, a triangular or uniform distribution can be defined over the range of each of the twenty risk variables. In the time program, a triangular distribution can be assigned to each activity duration to determine the effect on the whole project network. CASPAR then relies on Monte Carlo sampling for risk analysis processing. CASPAR supports NPV, IRR, and payback period. While CASPAR possesses some useful capabilities in terms of its network-based structure and the patterns of costs and revenues, its structure has several limitations. The model does not distinguish between the two phases treated in terms of the different calculation methods that usually apply to these two phases. Network activities can be designated as belonging to either of the two phases and assigned costs or revenues in an aggregated or unit cost/price form as described above. CASPAR structure does not model financing, revenue and maintenance methods. Further, the use of risk variables in the way it is described above means several risk elements (time and cost) are analyzed using the same range of variation or distribution of the risk variable that represent these elements. The CASPAR cost and time programs are not combined. 98 Since 1983 the United Nations Industrial Development Organization have been developing computer software for project appraisal which has resulted in the introduction of COMFAR III Expert (Computer Model for Feasibility Analysis and Reporting) for the financial and economic appraisal of industrial and non-industrial investment projects (UNIDO 1994). This program has been licensed to several users including development financing institutions, investment banks and industrial development corporations in 120 countries. COMFAR models the planning horizon of a project in two phases - a construction phase and production phase. Project costs are defined as sub-items under two categories: fixed investment costs and production costs. Project revenues are defined as sub-items/products under a sales program category. Cost and revenue sub-items have the same duration as of the phase to which they belong. A unit cost estimating method is used for all sub-items where quantities and costs/prices are defined at discrete periods during the duration of the relevant phase. Unlike the other systems described so far, COMFAR provides for the inclusion of sources for project finance including equity, long-term loans and short term finance. The model has the capability to use up to 20 currencies for cost/revenue calculations. COMFAR lacks three important elements which are essential for functionality and generality. First, it does not have a network structure that can define logic connecting the various cash flow streams. Second, it does not address the spectrum of calculation methods used in the estimation of projects. Third, it does not allow probabilistic risk analysis to be carried out on the variables of the model - only sensitivity analysis can be performed. Despite their weaknesses, CASPAR, COMFAR, BP's models and the World Bank's models represent distinguished systems for the economic appraisal of capital investment projects. Explained in the literature, however, are other works that contain a general description of 99 economic models with an implementation of sensitivity and/or risk analysis attached to the models. These applications included to large extent crude or aggregated models, with the goal being to explore the behavior of, or gain insights on, some specific problem types or areas. For example, Wahdan et al. (1995) and Wahdan (1995) developed an economic model for the analysis of public private partnership projects. The model had a pre-defined set of work packages with which the project economic structure must fit, it used an aggregated cost method, it recognized a demand calculation method for revenue calculation, it did not support a network structure, and it included a preliminary loan calculation but it did not support financial calculation methods. Ngee et al. (1997) explained the development of an automated mechanism for dynamic negotiation between government and developers. The mechanism is a spreadsheet that uses crude estimates of cost, concession period, and level of tariff. The purpose of the mechanism was to simplify project cash flow calculations during the final negotiation and reach a balance between risk and return. Salem and Ariaratnam (1999) developed a decision support system for life cycle cost analysis of construction/rehabilitation road projects. The model had a highly crude/coarse structure, did not distinguish between life cycle processes and calculation methods, did not support a network structure, and did not support indicators other than life cycle cost. Other models introduced to date for specific purposes include a cash flow model for risk analysis of power plants (Chee and Yeo 1995), a model for the evaluation of oil and gas projects (Skjong and Lereim 1988), and a model for the economic evaluation of chemical plants (Westerterp and Vrijland 1984). To large extent, the above models and systems can be categorized as semi-detailed models in which the differentiation between project phases was not clear and the methods used in calculations (e.g. financing and demand) were largely crude and semi-detailed methods. 100 3.4 Characteristics of the Proposed Model and Support System To satisfy the objective of this thesis of building a generalized economic model and support system for the risk analysis and evaluation of capital investment projects necessitates that the concepts elaborated upon previously should be employed in order to obtain a model which can provide a realistic representation of these projects. Therefore, the proposed generalized model and support system should recognize the following characteristics: 1. Differentiation between project attributes in rights, obligation and liabilities Regardless of the number of phases required to represent a project life cycle, the model should be able to distinguish between the various requirements of a project during its life cycle. 2. Recognition of attribute characteristics: Multiple Operations Each attribute should be capable of being represented in one or more phases in a project life cycle. Each phase should be represented by one or more operations or tasks. These tasks should be logically linked - both internal to a phase, and between phases, using a network structure. 3. Recognition of attribute characteristics: Methods The structure of the model should be able to include crude, semi-detailed and detailed calculation methods in its treatment for the various attributes. Crude and semi-detailed methods should be structured so that the model can address several types of projects in case the sector-specific detailed methods required for the project at hand are not supported by the model. Detailed methods of project financing and construction cost estimating should be included in the model. 101 4. Flexibility of the model structure The flexibility required of the model structure can be described as follows: • The ability to formulate any project alternative or scenario with any number of operations and methods, or indirectly with any number of phases; • The ability to model several performance measures which would assist different types of decision makers; • The ability to link together project operations; • The ability to support hierarchical representation (aggregation levels) of costs/revenues; • The ability to model project variables over time; and • The ability to perform sensitivity and risk analyses on any performance measure. 5. Effective risk analysis framework The risk analysis framework should support: • The ability to model the risk/uncertainty of any variable in a way that reflects the available amount of information about the variable in question; and • The ability to derive the statistics and probability distribution of performance measures. 6. Flexible decision support system The decision support system represents the implementation of the generalized model and risk analysis framework. The major capabilities of the system should include: (1) a flexible user interface; (2) the ability to formulate any alternative or scenario; and (3) the ability to maintain, manipulate and reproduce data and results for any alternative. 