Risk in the front end of Megaprojects

Risk in the Front End of Megaprojects The RFE Working Group Report Editor: Ivana Burcar Dunović Authors: Rafaela Alfalla-Luque, Ivana Burcar Dunović, Aldo Gebbia, Ana I. Irimia-Diéguez, Miljan Mikić, Marisa J.G. Pedro, Álvaro Sánchez- Cazorla, Joao de Abreu E Silva, Konrad Spang, 2 COST – European Cooperation in Science and Technology is an intergovernmental framework aimed at facilitating the collaboration and networking of scientists and researchers at European level. It was established in 1971 by 19 member countries and currently includes 35 member countries across Europe, and Israel as a cooperating state. COST funds pan-European, bottom-up networks of scientists and researchers across all science and technology fields. These networks, called ‗COST Actions‘, promote international coordination of nationally-funded research. By fostering the networking of researchers at an international level, COST enables break-through scientific developments leading to new concepts and products, thereby contributing to strengthening Europe‘s research and innovation capacities. COST‘s mission focuses in particular on: + Building capacity by connecting high quality scientific communities throughout Europe and worldwide; + Providing networking opportunities for early career investigators; + Increasing the impact of research on policy makers, regulatory bodies and national decision makers as well as the private sector. Through its inclusiveness, COST supports the integration of research communities, leverages national research investments and addresses issues of global relevance. Every year thousands of European scientists benefit from being involved in COST Actions, allowing the pooling of national research funding to achieve common goals. As a precursor of advanced multidisciplinary research, COST anticipates and complements the activities of EU Framework Programmes, constituting a ―bridge‖ towards the scientific communities of emerging countries. In particular, COST Actions are also open to participation by non-European scientists coming from neighbour countries (for example Albania, Algeria, Armenia, Azerbaijan, Belarus, Egypt, Georgia, Jordan, Lebanon, Libya, Moldova, Montenegro, Morocco, the Palestinian Authority, Russia, Syria, Tunisia and Ukraine) and from a number of international partner countries. COST‘s budget for networking activities has traditionally been provided by successive EU RTD Framework Programmes. COST is currently executed by the European Science Foundation (ESF) through the COST Office on a mandate by the European Commission, and the framework is governed by a Committee of Senior Officials (CSO) representing all its 35 member countries. More information about COST is available at www.cost.eu © University of Leeds Published by the University of Leeds, April 2015 ISBN 978-0-9576805-4-8 While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission. 3 Content Content 3 Background And Aims Of The Working Group 4 What The Working Group Did (Research Methodology) 6 Results Of The Working Group 12 Literature Review 12 Case Study Analysis Results 37 Reflections on EU Funded Major Projects 48 Conclusions and Recommendations 56 Papers And Other Outputs Produced By The Working Group 57 References 59 4 Background and Aim of the Working Group The Evolution of the Working Group The ‗Risk in the Front End‘ (RFE) Working Group idea resulted from a joint interest that arose both from the work of the Transportation and Cross Sectorial Working Groups in the MEGAPROJECT COST Action. Both working groups looked at issues related to risks associated not only with budget and schedule overflows but also related to the socio-economic impacts of projects. Specifically, most transportation and urban renewal megaprojects are marketed and sponsored based on their positive socioeconomic impacts. Thus, there are also risks associated with the projects not achieving the forecasted impacts. Both the Transport Working Group and the Cross-Sectoral Working Group expressed concerns about how and when these sort of risks should be dealt with. The concept of risk and concerns about when should the different risks should be considered and managed, is fundamental to the focus of the RFE Working Group. In addition, the way that risk was managed in practice in megaprojects and the possible gaps that exist between practice and academic research were formed part of the aims of the RFE Working Group. Finally, the issues related to risk evaluation in the front end and in project evaluation in general, lsuch as Cost Benefit Analysis (CBA), were strong motivations to constitute this RFE Working Group The RFE Working Group was initiated in a Management Committee Meeting in Dubrovnik on 30th September 2013 with the following aims: i) To review existing literature concerning risks in megaprojects ii) To identify the main issues in the common experience of cases in the MEGAPROJECT portfolio relating to risks at the front end. iii) To clarify the differencebetween risk identification at the front end of the megaprojects and the risk at the front end of the projects as a whole iv) To demonstrate the possible ways of dealing with risk in the evaluation of megaprojects in the front end The launch of the RFE Working Group was followed working meetings in Brno (14th February 2014.), Burgas (06th-07th July 2014.), Liverpool (11th-12th July 2014.), Kassel (17th November 2015.) and Zagreb (6th-7th February 2015). 5 The RFE Working Group was led by Ivana Burcar Dunovic. The membership of the RFE Working Group was as follows: + Milen Baltov – Bulgaria + Rafaela Alfalla-Luque – Spain + Miljan Mikic – Serbia + Joao Abreu E Silva – Portugal + Marisa Pedro – Portugal + Alex Stamov – Bulgaria + Konrad Spang – Germany + Vit Hromadka – Czech Republic + Zhen Chen – UK Further support was also received from: + Jana Korytarova – Czech Republic + Ana Irimia Dieguez – Spain + Alvaro Cazorla – Spain + Alison Hood – UK + Aldo Gebbia – Italy + Nigel Smith – UK + Marco Kuemmerle – Germany + Camelia Michaela Kovacs – European Commission 6 What the Working Group Did (research methodology) Introduction Risk management processes must be carried out continuously during the entire life cycle of a project, especially when making decisions. The level of uncertainty is known to be the highest at the beginning of a project, with a tendency to decrease towards the end of a project as knowledge about the project increases together with committed values, which is why management is most effective in the early phases of the project life cycle (Table 1). Table 1 Project risks and life cycle (Wideman, 1992) Table 2 The relationship between project risk and the level of knowledge about a project (Solomon, 2006) 7 Based on knowledge on change of risks and knowledge about the project over time, it is possible to establish their inter-relationship.(Table 2) Those statements apply only if the project is done and managed in the right way. Megaprojects (also known as ‗large-size infrastructure projects‘) have been defined as significant activities characterized by a multi-organization framework, producing relevant social impacts (Aaltonen, 2011; OMEGA, 2012). They are characterized by extreme complexity (in technical and human terms), high risk and uncertainty (concerning demand and cost estimations) and poor performance (Boateng et al., 2012; Priemus, 2010). The most common definition within experts and researchers is the concept of a large-scale investment project, typically costing more than EUR 0.5 billion (COST Action TU1003; Fiori and Kovaka, 2005) with colossal use of resources (money, human, equipment…) (Kardes et al., 2013), which frequently leads to cost overruns (Boateng et al., 2012; Flyvbjerg et al., 2003; Han et al., 2009). As a result, megaprojects should be defined in their construction management context: activities, resources, budgets and deadlines. Fiori and Kovaka (2005) present other four key-characteristics of megaprojects: extreme complexity, increased risk, lofty ideals, and high visibility. When these characteristics are skillfully managed, results in a successful megaproject, but the wrong combination can lead to a disaster. These features provide a guideline for planning and construction of megaprojects. They are not mutually exclusive, or they aren't hierarchically ranked. Rather, change in one characteristic drives changes on others. As result, they must be examined individually and in relation to each other. Both, Flyvberg (2005) and Miller and Hobbs (2005) emphasize that one of the main aspect is long, complex and expensive front-end which impact project management performance more than the management of the engineering, procurement, and construction phase. The development of the project during the front-end phase was shown to be a time-dependent, non-linear, and iterative process, during which the project was formulated, tested, challenged, and reformulated through a series of episodes, during which unforeseen risks and issues emerge in successive episodes and must be managed. (Miller and Lessard, 2000) On the other hand, statistical evidence from Flyvbyerg research (2005) shows that unplanned events were not taken into account and thus the budget and other reserve funds were insufficient. Since megaprojects involve substantial financial investments and commitment, and starting a wide set of socioeconomic effects, the decisions that are taken at early project development stage are of great importance. These decisions emerge under an environment of uncertainty. It is necessary have a flexible management since uncertainty is associated with vagueness, ambiguity and contradictions. This is linked to the lack of clarity due to missing data, incomplete and inaccurate detail related to the structure, the working and framing assumptions, known and unknown sources of bias, limited control of relevant project players and ignorance on how much effort is worth to clarify the situation. These projects are planned and 8 constructed with a professional culture of closed systems thinking which has a tendency to minimize risks and uncertainties. Understanding those developments is critical to evaluate what is a successful megaproject. IMEC research project ((Miller and Lessard, 2000); (Miller and Hobbs, 2005)) showed that shaping large engineering projects will be greatly influenced by compromises, external influence (pressure), long duration, great political pressure, complex regulatory framework. The anchoring of projects to institutional frameworks was one of the most critical aspects of these projects and for that were much better able to withstand and survive the impacts of emergent uncertainty. The projects were exposed to many different types of risk, several of which are not typically taken into account by traditional project management methods. Among the most important sources of risk are: Governments reneging on commitments, slowly materializing or insufficient markets, and social and political challenges to legitimacy. The level of uncertainty was extremely high, partly because of the large number of potential sources of risk, the projects' visibility, and their innovativeness. The length of time required for project development and anchoring increased the projects' exposure to emergent risk. Each project encountered an average of four unforeseen and potentially catastrophic events during their long life cycles. The time for a transport megaproject to mature is usually long; it can take several decades from the first idea or draft plans to the beginning of operation. As a result, it‘s common that changes might occur, in the economic, political, legal and regulatory and technological contexts (Bruzelius et al., 2002), during this extended period. These changes are also related with the transportation system (e.g. new operators or transport modes that might start to operate, new transport related