Forecasts are the foundation of all planning—whether implicit or explicit, qualitative or quantitative, objective or subjective, based on simple algorithms or complex statistical models—and are used in decision making. Because forecasts are used in planning and decision making they are of crucial importance in large capital investment projects to determine economic viability and project financeability. In Chapter 3 the record of project financing indicates that project failures are frequently due to forecast failures. To address this problem, we use strategic decision forecasting which is a method developed by Long Range Planning Associates to develop long‐term forecasts for large capital investment project decisions.

Forecasting for infrastructure projects requires strategic‐decision forecasts that are 15 to 30 years out, that are collaborative, and integrative of qualitative and quantitative methods. The focus of strategic forecasting is on modeling future project operations, uncertainty and risk identification and mitigation, and scenario development and planning to generate cost and revenue forecasts. Megatrend, industry trends, economic environmental‐factor analysis, industry and competitive analysis, demand analysis, and forecast ownership and process management are of equal importance to generate the best possible forecasts.

In large project development, nobody knows with certainty what may happen 10 to 15 years hence. If we cannot say with certainty about future events and the outcome of the project value creation, then what is the purpose of an infrastructure project forecast? First and foremost, the process of developing a forecast entails an independent, critical, and objective assessment of the project company's operations. It introduces discipline and reduces uncertainty in decision making and helps identify sources of risks in planning for the project. Furthermore, it helps to identify enabling scenarios for project success and produces a more effective planning and decision making environment. More importantly, forecasts are used as flight simulators or a tool to create the future today.

The sections that follow describe the basics of strategic‐decision forecasting for project finance starting with the definition of a forecast. Section 8.1 clears misconceptions about what is a good forecast. What needs to be forecasted for project finance deals and sources of forecasts outside of the sponsor organization are cited in Section 8.2. The important topics of ways of developing forecast assumptions and conducting sanity checks are treated in Section 8.3, while Section 8.4 introduces the intricacies of the project‐forecasting process.

Section 8.4.1 discusses what is involved in situational analysis and the benefits from it, trailed by the function of environmental and host country assessment in Section 8.4.2. The importance of assessing the effects of megatrends and subtrends are discussed in Section 8.4.3, and Section 8.4.4 addresses the issue of forecast stewardship and accountability. The meaning, purpose, and use of project demand analysis and the selection of a forecasting approach are covered in Section 8.5. Some common forecasting methods and techniques are briefly mentioned in Section 8.6, and the value of sanity checks of assumptions and scenarios selected is presented in Section 8.7. The discussion of forecasting for infrastructure project finance would be incomplete without an examination of the causes and consequences of forecast failures and that is the subject matter of Section 8.8. This chapter ends with Section 8.9, which deals with forecast monitoring and forecast realization planning—the latter is missing from the majority of even professional forecasts.

8.1 WHAT IS A GOOD FORECAST?

A forecast has many definitions and its function depends on its use as well as the skills of forecasters. The following are some views of what is a forecast:

1. An estimate of future performance based on a systematic approach using data, quantitative and qualitative tools, past performance, industry knowledge, and analogs from similar experiences
2. A statement about the future that for decision‐making purposes is a single number that is always wrong
3. A prognosis based on the process of identifying actions and events that affect a project
4. The art of translating qualitative and quantitative information used by models to predict future outcomes
5. The function that brings together data, business experience, and forecasting techniques to estimate what will happen at different future dates

Our definition of a forecast is all of the above, but more importantly, it is an independent, critical, and objective assessment of a project opportunity to guide decision making. It is, also, a tool to create reliable intelligence and the basis of sound decision‐making to manage a project effectively. In the hands of professional strategic‐decision forecasters, it is a way to create the future today and help run project development more effectively, thereby creating a competitive advantage for the project company.