102 By satisfying the foregoing characteristics the proposed generalized economic model and risk analysis framework would contribute for improved quality in decisions for capital investment projects. This would fill the gaps in the current state-of-the-art systems and models as explained earlier in the chapter, namely: • the inability to distinguish between the requirements of the different project phases as to the functional estimating and forecasting methods used in each phase (e.g. construction estimating methods and demand forecasting methods); • the inability to provide a link between the phases that might be required in the calculation of performance measures (e.g. a link between financing drawdown and capital expenditure required, and a link between O&M costs and project demand); • the inability to aggregate costs into groups for further analyses (e.g. to determine deterministically and probabilistically the total cost of a group of work packages); • the lack of any terms and methods (e.g. drawdown methods, repayment methods, interest rate types, debt fee calculations) for the financial instruments mainly used in financing capital investment projects (e.g. syndicated term loans and private placement bonds, floating rate notes); and • the inability to model the uncertainty of individual project variables in a comprehensive manner (as in CASPAR) and the inability to use different probabilistic methods in the modeling process (i.e. not only using full probability distributions). The following three chapters elaborate on these characteristics and describe the design of a generalized economic model and risk analysis framework, and their implementation in the form of a decision support system called Evaluator. 103 104 Chapter 4 Generalized Economic Model 4.1 Introduction This chapter presents the design of a generalized economic model for the analysis and evaluation of capital investment projects. The design is built on the characteristics required of an economic model for project appraisal as outlined in the previous chapter. The model is a multipurpose hierarchical time function structure. The model structure integrates the properties and methods of the various industries and markets in capital investment projects through what is introduced herein as components, classifications and shape functions. It is hoped with the structure of the generalized model that enables detailed modeling of capital investment projects to provide a contribution to the economic modeling of such projects that can fill the gaps of the previous models. A general description of the concepts and structure of the model is given in the next section. The four sections that follow it provide a detailed description of the components of the model structure. Model performance measures are explained in the final section. Equations that are part of the general model are numbered sequentially. Other equations numbered with the letter E as a prefix are used for explanatory purposes and are not part of the model. The risk analysis framework and decision support system (DSS) are introduced in chapters 5 and 6 respectively. However, as the DSS represents the implementation of the model and framework, the dialogue (interface) part of the DSS which deals with the input and output of data and results is used in this chapter to illustrate details of the economic model components developed herein. 105 4.2 Generalized Economic Model 4.2.1 Model Structure and General Concepts A project cash flow can be used to represent any of (1) "expenditures" for the design, construction, maintenance, replacement or operation of a project; (2) "revenues" earned from operation, subsidies, or collateral raised from side businesses to the project (e.g. leasing air space in a project right-of-way); or, (3) "financial funds" advanced, or repaid, for the project. The basic concept behind a cash flow is a relationship between time and money which can be expressed as a function,/̂  (t, x), where t is time and A: is a vector that contains references to the variables in the cash flow. The value obtained by a cash flow function at any time t depends on (1) the methods used in the calculation of that cash flow using the x vector and (2) the value at time t of each variable in x. In cost/revenue calculations, infrastructure cash flows may share common properties and common methods. Properties are specifications such as type of cash flow (e.g. construction cash flow and maintenance cash flow) while methods deals with computational procedures (e.g. revenue estimating methods). These properties and methods can be used to form classifications of cash flows, with each classification having its own particular properties and methods. A new cash flow formed for an economic analysis can become a member of a particular classification, and thereby inherit the properties and methods of the classification. A classification can be represented by a model function,̂  (t, X) where t is time and X is a matrix of all the x vectors in the classification. By forming classifications, the following benefits can be achieved: • Classifications provide an avenue to model the eight attributes of the rights, obligations and liabilities dimensions of a project and to differentiate between these dimensions. An 106 attribute can be represented by a classification and consequently the properties and methods of the classification can be used to represent the attribute operations and methods. For example, a development attribute can be represented by a classification that includes those properties and methods commonly found in design and construction operations. Consequently, a construction operation (e.g. work package or activity) can be represented by a cash flow of the "development classification" and estimates for that operation will be made using those properties and methods of the classification. • The whole life cycle of a project can be represented by the classifications representing the eight attributes. Differentiation between life cycle phases can be reasonably made since each phase would be represented by one or more classifications. For example, the engineering phase in a two-phase life cycle (e.g. engineering and operation) would be represented by cash flows of the finance and development classifications that represent the finance and development attributes. • Any number of operations (e.g. construction work packages or revenue streams) in a project life cycle can be introduced to an economic model since the addition or deletion of an individual operation, which is represented by a cash flow that inherits its properties and methods of its classification, will not affect the other parts of the model. • Methods of a particular classification can include the crude, semi-detailed and detailed methods used in the calculation or estimation of a relevant attribute (e.g. financing). Therefore, a classification can have a specialized as well as generalized representation covering the attribute calculation methods. The structure of the generalized economic model is built on classifications as defined above. The model structure as shown in Fig. 4.1 is represented by four inter-related components: capital expenditure (CE); operation and maintenance (OM); revenue (RV); and, financing (FN). A 107 component is the "physical" representation of a particular classification in terms of its properties, methods and X matrix, and corresponds to a classification function f (t, X). Therefore, the generalized model includes four classifications representing the attributes of the requirement structure. As shown in the sketch below, the capital expenditure, operation and maintenance, revenue, and finance classifications represent the development, operation, revenue, and finance attributes respectively. Methods for the possession, general liabilities and tax attributes are represented in the capital expenditure and operation classifications assuming they are represented by crude and semi-detailed methods. The risk analysis framework handles the risk attribute separately. Classifications Requirement Structure Attributes Capital Expenditure • Development, Possession, Liabilities, Tax Operation & Maintenance • Operation; Possession, Liabilities, Tax Revenue • Revenue Finance • Finance Each of the four components is considered to have basic "constructs" that are called work packages for the CE component and streams for the other components. A construct, i, Figure 4.2, is the "physical" representation of the inherited properties and methods and the x vector that is part of X, and represents a cash flow function/̂ /, x.).. Any number of constructs can be added to a component. Since each construct's cash flow belongs to the component classification, cash flow calculations for the construct can be formed using any of the methods (crude, semi-detailed or detailed) that are included in the component classification. This provides for the component cash flow function,/̂ , (t, X), to be formulated or integrated from the cash flows of the constructs in that component, knowing that each construct may be different than others in the component in terms of properties' value and method of calculation. This provides for a flexible model structure that is not affected by the addition of deletion of any construct to a model component. 