technologies, changes in pricing structures). Changes of project configuration and scope and, consequently, changes on cost, lifecycle and traffic forecasts might also occur. Their complexity has raised the attention of several stakeholders and usually triggers disagreements emerging under an environment of uncertainty (Curtis et al., 2012). Numerous transport megaprojects have not been successfully delivered on-time and on-budget resulting in a negative image of the transportation sector. Several experts have noted that the costs are usually significantly underestimated and traffic estimates are systematically and significantly overestimated in such projects (Van Marrewijk et al., 2008; Williams et al., 1995). Actually, in such projects is common to observe cost overruns of a magnitude around 50-100% in fixed prices and, sometimes, higher than 100% (Bruzelius et al., 2002). Also, traffic forecasts are usually off by 20-70%. Indeed, in transport megaprojects, rail based projects tend to be more overestimated than road projects and the project viability is often very optimistic. Ninety percent of the transport infrastructure projects, from 20 countries, studied by Flyvbjerg and colleagues (Flyvbjerg, 2007; Flyvbjerg et al., 2003; Flyvbjerg et al., 2005) produced very disappointing outcomes indicating a large element of uncertainty and risk, with rail projects being systematically more overestimated than road projects. Furthermore, Flyvbjerg et al. (2003) demonstrated that rail projects have the largest cost increase (actual costs are on average 45% higher than estimated costs), followed by bridges and tunnels (34%), and finally roads (20%). The same authors also showed that are common the 9 existence of construction cost in transport projects and exist athwart different project types, different continents and different historical periods. The authors concluded that decision-makers should be worried about long implementation phases and sluggish planning and implementation of large transport infrastructure projects. In addition, in rail projects, this over-estimate cost seems to be more obvious in developing countries than in North America and in Europe. However, cost underestimation seems been explained by strategic misrepresentation (i.e. lying). Bruzelius, Flyvbjerg and colleagues (Bruzelius et al., 2002) attribute the biases to the project proponents, which aim to presenting it in more optimistic way in order to guarantee its approval. Politicians may also promote the projects‘ approval by being overly optimistic even if more precise forecasts can be estimated (Bruzelius et al., 2002). At the end, Flyvbjerg and colleagues (Flyvbjerg, 2007; Flyvbjerg et al., 2003; Flyvbjerg et al., 2005) refer that the disappointing results of megaprojects come from mainly:  Megaprojects always involve the intersection of risk, democracy and power. Political and regulatory authorities normally define parameters and goals to suit their own ends, but frequently forget the transparency concept in regulatory structures that affect the viability of a project.  Megaprojects cannot be planned and executed in a predictable world where cause-effect are evident. Political interference and changing in governments make imperfect knowledge environments on executing megaproject developments.  Megaprojects undertaken in emerging economies, for example, face poor prospects for more transparent stakeholder involvement, efficient and effective public sector risk analysis, and government bodies in emerging economies often lack the institutional capacity and depth to perform proficient risk assessments. Research approach and methodology Case study methodology The first stage of the research is current literature analysis to establish research gaps. The literature review was done in three phases. First phase was bibliometric analysis of risk management in megaprojects, followed by bibliometic analysis aiming at identifying emerging topics and research gaps in risk management in projects and megaprojects. Third phase of literature review aimed at clarify the meaning of term ―risk in front end‖ analyzing risk and uncertainty. The second stage of research is case analysis. This research aims to analyze the data gathered in the questionnaire proposed by the RFE Working Group (WG) in order to understand within the case studies selected, what the common experience is of risk in megaprojects; how the risk has been managed in different megaproject case studies and develop some theoretical framework. In order to achieve this, the 16 Table 3 Literature reviews in RM in megaprojects Paper Objective Methodology Zhang, 2011 To locate the position of past studies of project risk between the two schools of risk analysis (risk as an objective fact and risk as a subjective construction) and to help the understanding of their basic assumptions, viewpoints, and tendencies. To review the papers published in the International Journal of Project Management and in the Project Management Journal in the period 1999-2009 that include the word "risk" in their title, abstract and/or keywords. Rezakhani, 2012 To develop an extensive literature survey in risk modeling and analysis methods with main focus on fuzzy risk assessment. To analyse papers published in the topic. There is no specification about methodology or database and journals analysed. Irimia- Diéguez et al., 2014 To establish the state of the art in risk management in megaprojects, to systematize the risks studied in the literature, and also to identify potential areas of further research. A systematic literature review of major databases (WoK, Scopus and ABI/Inform) from 2000 to March 2013 using the keyword "risk" in combination with "megaproject" or "mega project" or "big project" or "complex project" or "large project". Lehtiranta, 2014 To address how well the body of knowledge on multi-organizational RM corresponds to a state-of-art understanding on project RM and to identify which gaps need to be addressed in future research. To analyse the papers published in four top journals representing general project management (International Journal of Project Management and Project Management Journal), construction project management (Journal of Construction Engineering and Management), and software Project management (IEEE Transactions on Software Engineering) within the thirteen-year period from 2000 to 2012. Taroun, 2014 To review the literature of construction project risk modelling and assessment To analyse papers published in academic journals specialised in construction management, project management, risk analysis, and management science. The databases utilized were: Science Direct, Web of Science, ABI-Inform (Proquest), Business Source Premier (EBSCO), Emerald, and Sage Management & Organization Studies. The keywords used were "project risk", "construction risk", "risk analysis", "risk assessment", "risk modeling" and "RM". Zhang, 2011 To locate the position of past studies of project risk between the two schools of risk analysis (risk as an objective fact and risk as a subjective construction) and to help the understanding of their basic assumptions, viewpoints, and tendencies. To review the papers published in the International Journal of Project Management and in the Project Management Journal in the period 1999-2009 that include the word "risk" in their title, abstract and/or keywords. Rezakhani, 2012 To develop an extensive literature survey in risk modeling and analysis methods with main focus on fuzzy risk assessment. To analyse papers published in the topic. There is no specification about methodology or database and journals analysed. Irimia- Diéguez et al., 2014 To establish the state of the art in risk management in megaprojects, to systematize the risks studied in the literature, and also to identify potential areas of further research. A systematic literature review of major databases (WoK, Scopus and ABI/Inform) from 2000 to March 2013 using the keyword "risk" in combination with "megaproject" or "mega project" or "big project" or "complex project" or "large project". 17  The main risk studied (42.31%) is the construction risk, mainly in the form of cost and project schedule overruns. The risk related with clients and society (14.10%), due to the return on investment as well as the impact of the megaproject on society, is also a major factor. Risks from force majeure, and those related with workers are seldom studied. A large proportion of the studies analyses risks under a general focus; namely, there are 36 references (that represent 46.15% of total papers) which fail to detail any specific risk whilst 42 papers identify specific risks. In addition, if the type of risk studies in megaprojects are analysed in depth per sector, the lack of research in various sectors can be observed (e.g. aeronautic or refinery). Rail and road are the sectors where more types of risk are analysed whilst aeronautical papers focus on the construction risk.  Related to the methodology employed by the decision makers to deal with risk, a total of 27 references (32.5%) focus on this topic. The literature review has found no evidence of the existence of a single set of model RM in megaprojects. Instead, there are a variety of proposals supported by different tools and/or variables; that is, all the references propose their own model or tool to handle risk. The papers are usually focused on one phase of the RM process (risk identification, risk assessment or risk mitigation) although 9 references propose models to handle risk throughout the whole RM process. In summary, from the bibliometric analysis point of view, it can be pointed out that the number of papers in this field has been increasing in recent years; consistent with the importance that this topic has assumed. Numerous theoretical/conceptual papers (almost 30% of the total under study) have been identified. The most common type of empirical studies is that of case studies, whereby, in general, just one single case is presented. There is a lack of empirical studies, that provide an in-depth analysis of the various aspects of this process during the different life cycle phases, and longitudinal studies analysing the evolution over time of the RM models and their results. More research, in general, and more detailed case Paper Objective Methodology Lehtiranta, 2014 To address how well the body of knowledge on multi-organizational RM corresponds to a state-of-art understanding on project RM and to identify which gaps need to be addressed in future research. To analyse the papers published in four top journals representing general project management (International Journal of Project Management and Project Management Journal), construction project management (Journal of Construction Engineering and Management), and software Project management (IEEE Transactions on Software Engineering) within the thirteen-year period from 2000 to 2012. Taroun, 2014 To review the literature of construction project risk modelling and assessment To analyse papers published in academic journals specialised in construction management, project management, risk analysis, and management science. The databases utilized were: Science Direct, Web of Science, ABI-Inform (Proquest), Business Source Premier (EBSCO), Emerald, and Sage Management & Organization Studies. The keywords used were "project risk", "construction risk", "risk analysis", "risk assessment", "risk modeling" and "RM". 