Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Benefit-Oriented Modelling for Project Appraisal, Selection,
Monitoring, and Performance Judgement
Elham Merikhi
Master of Philosophy
2018
Benefit-Oriented Modelling for Project Appraisal, Selection, Monitoring,
and Performance Judgement
By
Elham Merikhi
A thesis submitted for the degree of Master of Philosophy
of The Australian National University
February 2018
@Copyright by Elham Merikhi 2018
All Rights Reserved
ii
SIGNED STATEMENT OF ORIGINALITY
The work presented in this thesis is, to the best of my knowledge, my own work, except
as acknowledged in the text and declaration statement. The material has not been
submitted, either in whole, or in part, for a degree at this or any other university.
Elham Merikhi
iii
DECLARATION FOR A THESIS BY COMPILATION
In accordance with The Australian National University Program and Awards
Status-Research Awards Rule 2017, I hereby declare that this thesis contains no material
which has been accepted for the award of any other degree or diploma at any university
or equivalent institution and that, to the best of my knowledge and belief, this thesis
contains no material previously published or written by another person except where due
reference is made in the text of the thesis.
This thesis includes two papers that are going to be submitted in two Top Tier
Journals. The core theme on the thesis is the modelling of project evaluation and selection.
The papers are the result of original research conducted by the candidate during the course
of the study in the MPhil Program. The ideas, development and writing of all the papers
in the thesis were the principal responsibility of myself, the candidate, working within the
Research School of Management under the supervision of A/Prof Ofer Zwikael (Primary
supervisor and Chair of the supervisory panel) and A/Prof Arik Sadeh (Associate
supervisor of the supervisory panel).
In papers development, the candidate is the primary author and contributed greater
than 50% of the content. A substantial portion of the initial drafts of both papers were
written by the candidate and any subsequent editing with the guidance of the supervisory
panel was performed by the candidate. Among these papers, the first one is an extension
of a submitted paper to the 78th annual meeting of the Academy of Management (AOM),
and the second one is an extension of an accepted paper in the 77th annual meeting of the
AOM, presented at Atlanta, GA in August 2017.
iv
PAPERS’ DETAILS
This is an MPhil thesis by compilation in which the following two papers as Paper 1
and 2 (i.e. Chapter 2 and 3 respectively) are elaborated
Paper 1 (Chapter 2)
Initial explanation This paper is an extension of a submitted conference
paper with the following characteristics.
Full title
An integrated benefit-oriented project evaluation
framework: appraisal, monitoring and performance
judgement
Authors Elham Merikhi; Ofer Zwikael
Candidate’s contribution (%)
The candidate/primary author contributed greater
than 50% to the paper’s content and was responsible
for the following:
Reviewing the literature and extracting the
research gap
Proposing the model
Writing the substantial sections of the
manuscript
Addressing the co-author’s comments and
editing manuscript accordingly
Conference This paper was submitted to the 78th annual meeting
of the Academy of Management (AOM)
Volume / Number / page N/A
Status Submitted
Date submitted: 9 January 2018
v
Paper 2 (Chapter 3)
Initial explanation
This paper is an extension of one accepted and
presented conference paper with the following
characteristics.
Full title Customizing modern portfolio theory for the project
portfolio selection problem
Authors Elham Merikhi; Ofer Zwikael
Candidate’s contribution (%)
The candidate/primary author contributed greater
than 50% to the paper’s content and was responsible
for the following:
Reviewing the literature and extracting the
research gap
Proposing the mathematical model
Writing the substantial sections of the
manuscript
Addressing the co-author’s comments and
editing manuscript accordingly
Conference 77th annual meeting of the Academy of Management
(AOM)
Volume / Number / page 2017, 1, 13178
Status Accepted and presented at Atlanta, GA in August
2017
Date accepted: 17 March 2017
Candidate’s Declaration
I declare that the above-mentioned papers meet the requirement to be included in the
MPhil’s thesis.
Candidate name Candidate Signature Date
Elham Merikhi
15 February 2018
vi
ACKNOWLEDGEMENTS
For the exceptional supervision, expert advice, encouragement, and unrelenting
confidence in me, my utmost respect and appreciation goes to A/Prof Ofer Zwikael. He
merits special recognition for his outstanding contributions to my development as a
scholar. I am fortunate to have had the chance to work with him and will always be
grateful for his generous guidance and provision growth opportunities throughout my
candidature. My heartfelt thank goes to A/Prof Arik Sadeh for his valuable comments and
guidance in developing this thesis.