108 a u > 3 •o c o D. X w Financing Component D e b t S t r e a m s Project Start < • Revenue Component R e v e n u e S t r e a m s Capital Expenditure Component Work packages < • Operation & Maintenance Component OM Streams Figure 4.1: Economic model - Interrelated components and basic constructs Construct (Work package/Stream) Properties Methods Figure 4.2: A construct has properties and methods Let m, p and q be the number of constructs in the relevant component and i be a construct number in the component, then: m i=\ (4.1) fRy (t,XRv)-^fc it ,XRV .)t i=\ (4.2) f()M & XOM ) - ^' ^ / ^ i i=l (4.3) 109 In the above formulations and throughout the model, t is global time defined in terms of global/ project time unit (GTU) and referenced to project start (Fig. 4.1), and f is construct local time defined in terms of construct local time unit (LTU) and referenced to construct start. Time conversion from t in the left side of the equation to f in the right side is made accordingly and will be described below. The subscripts CE, RV, and OM of x, X and construct cash flow function refer to a classification/component type and have been omitted later for clarity with x and X, as has the subscript / which refers to a specific construct/cash flow in a component. The FN component has its own constructs, properties, methods and X matrix. However, because of the discrete nature of the flow of financial funds, the cash flow of the z'-th debt stream is represented by an information vector TRIFj instead of a cash flow function/™ (t, xj.. The concept of cash flow classifications of properties and methods, which is behind the formulation of components and constructs, represents the main concept behind the structure of the generalized economic model. Each of the four components and their classifications will be described later in detail. The following subsections, however, describe a set of properties and methods that are common and essential to all components. Unless otherwise noted throughout the model formulations, all flows and usage rates are assumed to be continuous. However, allowance is made for discrete flows in the four model components particularly the FN component. Discounting of continuous and discrete cash flows is performed using continuous compounding (Remer et al. 1984; Tanchoco et al. 1981). Continuous compounding gives flexibility in dealing with any cash flow pattern (Park and Sharp 1990) which is suitable for dealing with the various types of cash flows in capital investment projects. 110 4.2.2 Component Common Properties Time Unit Characteristics Time characteristics of a project are defined in terms of several properties. Two time unit properties are defined and each can assume any value of the following time periods/intervals: one-month, three-month ("quarter), six-month (semi-annual) or twelve-month (annual). The first time unit property refers to project time, t, that uses project/global time unit (GTU) which can be used to refer to project duration, a constructs1 early start time, and elapsed times after project start. The second time unit property refers to local time of a construct, t\ which uses local time unit (LTU) to define the duration of the construct or to delineate the time scale for the cash flow pattern of the construct. With LTU each construct can have its own value of the time unit properly. For example, a revenue stream can be defined in annual periods while another revenue stream can be specified in month periods. The CE component, however, is an exception where the same time unit value must be shared by all of the CE constructs, i.e. work packages, for scheduling purposes. Reference Characteristics Since continuous cash flow patterns are used in the generalized model, a reference property is used to define the origin of calculations for the cash flows. This property can take project (global) start or construct (local) start as a reference value for calculation. Nearly all the model variables that will be defined later have an attached project or local reference. When a variable in a construct has a continuous function with time that uses project start as a reference, then GTU is used as the time unit for the time scale of the function. When the reference is local, i.e. construct start, then LTU is used. The reference used for all model variables will be detailed later when describing the model components. Ill Since several time unit values are involved in model calculations, conversion from GTU to LTU and from LTU to GTU is required. The conversion is made simple by a conversion factor that divides the month equivalent of the first time unit by the month equivalent of the second time unit: Global to local time conversion (GtoL) = GTU (in months) / LTU (in months) (E4.1) Local to global time conversion (LtoG) = LTU (in months) / GTU (in months) (E4.2) As an example, consider converting from a project time t defined in annual periods to a construct local time f defined in quarters. Then GtoL would be equal to 12/3 = 4. Where ESc is construct start time in GTU, then the construct local time is /' = (t-ESc) -GtoL (E4.3) In another example that considers the reference property as well as time unit property, a work package construct may use month-period as a LTU and the work package starts at ESc = 2.5 years (GTU in years) as shown in the sketch below. If in the construct a unit price of material $/m3 is defined as a continuous function with global reference and global time unit, then the spot rate of the unit price of material will be determined at all times by converting LTU to GTU, where LtoG = 1/12. Unit price $/m TTTTTTTTfrTTf̂ time in years (GTU) project start time in years (GTU) ESc = 2.5 years 6 months (LTU), Work package #101 Therefore, the unit price of material at a time f = 6 months will be obtained by setting the unit price function at that time to 3 years using t = f • LtoG + Esc (E4.4) 112 Network Characteristics A project operation can be described by a time span (duration), logic and lead/lags time with other operations, therefore duration, lead/lag and logic represent three properties included in the generalized model for component constructs. These three properties integrate together to provide the generalized model with a network-based structure. These properties delineate the domain of application for the cash flow of a relevant construct where the cash flow time function has a value only within the construct duration. These properties are further detailed later. Contract Characteristics Since capital investment projects may be procured through the alternative PPP delivery system, a contract period is usually defined for developers in the concession or development agreements as explained in Chapter 2. This contract period can be referenced to start from the signing of contracts or from an established future date. Therefore, contract period and its contract reference are another two properties in the generalized model. For example, the Northumberland Strait Crossing has a 35 year concession period starting from "Date Certain" (construction completion). These properties define the domain of application for the RV and OM streams to be limited to the end of the contract period, if it is given (e.g. 35-year concession); otherwise calculations consider the whole duration of each cash flow (e.g. 45-year revenue stream). Figure 4.3 illustrates an interface window for project identification in Evaluator showing some of the above properties. The list box in front of "Project/Global Time Unit" allows the choice of a time period as a GTU for the project. "Contract Duration" is illustrated with a "Contract Duration Reference" to satisfy the properties under the contract characteristics above. Figure 4.4 illustrates the use of LTU for the CE component along with information explaining that the constructs of the other components will have their own time units. 