18 studies and survey studies, in particular, are required in order to improve megaproject management and risk process management. As Hartono (2013) states, there is a limited utilization of project risk models, tools, and methods which were developed on the basis of normative decision theories. Nevertheless, the number of models, methods and tools settled by researches and practitioners is growing, although it is not followed at a similar rate by the adoption of them by project practitioners. Another important issue is to identify those topics of interest in the field of RM in megaprojects. A word count analysis of the whole text of the papers selected by Irimia-Diéguez et al. (2014) shows that the word ―risk‖ appears 6,506 times, whilst other frequent words are ―complex‖ (and similar topics as complexity) which appears 25.9%; ―sustainability‖ (including environmental, ecologic,…) with 20.8%, ―governance‖ with 15.9%, and finally, ―stakeholder‖ and similar words with 11.1%; all the percentages are expressed with respect to the number of times that risk appears (100%). Nevertheless, if the word count analysis is performed just within the keywords, the word ―complex‖ (and its variations) appears in 20.7% of the papers, whilst the percentage decreases to 4.9% in the case of ―sustainability‖ and similar words, to 2.4% for ―stakeholder‖ and similar words, and lastly only 1.22% of the papers include the word ―governance‖ as keyword. In consequence, the mentioned topics can be found in, although they have not included as keywords of the paper. Perhaps, the origin of this situation could be that some of these topics may be considered as emerging lines of research on risk in megaprojects and there is a scarce literature focused on them. Therefore, a more detailed bibliometric and content analysis is developed in the next sections. Evolution of the emerging topics In order to analyse the evolution of the identified topics (stakeholders, governance, complexity, and sustainability), three major academic journals in Project Management have been selected: International Journal of Project Management (IJPM), Project Management Journal (PMJ), and International Journal of Managing Projects in Business (IJMPB). Our choice of the journals is guided by previous papers which performed literature reviews in project management topics (Littau et al., 2010; Zhang, 2011; Lehtiranta, 2014). A first search of the analysed topics (sustainab* or green or environ* or ecologic* or governance or complexit* or stakeholder) and the words ―risk‖ and ―megaproject‖ or ―complex project‖ or ―large project‖ in the title, keywords and abstract of the articles published until February 2015 in the three journals selected papers was performed. Due to the scarce number of papers found, a second search with the word ―project‖, instead of ―megaproject‖ or ―complex project‖ or ―large project‖, was then executed. We identified 101 papers. Subsequently, the abstracts and contents were assessed for their suitability to the present study. After selecting those papers which really analysed the topics considered, 30 papers were finally found. The number of papers identified is mainly concentrated in the last four years (2010- 2014), although some of the topics are treated in the journals sporadically since 1995. As shown in Figure 1, the International Journal of Project Management (IJPM) in the journal that contains more papers about these topics (53.3% of the total of 30 identified references). Instead, we have 36.7% of Project of Management 19 Journal (PMJ) and 10% of International Journal of Managing Projects in Business (IJMPB). The three journals study all these topics, with the exception of sustainability. As shown in Figure 2, IJPM contains a higher proportion of papers on stakeholders. As can be observed in Figure 3, stakeholder was the most frequently topics investigated, although, since 2009, governance and complexity has also picked as a topic of interest. Finally, sustainability is the least studied issue (only one item identified). 20 The pioneer in publishing a paper on these topics, namely governance and complexity, was PMJ, it was Scott (1996). PMJ also contains the first paper identified on the topic stakeholders (Piney, 2003). The reason should be that PMJ was the first journal, starting in1970 and was indexed in 1985 in ABI, quite earlier than IJPM which began in 1983 and was indexed in 1992 in ABI, whilst IJMPB appeared and was indexed in 2008 in ProQuest,. Most of the studies focus on addressing these topics for projects in general. There are a small number of papers about complex projects, large projects or megaprojects (28.21%). As shown in Figure 4, only 11 papers specified megaproject in their analysis and most of them are focused on governance. Most of the papers analyzing the topics complexity and stakeholders refer to projects in general. In spite of the topics selected being closely related, in the abstract analysis, only 9 of the 30 references (30%) considered two topics simultaneously (Table 2). The remaining papers (21 references, 70%) focus on just one topic. Among the topics most frequently studied together are governance and stakeholders (4 of 30 papers), and governance and complexity (3 papers). Table 3 shows the number of papers focused on the different topics. 21 Table 2. Number of topics analysed by type of project No. topics analysed Megaproject, Complex project or Large project Projects (in general) Total No. % 1 3 18 21 70% 2 4 5 9 30% Total 7 23 30 100% Table 3. Topics of the papers Sustainability Governance Complexity Stakeholders Total Sustainability 0 1 1 Governance 1 3 3 4 11 Complexity 0 3 6 1 10 Stakeholders 0 4 1 12 17 Total 1 11 10 17 39 Since the number of papers focused on these topics is very low, an additional search (similar to the previous search but deleting the word ―risk‖) was performed. The purpose was to check if these topics are being analysed in the literature of Project Management although the analysis is not related with risk. The results show 777 papers found (that means an increase of the 769% in relation with the 101 papers found with the word risk included). It implies that these topics have a greater consolidation in the field of project in general than in the area of risk management in particular. An analysis of the emerging topics 1.1 Stakeholders A good implementation of project risk management is believed to be one of the leading factors attributable to project success and hence companies' long-term success (Hartono et al., 2014). In the relationship between risk management and project success, key elements are stakeholder perception of risk and success and stakeholder behaviour in the risk management process (de Bakker et al., 2010). According to Millar (2007), the most important uncertainty management issues are usually related to objectives and relationships between the key stakeholders, particularly the internal stakeholders and especially within the ‗project owner‘. In the same way as for any project, megaprojects represent a significant challenge to the stakeholders (Fiori & Kovaka, 2005). These stakeholders could be defined from different perspectives (Littau et al., 2010). From our approach, stakeholders are individuals, groups or institutions with an interest in the project, who can affect the outcome (Boddy & Paton, 2004), whereby it is understood that stakeholders develop an active role in functions such as the assumption and management of a certain type of risk. A more comprehensive definition of stakeholders can be found in Winch (2010), who describes them as those actors which will incur a direct benefit, or loss, as a result of the project. This author provides two classifications of stakeholders in order to aid the analysis and their management, namely internal and external stakeholders. Internal stakeholders are in legal contract with the client and can be categorised to those clustered around the client on the demand side and 22 those on the supply side. External stakeholders are comprised of private and public actors. The private actors are from the local residents, landowners, environmentalists, and archaeologists, whereas the public actors are from regulatory agencies, and local and national government. The internal stakeholders will largely be in support of the project and external stakeholders may be in favour, against, or indifferent (Takim, 2009). Particular challenges are presented by megaprojects since they usually are ‗multiowned‘ projects, where more than one organization shares ultimate control over fundamental aspects of the megaproject. In these contexts, key issues are governance arrangements and the allocation of risks and rewards so as to create and maintain incentives for cooperative behaviour as the project progresses (Millar, 2007). In many cases, risk sharing makes sense because most project risks commonly concern project participants (Tang et al., 2007). Since stakeholders should be concerned about and may be affected by the risks of the project, the consensus is that risks should be allocated to the party that is in the best position to manage them (e.g., International Organization for Standardization ISO, 2009). According to the World Bank the allocation of risks should be made according to two criteria: (a) the risk should be allocated to the stakeholder best able to manage the risk outcome and (b) the risk should be borne by the stakeholder best able to handle the risk at the lowest cost (Global Development Finance. World Bank, Washington, D.C. (2007), cited in Vassallo et al., 2012). Therefore, those risks that can be assumed by the megaproject may be managed by their own, and those which are not affordable should be transferred by contract to several stakeholders in order to best control the risk management. This crucial issue needs to be analysed carefully once the various risks have been identified in the megaproject. Identifying and allocating risks to those stakeholders best able to manage them is crucial in megaproject management (Beidleman et al., 1990). Risk allocation refers to a primary measure of assignment between the projects‘ direct participants‖ (Bing et al., 2005). Risks are usually allocated between two parties (public and private sector) or three parties (public sector, private sector and end-users). Risk allocation should be based on a balance of parties' interests and should ―distribute liability associated with risk events to proportionally distribute the possible prospect loss or gain of project‖ (Khazaeni et al., 2012). Risk allocation has a direct and important bearing on the financial cost of the project. When the risks allocated to the private sector are very high, the financial cost of the project becomes significant and can threaten the ultimate financial feasibility of the project. Conversely, too much risk retained by the public sector might not encourage the private sector to perform properly and might end up proving to be too costly for the public budget in the future (Vassallo et al., 2012). Therefore, an adequate risk allocation profile requires striking the right balance between risks retained by the public sector and risks transferred to the private sector. Ke et al. (2010) state that risk allocation is highly related to the unique social, economic, legal situation of the country. An example of matrix of risks allocated to stakeholders is shown in Table 4, where the risk that could be assumed by each stakeholder is identified. This table has been elaborated from the references analyzed (e.g. Beidleman et al., 1990), other examples of risk allocation (Bing et al., 2005; Ke et al., 2010), and our previous experience. Those other risk allocation matrices only differ between public and private sector. Our focus is broader and considers the six main stakeholders to be 23 found in megaprojects, excluding end-users. This matrix may not be very meaningful due to the different features of each stakeholder. Table 4. Risk allocation to stakeholders Stakeholders Public sector Management company Construction company Shareholders Financial institution Consultants Type of risk 1. Design     2. Legal/political     3. Contractual    4. Construction  5. Operation   6. Labour   7. Clients/users/society   8. Financial/economic     9. Force Majeure    Hartono et al. (2014) identify significant gaps of risk-related concepts between project stakeholders' perspectives and the rational assumptions of the normative decision theories (e.g., risk is widely viewed by practitioners from the negative domain while the rational theory would suggest a more neutral perspective of risk). Another research gap is pointed out by Loosemore (2010) who discusses how multimedia technology can be used to effectively engage stakeholders in the management of risk in projects and in business. The author draw attention to explore the pedagogical advantages of multimedia in helping organisations develop a risk management culture as future research needs. 