For the financial support and professional development sponsorship, I am highly
grateful to A/Prof Ofer Zwikael, Prof George Chen, Prof Byron Keating, CBE HDR
administration office- particularly, Mrs. Julie Fitzgibbon and Ms. Sarah Woodbridge- the
Australian department of Education and Training-Endeavour Postgraduate Scholarship
and the ANU College of Business and Economics.
For the excellent service, leadership, and support throughout my candidature, I
am grateful to the Research School of Management faculty and administration. For the
wonderful friendship, kindness, and constant reassurance, thanks to all my friends in the
Research School of Management.
My heartfelt thanks go to my siblings, Neda and Mehdi, for their love and
generosity and above all, I owe a debt of gratitude to my parents, Hamid and Mahnaz, for
their unconditional love and unyielding support in all of my endeavors.
vii
DEDICATION
This thesis is dedicated to my beloved GOD
“Thank you for all your unconditional love and support”
viii
ABSTRACT
Benefit-Oriented Modelling for Project Appraisal, Selection, Monitoring, and
Performance Judgement
A thesis submitted in partial fulfillment of the requirements for the degree of
Master of Philosophy
The Australian National University, 2018
Primary supervisor and Chair of the supervisory panel:
A/Prof Ofer Zwikael
Associate supervisor of the supervisory panel:
A/Prof Arik Sadeh
Although the realization of benefits is the main reason projects are funded, current project
evaluation (i.e. appraisal, monitoring and performance judgement) frameworks underplay
the importance of benefit realization, because they focus on delivering project’s outputs
on time, budget and to specifications. Moreover, the integration among the various
evaluation frameworks is poor because different evaluation tools are used at various
project phases. Altogether, to address the research gap, there is a need to develop an
ix
integrated project evaluation framework that takes benefits into consideration and can be
used throughout all project phases. Furthermore, because a project with very attractive
expected benefits-and consequently a very attractive expected return- but high risk might
expose the organization to a large loss, evaluation frameworks should consider the
combined analysis of projects’ returns and risks, as supported by utility theory.
In addition, after project appraisal, because resources are limited, organizations need to
select the best set of projects, i.e. project portfolio. Modern portfolio theory (MPT) in
(financial) portfolio selection problem suggests that in addition to the expected return and
risk, “risk interdependencies among projects” should also be considered, because they
lead to achieve a project portfolio with lowest level of risk for the same level of return.
However, there is gap in the literature as it is underdeveloped in providing a clear
approach to estimate risk interdependencies among projects.
In order to close these gaps, two papers are developed in this thesis: paper 1 tailors utility
theory to propose an integrated benefit-oriented framework for project evaluation on the
basis of its return and risk; and paper 2 tailors MPT to propose a risk-return mathematical
model for project portfolio selection. It is expected that the proposed project selection
model and evaluation framework improve the quality of portfolio selection decision in
organizations and help them ensuring the realization of projects’ intended benefits.
x
TABLE OF CONTENTS
SIGNED STATEMENT OF ORIGINALITY ........................................................................... ii
DECLARATION FOR A THESIS BY COMPILATION ....................................................... iii
PAPERS’ DETAILS ................................................................................................................... iv
ACKNOWLEDGEMENTS ....................................................................................................... vi
DEDICATION............................................................................................................................ vii
ABSTRACT ............................................................................................................................... viii
TABLE OF CONTENTS ............................................................................................................ x
LIST OF TABLES ..................................................................................................................... xii
LIST OF FIGURES .................................................................................................................. xiv
CHAPTER 1: INTRODUCTION .............................................................................................. 1
REFERENCES ..................................................................................................................... 6
CHAPTER 2: Paper 1 ................................................................................................................ 9
An Integrated Benefit-Oriented Project Evaluation Framework: Appraisal, Monitoring and
Performance Judgement ............................................................................................................ 9
CO-AUTHOR AUTHORIZATION FORM .................................................................... 10
ABSTRACT ........................................................................................................................ 11
1. INTRODUCTION ........................................................................................................ 12
2. LITERATURE REVIEW ............................................................................................ 15
2.1. Existing frameworks in project evaluation ........................................................ 15
2.2. Utility theory ......................................................................................................... 18
2.2.1. Principles ............................................................................................. 18
2.2.2. Indifference curves ............................................................................... 20
2.3. The application of utility theory in project evaluation ......................................... 21
3. THE DEVELOPMENT OF PROJECT EVALUATION FRAMEWORK ............ 23
3.1. Project evaluation map ........................................................................................ 23
3.1.1. The dimensions of the project evaluation map ......................................... 23
3.1.2. Indifference curves ............................................................................... 30