113 f i Evaluator[EX-001] File Project Analysis Window Help 7[nrx| « l Project Properties Identification j Components Identification Code Title hwy1 01 Highway 101 Comments Owner Highway 101 is a transportation facility under the BOOT arrangement. * Financed by 80/20 debt equity ratio. Concession for 20 years. Debt (financed via syndicated loan with 15 years term, based on LIBOR. y j Private Transportation Corporation, PTC. Consultant Consulting Group Inc. Project Time Frame Project/Global Time Unit (GTU) Project Start Date PSD (Time Zero) (mm/dd/yy) annual (12-month) 3 ] 1/1/00 Contract Duration (e.g. concession period) in GTU Contract Duration Reference | Specific Work Package Work Package Code 20 CEidl Annual Nominal Discount Rate Discount all cash flows to time? (in GTU, time zero is the default) 0.2 Figure 4.3: Project identification and common properties 114 #i. Evaluator [EX-001] mmm 1-1--1 File Project Analysis Window Help jjj £fc Project Properties BHHQlnjjcj Identification Components Capital Expenditure Component Start period of the first/start work package in the Capital Expenditure Component Network, in GTU from Time Zero annual (12-month) Same for all materials Time unit for all work packages (LTU) Inflation strategy Profit margin (decimal) JO ______ Overhead rate (decimal) m Revenue and O&M Components The start of each stream in both components will be linked to project start or to another stream in the relevant component Time unit (LTU) for each stream of both components can be made unique. Inflation strategy for revenue I Same for all streams Inflation strategy for O&M " 3 Unique for each stream Finance Component Each of the finance streams will have its own properties. " 3 3 Figure 4.4: Component properties 115 4.2.3 Components Common Methods: Shape Functions Component classifications have two categories of methods: component-specific and common methods. Component-specific methods represent a group of methods unique to each classification type and will be described later for each component. The "general" methods are a number of methods common to all classifications and are described herein as "shape functions", fs (f\ y), where t" is either local or global time andy is a vector of sub-variables. Shape functions represent patterns of change of a variable over time. In the generalized economic model, any variable Z in a cash flow vector x can be represented by a shape function and described by function notation, Z(t \ y) or Z(f\ y). In model formulations, y will be dropped for clarity. With f the variable is expressed in terms of a local time unit and local reference. With t" the variable can have either local time (t" = f) or global time (t" = t) that is defined in terms of a global time unit and global reference. Some variables in the model, as explained later, are local only, such as productivity and work quantities; others such as unit costs, inflation rates and floating interest rates may experience both local and global domains. The main use of shape functions in the generalized model is to give all model variables, particularly those representing rates, the ability to change over time. This allows improvements to economic analyses and economic models that mainly express variables as average fixed values over time, as explained in the previous chapter. The other use of shape functions is to represent semi-detailed calculation methods. Semi-detailed methods were described before as methods that can represent aggregated sums of cost/revenues spread over time according to a specific pattern such as a uniform flow of expenditures. 116 To maintain generality, a total of 32 shape functions in two categories "rate functions" and "area functions", as described below, have been included in the economic model. These functions may experience several characteristics or forms such as step, ramp, decay, and exponential growth. Rate Functions Rate functions include a number of time-related functions. The mathematical expressions for these functions are described in Table 4.1 along with a description of the sub-variables or parameter variables that are included in the y vector of the shape function. Some of these functions have been used by others to describe general cash flow patterns in economic analysis (Remer et al. 1984; Tanchoco et al. 1981; Almond and Remer 1979; Park and Sharp-Bette 1990). The value assumed by any single variable (e.g. inflation) in a construct cash flow calculation at a given time is obtained using its shape function expression and depends on the local/global domain of the variable. It is worth mentioning that the mathematical forms of shape functions may depend on several sub-variables. For example, the "step uniform" pattern could be defined using nine parameter variables where five rates can be defined along with four time limits as shown on the step uniform figure in Table 4.1. This provides for a refined level of detail in modeling any single variable in the generalized model. Area Functions Area functions are functions that model situations in which the total value of a variable is constrained regardless of time or of how the variable will change over time. Area functions can only use a local reference as compared to rate functions that can use global/local reference. 117 S 1 ^ 1 c u a \£ CU 03 co fi O CP -5 a a CU o o CN 9n[BA c fi s fi <u O ^ CP c j CP ^ 43 " fi o s-l a + co c3 CP fi "! o  vi o d o - " X o CD O O ON J x J cb i-) 3n[BA q co" CP <p CP <p" 13 > t-c o co' CP •a <-l fi CP fi 03 fi CP > . 4=i e3 +-> O ^ i cS ^ <4-l CO * 2 O -fi o fi fi w <p £ c 2 t—1 ^ CO CP ccj SH CP o > CP ÎH CO H V VI m H 2 VI o <4H ' f i OH CP 00 H Al in C/J c _ CN co E—i E—1 E—1 v v V VI VI r-H CN E—i H i CN m &i K ^ H H CN H h in cn B 00 CP CP 4=1 CO CP _ f i > CP 73 ;-< in o co CP > CP £1  o CN    •3- The mathematical expressions of area functions, Table 4.2, are derived by integrating the relevant rate function over a total duration b, which represents the duration of a construct, and equating the results to the total value of the variable over that duration, i.e. Qtr- fs(t,y)dt 0 (E4.5) For example, the linear function L [= f (t, y)] is derived as follows. Letting Qs and Qr be the start and rate values, respectively and letting Qt be the total value of the variable, i.e. fs(t,y) = Qs + Qr-t (see Table 4.1) (E4.6) Qt , Qs-b , Q r ' b l 2 (E4.7) ^ , x -Qr-b2 + 2-Qt n L(Qt,Qr,b,t) := -~ —^ + Qrt (see Table 4.2) 2 ' b (E4.8) The following example helps explain the use of rate and area functions. If 10000 mhrs are needed to finish a work package in 50 days, then a variable for labor input can be described as 200 mhrs/day (25 laborers/day) using a "uniform" rate function (see the solid line in Figure 4.5) which means that if the duration is extended to 60 days then another 2000 mhrs would be needed. Alternatively, using the "uniform" area function then 10000 mhrs will be fixed; that is 166.67 mhrs/day with 60 days - see the dashed line in Figure 4.5. Further, if a minimum number of laborers is needed in a day and the work force builds up as work progresses, then a "Normal" area function can be used with 10000 mhrs as a total value, 2000 mhrs as a base value (i.e. at least 4 laborers/day in 8-hrs day), and 60 days as duration. Thus, 37.024 mhrs/day (5 laborers), 107.045 mhrs/day (13 laborers), 355.89 mhrs/day (44 laborers) will be assigned for the 1st, 2nd and 3rd day respectively, Figure 4.5. Figure 4.6 illustrates how the shape functions have been implemented into the system. Material inflation in a work package is described using Step Uniform with 5 rates and global reference. Area functions have the word "Total" attached to their names in the system. 