1.2 Governance Governance is a growing area of interest for management and organizational researchers and theorists, although there is scarce literature about its role and impact on projects and risk management of projects; and even less is known about the systemic impact of project governance, that is, how governance and project systems have a reciprocal impact (Pitsis et al., 2014). In this field, there is still work to be done in specifying how projects are distinct from other forms of organizing conceptually and philosophically (Morris, 2013). Pitsis et al. (2014) point out that existing theoretical perspectives offer many opportunities further to explain the tensions, challenges and opportunities inherent in project governance, making it a ripe and vibrant field of research, theory and practice. Nevertheless, according to Sanderson (2012), governance has become an increasingly popular theme in the project management literature. This fact reflects a widening of focus away from the purely technical and operational tasks that need to be fulfilled to deliver project outcomes, to encompass a much greater interest in how interactions between the multiple actors responsible for undertaking those tasks are organized and coordinated (see, Atkinson et al., 2006; Clegg et al., 2002; Flyvbjerg et al., 2003; Miller and Lessard, 2000; Pitsis et al., 2003; Pryke and Smyth, 2006; van Marrewijk et al., 2008; Winch, 2001, 2009). Governance mechanisms refer to processes of institutional, market or network organization through legal, normative, discursive or political processes (Bevir, 2013). In its broadest definition good governance can be thought of as how individuals, groups, organizations, societies, nation states are held accountable not only for outcomes but also ethical behaviors (Clegg et al., 2011). Much of the recent 24 governance literature focuses on the governance of organizational relationships (networks, collaboration and partnerships for example) pertaining to projects (Clegg et al., 2002). Risk governance is defined by IRGC (2005) as the identification, assessment, management and communication of risks in a broad context. It includes the totality of actors, rules, conventions, processes and mechanisms concerned with how relevant risk information is collected, analysed and communicated, and how and by whom management decisions are taken and implemented. IRGC`s approach states that risk governance is context-specific. A range of factors (the nature of the risk itself, how different governments assess and manage risks, and a society‘s level of acceptance or aversion to risk, among others) means that there can be no single risk governance process. The framework is therefore deliberately intended to be used flexibly. In this line, Cui & Olsson (2009) suggest that the more uncertainty the project has, the less likely it is to anticipate project flexibility that is to be applied in the future. Nevertheless, flexibility cannot substitute the need for governance and change control to deal with unplanned change requests (Gil & Tether, 2011). In accordance with Pitsis et al. (2014), the design, execution, management and close out of contemporary complex projects occur in contexts of unparalleled uncertainty, making it difficult to govern these projects in line with intended and anticipated strategic objectives and imperatives. Projects must be managed dealing with challenges posed by ―uncertainty in ecological, social and economic sustainability; ambiguity arising from advances in the technological means of communications; shifting geopolitical power relations that bring both challenges and opportunities, and at the same time the governance of these projects must be able to attract and retain people who are not only skilled and knowledgeable in all technical matters relating to projects but also able to adapt to turbulence in the operating environment‖. The literature tends to treat governance issues as being static (Miller & Hobbs, 2005), but megaprojects can rarely be treated within the context of a single organization, since their project development processes and environments are dynamic. The governance of large complex projects requires governance regimes that are themselves dynamic—that can change themselves to adapt to the emerging context. The governance regimes must adapt to the specific project and context, deal with emergent complexity, and change as the project development process unfolds. Learning to manage project governance regimes is difficult for organizations that are not involved in great numbers of large complex projects. The framework based on the progressive shaping of the project through the project development life cycle is designed to help overcome this dilemma (Miller & Hobbs, 2005). Based on a re-examination of a study of 60 large capital projects and interviews to practitioners, these authors proposed the following lessons: (a) The management of front end phase is critical and shows significantly more impact on project performance than the management of the engineering, procurement, and construction phase. (b) The anchoring of projects to institutional frameworks is one of the most critical aspects of project governance. (c) In large complex projects the governance relationship is very rarely a binary relationship between one project sponsor and one governing body. The interactions among the different groups of stakeholders involved in project 25 development, approval, and delivery can best be represented as a dynamic social network. (d) Megaprojects are exposed to many different types of risk, several of which are not typically taken into account by traditional project management methods. Governments reneging on commitments, slowly materializing or insufficient markets, and social and political challenges to legitimacy are among the most important sources of risk. The level of uncertainty is extremely high, partly because of the large number of potential sources of risk, the projects‘ visibility, and their innovativeness. The length of time required for project development and anchoring increased the projects‘ exposure to emergent risk. (e) The development of the project during the front-end phase is time-dependent, non-linear, and iterative process, during which the project is formulated, tested, challenged, and reformulated through a series of episodes. Unforeseen risks and issues emerge in successive episodes and must be managed. (f) Projects and their contexts vary so greatly that no one strategy is appropriate to all cases. However, a strong correlation was found between the variety of strategies deployed, or strategic depth, and project performance. The need for strategic flexibility is in-line with the episodic nature of the project development process and the uncertainty as to the nature of the challenges and risks in future episodes. The need for strategic flexibility creates a strategic planning paradox in that being well prepared for the requirements of early episodes may result in inadequate preparation for later episodes and the associated emergent risks. Project development requires a rich and varied pool of strategic resources and the flexibility to adapt to emergent situations. (g) The capabilities of the project sponsor/developer had an important impact on the way the project unfolded and ultimately on performance. Strong sponsors showed: integrative business perspective, ability to evaluate complex systems from multiple perspectives, relational and coalition-building competencies, political and negotiating skills, resources necessary to support long development processes, possibility of diversifying risk through a portfolio of projects or other activities, and will to abandon bad projects. (h) High performing projects are subjected to intensive scrutiny. The project sponsor plays an important role in ensuring that projects are scrutinized. The involvement of other stakeholders with diverse interests and perspectives in a governance structure that encouraged scrutiny also contributed to the development and delivery of feasible projects. Sanderson (2012) critically discusses different explanations for the performance problems exhibited by many megaprojects, and examines the proposed governance solutions. This author concludes that governance in megaprojects should make forms of organization designed ex ante, and should ignore spontaneous microprocesses of organizing emerging ex post. Identification of this gap adds support to calls by projects-as-practice researchers for a broadening of research to encompass the actuality of projects. A new line of enquiry within this broad projectsas-practice agenda is suggested. This author agrees with the general argument that research on projects ‗should spend less time looking at strategic planning and more time researching everyday organizational life‘ (Pitsis et al., 2003), and supports similar calls for a greater focus on the ‗actuality of project based working and management‘ (Cicmil et al., 2006) to stimulate a more reflexive and developmental approach to understanding project performance. 32 Previously demonstrated definitions of risk can be viewed from two aspects - epistemological, which includes people‘s epistemological assumptions and project and ontological, which considers origin of uncertainty and risks. Epistemological aspect Where the level of uncertainty is concerned, in a project there are three types of data: known knowns, known unknowns and unknown unknowns. Ward and Chapman (Chapman and Ward, 2002) define known unknowns and unknown unknowns as explicit or implicit assumptions or conditions which, if not assigned a value, can have uncertain, significant consequences. Chapman and Ward assume that uncertainty is the lack of certainty, visualized in the form of variability and ambiguity that cannot be completely separated. (Chapman and Ward, 2002) Table 4 Illustration of uncertainty according to Chapman and Ward Table 5 Illustration of uncertainty according to Vose Such an interpretation of uncertainty can be compared with Vose's concept that there are "two components of our inability to accurately predict what the future holds: these are variability and uncertainty," and he relies on a quote by Sir David Cox: "Variability is a phenomenon in the physical world that is measured, analysed and explained, if necessary. In contrast to uncertainty that is an aspect of knowledge." (Vose, 2005) The characteristics of variability are that it is an effect of chance (luck) and that it is a function of the system, such that it cannot be reduced through study and measurement but only by changing the physical system. Authors also call it "aleatory uncertainty" (Williams, 2002) or stochastic variability. Overall uncertainty comprises variability and uncertainty which is a reflection of the lack of the estimator's lack of knowledge in respect of the parameters and characteristics of the physical system, and is also called "epistemic uncertainty" (Williams, 2002) or "fundamental uncertainty" (Vose, 2005). In a critical review of risk, uncertainty and governance in mega projects Sanderson (2012) emphasises the importance of considering people‘s ‗fundamental epistemological assumptions about decision-maker cognition and about decisionmaker views on the nature of the future (risky or uncertain)‘. He identifies the following