Project revenue forecasts drive the calculation of benefits and along with cost estimates they form the basis of major‐impact investment decisions. For that reason, revenue forecasts must be as reliable as possible. But what are good forecasts? Good forecasts are those that:

1. Incorporate external operating environment, market, industry, and trend evaluations and quantify their impact on project performance
2. Are derived from sound forecasting function ownership and management
3. Have clear, balanced, well‐thought‐out forecasting processes and are easy to explain
4. Create well considered, tested, and reasonable set of assumptions
5. They are based on the 4C principle of strategic decision forecasting: Unimpeded 360‐degree communication, coordination, cooperation, and collaboration
6. Identify and evaluate key project risks through assessment of alternative future state scenarios
7. Are in line with similar project experiences and build confidence in using them to make management decisions
8. Come with instructions and guidance on their use, their shortcomings and limitations, and their implications
9. Have the support of the entire project team and the backing of senior managers responsible for project decision making
10. Are reality based and not influenced by optimistic and wishful thinking or personal agendas

Some revenue forecast users and decision makers judge the quality of forecasts based on measures such as mean absolute percent error. But such measures are of little use in the case of 20‐ to 30‐year forecasts where the structure that generates the forecast changes. When forecasts include management adjustments or are numbers picked from a range of model simulations, subjective judgement enters the picture and then one cannot really talk about a forecast error or quality of the forecast; only about the quality of management judgement applied.

8.2 WHAT TO FORECAST AND SOURCES OF FORECASTS

The first activity in forecasting is to determine what needs to be forecasted followed by identifying what drives the forecasts of demand and what pricing driving factor data are available. Forecasts used in infrastructure development start to get created in the prefeasibility study and undergo updating and enhancement as the process moves through the project development stage and get finalized by the financing stage.

8.3 FORECAST ASSUMPTIONS

The development of assumptions is a crucial activity in large project financings because assumptions underlie and drive project planning, forecasting, budgeting, decision making, and financing. Due to the particulars and complexity of infrastructure projects, the assumption set is project specific and an art based on:

1. Broad project team experience and understanding of the functioning of the industry
2. Review of the literature of business cases and on specific infrastructure project financing issues
3. The collection of expert opinions of project participants and market research information
4. Engineering studies of earlier and competitor projects and experiences
5. Review of assumptions negotiated and used in earlier project contracts
6. Shared knowledge of host country ceding authority personnel based on earlier projects
7. Management and key stakeholder judgment conditioned by the company's strategic objectives and financial priorities
8. Distilled learning from past project information memoranda and post mortem analyses

Once the project definition is done and the company's operating environment is evaluated, the development of the assumptions comes into focus and the key project‐team activities then determine:

1. Impacts of megatrends and industry trends on the project company's consumer or user demand
2. What factors that impact pricing need to be known, what are knowable and unknowable, what are controllable and uncontrollable
3. The evolution of the original pricing of the project company's output over the project's lifetime
4. Current industry capacity and historical growth versus needed capacity
5. Numbers of consumers or users of the project company's output and their consumption patterns
6. Adequacy of project technology and efficiency to meet customer needs and expectations
7. Changes in user‐demand patterns over the project lifetime

Assumption development is an evolutionary process that researches, evaluates, and selects information that is applicable to the project at hand and determines if there are variants of existing information to establish that are more verifiable and reasonable. The question now is: How does one go about creating well thought out and reasonable assumptions? The first and easy response is to rely on engaged advisors and consultants for help. The alternative is to research analogs and information from industry publications, to seek out the views of host government agencies' managers, and to obtain guidance from unilateral and multilateral institution publications and officials.

Another approach is to establish a basis through benchmarking by using publically available information, competitive analysis, and interviews with industry participants. These efforts lead to the development of the initial set of assumptions that are recorded with the name or source of the assumptions and the date associated with them. But, the analyses and evaluations of the feasibility study and the due diligence provide tips and guidance to update the initial assumption set. Also, it is a good idea to consolidate the inputs from different sources into one view of assumptions and subject those assumptions to scrutiny and sanity checks. In every case, when the assumption set is fairly stable, it is crucial to obtain project stakeholder consensus and buy‐in on assumptions underlying the financial model and obtain political support.