125 s IMI c o c cu al l I * CO fi o o a <D < CN 4°. 03 H 43 c s-CU +•> PH s 2 o o CN 3 I o 'ii >-fi T3 <D _ f i (H O CD +3 ^ a *43 CD s -4-» o o d o m Y O (fi ^ o ' * o w *o g n j B A 43 s-T • r fi CD O J 13 I CC3 S-H fi T3 i-i O CD ' ' O (-1 Cd •4—» O - | C N £ a a 43 CD fi H V VI H C O t—1 I H V -4—» VI o ' f i OH CD H—» 00 VI E—1 uo < CN H V ~ 4H' H • CN H CN CO c C O E—i V 43 H CO*" H CN h in" < co < -1 •< CN H C O H C O 00 CD (D 43 CO CD 13 > <D uo < o fl o "el o CO CD 13 > CD s ccj -fi o £ <c H    o S T3 <D C • —* o o GO a o • i—i +-» o 0) IMI a o • a \S a o- a CU •«-< CM VI VI o o H O o V VI o I a (N CN i 3 a CD a CD a + 91 oi CN O 3 I H 43 a CN I N a CN CD C/3 5 1 § § 2 g CD X> & CU 6 a. a. H a. ' I'd o N CD S-i H s o CN o .1 ,1 ,t  ,2 ,5 ,t  fN IV) o o O o o O o o Y o + o o'oc - C Q -|- CQ CQ cu fi fi o o co fi o o fi fi pH <D «! CN CU t—I M l c o CJ C \& ii All « I-* I I CN fi. I CN fi. fi. I CM fi. I 3 fi. I CN + CN fi. CN CD 43 CO / — C O •4-> u< o « fi CN CD T 3 13 fi <U > co CN fi. + fi. O q £ « 2 fi o fi CU a O +J cu O S o tJ cu S OH » " co <<vs CN l-i co fi-<3 £ S3 CU H-< •<-< cU -4-» IPH cd H-» CU 20 30 40 time, days 50 60 Figure 4.5: Shape functions Uniform, Uniform (a), Normal I (a) B> Evaluator [Highway] - [Material] - Of X | ft • Eile Eroject Analysis Vtfndow Help _ | S | X | Material Calculation Method Material Cost Inflation Value of First Rate/Area] Deterministic • Deterministic _] Value of Second Rate/A j Deterministic Time to Start Third Rate/Area Choose Trend Method Step Uniform I (5 Rates) Exponential I Exponential II Exponential III Logarithmic Growth I Growth II Power Function Reciprocal Function — 0.05 I Unit Cost & Quantity: C(t)- u(t).Q(t) Inflation Time Reference Start of Project • _J j Deterministic • 1 M |5 Value of Third Rate/Area | Deterministic _ j u |0.065 r Time to Start Fourth Rate/Area | Deterministic M |8 Value of Third Rate/Area | Deterministic _ ] V- JO.OS Time to Start Fifth Rate/Area | Deterministic u |l4 Value of Fifth Rate/Area | Deterministic jrj ^ |0.1 Figure 4.6: Shape function and global reference of material inflation of a work package 132 4.3 Capital Expenditure Component Before a capital investment project becomes a productive asset, i.e. before start of operation, a large sum of money must be put in place to develop the project. This sum of money is called the total investment cost, which is composed of three components: fixed-capital investment, pre- production expenditure and working capital investment (Humphreys 1991; Peters and Timmerhaus 1991; Behrens and Hawranek 1991; Frohlich et al. 1994). The first two cost parts represent the capital required to develop, construct and equip the project. The foregoing authors explained in detail several cost items and categories of costs for these two components of the total investment cost, using the following classification (Behrens and Hawranek 1991): 1. Land (e.g. purchase and transfer, legal charges) and site preparation and development 2. Civil works, structures and buildings (including engineering and design costs) 3. Plant machinery and equipment 4. Incorporated fixed assets (e.g. industrial property rights and technical know-how, patents and license fees) 5. Pre-production expenditures (e.g. business set-up costs, human resources, pre- investment studies, project and site management, marketing costs) Working capital investment represents the amount of capital needed to get the project started and meet its current obligations. Humphreys (1991) identified the following working capital items: 1. Raw materials inventory 2. Work-in-progress inventory (semi-finished goods) and finished-products inventory 3. Supplies for product manufacture 4. Taxes payable 5. Accounts receivable 6. Cash or equivalent on hand for salaries, wages, etc. 7. Accounts payable. The different types of cost items described above for the total investment cost cover the costs needed to meet the requirements of the development, possession, general liabilities, and tax attributes as explained previously in Chapter 2. The capital expenditure component (Figure 4.1) is the vehicle through which cost items of the total investment cost can be estimated and modeled for cash flow analysis. 1 3 3 4.3.1 C E Component Structure and Properties The structure of the CE component is made of constructs called work packages linked together via the critical path method (CPM) with finish-to-start relationships and lead/lag times (Clough and Sears 1991). A work package possesses time and cost characteristics using which the cost of the work package can be estimated, distributed over a period of time and presented for cash flow analysis. In a coarse representation, the total capital expenditure cost can be defined in a few number of work packages (e.g. three) linked together by CPM as shown in level 1 in Figure 4.7. Alternatively, at a higher level of detail, the total capital expenditure can be defined in several number of work packages detailing all items of cost, and linked together by CPM as shown in level 4 in Figure 4.7. The cost of any group of work packages can be combined to provide an aggregate or lump sum cost of the group; two or more groups can further be combined to yield a higher lump sum. The CE component provides four group/aggregation levels as shown in Figure 4.7. Detail/coarse representation, however, can better be defined in terms of the type of cost estimating method selected for cost calculation of a work package, as detailed later. The CE component has time unit, network and reference properties, which can be inherited by any new work package where each work package may have its own values of such properties. The same time unit property, however, must be shared by all work packages. If the component uses, for example, a year as a time unit, then all work packages must use the same time unit (LTU). While CE component work packages use a LTU, the early and finish times of work packages, ESc and EFc, and defined in global time units GTU (see Eqs. E4.1, E4.2). Figure 4.8 shows three windows that explain the implementation of all such properties in the support system. Through the "capital expenditure" window any work package can be added or deleted. The "scheduling" window shows the duration, predecessors, and lead/lag times for work package CeidS. Work package time unit is shown in the "Project Properties" window. 134 in cn  Any item of cost or expenditure must be attached or made within a CE work package. Each work package represents a cash flow function^ (t, x) where x refers to the work package variables that are used in cost and cash flow calculations. These variables can be categorized under material, labor, equipment, indirect or sub-contracting, capital forecasting, discrete, inflation, and network variables. The variables that are price-variables (e.g. unit costs and wages) can have their reference property set to either a global or local reference; inflation variables can be local, global or null (i.e. no inflation); capital forecasting variables and non-price-variables (e.g. quantity and productivity) have a local reference only; however, discrete and network variables may have no reference. Table 4.3 summarizes some of the properties of these variables. Figure 4.6 shows the specification of inflation of material cost of a work package using a global reference system. Table 4.3: Properties of work package variables Variable Definition Shape Functions Time Reference Function Input" (e.g.) Function Output (e.g.) M(0 Total material cost per time unit Area/rate Local $ or $/day $/day c (n . Unit cost of material Rate Global/local $/m3 $/m3 Q(0 Quantity per unit of time Area/rate Local m 3 or m3/day m3/day u/O Labor usage/input per time unit Area/rate Local mhrs or mhrs/day mhrs/day P,(0 Labor productivity/ Rate Local m3/mhr m3/mhr wfn Labor unit cost per labor time unit Rate Global/local $/mhr $/mhr Labor unit cost per unit of Rate Global/local $/ m 3 $/ m 3 production S(t) Subcontracted/indirect cost Area/rate Local $ or $/day $/day x(f) Capital forecasting A/R/other Local $ or $/day $/day Inflation, 6 variables Rate Global/local/ N / A N / A No reference * Variables subscripted to / (labor) have equivalents referring