categories with regard to the assumptions about decision-maker views on the nature of the future: 33 Of key importance is Sanderson‘s emphasis on the consideration of people‘s epistemological assumptions when considering the management of risk and uncertainty. The perspective is that there is a continuum between the two concepts depending on the degree of knowledge and calculation (Sanderson, 2012). Risk involves situations where the probability of outcomes is ‗known‘, while uncertainty is the opposite (i.e. when the probability of outcomes is not known). All authors separate variability and uncertainty due to the need for different modelling, which is the reason that in this study these terms are maintained separately, not only because of differences in modelling but because of different methods of management as well as a different share in management during the life cycle, which will be elaborated below. The previous analysis Burcar Dunović (2012) defined three areas of management of uncertainty, of which only one relates to traditional risk management. It starts with planning with variabilities, which is supplemented by the management of risks and uncertainties. At different phases of project the participation of specific elements of the management of uncertainty will not be equal, and depends on the degree of development of the project. Knowing the types of uncertainty, is not only important for the selection of a method of assessment and risk modelling, it is very important for the selection of a strategy and plan for dealing with risk. Based on this explicated look at managing uncertainty it can be concluded that there are even two areas of managing uncertainty associated with the level of knowledge about a project (Table 6). Risk/Uncertainty Category Decision-Makers View Risk Category 1: a priori probability The decision-makers view is that they are able to assign objective probabilities to a known range of future events on the basis of mathematically ‗known chances‘, e.g. the probability of throwing a six when a perfect die is 1 in 6. Risk Category 2: statistical probability The decision-make‘s view is that they are able to assign objective probabilities to a known range of future events on the basis of empirical/ statistical data about such events in the past e.g. the probability of being involved in a fire. Uncertainty Category 1: subjective probability The decision-makers view is that they face a known range of possible future events, but lack the data necessary to assign objective probabilities to each. Instead they use expectations grounded in historical practice to estimate the subjective probability of future events. Uncertainty Category 2: socialised The decision-makers view is that they face a situation in which the nature and range of future events is unknown, not simply hard to understand because of a lack of relevant data. The future is inherently unknowable, because it is socially constructed and may bear little or no relation to the past or the present. 34 The share of specific areas of uncertainty management change over time, and Table 7 illustrates the share of a specific area in relation to the overall management of uncertainty. Uncertainty decreases during the project as knowledge about the project grows, and their behaviour can be illustrated with the graph below (Table 8). Table 6 Illustration of the field of management of uncertainty Thus, uncertainty and risks in a project change depending on the degree of development of the project, i.e. the definition of the project, with which the level of knowledge about the project increases. In order to achieve the project objectives the project management process must ensure that the level of knowledge about the project and the definition of the project continually increase in order reduce uncertainty and risks. However, the uncertainty of a project is not only related to the level of knowledge about the project and the project definition. External changes as well as changes within the project will also affect the uncertainty of the project. Table 7 Relative share of areas of management in the management of uncertainty over time Management of variabilities Mananagement of risk Management of unknowns Management of uncertainty Known knowns Known unknowns Unknown unknowns Variability Ambiguity Management of lack of knowledge Management of risks Planning with variability 35 Table 8 Areas of uncertainty management over time Ontological aspect Analyzing the origin of risk impact there is different models of risk. Burcar Dunović (2009) analysed different models with purpose to find appropriate model for developing risk register for risk and knowledge management in construction projects. During the research on risk management in construction projects Burcar (2005a) and Burcar Dunović and Radujković (2009) understood that the main problem to project risk from issue and they cannot differ what is risk and what is source. The literature review showed that risks and models which are representing risks are understood very differently. There are no clear descriptions what risk components and characteristics are. Source and cause are treated as synonyms, as well as consequence and impact. When we talk about risks in construction projects, source and cause can be treated as synonyms which why we casted out cause from our concept. Consequence and impact, on the other hand, have absolutely different meaning and function in this risk model – consequence is component and impact is its characteristic. Model developed for designing risk register system Burcar Dunović et. Al. (2013) treats event/risk as the central part of the model (Error! Reference source not found.). It represents an uncertain occurrence, action or event the occurrence of which causes a consequence. 36 Table 9 Risk components and characteristics ource urcar unovi et al. (2013). The source of risk is defined as a condition of area of human activity or a natural or other phenomenon, from which the risk is generated, or from which the possibility that the risk might occur is generated. It exists either in the project or outside of the project, and does not have a variable characteristic. Its important characteristic is the owner i.e. the participant because of which a source has come to exist. The consequence is the condition, occurrence or event that has occurred precisely because of the occurrence of an event/risk, and which affects the success of the project, i.e. the project objectives, through the risk impact. Significant characteristics of the impact are the nature, size and place that define in which way the risk will impact the project and its objectives, i.e. to what extent and on what part of the project, WBS or activity. The next component is the driver that can be either an event, occurrence or a change of condition, which leads to activation of the risk mechanism, i.e. which initiates transformation of risk into actual event. Through its actual occurrence, the risk stops being a risk and becomes actual event or problem to be dealt with. The risk can be described as a mechanism in a latent state that needs driver to be activated. In addition to the source, the driver and the consequence must also have the owner i.e. the stakeholder to which the two events or occurrences are related. The time and probability of occurrence are the characteristics that are most often related to risk in general, but are related in this model to the event, which is regarded as the central component of risk. Both components could also be related to the occurrence of the driver being an event that activates the risk mechanism, while the time of occurrence could also be related to the consequence. Correlation is very important part of risk model which characterize the interaction between two risks. In developing this model our intention was to link all risk characteristics with risk components, but in case with correlation we made the exception defining it on risk level instead of component level. Correlation between two risks can include any component of each risk and they together with their interactions are used to describe and define correlation type and mechanism. 37 Vast of research papers dealing with risks result with risk factors that affecting decision-making and which have been classified into two groups: internal and external factors. For example, internal includes ‗age‘, ‗education level‘, ‗work experience‘, and ‗personality type‘, whereas external factors embrace ‗company culture‘, ‗country culture‘, ‗regulation‘, ‗socio-economic condition‘ and ‗geography‘ (Luu et al., 2008). Others associate internal factors with aspects related to the project as relationship among stakeholders as well as the technical and operational sides (Bing et al., 2005; Haidar and Ellis, 2010; Kendra, and Taplin, 2004). On the other hand, the external factors are not related to the project and have an indirect impact on the project success. These are associated with political, economic, natural, social and environmental changes (Bing et al., 2005; Haidar and Ellis, 2010). All differences are due to the megaproject subject (transport, business, industrial…). On, the other hand, Burcar Dunović at all. (2013) suggest classification of risk sources in risk breakdown structure, classifying natural, economic, political, legal, social in external and management, human factor, technical, procurement and contractual as internal. Chen at all. (2009 and 2011) classified risks in developmental phases of megaprojects in STEEP categories social, technological, economic, environmental and political used by Boateng and all (2012) for System Dynamics (SD) modeling for social and environmental (SE) risk management during megaprojects development. Irimia-Dieguez at all (2014) classified risks in megaprojects in 9 mail categories: design, legal and/or political, contractual, construction, operation and maintenance, labour, clients/users/society, financial and/or economic and force majeure. Looking from aspect of origin of risks, Atkinson et al. (2006), define three key-areas of uncertainty: i) uncertainty linked to estimations (of cost, schedule and demand); ii) uncertainty associated with project parties (related to infrastructure management) and iii) uncertainty regarding to project lifecycle stages (related with the failure of thoroughly carrying out the design and planning stages). As a result the project proceeds with insufficient detail and specifications. These uncertainties have negative effects on costs, schedule and performance of the projects because they lead to additional design and planning tasks during project implementation. In fact, the project construction phase is frequently the most critical part as far as risks are concerned. During this stage, substantial funds are spent without any project‘s cash flow and technical and economic viability could be compromised. Besides, the construction could be negatively affected by environmental changes as natural disasters and/or geological unforeseen conditions due to geology. Chapman and Ward (2003a) present 5 areas where uncertainty is prevalent: i) Variability associated with estimates. a. Lack of clear specification what is required b. Novelty or lack of experience of this particular activity (project) 38 c. Complexity in terms of number of influencing factors and interdependencies d. Limited analysis of the processes involved in the activity (project) e. Possible occurrence of particular events and conditions that might affect the activity (project) ii) Uncertainty about the basis of estimates. a. Who produced estimates b. What form they are in c. Why, how and when they are produced d. What resources and experience are based on e. How they take into account ―known unknowns‖, ―unknown unknowns‖, ―bias‖ iii) Uncertainty about design and logistics. a. Uncertainty about project deliverable b. Uncertainty about process of delivery iv) Uncertainty about objectives and priorities. a. Uncertainty about objectives b. Uncertainty about priority of objectives c. Trade-offs/compromises d. v) Uncertainty about the fundamental relationships between project parties a. Specification of responsibilities b. Perceptions of roles and responsibilities c. Communication across interfaces d. Capability of the parties e. Formal contractual conditions and their effects f. Informal understanding on top of, or instead of, formal contracts g. Mechanisms for coordination and control The complexity of a project, along with the level of uncertainty, is the characteristic most commonly associated with large-scale projects (Burcar Dunović et al., 2014). In the literature one can find various types of relationships between complexity and risk, i.e. uncertainty, which can be categorized in the three groups:  Uncertainty and complexity are independent characteristics (Clegg et al., 2002),(van Marrewijk et al., 2008),  Complexity is compounded by uncertainty (Williams, 2002)  Project complexity is the source of uncertainty in project. (Danilovic and Browning, 2007),(Secretariat, 2007) Looking at this relationship from epistemological aspect it is best described with Williams, while taking into account ontology of risks the third statement is the most appropriate. Therefore, complexity can be considered both, compounded by uncertainty and as source of uncertainty. 