3.2. Project appraisal ................................................................................................... 31
3.2.1. The effect of risk mitigation actions on attractiveness contours ................. 32
3.3. Project monitoring ............................................................................................... 33
xi
3.4. Project performance Judgement ......................................................................... 34
4. A NUMERICAL EXAMPLE ........................................................................................ 36
4.1. Project appraisal ................................................................................................... 36
4.2. Project monitoring ............................................................................................... 39
4.3. Project performance judgement ......................................................................... 40
5. CONCLUSIONS ............................................................................................................. 41
REFERENCES ................................................................................................................... 43
CHAPTER 3: PAPER 2 ........................................................................................................... 57
Customizing Modern Portfolio Theory for the Project Portfolio Selection Problem ......... 57
CO-AUTHOR AUTHORIZATION FORM .................................................................... 58
ABSTRACT ........................................................................................................................ 59
1. INTRODUCTION ........................................................................................................ 60
2. LITERATURE REVIEW ............................................................................................ 63
2.1. The definition of “project portfolio selection problem” (PPSP) ...................... 63
2.2. PPSP models in the literature .............................................................................. 64
2.3. Modern Portfolio Theory (MPT) ........................................................................ 66
2.4. PPSP-MPT models ............................................................................................... 69
2.5. Limitations of the existing PPSP-MPT models .................................................. 71
3. DESIGN PRINCIPLES ............................................................................................... 72
4. MODEL DEVELOPMENT ........................................................................................ 76
4.1. Model assumptions ............................................................................................... 76
4.2. Threats and Opportunities Identification .......................................................... 77
4.3. The customized approach to estimate MPT’s parameters in PPSP ................. 78
4.4. Benefit, cost and technical interdependencies among projects ........................ 85
4.5. The optimization model ....................................................................................... 88
5. A NUMERICAL EXAMPLE ...................................................................................... 90
6. CONCLUSIONS .......................................................................................................... 98
REFERENCES ................................................................................................................. 101
CHAPTER 4: CONCLUSIONS ............................................................................................ 122
xii
LIST OF TABLES
Chapter 1: Introduction
-
Chapter 2: An Integrated Benefit-Oriented Project Evaluation Framework:
Appraisal, Monitoring and Performance Judgement
Table1 The differences between the approaches used by utility theory in asset
evaluation with those should be used in project evaluation
Table 2 Project E’s risk register in appraisal stage
Table 3 The return and cost of project “E” in different potential situations in
appraisal stage
Table 4 Candidate risk mitigation programs for project “E” in appraisal stage
Table 5 The updated risk register of project “E” at t=9 in monitoring stage
Chapter 3: Customizing Modern Portfolio Theory for the Project Portfolio Selection
Problem
Table 1 A comparison of PPSP definitions in the literature
Table 2 The similarities between PSP and PPSP
xiii
Table 3 The differences between the approaches used in PSP with those should be
used in PPSP
Table 4 Project 1’s risk register
Table 5 Project 2’s risk register
Table 6 The return of project 1 in different potential situations
Table 7 The estimated returns, risks and costs relevant to 15 project proposals
Table 8 The returns of projects 1 and 2 in their common potential situations
Table 9 Technical interdependencies among 15 project proposals
Chapter 4: Conclusions
-
xiv
LIST OF FIGURES
Chapter 1: Introduction
Figure 1 Project evaluation and selection in different projects’ phases
Chapter 2: An Integrated Benefit-Oriented Project Evaluation Framework:
Appraisal, Monitoring and Performance Judgement
Figure 1 The indifference curves in the return-risk plane
Figure 2 The project evaluation map
Figure 3 The project evaluation map of the numerical example
Chapter 3: Customizing Modern Portfolio Theory for the Project Portfolio Selection
Problem
Figure 1 The efficient frontier of risky assets with the optimal capital allocation
line, CAL (P), and P as the optimal portfolio
Figure 2 The design principles of project portfolio’s attractiveness evaluation
Figure 3 The efficient frontier of the numerical example
Chapter 4: Conclusions
-
1
CHAPTER 1: INTRODUCTION
2
Organizations fund projects to enhance their performance (Lewis, Welsh, Dehler,
& Green, 2002). To reach project success, projects should be systematically evaluated in
three stages (Frechtling, 2002; Lee, Son, & Om, 1996): (1) Ex-ante evaluation, i.e. project
appraisal, which is conducted in the project’s initiation phase to see what goals are
targeted; (2) Interim evaluation, i.e. project monitoring, which is applied through the
project’s planning, execution and benefit realization phases to find whether the intended
goals are being met; and (3) Ex-post evaluation, i.e. project performance judgement,
which is conducted after the project’s benefit realization phase to see whether the intended
goals have been met. Figure 1 depicts these three evaluation stages, i.e. appraisal,
monitoring and performance judgement and their position through the four project’s
phases (Zwikael & Meredith, 2018).