to equipment and are subscripted to e Depending on the type of function, arguments to a function may vary, e.g. total value or value per unit of time. 137 As explained previously, each work package has a cash flow vector x that refers to the above variables; and typically, each work package may have its own values for such variables. Economic analysis may be performed, however, with the assumption that values of some variables are the same throughout the calculation. For example, a variable for concrete unit price CJt") may be considered to have the same value (e.g. 1000 $/m3) across some work packages and different values (e.g. 2000 $/m3) for others. Sensitivity analysis on total cost of construction with respect to the 1000 $/ m3 concrete unit price will need to check on which work packages have this value of this variable before performing such analysis. For this purpose, the CE classification has three properties that can be set to determine the treatment of the CE variables. The first property refers to common variables where if a variable is common then all work packages that use this variable will use the same value and shape function method in all cash flow calculations - see Figure 4.9. The second property refers to special variable where only those identified work packages that have this special variable will have the same value and shape function method in their cash flow calculation - see Figure 4.9. The third property refers a treatment of inflation variables. The CE classification provides several strategies to deal with inflation as shown in Figure 4.10, where except for the unique treatment, the same value and shape function method will be used according to the treatment: • Unique: each work package will have its values of the inflation variables; • Same-for-material or -labor or -equipment; • Same-for-indirect or -discrete or -capital forecasting; • Same-for-all-work packages; • Same-for'-proj'ect: all three component CE, RV and OM will have same value and shape function method across all project work packages and streams. 138 g T3 cn O B ° o TJ O <d "—, ^ CD s= «5 8 CO • O en * s i , p co o 3 HI D CJI„ TJ -5: I | p g £ ? 3 0- E E o o o .9-.Q- J] J 3 3 e O d c r c r _J _ J _1 LU LU 1) t i l n ai o "5 o co •- £ -3 « ^ ° 2 O CL .« c -) in cd 3 £ O Q Q) U- ai 13 i f Q O o o o o p o E l LU 2 9 .9 .9 .9 .9 .9 c cd (0 CO CD CD c c c c c c en 0 5 S d a ! ^ ^ J S U J W Q X © co ro _co eg © oi c5 "aj "55 ai ai CO LX UJ L O -z Si .fi S > 0 o O CO 2 o O TS co in o a w ' i ^ > S a £ Q o O co O o O 3 1 1 I £ 0) I 2 L <3 L L L  4.3.2 C E Component Methods and Cash Flow Formulation The CE component performs all of its cost calculations using semi-detailed (preliminary), detailed methods (Clough and Sears 1991; Humphreys 1991; PMI 1996; Hendrickson and Au 1989), and crude methods as represented in the CE classification. The cost of a work package can be distributed over its duration in two ways. The first way is to distribute cost according to a selected pattern of the shape function. The second is indirect where the variables comprising the cost estimate are allowed to change over time using shape functions and the distribution pattern of the total cost of the work package is derived from its variables. Therefore, while a work package cash flow function represents the expenditure flow needed to produce a quantity of work, it represents a cost estimating method as well. The following sub-sections describe how a work package cash flow function/̂  (t, x) is derived using the methods in the CE classification. C E Semi-detailed Methods Semi-detailed methods represent preliminary methods that arrive at a cost estimate based on experience, judgment and historical records. In the CE classification, semi-detailed methods are referred to as "gross" methods that distribute the total cost of a work package over its duration using a particular chosen pattern without regard to the individual items (e.g. material, labor, equipment) that contribute to total cost estimate. A cash flow function of a CE work package /' can be described as follows: fCcE(t\x) = X(f).e^ ° x { m (4.4) where f and t" are as defined before; X(f) is a constant dollar capital expenditure variable, explained below, and has local reference; and, 0x(t") is an inflation variable for X and can have global/local reference and can take any form of rate functions in Table 4.1. 141 Semi-detailed methods representing X(f) can be categorized in two parts. The first part of the methods is represented by the area functions described in Table 4.2. Therefore, the total cost of a work package must be defined in advance where the shape/pattern of the chosen area function is used to distribute the total cost over the work package duration. The following are examples that help to explain the use of area functions in modeling capital expenditure. A preliminary estimate may distribute the cost of a project over its duration using a distribution or loading profile that divides the duration of a project into a number of periods where each period is loaded with a percentage of the total cost. This loading scheme was used by Texas TGV Corporation for their proposal for Texas High Speed Rail ("Texas" 1991). The corporation loaded the costs for one phase of the project as 5%, 10%, 15%, 40% and 30%o over five years for engineering costs; 60% and 40% for right-of-way; 20%, 40%) and 40%) for fixed capital; 10%>, 50% and 40% for rolling stock; and 40% and 60% for the cost of stations. Similar loading profiles called disbursement profiles were explained by Bacon and Besant-Jones (1996) and Merrow et al. (1990) from the World Bank, see Table 4.4. The above two loading profile examples can be readily modeled by the step uniform area function in Table 4.2. However, the step uniform area function as shown in Table 4.2, which is implemented in the generalized model, can have up to five areas or totals representing the cost in a work package. If the total cost is to be divided in more than five areas, then another one or more work packages will have to be defined. For example, to model in the generalized economic model the loading profile that has 11 disbursements in Table 4.4, three work packages will be needed and linked together by finish-to-start relationships with zero lead times. 142 Table 4.4: Disbursement profiles for project cost in current price terms, Merrow et al. 1990 Annual Disbursements (of totals) in years Implementation periods (years) 11 10 9 8 7 6 5 4 3 2 1 0.02 0.03 0.03 0.04 0.05 0.10 0.15 0.20 0.25 0.55 2 0.03 0.04 0.07 0.10 0.15 0.20 0.25 0.30 0.50 0.65 3 0.06 0.08 0.12 0.15 0.20 0.25 0.35 0.40 0.25 4 0.01 0.12 0.15 0.25 0.25 0.25 0.20 0.10 5 0.20 0.20 0.20 0.20 0.20 0.15 0.05 6 0.20 0.20 0.20 0.15 0.10 0.05 7 0.15 0.15 0.12 0.07 0.05 8 0.10 0.10 0.08 0.04 9 0.08 0.05 0.03 10 0.04 0.03 11 0.02 S U M 1.00 1.00 1.00 1.00 . 1.00 1.00 1.00 1.00 1.00 1.00 The second part of the semi-detailed methods for X(f) represents CE component-specific methods. These methods represent a number of capital forecasting cash flow models that were developed based on historical data from building and civil projects. Using these models, total project cumulative cash flow is described by an S form or pattern. For this case, total work package/project cost and duration are specified along with a number of parameters. The CE classification includes five of such models referred to in the literature among a body of work in capital expenditure modeling. The models are based on the logit transformation and s-curve formulas described in Table 4.5 (Kenley and Wilson 1986, 1989; Hudson 1978; Berny and Howes 1982; Miskawi 1989; De La Mare 1979). Capital forecasting models in Table 4.5 can be considered as shape functions that have a vector of variables y, which represents the parameter variables in each of the five models. The expressions provided in Table 4.5 are the cumulative forms of the five models. The first derivative of these forms represents the expenditure function for each capital forecasting model, which usually takes a bell shape.   