39 Case study analysis results Case Studies Profiles Megaproject 1: Offshore Platform EPCI in the Mediterranean Sea Due to sensitivity of project data the project cannot be named. Megaproject 2: Sava Zagreb, The River, Croatia Project ―Zagreb on the Sava River‖ is a multifunctional program of regulation, protection and utilization of river Sava from Slovenian border to the town of Sisak. The experts have been dealing with the regulation of Sava River for some decades now. It all started with a big flood in Zagreb in 1964. Several multifunctional concepts have been made throughout the years, including power plants, trying to resolve flood protection problems. Probably one of the reasons why none of the concepts ever started construction is that there wasn‘t a management model which would gather, coordinate and manage all Program stakeholders. In 2012 a new company was established as a subsidiary of HEP Group (Croatian Energy Utility Company) to manage the project. The project manager created a model that puts together stakeholders at one side and experts council, as verification body, on the other side, connecting them through the operational team. Zagreb on the Sava River is a long-term sustainable solution to the problems related to the Sava River and the hinterland area of the Slovenian border to Sisak, and the project benefits are environmental, social and economic. Potentials and benefits of the project will be realized in water management, transportation, energy and space and will enable long-term sustainable development of the area. From the WBIF Program the Project management company received a grant funds in the form of Feasibility, Environmental and Social Impact Study. It will evaluate three different solutions/concepts and will select the most acceptable one. The project is still under development, with current estimated budget of € 1.4 billion and a project estimate at completion of 15 years. Megaproject 3: Danube Bridge 2 – Combined rail/road bridge between Bulgaria and Romania at Vidin-Calafat Megaproject 4: FERTAGUS Train Concession – Railway Axis North/South Lisbon, Portugal This railway service is known as FERTAGUS Train - Railway Axis North/South (Lisbon) linking the North and South railways of the Lisbon Metropolitan Area through the bridge ―Ponte 25 de Abril‖. The network started into operation in 1999 with a total of 14 stations and 54 km length. The whole project comprises buses and light rail services, several viaducts, renovation of the existing line and construction of elevated crossings. FERTAGUS rolling stock is characterized as a suburban train service, providing the operation between Lisbon and the South municipalities of the Tagus River. 40 The Project has been sponsored by the Ministry of Finance, Ministry of Public Works (through IMTT, Land Transport Regulator) and local/regional municipalities (the initial investments and infrastructure construction were 100% Government budget). The contract was awarded in competitive tendering procedure to FERTAGUS (the private company) operate the service in 30 years of concession period. The contract includes the supply of equipment and rolling stock, operation and maintenance. In 2005 there was a contract renegotiation and was done a public-private partnership between the operator and the State. The contract was again renegotiated in 2010. Although the contract value is around EUR 132.1 million, the total costs of the project are around EUR 632.1 millions, including EUR 250 million for bridge upgrades; EUR 255 million for infrastructure construction of South line Coina/Setubal (FERTAGUS/State, 2005; Gomes, 2012; Tribunal de Contas, 2002; REFER EP, 2004). Thus, this project is a transport infrastructure megaproject with a huge impact for the Portuguese railway transportation sector since it provides a new (first) link between the North and the South of the country. The interview partner was the chief project manager of FERTAGUS. Megaproject 5: Industrial Zones, Bulgaria Megaproject 6: Higway A1, Croatia The A1 motorway (Croatian: Autocesta A1) is the longest motorway in Croatia, spanning 478.9 kilometers. As the route traverses rugged mountainous and coastal terrain the route, completed as of 2014, required 376 bridges, viaducts, tunnels and other similar structures, including the two longest tunnels in Croatia and two bridges comprising spans of 200 meters (660 ft) or more. A motorway connecting Zagreb and Split was designed in the early 1970s, and a public loan was started in order to collect sufficient funds for its construction. However, due to political upheavals in Croatia and Yugoslavia, construction of the motorway was cancelled in 1971. After Croatian independence and conclusion of the Croatian War of Independence, efforts to build the motorway were renewed and construction started in 2000. The Zagreb–Split section of the route was completed by 2005, while the first sections between Split and Dubrovnik opened in 2007 and 2008. Construction costs incurred over 3 billion euro. Data are gathered for this research interviewing project managers responsible for each part of highway which were mainly involved in construction phase which were constructed on the route Zagreb-Split. Megaproject 7: City Tunnel Leipzig, Germany A very big (mega project) double tube railway tunnel project of the Deutsche Bahn (German Federal Railway) in the City of Leipzig of about 1 billion €. All Regional trains can now cross the town directly, which reduces the running time up to 40 min. The tunnel is in operation since December 2013. It is 1,5 km long (9 m diameter) and includes 4 underground stations. He was ―built‖ partially by 2 tunnel shield driving machines (TBM) and partially by mining technique. The ground situation a cross the city and under the houses was a big challenge. Another challenge was the mixture of national, regional, European and private financing of the project. The overall costs were 960 Mio €, with the following financial partners: State Saxonia (52%), European Union (23%), Federal Republic of Germany (22%), Deutsche Bahn (2%) and City of Leipzig (1%). Planning took about 15 years and construction 41 10 years. The interview partner was the project director of the project management company, who was doing the project owners job for the Deutsche Bahn – in this case it was the DEGES, a German PM company, which is normally doing road projects. Megaproject 8: VDE 8 (HSR Berlin–Nurnberg), Germany A very large railway project - may be the largest in Germany at all. It is the 515 km long High-speed Line (HSL) from Berlin to Munich of about 10 billion €. The projects consists partially of renewing an existing railway line (288km) for 200 km/h and a completely new line (227km) for 300 km/h. When all parts will be in operation, the travel time from Berlin to Munich will be reduced from 7 to 4 hours. As the HSL crosses in large parts German secondary mountain it contains a high percentages of tunnels (about 60km) and large bridges (35km). Planning time is about 23 years and construction time is about 20 years (design and realization was overlapping for individual sections from the beginning). Interview partner was the project director of the project management company – a subcompany from the Deutsche Bahn - who is doing the project owners job for the Deutsche Bahn. Megaproject 9: Sofia Tech Park, Bulgaria Quantitative Analysis Megaproject quantitative template consists of an Excel spreadsheet with data about megaproject cases. A total of 35 questions (149 items) have been codified and analyzed. Each case contains information about respondent data, project data and risk management. Respondent Data This section is regarding to the respondents information: gender, how many years of experience they have in project management, how many megaprojects they have already participated, if they have any education/qualifications in Risk Management and so on and so forth. The results are presented with following graphs and tables. 48 Parties involved and allocation of responsibilities MP [#] Frequency [%] RM facilitated and involving stakeholders beyond the core project team 2 22% RM facilitated throughout the core project team 5 56% Specific functions with limited roles 0 0% Scattered, ad hoc and left to individuals 2 22% Other 0 0% Total 9 100% How was risk management documented? Frequency [%] Documentation reported and updated through whole lifecycle 56% Limited documentation 33% No documentation 11% Was there formal process in place to identify risk owners and to empower them for effective risk treatment? MP [#] Frequency [%] Yes 4 44% NO 5 56% Total 9 100% Thirty per cent of the megaprojects analysed have limited documentation on risk management process; 56% have, in fact, risk management documented in reports and updated through whole lifecycle of the project. 