Figure 1
Project evaluation and selection in different projects’ phases
In each of the three stages of project evaluation, literature has traditionally focused
on the evaluation of output delivery (Atkinson, 1999). However, evaluation with a focus
on outputs delivery, i.e. output-oriented project evaluation, is insufficient because
3
delivering an output efficiently does not necessarily imply benefits realization for the
funding organization (Winter & Szczepanek, 2008). Accordingly, over the past few years
scholars have paid more attention to the realization of projects’ benefits. However, of
three stages of project evaluation, benefit-oriented evaluation frameworks have been
developed for only one (e.g. Angus, Flett, & Bowers, 2005; Zwikael & Smyrk, 2012) or
two (e.g. Barclay & Osei-Bryson, 2010) stages. As a result, these evaluation frameworks
are dispersed and incoherent, which means a framework applied for one stage of project
evaluation is useless for the others. Using a different framework for each stage of project
evaluation can also lead to inconsistency and waste of resources, e.g. time and money. To
address this research gap, there is a need to have a coherent evaluation framework
throughout the entire project’s life rather than having different incoherent frameworks for
each evaluation stage. Furthermore, in such a framework, the combined analysis of
projects’ benefits-and consequently projects’ returns- and risks is important because a
project with a very attractive expected return but high risk might expose the organization
to a large loss, whereas a low-risk project might secure the organization a lower but more
certain return (Hillson, 2002; Sefair, Méndez, Babat, Medaglia, & Zuluaga, 2017).
Altogether, paper 1 (chapter 2) of this thesis develops a coherent integrated project
evaluation framework which takes into account both projects’ returns and risks, from the
early phase of projects until after their completion. Considering the similar core concepts
between asset evaluation and project evaluation, in paper 1, I use principles of utility
theory, a well-acknowledged theory for asset evaluation in Finance discipline, and tailor
it to develop an integrated benefit-oriented project evaluation framework. Accordingly,
4
paper 1 aims to answer the following research question: “How can utility theory’s
principles be applied in the project context to develop an integrated benefit-oriented
project evaluation framework?”
Furthermore, after project appraisal, i.e. at the end of projects’ initiation phases
depicted in Figure 1, because resources are limited, not all suitable for funding project
proposals can be funded (Carazo, 2015). The selection of subset of projects to be funded
to optimize organizational performance given limited resources is known in the literature
as the “project portfolio selection problem” (PPSP) (Li, Fang, Tian, & Guo, 2015; Shou,
Xiang, Li, & Yao, 2014). A major limitation with most PPSP models is disregarding risk
interdependencies among projects. Risk interdependencies occur when two or more
projects have greater or lower risks if carried out simultaneously than if they were
accomplished at different times (Mutavdzic & Maybee, 2015). Such interdependencies
arise over time from overall social and economic changes and can affect multiple projects
(Gear & Cowie, 1980; Guo, Liang, Zhu, & Hu, 2008). Disregarding such
interdependencies, i.e. assuming that projects are independent, can leads to funding an
inappropriate project portfolio as the result of underestimating or neglecting an
appropriate project portfolio as the result of overestimating the risks of project portfolios.
Accordingly, literature (e.g. Gear & Cowie, 1980; Guo et al., 2008) calls to add risk
interdependencies among projects into the PPSP optimization model. For this purpose,
having considered the similar core concepts between “(Financial) portfolio selection
problem” (PSP) and PPSP, e.g. return, risk and risk interdependency, in paper 2 (Chapter
3), I use principles of Modern Portfolio Theory (MPT), a well-acknowledged theory to
5
PSP in Finance discipline, and tailor it to develop an optimization model of PPSP.