8.4 PROJECT FORECASTING PROCESS

The work of Flyvbjerg, Garbuio, and Lovallo (2014) confirms that there are psychological biases that create cognitive delusion in forecasting for large capex infrastructure projects. They suggest that there is a sponsor/developer tendency to overstate completion times compounded by management bias originating with initial estimates and assumptions. They claim that there are misdirected incentives that result in different project outcomes than those preferred by senior management. Furthermore, there is evidence that in some industries inadequate cost forecasts are due to the behavior of some project participants and not due to lack of insight or forecasting techniques (VanWee, 2007).

A good starting point of the discussion of revenue forecasting for an infrastructure project financing is Figure 8.4. It begins with a situational analysis and ends with a final forecast used for funding decisions. In the subsections that follow we address the process components but focus on the situational analysis, the environmental and host country analysis, the effects of megatrends and subtrends on demand, and the issue of forecast stewardship and responsibility.

8.5 PROJECT DEMAND ANALYSIS

Demand analysis is a multidimensional investigation that leads to ability to forecast demand and project revenue. A good demand analysis gives insights on how to shape the future levels of demand and what scenarios, assumptions, and project company business plan are appropriate, and what planning is needed for project development, inventory control, and corresponding costs and revenues. In infrastructure projects, demand analysis is also used to determine the required capacity to meet production of goods or services customers or users are willing to buy. The topics discussed in the forecasting process sections that follow are some key considerations of demand analysis and forecasting for infrastructure projects. Other key components of demand analysis are research, benchmarking and data collection, validation, and evaluation.

The research and benchmarking part of demand analysis is a subset of the broader project development effort and it is primarily concerned with:

1. Reviewing the infrastructure forecasting literature, both scholarly and applied, for innovations in approach and forecasting techniques
2. Obtaining valuable data, project statistics, and other useful information
3. Examining case studies to identify comparable projects for applicability of analogy forecasting
4. Identifying best‐forecasting practices well suited for different types of projects
5. Getting insights from earlier experiences on how to approach a project's forecasting problem and develop the right solutions
6. Determining what analyses may not yield good results and avoiding them
7. Finding out competitors' forecasting processes and methods and learning from them
8. Evaluating the findings, summarizing them, and creating a set of benchmarks to serve as guidelines to forecasting solutions for current projects

The data collection, validation, and evaluation part of demand analysis focuses on:

1. Determining what needs to be known and what does not and what is knowable and what is not
2. Identifying what specific qualitative and historical or cross‐sectional quantitative information is needed
3. Determining the key driving factors of demand and pricing and the direction of influence
4. Separating out the uncontrollable factors and concentrating on controllable factors
5. Branding uncontrollable factors as potential sources of risk and make them known as such to the project team
6. Determining what can serve as proxies for important data that is missing or not available
7. Validating data collected for reasonableness, quality, and consistency with comparable project data using the tools of data analysis
8. Developing cross‐sectional and historical data plots, understanding causes of outliers, calculating growth rates, checking for correlations, and identifying feedback effects
9. Assessing the extent to which the project company's products or services meet consumer and user desires and tastes
10. Evaluating the scale of consumers or users willingness and ability to pay for the project's output at planned pricing levels
11. Investigating if the demand and pricing driving factors are forecasted out and if their quality is acceptable to create reliable project revenue forecasts
12. Preparing data and information collected for use in modeling and forecasting demand along with a report summarizing the findings of data validation and evaluation

Different types of projects require different forecast solutions and have different data needs. However, a sample of commonly used data in‐demand analysis, modeling, and forecasting includes:

1. Host country population data and historical growth rates
2. Educational levels and demographic data and distributions by age, region, etc.
3. Per capita income, disposable income, and income distribution
4. Host government subsidies, and credits for the project company's products or services
5. Industry capacity historical data and usage or purchase data, and pricing data for an existing own or competitor company or, at least, a comparable company
6. Market‐size assessment, market research data, and observed market shifts
7. Historical data for project‐company production inputs and materials and labor rates
8. Information on consumer or user needs and the extent to which they are currently met
9. Information about what factors may detract from future demand for the project company's output
10. Project equipment, technology attributes, and output quality and reliability versus those of competitors or other projects
11. Industry capacity gaps; that is, existing versus needed capacity to produce output to meet the projected demand
12. Evidence of past price change and ROI flexibility on the part of the industry's regulatory agency