Figure 4.11 shows that semi-detailed methods, called "holistic" in the interface, can be used for cost calculation of a work package. In the "capital forecasting" window, the De La Mare capital forecasting method is used and therefore two parameters must be specified as in Table 4.5, the first is the total cost of the work package and the second is the shape factor p. The order of these parameters is the same order used to explain the variables in they vector in Tables 4.1,4.2, and 4.5. As shown on Figure 4.11, capital forecasting has no reference as it is considered local by default. 146 C E Detailed Methods The second category of CE classification methods represents detailed cost estimating. An item of work is decomposed into its basic four elements - material, labor, equipment and subcontracted/indirect costs. The cost of the work item is then determined by aggregating the costs of its elements. Detailed estimating is generally referred to as definitive or bottom-up estimating (Clough and Sears 1991; Humphreys 1991; PMI 1996). Each of the basic elements may have its own methods of calculation; the CE classification, therefore, contains several methods for each of the four elements, as detailed below. Equation 4.4 describes how a work package cash flow function is derived using semi-detailed methods for preliminary estimating. A cash flow function, however, for a decomposed estimating approach a CE work package can be written as follows: + fcs{t\x)-e (4.5) where f and t" are as defined before. 0 (t), 90 , 6 (t) , and 6(t) are inflation variables for the money-related variables (described below) in the material, labor, equipment and subcontracted elements respectively and they can be represented by any form of rate function in Table 4.1. The functions / (t\ x),f. (t\ x),f (t\ x) and / (f, x) are for cash flow calculations of the material, labor, equipment and subcontracted items respectively. The CE classification methods used for each of these functions are described below. All the variables in these functions are modeled by shape functions. 148 a) Material Estimating Methods The material element^ (t\ x) of the decomposed estimate of a work package can be estimated through semi-detailed and detailed methods. Semi-detailed material estimating is where gross total material cost is defined and distributed over the work package duration using an area function or where gross total material cost per unit of time is defined using a rate function over the duration of the work package. M(f) as shown in Eq. (4.6) below represents material cost gross estimating methods, uses area/rate functions and a local reference system. Detailed material estimating corresponds to unit cost estimating methods. These methods depend on the use of unit cost/price C (t") and work quantity Q(f) to determine the total material cost of a work item. Alternatively, the quantity of work can be obtained through the labor input Uff) required by the work item and the productivity of labor P{(f). Both methods are described in Eq. (4.6) below. C (t") can have a local/global reference while Q(f), Ut(f) and P ff) can only have a local reference. C (t") and Pff) are modeled by rate functions while Q(f) and Uff) are modeled by rate/area functions. A work package material cost cash flow function can be described as follows: / (t\x) = cm M(f) Cm(r)-Q(0 (4.6) [cm(n-Pi(n-ui(n where M(f) the total cost of material per unit of time, CJt") is the unit cost of material (e.g. $/m3); Q(f) is the work package scope/quantity placed per unit of time; Uff) is the labor usage/input per unit of time (e.g. mhrs/day); and P{(f) is the labor productivity in placing a unit of quantity (e.g. m3/mhr). 149 For example, if a work package needs 500 m3 of concrete to be placed in 10 days, and where the price of 1 m3 is $1000, then the material cost of the work package is $500,000. Assume that the price will remain constant during the work package duration, i.e. uniform rate function (Table 4.1). Further, assume that large quantities will be placed in the first four days and the placement rate will tail off over the remaining duration. This placement profile can be represented by an exponential III area function (Table 4.2). Then the material cash flow function Eq. (4.6) will be as shown in Figure 4.12. If it is assumed that the placement rate was constant over the work package duration, a uniform area function would be used and the cash flow function would be the dashed line in Figure 4.12. In both cases, the total cost by the end of the work package is $500,000. However, their discounted values would be different. fus(D5' t : time Figure 4.12: Material cost cash flow function b) Labor and Equipment Estimating Methods Both the labor cost/^ (t\ x) and equipment cost/̂ g (f, x), as defined in Eq. (4.5), use similar calculation methods and variables. These methods, described in Eq. (4.7) below, fall into three categories: gross estimating methods, cost per unit of time and cost per unit of production. The gross labor methods determine the labor cost in a work package in terms of total labor cost that is distributed over the work package duration using a particular pattern or in terms of gross total labor cost per unit of time. Gross methods are represented by the area/rate functions and modeled by Hff) in Eq. (4.7) using a local reference system. 150 The second category determines labor cost per unit of time in terms of labor unit cost per unit of time Wfi") (e.g. hourly labor-wage rate) multiplied by the labor usage/input U{ (f) (e.g. number of man hours per unit of time). Alternatively, where rate of work Q(f) and productivity of labor P/O are available, U{ (f) can be evaluated (i.e. Q(f) = Ul (f)- P/O) as shown in Eq. (4.7). Wt(t") can have a local or global reference while Uff), Q(f) and P (f) can have only local reference. Wft") and Pare modeled by rate functions while Q(f) and Uare modeled by rate/area functions. The third category for labor cost determines labor cost in a work package in terms of labor unit cost Cft"), i.e. labor cost per unit of production, and rate of work Q(f). The quantity of work can be obtained alternatively through the labor input Uff) required by the work item and the productivity of labor P(t) as shown in Eq. (4.7). Cft") can have either local or global reference while Uff), Q(f) and Pff) can only have a local reference system. Cft") and Pff) are modeled by rate functions while Q(f) and Uff) are modeled by rate/area functions. Using the above three categories of methods, labor cost cash flow function can be expressed as follows: Wiin-QWW) (4-7) Q ( O - 0 ( O /, (t\x) = cl where, Hff) is a total labor cost per unit of time; Wft") is unit labor cost per unit of labor time (e.g. $/mhr); Uff) is labor usage per unit of time (e.g. mhrs/day); Q(f), is work package scope/quantity placed per unit of time; Cft") is labor cost per unit of quantity; and Pt(f) is labor productivity in placing a unit of quantity (e.g. m3/mhr). 151 The expressions for the equipment component make use of the same cost estimating methods just described for the labor component. Thus, the equipment cash flow function^ (t \ x) is similar to the labor cash flow function in Eq. 4.7, but with all variables subscripted to e. c) Subcontracted/Indirect Estimating Methods Subcontracted costs are expenditures made to subcontractors for performing a specific portion of the project. Indirect costs generally represent those expenditures made to cover overheads (e.g. job and office), supervision, construction expenses (e.g. utilities, temporary facilities, permits, taxes), and contingencies (Humphreys 1991; Peters and Timmerhaus 1991). Although these costs are general project expenses and usually not assignable to a specific work package they can be represented in the generalized model from within work packages. This provides for these costs to be either reflected as part of a direct cost work package, or to stand alone as if one or more work packages dedicated solely to indirect costs. The subcontracted/indirect element of a work package cash flow is represented by gross methods that determine the estimate in terms of cost distributed over the work package duration in a particular pattern or in terms of gross total cost per unit of time. Gross methods are represented by the area/rate functions and modeled by S(t) using a local reference system. Therefore, the subcontracted/indirect cash flow of a work package is represented as fcs(t\x) = S(0 (4.8) 152 C E Crude