Accordingly, the paper 2 aims to answer the following research question: “How can MPT
principles be applied in the project portfolio context to improve the optimization model
of PPSP?”
This thesis contributes to the literature in two ways: (1) it proposes an integrated
benefit-oriented framework for project evaluation, and (2) it tailors MPT and utility
theories to make it suitable for investment decisions and evaluation that may also include
projects. Furthermore, in practice, this study enhances the quality of project appraisal,
selection, monitoring, and performance judgment in organizations and helps them
ensuring the realization of projects’ intended benefits.
The rest of this thesis is organized as follows. Chapter 2 is assigned to the
explanation of paper 1. Paper 2 is presented in Chapter 3 and finally, the conclusions are
drawn in Section 4.
6
REFERENCES
Angus, G. Y., Flett, P. D., & Bowers, J. A. 2005. Developing a value-centred proposal for
assessing project success. International Journal of Project Management, 23(6): 428-
436.
Atkinson, R. 1999. Project management: cost, time and quality, two best guesses and a
phenomenon, its time to accept other success criteria. International journal of project
management, 17(6): 337-342.
Barclay, C., & Osei-Bryson, K. M. 2010. Project performance development framework: An
approach for developing performance criteria & measures for information systems (IS)
projects. International Journal of Production Economics, 124(1): 272-292.
Carazo, A. F. 2015. Multi-criteria project portfolio selection. In J. Zimmermann & C.
Schwindt (Eds.), Handbook on project management and scheduling vol. 2: 709-728.
Switzerland: Springer International Publishing.
Frechtling, J. 2002. The 2002 User-Friendly Handbook for Project Evaluation.
Gear, T. E., & Cowie, G. C. 1980. A note on modeling project interdependence in research
and development. Decision Sciences, 11(4): 738–748.
Guo, P., Liang, J. J., Zhu, Y. M., & Hu, J. F. 2008. R&D project portfolio selection model
analysis within project interdependencies context. In Industrial Engineering and
Engineering Management, 2008. IEEM 2008. IEEE International Conference on (pp.
994-998). IEEE.
7
Hillson, D. 2002. Extending the risk process to manage opportunities. International Journal
of project management, 20(3): 235-240.
Lee, M., Son, B., & Om, K. 1996. Evaluation of national R&D projects in Korea. Research
Policy, 25(5): 805-818.
Lewis, M. W., Welsh, M. A., Dehler, G. E., & Green, S. G. 2002. Product development
tensions: Exploring contrasting styles of project management. Academy of
Management Journal, 45(3): 546-564.
Li, X., Fang, S. C., Tian, Y., & Guo, X. 2015. Expanded model of the project portfolio
selection problem with divisibility, time profile factors and cardinality constraints.
Journal of the Operational Research, 66(7): 1132–1139.
Mutavdzic, M., & Maybee, B. 2015. An extension of portfolio theory in selecting projects to
construct a preferred portfolio of petroleum assets. Journal of Petroleum Science and
Engineering, 133: 518-528.
Sefair, J. A., Méndez, C. Y., Babat, O., Medaglia, A. L., & Zuluaga, L. F. 2017. Linear
solution schemes for Mean-SemiVariance Project portfolio selection problems: An
application in the oil and gas industry. Omega, 68: 39-48.
Shou, Y., Xiang, W., Li, Y., & Yao, W. 2014. A multiagent evolutionary algorithm for the
resource-constrained project portfolio selection and scheduling problem. Mathematical
Problems in Engineering, 2014: 1-9.
8
Winter, M., & Szczepanek, T. 2008. Projects and programmes as value creation processes:
A new perspective and some practical implications. International Journal of Project
Management, 26(1): 95-103.
Zwikael, O., & Meredith, J. 2018. Who’s who in the project zoo? The ten core project roles.
International Journal of Operations & Production Management (forthcoming).
Zwikael, O., & Smyrk, J. 2012. A general framework for gauging the performance of
initiatives to enhance organizational value. British Journal of Management, 23(S1).