49 The financial risk assessment is presented in megaproject feasibility studies in all the answers (5 megaprojects). Sixty seven per cent used a mix of methodologies; sensitivity analysis (60%) and scenario analysis (60%) are the most used. Three cases analysed present socio-economic assessment in their feasibility study. All used sensitivity analysis as methodology; 2 megaprojects also used scenario analysis as a methodology. Who was involved in risk assessment process? MP [#] Frequ ency [%] Project manager 5 56% Project team members 7 78% Consultants 2 22% Owner 2 22% Risk management specialist (internal) 2 22% Risk management specialist (external) 1 11% Possible contractors 2 22% Legal/regulatory stakeholders 2 22% Politicians 1 11% Other (Commercial/ Proposal Team) 1 11% Total 9 How data for risk assessment process were gathered? MP [#] Frequen cy [%] Historical data 5 56% Check list 4 44% Survey 3 33% Interviews with stakeholders 3 33% Past experience 6 67% Brainstorming 4 44% Workshops 3 33% Other 1 11% Total 9 100% If yes, which methodology MP [#] Frequency [%] Sensitivity Analysis 3 60% Scenario Analysis 3 60% Multi-criteria Analysis 1 20% Cost-Benefit Analysis 1 20% Total 5 100% If yes, which methodology MP [#] Frequency [%] Sensitivity Analysis 3 100% Scenario Analysis 2 67% Multi-criteria Analysis 1 33% Cost-Benefit Analysis 1 33% Total 3 100% 50 Formal projects documents with risk management results Megaproject [#] Frequency [%] Initial planning documents 1 11% General risk management 4 44% Status / Monitoring and technical 5 56% Financial 4 44% Audits 2 22% Quality 3 33% Others 2 22% At the end, usually, project managers and project team members are involved in risk management process. In fewer case studies analysed, contractors, consultants and owner are also involved. The data for risk assessment was gathered basically through the past experience and historical data. The megaprojects analysed have also a high heterogeneity of the formal project documents with risk management results: from initial planning documents to financial documents, audits and quality service. Qualitative Analysis This section aims to present a qualitative analysis of megaprojects case studies. Some questions in the survey (14, 15, 16, 29, 30 and 35) need to be analysed in a qualitative perspective, taking into account others variables. The analysis of the question 14 will generate a list of success criteria. We will obtain a list of constraints (question 15) and critical factors (question 16), and their impact on success criteria. Figure 1 explains the framework of the analysis. Basic components by which the risk mechanism is presented are the event/risk, source, driver, consequence, and impact. Each of these components has some characteristics by which it is described. Therefore, the risk is fully described through components and their characteristics. Reflection on EU funded major projects Risk Assessment on Major Projects In this part of the report we present brief insights from guidelines about the procedure and requirements for risk assessment as a part of proposals preparation for major projects intended to be co-financed by EU funds. Since the three EU funding programming periods mentioned in this part of the report (2000-2006, 2007-2013, 2014-2020) have different context and different set of rules in relation to major projects, the intention here was not to provide exhaustive crosscheck conclusions, but only to offer short reminding on what-and-why was and is necessary when assessing risk in front-end of major projects, as per EU legislation. Major project definition Definition of a major project that may be co-financed by European Regional Development Fund (ERDF) and Cohesion Fund (CF) is the following (Common 51 provisions regulation (CPR) No 1303/2013 (European Parliament, 2013), Article 100): An operation comprising a series of works, activities or services intended in itself to accomplish an indivisible task of a precise economic or technical nature which has clearly identified goals and for which the total eligible cost exceeds EUR 50 000 000 and in the case of operations contributing to the thematic objective under point (7) of the first paragraph of Article 9 (promoting sustainable transport and removing bottlenecks in key network infrastructures) where the total eligible cost exceeds EUR 75 000 000 (the 'major project'). Financial instruments shall not be considered to be major projects. Comparing to the previous programming period (2007 – 2013), the financial criteria for defining major project have been enlarged from 25 and 50 EUR million to 50 and 75 EUR million respectively. Cost and Time (non)Efficiency of Previous Major Projects Delivery EU Cohesion Policy Synthesis report published in March 2010 presented ex post evaluation of Cohesion policy programmes 2000-2006 financed by the European Regional Development Fund (ERDF) in Objective 1 and 2 regions. In chapter 3 (POLICY OUTCOMES AND EFFECTS IN MAIN POLICY AREAS) it is stated: The investigation, which covered the construction of roads, railways, urban transport systems, infrastructure for water supply and treatment and energy supply, revealed, first, that major projects were prone to have a high incidence of delays in completion and of cost overruns. Some threequarters of projects were subject to some delay, with the average time amounting to around 26% of the initially estimated period of completion. Just over half of the projects investigated exceeded their budget, with an average cost overrun of 21%. These figures, however, are not unusual and are very much in line with the incidence and scale of delays and cost overruns of major projects funded from national sources across the EU or in other parts of the world. For the most part, they do not vary systematically between countries or different types of project. For the projects examined, however, some differences were evident: • costs overruns were less of a problem in Germany then elsewhere and more of one in Poland (averaging 50% for the four projects investigated); • delays were particularly lengthy in Portugal (averaging 85% of the initial estimates for the 5 projects covered); • urban transport projects tended to be subject to larger cost overruns and delays than other types, average 45-50% in each case. It is difficult to draw conclusions about efficiency from this evidence, since there are a wide range of potential reasons for both cost and time overruns, many of which are outside the control of contractors or contracting authorities. However, they clearly indicate the importance of building in sufficient allowance for contingencies and delays in the planning and budgeting of projects. This tended not to be done adequately for most of the projects investigated. Indeed, in most cases, there was a bias towards optimism which is typical for large-scale infrastructure projects. 52 Major Projects Risk Assessment Procedure In the direction of the assessment of financial and economic risk and uncertainty in preparation of major projects that would be co-financed by Structural Funds and the Cohesion Fund, it can be seen that in all three programming periods covered by this study (2000-2020) of the ERDF and CF funds requirements for major project proposals, a Cost-Benefit Analysis (CBA) for major projects (including risks assessment part) was requested in the respective legislations for the corresponding programming period. For the programming period 2007-2013, Article 40(e) of Reg. 1083/2006 required the Member State (or the managing authority) to provide the European Commission with a CBA for major projects and two main reasons are given why CBA is required for major projects: 1) To assess whether the project is worth co-financing; for that, an economic analysis is required. If the project‘s economic net present value (ENPV) is positive, then the society (region/country) is better off with the project because its benefits exceed its costs. 2) To assess whether the project needs co-financing; for that, financial analysis is required: if the financial net present value (FNPV) of the investment without the contribution of the Funds (FNPV/C) is negative then the project can be co-financed. The Guidance on the Methodology for Carrying Out CBA for the Programming Period 2007-2013 (European Commission, 2006) has recommended two main steps to be undertaken in relation to risk assessment within CBA: 1. Sensitivity analysis, which aims to identify the project‘s critical variables. This is done by letting the project variables (input variables for financial and economic analysis) vary according to a given percentage change and observing the subsequent variations in both financial and economic performance indicators (FNPV / ENPV and Financial Rate of Return (FRR) / Economic Rate of Return (ERR)). Variables should be varied one at a time, while keeping the other parameters constant. 2. Risk analysis; assessing the impact of given percentage changes in a variable on the project‘s performance indicators does not say anything about the probability with which this change may occur. Risk analysis deals with this. By assigning appropriate probability distributions to the critical variables, probability distributions for the financial and economic performance indicators can be estimated. This enables the analyst to provide interesting statistics on the project‘s performance indicators: expected values, standard deviation, coefficient of variation, etc. It should be noted that while it is always possible to do a sensitivity analysis, the same cannot be said for risk analysis. In some cases (e.g. lack of historical data on similar projects) it may prove rather difficult to come up with sensible assumptions on the critical variables‘ probability distributions. In such cases, a qualitative risk assessment should at least be done to support the results of the sensitivity analysis. In the same guidance it is underlined that European Commission‘s ―Guide to Cost- Benefit Analysis of investment projects” should be considered as the main reference which can provide the reader with a thorough treatment of the subject of CBA procedure (including risk assessment part). 53 The Guide to Cost-Benefit Analysis of Investment Projects (European Commission, 2008), which should be looked together with the mentioned Guidance on the Methodology for Carrying Out CBA for the Programming Period 2007-2013 updates and expands the previous edition of the same Guide (2002), which in turn was the follow up of a first brief document (1997) and of a subsequent substantially revised and augmented text (1999). The objective of the Guide reflects a specific requirement for the EC to offer guidance on project appraisals, as embodied in the regulations of the Structural Funds (SF), the Cohesion Fund (CF), and Instrument for Pre-Accession Assistance (IPA). The Guide, as said in the document, should be seen primarily as a contribution to a shared European-wide evaluation culture in the field of project appraisal. It is stated in the Guide that risk, but not uncertainty, is subject to empirical measurement, and can be analyzed and possibly managed. Against this background, the Funds‘ regulations required a risk assessment for major infrastructure and productive investment projects (Article 40 1083/2006 EU Regulations). The recommended steps for assessing the project risk are: 1. Sensitivity analysis; the Guide suggests generally considering as ―critical‖ those input variables for which a 1% variation (positive or negative) gives rise to a corresponding variation of at least 1% in the financial/economic indicator base value. Variables‘ potential variability and it's impact on financial/economic output should be assessed through elasticity analysis in a qualitative and/or quantitative form. 2. Probability distributions for critical variables; process of assigning probability distributions to each of the critical variables, defined in a precise range of values around the best estimate, used as the base case, in order to calculate the expected values of financial and economic performance indicators. 3. Risk analysis; calculation of the probability distribution of the FRR or NPV of the project. For this purpose, the use of the Monte Carlo method is suggested in the Guide. The method consists of the repeated random extraction of a set of values for the critical variables, taken within the respective defined intervals, and then calculating the performance indices for the project (FRR or NPV) resulting from each set of extracted values. By repeating this procedure for a large enough number of extractions (generally more than a few hundred) one can obtain a pre-defined convergence of the calculation as the probability distribution of the FRR or NPV. 4. Assessment of acceptable levels of risk; Generally, a neutral attitude towards risks is recommended in the Guide, because the public sector might be able to pool the risks of a large number of projects. In such cases, the expected value of the ERR could summarize the risk assessment. In some cases, however, the evaluator or the proposer can deviate from neutrality and prefer to risk more or less for the expected rate of return; there must, however, be a clear justification for this choice (for example, a very large project in a small country). 