The output of the project‐demand analysis is used to decide what forecasting methodology to adopt, the kind of model specification that may be appropriate, and feed qualitative and quantitative information to models and scenarios in order to develop project forecasts. It is also used in the due diligence phase to validate funding needs and determine project financeability, in management forecast presentations, and in agreement negotiations. Just as importantly, it helps to identify the nature and sources of risks that underlie the forecast and quantify their impacts. Also, two important additional functions of demand analysis are to:

1. Develop ways to incorporate both qualitative and quantitative information from the country, industry, and trend analyses into predictive models
2. Help make adjustments to controllable driving factors and initiate actions to shape future project demand to reasonable expectations

8.6 FORECASTING METHODS AND TECHNIQUES

Generating revenue forecasts for project financing involves some kind of accepted qualitative or quantitative models or a combination of forecasts from both types of models. While forecasting models vary from simple deterministic relationships to qualitative and quantitative models to system dynamic models, our research findings suggest that roughly only a quarter of business decisions are based on analysis, evaluations, and substance (Triantis, 2013).This is a major factor in project failures to achieve expected performance and value creation.

Simple deterministic models are functions such as y = a*x and are based on engineering studies, past experience, and financial/accounting relationships and are useful in developing cost component projections. Also, for small projects, deterministic models using good assumptions may be appropriate to forecast costs and revenues, but they are wholly inappropriate and inadequate to use for large, greenfield, complex infrastructure project revenue forecasting.

Quantitative methods rely on sufficient past data to create relationships and models to forecast project revenue. They are statistical models based on internal company, industry association, and the host government authority's historical data and are appropriate to use in brownfield projects with adequate observations. Quantitative forecasting methods include the following types of models: Univariate time series, multivariate time series, various exponential smoothing models, and causal statistical models of high levels of complexity. Quantitative models do not use expert opinions and try to remove subjective judgement from forecasts.

For greenfield projects, the lack of historical data necessitates the use of qualitative methods, which includes:

1. Industry studies and market research
2. Delphi method and its variants
3. Sales force polling
4. Life cycle analogy
5. Panel of experts meeting face to face
6. Scenario building
7. System dynamics
8. Foresight maturity

The qualitative forecasting models of demand and project revenue use expert judgment for medium‐ and long‐term planning purposes. Industry studies provide insights about upper demand limits and market research is conducted to identify consumer or user needs and preferences, their ability and willingness to pay, and yield additional data on the size of the market, potential demand, and project revenue. Sales force and regional company office personnel polling is another method of generating useful information and insights on demand and pricing developments and forecasts. Also, face‐to‐face expert discussions are used to create and validate assumptions and project revenue forecasts.

When a project team expects that demand and revenue of a PPP project is likely to follow the typical growth, maturity, and eventual decline, life‐cycle analogy models are used, often with little consideration of underlying demand factors. Occasionally, these S‐curve kind of models can have their basic structure modified to incorporate external influences and judgment to generate revenue forecasts. S‐curve models are widely used in forecasting demand and revenue for new product or service introductions and do have some applicability in availability projects.

The Delphi method was introduced in the 1950s and is a commonly used method to create assumptions, relationships, and forecasts and to build consensus. It is an interactive and iterative method of forecasting based on the views and opinions of panel experts. Panel participants can provide their individual inputs on data, assumptions, and forecasts and after each iteration, these experts revise their judgments toward the mean responses. The Delphi steps end when a process stopper is reached; usually when the responses converge to a consensus view.

The scenario building and planning method is related to contingency planning and is used when there is uncertainty concerning pricing and demand for a new infrastructure project. The purpose of scenario building is to identify a few possible scenarios for revenue outcomes and to plan and prepare for responding when negative cases materialize. This method of forecasting is used not only in the early stages of project development, but also when project risks are not fully identified and mitigated or when there is uncertainty about external factors impacting project revenue. Because of the importance of scenario building and planning in forecasting infrastructure project demand, this topic is discussed in more detail later in this section.