9
CHAPTER 2: PAPER 1
An Integrated Benefit-Oriented Project Evaluation Framework: Appraisal,
Monitoring and Performance Judgement
11
ABSTRACT
The realization of benefits is the main reason projects are funded. However, current project
evaluation (i.e. appraisal, monitoring and performance judgement) frameworks underplay the
importance of benefit realization, as they focus on delivering project’s outputs on time,
budget and to specifications. Moreover, the integration among the various evaluation
frameworks is poor with different evaluation tools used at various project phases. Altogether,
there is a need to develop an integrated project evaluation framework that takes benefits into
consideration and is usable for all project phases. Furthermore, because a project with a very
attractive expected return but high risk might expose the organization to a large loss, such a
framework should consider the combined analysis of projects’ returns and risks. The
principles of such return-risk evaluation are well supported by utility theory. This paper
tailors utility theory to propose an integrated benefit-oriented framework for project
evaluation on the basis of its return and risk. It is expected that the proposed project
evaluation framework will help organizations ensuring the realization of projects’ intended
benefits. To demonstrate how to apply the proposed framework, we employ a numerical
example and report the results.
Keywords:
Project evaluation; project appraisal; project monitoring; project performance judgement;
utility theory; risk register; benefit management
12
1. INTRODUCTION
Projects are the fundamental drivers of enhanced performance from an individual
business through to national economies (Lewis, Welsh, Dehler, & Green, 2002). A successful
project can bring benefits to the organization, such as better utilization of resources and
enhanced productivity (Lientz & Rea, 2016). To reach project success, a conceptually
coherent framework is required to systematically evaluate projects in terms of what goals are
planned to meet and whether the goals are being or have been met (Frechtling, 2002). The
evaluation of projects is implemented continuously along the whole projects’ life in three
stages: ex-ante, interim, and ex-post evaluation (Lee, Son, & Om, 1996). Ex-ante evaluation,
i.e. project appraisal, is aimed to prioritize project proposals before one or several projects
are funded (Laursen, Svejvig, & Rode, 2017). Interim evaluation, i.e. project monitoring,
compares from funding until completion the ongoing performance of a project to its initial
goals, to understand what went right or wrong in order to improve the strategy or the
processes (Crawford & Bryce, 2003). Ex-post evaluation, i.e. project performance
judgement, measures the realized project performance at the end of the project’s life to judge
whether the initial goals have been achieved and enhance organizational learning in order to
achieve successful projects in future (Angus, Flett, & Bowers, 2005).
In each of the three stages of project evaluation, i.e. appraisal, monitoring and
performance judgement, literature has traditionally focused on the delivery of outputs, such
as products (Atkinson, 1999). For example, project performance judgement particularly
13
focuses on delivering outputs on time, on budget, and to a defined quality, which is often
articulated as adhering to the “iron triangle” (Andersen, 2010). Project monitoring uses
methods such as Earned Value Management to track whether the project is on track to achieve
each of the iron triangle dimensions (Ong, Wang, & Zainon, 2016). In project appraisal,
output-based methods such as ‘scope-schedule-cost plan’ are applied (Project Management
Institute, 2017).
However, evaluation with a focus on output delivery is insufficient because delivering
an output efficiently does not necessarily imply benefits realization for the funding
organization (Winter & Szczepanek, 2008). Thus, in a wider view on the management of
projects (Morris, 1997), we see a shift from a sole focus on output creation to a holistic focus
on both output creation and benefit realization (Winter, Andersen, Elvin, & Levene, 2006).
Accordingly, over the past few years scholars have paid more attention to the realization of
projects’ benefits to develop some frameworks for one (e.g. Angus et al., 2005; Zwikael &
Smyrk, 2012) or two (e.g. Barclay & Osei-Bryson, 2010) out of three stages of project
evaluation. However, these frameworks are dispersed and incoherent, which means a
framework applied for one stage of project evaluation is useless for the other stages. Using a
different framework for each stage of project evaluation can lead to inconsistency and waste
of resources, e.g. time and money. Thus, there is a need to have a coherent evaluation
framework throughout the entire project’s life rather than having different frameworks for
each evaluation stage.
14
Furthermore, in such integrated framework, the combined analysis of project’s
benefits- and consequently project’s return- and risk is important because a project with a
very attractive expected return but high risk might expose the organization to a large loss,
whereas a low-risk project might secure the organization a lower but more certain return
(Hillson, 2002; Sefair, Méndez, Babat, Medaglia, & Zuluaga, 2017). Altogether, there is a
need to develop a coherent integrated project evaluation framework which takes into account
both projects’ returns and risks, from the early phase of projects until after their completion.