Methods: Discrete Costs The preceding sections determined work package cash flow in terms of semi-detailed and detailed calculation methods that assumed expenditures were made as continuous functions of time. However, some expenditures may have to be made as discrete sums (e.g. procurement of major permanent equipment item or construction equipment). Therefore, the generalized model provides for discrete costs to be added to any work package, independent of whether the other cash flow component is formulated by semi-detailed or detailed methods. CE CE CE The discrete cost Dj is represented as follows: Let Dvy andDty represent constant dollar value and local time respectively for the j-th of n discrete costs in a work package, and let 9Jt), modeled by rate functions, be the discrete-cost inflation variable. Then, * CE r*=Dt. if 0 j is in local time (49) t*=mCjE +ESc if is in global time C E Component Cash Flow It is useful to summarize the flexibility offered in modeling CE component cash flows. Any work package cash flow function ,̂ (t, x) can be formulated either through a semi-detailed function using Eq. (4.4) or through a detailed function using Eq. (4.5). The detailed function is represented through material, labor, equipment and indirect cost components as per Eqs. (4.6,7,8). Both the semi-detailed and detailed formulation of a work package can be supplemented by the discrete formulation given in Eq. (4.9). 153 f (t, x) represents the essential information for any work package in the CE component. By aggregation, all work package cash flow functions integrate to form the CE component cash flow function as described in Eq. (4.1). Evaluator, the decision support system developed as part of this thesis, has implemented the foregoing cash flow formulation. Figure 4.11 depicts the semi-detailed cost estimating approach. Figure 4.13 illustrates the detailed cost representation, where the top menu in the window shows material, labor, equipment, and indirect cost menus. Figure 4.13 shows the quantity window as well, where total quantity is assumed to be distributed using a uniform area function with no time reference as it is considered as local by default with quantity. Figure 4.14 illustrates the material and labor cost calculation windows for a work package. The top bars in these two windows allow the selection of a material and labor calculation method from those described in Eqs. (4.6,7). According to the method selected in these two windows, appropriate tabs are shown to request the information required for the selected calculation method. Figure 4.15 shows the input windows for the indirect and discrete costs in a work package. Note that each of the material, labor, indirect, and discrete cost windows has a tab for describing the inflation function, which can be unique for each of them. In Evaluator, any number of work packages can be used to represent the capital expenditure of a project. However, in terms of Evaluator's performance, the reduction in the number of work packages and the use of semi-detailed methods contribute to faster processing of the performance measures. It is suggested to use the detailed methods for the elements of costs or work packages that require more analysis in the appraisal stage of a project. The more the number of work packages, particularly with the cost elements represented as risk variables as explained later, the more the time it takes Evaluator to do processing. 154 #i hvoluotor [Highway] iVork packages Scheduling Quantity Material Labor Equipment Indirect/Sub Costs Discrete Costs &dd New Delete Next Back Levels Variables k* Capital Expenditures - • X General Work package Identification WP Code |CEid3| Total Number of Streams 3 Comments Highway bridge abutment #2. Will start after finishing work on its piles (CEid2) and finishing abutment #1. J Cost Calculation Methods Calculate Work package Total Cost Using: 1 Decomposed Calculation 1\ Use the "Material". "Labor" and "Equipment" Menus to select methods to be used in cost calculations forthe work package. Note: value of work package quantity will be used only if you chose a method in the material, labor, or equipment menus that uses quantity. Note: tabs lor labor productivity and usage (of the Labor Menu) will be available in the Material Menu only if you select a material method that uses them. Quantity Parameter 1 Choose Trend Method I Uniform Total | Deterministic T | u 110000 • X 31 Figure 4.13: Work package cost calculation using decomposed methods and quantity window 155 flpi. hvaluator [Highway] Work packages Scheduling Quantity Material Labor Equipment |ndirect/Sub Costs Discrete Costs _ |n| x| A,dd New Delete Next Back Levels Variables Six Material Calculation Method | Unit Cost Labor Usage a Productivity: C(t) • u(t).P(t).L(t) ~ Material Cost | Inflation] Lab. Productivity | Lab. Usage | Choose Trend Method Time Reference Uniform I Parameter 1 - - I Deterministic T]:f.[500 • | | Start of Project V ] Labor Cost Calculation Method | Wage Cost & Labor Usage: Of) - w(t) Lit) Lab Wage ] Lab. Usage | Lab. Inflation | Choose Trend Method _] Time Reference Parameter 1 Linear ~y| | Start of Project j Deterministic Parameter 2 Deterministic ~T] u|0.05 Figure 4.14: Work package material and labor cost window methods 156 Figure 4.15: Work package indirect and discrete cost methods 157 4.3.3 C E Component Aggregation Levels As explained early for the structure of a the CE component (section 4.3.1), a coarse or refined representation of capital expenditure cost can be modeled through the number of work packages used in the representation. However, the method used for cost estimating of a CE work package can contribute to how detailed or coarse the representation is. A work package can represent capital expenditure in a coarse level (e.g. total construction cost) using semi-detailed methods or in a refined level (e.g. cost of a single abutment) using detailed methods. The cost of a group of work packages can be combined/aggregated in what is called a CE area. The CE component supports four aggregation levels each of which can contain several CE areas. A work package can belong to only one CE area per level. Two or more CE areas in a level can be aggregated as well but in a higher level. This reflects the concept of a work breakdown structure (WBS) (Mansuy 1991). Therefore, with reference to Figure 4.7 and a model that only includes the work packages in the fourth level, using the cost of "Abut. #1" and other fourth level work packages, it is possible to aggregate and get cash flows for "abutments" in the third level, "bridge" in the second level, and "construction phase" in the first level. Using the work packages' cash flow function, the cash flow function of CE area A is: faCE(t,X)^fcCE(t\xA.)A. (4.10) /=i in which n is the number of work packages in the CE area A, A is a vector that contain work packages in the area, and t and f are as defined before. Figure 4.16 illustrates how the system implements the levels and shows how areas are defined in the CE component. Figure 4.17 shows how work package Ceidl can belong to any or all of the three aggregation levels as exemplified by Figure 4.7 before. 158 ••Evaluator [EX 001] Work packages Scheduling Quantity Material Labor Equipment Indirect/Sub Costs Discrete Costs Add New IsL*J Qelete Next Back Variables Close • X Expenditure Levels Capital expenditures are made within work packages. A group of work packages is called an area. An area aggregates/combines the cost of its work packages. Two or more areas can be grouped together in another higher level area. Four levels of aggregations are supported. By defauluhe first level is the work package level. A higher level area combines the cost of a lower level area if a work package is assigned to both areas. Level One Areas in level one can represent project phases such as pre-design, design, construction, etc. Each work package may belong to one of these areas. (Area no. 2 MM Level 1 Areas U S Construction Phase Write down the title for the area number attached. Level Two Areas in level two can represent intermediate project units, e.g. substructure, superstructure, etc. Awork package may belong to one of these areas. [Area no. 1 Bridge |<MI Level 2 Areas Hw] Write down the title for the ar