System dynamics models have been widely used in engineering, social sciences, and the military and recently in forecasting for infrastructure projects (Bala, Arshad, and Noh, 2016). But what are system dynamics models? They are a modeling technique that is capable of building a replica of the complex structure of a project's demand and revenue using assumptions, data, mathematical functions, and diagrams. Their structure allows for introducing quantitative and qualitative new factors, changes, and shocks to the project company operations and its revenue structure. In our judgment, they are well suited for large infrastructure project finance forecasting needs and evaluations.

System dynamics models are well suited for forecasting demand and revenue, especially in PPP projects because of their ability to integrate many different factors, external influences, assumptions, data, and relationships into one place. Their structure requires expertise in handling the interactions between the multitude of demand, pricing, and supply influence factors and the risks associated with these factors. The scenario development and simulation capabilities of system dynamics models shed light on the impact of each factor involved and their unique sensitivity analysis and simulation capabilities to generate project revenue forecast ranges for senior managers to have a higher comfort level in making decisions. They are also useful in determining the costs and benefits associated with a risk factor and whether to insure against it or absorb its impact.

The development of system dynamics involves the creation of conceptual representations of demand and revenue, which includes the following key elements:

1. Identifying the key influencing factors using any of the techniques mentioned earlier to anchor the model to a sound foundation
2. Creating causal loop diagrams that require knowledge of the host country's environment, the project's industry structure, and the project company's customers or users to assign direction of influence and feedbacks
3. Validating causal loop diagrams by expert analysis of the system dynamics model and forecasts generated by expert analysis and knowledge

Another approach to generate forecasts and make decisions in complex project structures that is beginning to get traction is foresight maturity models, which are particularly useful in guiding project teams when applied in the early stages of project development. Foresight maturity models help project teams to define desired, probable, planned, and creatable futures using data, qualitative inputs, and scenario building. In the presence of uncertainty, the foresight maturity technique helps to create strategies and plans to enhance project success rates by generating a baseline and plausible scenarios and limits.

In its simplest form, the basic structure of the foresight method consists of the following elements:

1. Gathering inputs from the situational analysis, expert views, and the project feasibility study
2. Assessing the inputs assembled, which helps in understanding what seems to be taking place and determines the project's future and the revenue stream
3. Interpreting the results of the assessment to determine what factors are indeed influencing project revenue
4. Exploring or prospecting what alternative project revenue outcomes may materialize
5. The actions the project team may need to be taking at this stage for the planned future revenues to be realized
6. Evaluating the outputs and conclusions and then developing a strategy and recommendations of what the project team needs to do and how to do it to ensure that revenue forecasts materialize

Companies with broad experience in strategic decision forecasting, which includes infrastructure revenue forecasting, usually employ more than one of the forecasting techniques and models discussed. However, better project revenue predictions are created by combining forecasts that come out of both qualitative and quantitative methods. The advantage of this practice is that it incorporates elements from different knowledgeable perspectives, which usually produce more easily accepted forecasts.

Every sound forecasting methodology, whether quantitative or qualitative, involves scenario building and the development of project forecasts. Scenarios are schemes, concepts, sketches, outlines, representations or plans of the sequence of events, their timing, and what happens when decisions are made and begin to get implemented. They are models of assumed or expected sequences of decisions, inputs, actions, reactions, and events constructed for the purpose of capturing their effect on a target variable. Namely, they show how a hypothesized chain of events leads to future states in a structured way of seeing beyond the current state, creating descriptions of future states, and describing how they unfold.

Scenarios are also used to explore wild card possibilities and black swans and quantify their impacts. In defining possible futures, scenarios help project‐team members to understand the time‐ordered events and causation from the current stage to the project implementation, and to create strategies and options to deal with uncertainties. They build “flight simulators” to create learning and sound project implementation strategies by clearly articulating the events and processes generating the future states, simulating them, and answering critical questions. Good scenarios are stories of plausible, divergent but deterministic futures, and they capture project‐team biases and different points of view. Nevertheless, they can help to see vividly what drives the business and what is required to achieve the objectives of a project, to stimulate discussion, to question assumptions and the model of business operations, and to increase project‐team effectiveness.