54 5. Risk prevention; a typical source of forecasting mistakes in project appraisal is optimism bias, i.e. the demonstrated systematic tendency for project appraisers to be over-optimistic about the estimation of the key project parameters: investment costs, works duration, operating costs and benefits. To minimize the level of optimism bias, specific adjustments in the form of increased cost estimates and decreased, or delayed, benefit estimates should be made. Such adjustments should not be seen as a substitute for risk assessment, but rather as a more accurate basis on which to develop risk analysis. In the Guide additional topics have been considered:  The switching value – indicates what percentage change in the variables would make the NPV (economic or financial) equal to zero.  Scenario analysis – the specific form of sensitivity analysis. As opposed to sensitivity analysis, scenario analysis studies the combined impact of determined sets of values assumed by the critical variables. Combinations of ‗optimistic‘ and ‗pessimistic‘ values of a group of variables define the optimistic and pessimistic scenarios. Project performance indicators are then calculated for each combination. Major Projects Risk Assessment in the Legislation for the Programming Period 2014-2020 While waiting for the new edition of Guide to CBA of investment projects (which will be built on previous version), it is presented here how CPR No 1303/2013 (European Parliament, 2013) and Commission Implementing Regulation (EU) 2015/207 (European Commission, 2015) define procedure of risk assessment in CBA. In the Article 101 of CPR No 1303/2013, it is, i.a., stated: Before a major project is approved, the managing authority shall ensure that the following information is available: … (e) a cost-benefit analysis, including an economic and a financial analysis, and a risk assessment; (f) an analysis of the environmental impact, taking into account climate change adaptation and mitigation needs, and disaster resilience; … (h) the financing plan showing the total planned financial resources and the planned support from the Funds, the EIB, and all other sources of financing, together with physical and financial indicators for monitoring progress, taking account of the identified risks; Commission Implementing Regulation (EU) 2015/207 states: The cost-benefit analysis, including an economic analysis, a financial analysis and a risk assessment is a prerequisite for the approval of a major project. A methodology for carrying out the cost-benefit analysis should be developed based on recognized best practices and with a view to ensuring consistency, quality and rigour, both in carrying out the analysis and in its assessment by the Commission or independent experts. The cost-benefit analysis of major projects should show that the project is 55 desirable from an economic point of view and that the contribution from the ERDF and the Cohesion Fund is needed for the project to be financially viable. Under heading 2.4. of Commission Implementing Regulation (EU) 2015/207 risk assessment procedure has been defined. As set out in Article 101(1)(e) of Regulation (EU) No 1303/2013, a risk assessment must be included in the CBA. Risk assessment enables the project promoter to better understand the way the estimated impacts are likely to change should some key project variables turn out to be different from those expected. A thorough risk analysis constitutes the basis for a sound risk-management strategy, which in turn feeds back into the project design. Particular attention should be paid to climate change and environmental aspects. According to Commission Implementing Regulation (EU) 2015/207, the risk assessment shall comprise two steps: 1. ensitivity analysis, which determines the ‘critical’ variables or parameters of the model i.e. those whose variations, positive or negative, have the greatest impact on the project's performance indicators, shall take the following aspects into consideration: — the critical variables are the ones whose 1 % variation results in more than 1 % variation of the NPV; — the analysis is carried out by varying one element at a time and determining the effect of that change on the NPV; — the switching values are defined as the percentage change the critical variable should assume to make the NPV equal to zero; — scenario analysis allowing the study of the combined impact of determined sets of critical values and in particular, the combination of optimistic and pessimistic values of a group of variables to build different scenarios, which may hold under certain hypotheses. 2. Qualitative risk analysis including risk prevention and mitigation, which shall include the following elements: — a list of risks to which the project is exposed; — a risk matrix showing for each identified risk: — the possible causes of failure, — the link with the sensitivity analysis, where applicable, — the negative effects generated on the project, — the ranked (e.g. very unlikely, unlikely, about as likely as not, likely, very likely) levels of probability of occurrence and of the severity of impact, — the risk level (i.e. combination of probability and impact); — identification of prevention and mitigation measures, including the entity in charge of preventing and mitigating the main risks, standard procedures, where appropriate and taking into account best practices, where possible, to be applied to reduce risk exposure, where considered necessary; — interpretation of risk matrix including an assessment of the residual risks after the application of prevention and mitigation measures; — In addition the risk assessment may, where appropriate (depending on project size, data availability), and should, where the residual risk exposure is still significant, include the probabilistic risk analysis, which involves the following steps: 1. Probability distributions for critical variables informing about the likelihood of occurring a given percentage change in the critical variables. Computing the probability distribution of critical variables is necessary to carry out a quantitative risk analysis. 56 2. Quantitative risk analysis based on Monte-Carlo simulation, providing probability distributions and statistical indicators for expected result, STD, etc. of project financial and economic performance indicators. What is more detailed in regulation for this programming period comparing to the previous is the list of the main risks per sector to be taken into account in the risk assessment, set out in the implementing regulation. In order to assist the project promoters in preparation of qualitative risk analysis, EU Member States are encouraged (if they consider this appropriate and/or feasible) to develop national guidelines on valuation of certain standard project risks, and list of mitigation and prevention measures across sectors. On Figure 1 the required info about risk assessment as a part of submission of the information on a major project proposed for co-financing by ERDF / Cohesion Fund is presented. It may be seen that detailed info is required about methodology of risk assessment, sensitivity analysis, risk matrix and risk mitigation strategy and measures. 57 Required info about risk assessment within CBA for major project (European Commission, 2015) 64 European Parliament, Council of the European Union, 2006. Regulation (EC) 1083/2006 laying down general provisions on the European Regional Development Fund, the European Social Fund and the Cohesion Fund and repealing Regulation (EC) No 1260/1999. European Parliament, Council of the European Union, 2013. Regulation (EU) No 1303/2013 laying down common provisions on the European Regional Development Fund, the European Social Fund, the Cohesion Fund, the European Agricultural Fund for Rural Development and the European Maritime and Fisheries Fund and laying down general provisions on the European Regional Development Fund, the European Social Fund, the Cohesion Fund and the European Maritime and Fisheries Fund and repealing Council Regulation (EC) No 1083/2006. 65 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of XXXXXXX Title of book First published in hardback 2014 First published in paperback 2014 © XXXXXXXXXXXXX ISBN 01234 5678910 (hardback) 01234 5678910 (paperback) RRP: $XX.XX APM 2006. APM Body of Knowledge, 5th Edition, Association for Project Management (APM). BOATENG, P., CHEN, Z., OGUNLANA, S. & IKEDIASHI, D. 2012. A system dynamics approach to risks description in megaprojects development. Organization, technology and management in construction: an international journal, 4. BURCAR DUNOVIC, I. 2012. Risk management in large infrastructure projects. PhD, University of Zagreb. BURCAR DUNOVIĆ, I., RADUJKOVIĆ, M. & VUKOMANOVIĆ, M. 2013. Developement and implementation of Risk register system for construction projects. Građevinar Journal of Croatian ociety of Civil Engineers, 65, 23- 35. BURCAR DUNOVIĆ, I., RADUJKOVIĆ, M. & ŠKREB, K. A. 2014. Towards a New Model of Complexity – The Case of Large Infrastructure Projects. Procedia - Social and Behavioral Sciences, 119, 730-738. BURCAR, I. 2005a. Risk register structure for construction projects. M.Sc., University of Zagreb, Faculty of Civil Engineering. BURCAR, I. 2005b. Risk register structure for construction projects. M.Sc., University of Zagreb. BURCAR, I. & RADUJKOVIĆ, M. Risk model for construction projects Risk register system. In: CERIĆ, A. & RADUJKOVIĆ, M., eds. Construction facing worldwide chalenges - Joint 2008 CIB W065/W055 Commissions Symposium Proceedings, 2009 Dubrovnik. Faculty of Civil Engineering, University of Zagreb, 1050-1059. CHAPMAN, C. & WARD, S. 2002. Managing Project Risk and Uncertainty: A constructively Simple Approach to Decision Making, West Sussex, John Wiley & Sons. CHAPMAN, C. B., WARD, S. & WARD, S. C. 2003. Project Risk Management: Processes, Techniques and Insights, Wiley. CLEGG, S. R., PITSIS, T. S., RURA-POLLEY, T. & MAROSSZEKY, M. 2002. Governmentality Matters: Designing an Alliance Culture of Interorganizational Collaboration for Managing Projects. Organization Studies (Walter de Gruyter GmbH & Co. KG.), 23, 317. 66 DANILOVIC, M. & BROWNING, T. R. 2007. Managing complex product development projects with design structure matrices and domain mapping matrices. International Journal of Project Management, 25, 300-314. FLANAGAN, R. & NORMAN, G. 1993. Risk Management and Construction, Blackwell Scientific. FLYVBJERG, B. & BANK., W. 2005. Policy and planning for large infrastructure projects problems, causes, cures. Policy research working paper 3781. World Bank. HILLSON, D. & SIMON, P. 2007. Practical Project Risk Management: The Atom Methodology, Management Concepts, Incorporated. IRIMIA-DIÉGUEZ, A. I., SANCHEZ-CAZORLA, A. & ALFALLA-LUQUE, R. 2014. Risk Management in Megaprojects. Procedia - Social and Behavioral Sciences, 119, 407-416. KERZNER, H. 2009. Project Management: A Systems Approach to Planning, Scheduling, and Controlling, John Wiley & Sons. MILLER, R. & HOBBS, B. 2005. Governance Regimes for Large Complex Projects. Project Management Journal, 36, 42-50. MILLER, R. & LESSARD, D. R. 2000. The Strategic Management of Large Engineering Projects: Shaping Institutions, Risks and Governance, Massachusetts, Massachusetts Institute of Technology. PMI 2000. A Guide to the Project Management Body of Knowledge (PMBOK), Upper Darby, PA., Project Management Institute (PMI). SECRETARIAT, T. B. O. C. 2007. Standard for Project Complexity and Risk. SMITH, P. G. & MERRITT, G. M. 2002. Proactive Risk Management: Controlling Uncertainty in Product Development Productivity Press. SOLOMON, M. 2006. Pmp Exam Cram 2, 2/E (With Cd), Pearson Education. VAN MARREWIJK, A., CLEGG, S. R., PITSIS, T. S. & VEENSWIJK, M. 2008. Managing public–private megaprojects: Paradoxes, complexity, and project design. International Journal of Project Management, 26, 591-600. VOSE, D. 2005. Risk Analysis - A Quantitative guide, West Sussex, John Wiley and Sons. WIDEMAN, R. M. 1992. Project and Program Risk Management: A Guide to Managing Risk and Opportunities, Dexel Hill, PA, Project Management Institute (PMI). WILLIAMS, T. 2002. Modelling Complex Projects, West Sussex, John Wiley & Sons.