Scenario building is a structured approach to predict future project revenue by assuming a series of alternative possibilities instead of forecasting the future on the basis of extrapolated historical or analog data alone. Scenario planning, also called scenario thinking or scenario analysis, is a corporate planning method employed in making strategic decisions and long‐term plans. It is an adaptation of the methods used by military intelligence and strategic planning that relies on model‐simulation tools and controllable factors to manage the future.

When other methods of forecasting are not appropriate, scenario development and planning is a practical tool used to solve strategic‐decision problems and create future‐state forecasts. It is a method to learn about the future by understanding the impact of the most uncertain and important forces driving the business. Scenario development is based on the belief that strategic‐decision forecasters are not at the mercy of fate; instead, they use this method of envisioning future states to incorporate them into scenario models to be simulated. The common steps involved in part one of scenario development include:

1. Starting with an accurate description of the project attributes and defining the project's environment, the context of the decision to be made, and its major objectives
2. Defining the scope of each scenario and brainstorming on the megatrends and the host country's external environment driving forces surrounding the project
3. Developing major assumptions about timing, causality, and the strength of relationships, gathering information, and evaluating industry and market trends and structural changes
4. Engaging an independent consultant to screen and provide suggestions on the driving forces and events and ensure objectivity and reasonableness of scenario building
5. Determining the extent to which scenario‐driving forces can be predetermined, projected, and fixed and how steady their influence is on the projected project revenue
6. Creating distinct and convincing stories based on the effect of driving forces and critical uncertainties and eliminating competing narratives
7. Weaving hypotheses and plots to the stories to fit the identified events and forces and creating, at the most, three or four plausible scenarios

In part two of scenario building, the forecasting team conducts sanity checks on the assumptions, actions, reactions, events and their timing, and on the processes that generate the future project states and their logic and performs the following activities:

1. Developing decision trees and influence diagrams based on modeling, simulations, and the Delphi technique and performing forecast analysis
2. Focusing on analysis of demand discontinuities of trends, cross‐impact analysis, and analog experiences to identify unexpected driving factors and uncontrollable events
3. Assessing the financial, human resource, competitive, operational, and strategic implications of each scenario and evaluating differences with project expectations
4. Comparing the scenario generated forecasts against the baseline projection and estimating the contribution of each driving factor to the additional value created by the project
5. Creating a system of early warning indicators to monitor each scenario's performance as it unfolds through time and adapting the scenario to fit circumstances that approximate reality
6. Using the decision‐selection matrix to identify, evaluate, and rate the economic value of each future state scenario and the likelihood of achieving them

8.7 FORECAST SANITY CHECKS

Forecasting infrastructure project revenue is an iterative and interactive process based on the considerations mentioned earlier. Forecast sanity checks are part of the process which entails reexamination, reevaluation, and revalidation of the following:

1. Project requirements and attributes, the extent to which it satisfies host government and consumer or user needs, and the political support it enjoys
2. Evaluation of the host country's PESTLED milieu and assessment of megatrend and industry trend impacts on the project
3. Information and data obtained from project participants and reliable external sources, advisors, and consultants
4. Checks of explicit and, sometimes, implicit assumptions about competition, the project company's output demand and pricing, and external future events
5. Input and guidance from industry experts, advisors, and international development institutions on project particulars
6. The conditioning effect of the risk tolerance of the project sponsor or developer and the host government ceding authority
7. The most likely scenario underlying the revenue forecast to validate its probability of occurrence and its applicability
8. Quantitative and qualitative models used to make revenue forecasts and their logical consistency compared with similar project experiences
9. Management or expert advisor adjustments to baseline forecasts to reflect their priorities
10. The strength of the negotiating position of the sponsor or the project originator in the project

Due to the vital importance of the revenue forecast to project valuation and its financeability, it is crucial that each of the above elements that affect the forecast be scrutinized, validated, and verified in an independent, objective, and balanced manner. Starting with project attributes and political support, and going down to the negotiating position of the project participants, the project team must determine whether the forecast foundations are sound and will withstand lender and investor scrutiny and the test of time. The second step in performing forecast sanity checks is comparison of forecast revenue growth rates with project‐team expectations and those of similar projects to determine the reasonableness of the forecast. The third component of validating the forecast is the scrutiny of the project feasibility study and the due diligence phases where all data, assumptions, relationships, influences, analyses, and evaluations are subjected to sanity checks.

A fourth step in project revenue forecast validation is a thorough review by project advisors and consultants and other project participants to determine how forecast consensus was obtained and verification and validation of key models and forecasts. The next step is a forecast sensitivity analysis to see how variations in each key driver impact the revenue forecast due to deviations from their baseline specifications. However, the deciding activity of reality checks on the project revenue forecast is performed by the lending institutions, potential investors, entities providing credit support, and parties assuming business risks. The next question then is: What happens when a forecast fails one or more of the sanity checks? If the forecast validation fails for small infractions and has small impacts, a more conservative forecast may be picked from the range of forecast simulations. If on the other hand major objections are raised with respect to key assumptions, forecast adjustments, and the like, a revisiting of the issue and reexamination of risks should take place so that the remaining risks are managed effectively in order to satisfy all project participants. Appropriate adjustments are made to the forecast if warranted by evidence and convincing arguments.

8.8 CAUSES AND CONSEQUENCES OF FORECAST FAILURES

The causes of infrastructure project finance failures are several, but erroneous revenue forecasts are among the key culprits. Project forecast failures are instances where a forecast misses the mark by cumulative amounts in the range of 20–30% over the project's lifecycle. At the foundation of forecast failures is the wishful thinking produced by sponsor or host government project objectives that have superficial strategic, portfolio, and operational fit assessment, and minimal sanity checks. The prevalence of the groupthink problem related to optimistic views of a project's prospects is compounded by the poor or nonexistent forecast ownership and management function, which causes projections to be driven only by positive sentiments. A lack of qualified, strategic‐decision forecasting skills and expertise, and late and restricted forecast organization participation in project evaluations are other causes of forecast failures.

Conflicting project participant views of future revenues sometimes lead to consensus views that are adopted so that the project can move forward. Inadequate preparations and planning for revenue forecasting, along with a lack of good data and expert advice, cause inappropriate assumptions to be created that, in turn, lead to erroneous forecasts. In some cases, an incomplete project environmental factor assessment and inadequate checking of facts and data lead to inadequacy and the failure of sanity checks, which allow the wrong model development and forecasting methods to be selected. Also, the lack of standardized infrastructure project revenue forecast processes and templates cause unnecessary iterations, often due to the inability to identify and quantify the driving factors and predict their future behavior.

Market research and other qualitative method data are often given undue weight but go unscrutinized and result in erroneous forecast adjustments. Inappropriate baseline scenario selection, no real sensitivity analysis, and ad‐hoc forecast model simulations lead to the selection of forecasts on the optimistic side of the simulation range of values. In organizations with limited strategic decision forecasting capabilities, forecasts are taken as gospel. They are misused and not applied as intended to guide the overall value created by the project. In changing environments and industry structure, static forecast validity cannot be judged, but the absence of forecast realization planning is a major fault of most project revenue forecasts. The issue of project forecast realization is discussed in the next section.

Project revenue forecast failures are judged by forecast errors and are manifested in the project cash flow values coming in short of expectations. However, there are a number of forecast failure impacts to assess before looking at cash flow and beyond measuring actual versus projected value creation. Poor forecasts not only lead to wrong project selection but, also, to missed opportunities to invest in other projects. Regardless of the cause, project forecast revenue failures lead to miscalculations of costs and benefits, but the impact is much more severe if poor demand analysis has failed to identify and quantify correctly the risks that were thought easy to mitigate.

Another costly time, human resource, and money impact is the expansion of sizable resource allocation to a project's development phase, which should have been rejected in the first place, but because of erroneous forecasts it was not. Addressing the miscalculation of project risk failures caused by poor forecast impacts requires additional time, project team effort, and cost expenditures. There are increased debt financing and insurance costs, the risk of the project license being suspended, and termination in the case of large forecast failures. Poor project performance, repayment delays, and less than expected returns to investors leave many project participants very unhappy with the consequence of the snowballing effects on the sponsor company's reputation and future profitability.

In addition, project forecast failures cause funding sources to question the integrity of the project and require additional credit support and enhancements. Forecast failures eventually lead to rating agencies downgrading the project rating, which leads to additional financing costs. Also, a meager, actual project performance versus forecast expectations affects the credibility of the project team and the project originator, as well as that of the host government agencies that approved the project. Lastly, the project forecast failures lead to damaged reputation of the sponsor or developer companies among project financing circles, and this is a lasting and costly effect to remediate.

8.9 FORECAST MONITORING AND REALIZATION PLANNING

Forecast monitoring and reporting is done on a regular basis and usually involves a variance analysis of forecast versus actual revenues. Variance analysis reports on forecast errors and factors believed to be causing those errors. However, focusing on and rushing to judge forecast performance is incorrect without examining factors, such as:

1. Whether the initial project attributes are still factual in the forecast period
2. All project participants made timely required human, financial, and in‐kind contributions
3. The quality of project implementation that was delivered met the required specifications
4. The validity of the project company's business plan, and its performance objectives and targets
5. The quality of the project company's operations management team

A proper forecast monitoring method first involves the creation of a forecast‐monitoring dashboard that is usually a part of the overall project‐monitoring system. The purpose of the forecast dashboard is to track in a systematic and effective manner the performance of the project‐company operations and forecasts to get insights on how to correct problems as they appear. Then comes identification of the factors to include on the dashboard to monitor and how to do it, but it is necessary that only key factors be included in the variance analysis.

A number of questions need to be answered before blaming individual project finance team participants for forecast errors, such as:

• Were project forecasts done simply to satisfy baseless decisions despite objections to adopted forecasts?
• Were forecasts created by project participants or were they outsourced?
• Did forecasts have the benefit of expert reviews, inputs, and recommendations of independent consultants?
• Did the forecasting models identify all key drivers and did these drivers behave in the forecast period according to expectations?
• What changes in the project's host country and its operating environment take place that were not accounted for in the forecast scenario?
• Can forecast variances be attributed to overoptimism permeating all decisions?
• Is there a strategy to address forecast failures and what corrective actions were taken?

The discussion now leads us to the subject of forecast‐realization management or adoptive‐response planning, which stems from the belief that with appropriate structuring, evaluation, and implementation a project's future can be shaped well before actual operations. The forecast realization planning model consists of the following elements that are reviewed at each variance analysis session:

1. Environmental assessments which include industry, market, and PESTLED analysis
2. Independent verification and validation of processes, assumptions, models, and scenarios used to generate forecasts
3. Measuring project performance, which requires creation of a forecast dashboard and an early warning system, monitoring and understanding forecast variances, and determining how to prepare an updated risk management and contingency plan
4. Review and evaluation of the current state of the project company's operations and its initial business plan
5. Assessment of forecast error implications, which includes identification and quantification of new risks in the horizon and determining what preparation is needed to prepare a response plan
6. Preparation of a tactical response plan that involves assigning resources to coordinate the project company's response actions, make the7P adjustments defined below when necessary, and changing the project‐company management team and providing support when warranted
7. Development of a strategic response plan that encompasses assessment of options to initiate, evaluation of restructuring and reorganizing the project company, undertaking new complementary support projects, revisiting corporate strategy and project objectives, and making adjustments where and when required

Development of the forecast realization plan begins with determining what needs to be done to determine whether underperformance is due to forecast or actual data failures. This requires the right mindset, an appropriate and correct approach, employing tried and tested processes, and objectively assessing the results of the plans created. The process of creating the forecast realization plan requires a clear understanding of causes and the need for action and analytics to be performed in order to identify and evaluate the implications of the results. Following that activity, project uncertainty and forecast risks are reevaluated and adaptive remedial planning takes place to come up with a change plan. Changes may include initiatives such as sponsor, developer, or host government agency actions; industry regulator intervention, or key project participant or third‐party additional credit support.