The principles of such evaluation framework are well supported by “utility theory”, as is
discussed next.
Utility theory is widely acknowledged in the Finance discipline to support decision
making under uncertainty, for example in asset evaluation, which trades off between asset’s
return and risk regarding the risk aversion of investors (Bell, 1995). A core principle of utility
theory is that each investor can assign a utility score to competing assets on the basis of their
expected returns and risks, which represents the welfare each asset has to the investor (Bodie,
Kane, & Marcus, 2014), to make a decision that maximizes the utility. This theory suggests
that the indifferent relation of assets’ returns and risks is transitive, and this leads to
equivalence classes of indifferent elements or, equally, to indifference curves (Luce, 1956).
Having considered the similar core concepts between asset evaluation and project
evaluation, i.e. return, risk and risk aversion, in this paper, we employ utility theory to
develop the integrated project evaluation framework based on projects’ returns and risks.
15
This paper contributes to the literature by proposing an integrated benefit-oriented framework
for all three stages of project evaluation. The paper also enhances the quality of project
evaluation in organizations and helps them ensuring the realization of projects’ intended
benefits. These contributions are achievable through answering the following research
question: “How can utility theory’s principles be applied in the project context to develop an
integrated benefit-oriented project evaluation framework?”
The rest of this paper is organized as follows. Section 2 reviews existing frameworks
in project evaluation, describes utility theory and discusses applying utility theory in project
evaluation framework. The new framework of project evaluation is proposed in Section 3.
Section 4 presents a numerical example to demonstrate how the proposed project evaluation
framework should be used. Finally, the conclusions are drawn in Section 5.
2. LITERATURE REVIEW
2.1. Existing frameworks in project evaluation
The prerequisite step to develop any framework for project evaluation is to clarify
what a project is. There are two common views in the literature to define a project (Angus et
al., 2005). The first is output-oriented view, exemplified by Project Management Institute
(2017): “A project is a temporary endeavour undertaken to create a unique product, service,
or result” (p. 4). Building on the output-oriented view, various frameworks exist for the three
stages of project evaluation, for instance: (1) Project appraisal - ‘scope-schedule-cost plan’
16
is applied (Project Management Institute, 2017); (2) Project monitoring - earned value, and
earned schedule are employed (Ong et al., 2016; Project Management Institute 2017); and
(3) Project performance judgement - the “iron triangle” (or triple constraint) test of project
success is the most conventional framework whereby performance is judged by the delivery
of project’s outputs fit-for-purpose, on time and within budget (Dvir & Lechler, 2004).
However, output-oriented evaluation frameworks ignore the realization of project’s
benefits (Ashurst, Doherty, & Peppard, 2008; Müller & Turner, 2007). Benefits are the flows
of value, i.e. the project’s desired end-effects derived from utilizing the project's outputs,
which is perceived as positive by the project funder, i.e. the senior manager who commits
funds and/or approves allocation of labor to the project (Laursen & Svejvig, 2016; Zwikael
& Meredith, 2018). In other words, although the underlying rationale for all projects is that
they seek specific target benefits (Dvir & Lechler, 2004), benefits do not appear among the
project evaluation criteria used in the output-oriented view (Zwikael & Smyrk, 2012).
To address the limitation of the project’s output-oriented view, the benefit-oriented
view was introduced, where a project is a unique process undertaken to achieve target
benefits (Zwikael & Smyrk, 2012). Building on the benefit-oriented view, a number of
frameworks exist for the three stages of project evaluation, for example: (1) Project appraisal
- Remer and Nieto (1995) discuss some return representatives in project appraisal such as
“net present value”, “payback period”, and “cost-benefit analysis”; (2) Project monitoring -
Sapountzis, Yates, Kagioglou, and Aouad (2009) focus upon the requirements to manage
17
change, tangible and intangible benefits in primary healthcare infrastructures. Ashurst et al.
(2008) propose a conceptual model of a benefits realization capability drawing on the
resource-based view of the firm; and (3) Project performance judgement - Atkinson (1999)
suggests a square route model to elaborate our understanding of success criteria in projects
with dimensions such as benefits for organization and community. Ashley, Lurie, and
Jaselskis (1987) refer to project success as having results much better than expected or
normally observed in terms of cost, schedule, quality, safety and participant satisfaction. De
Wit’s (1988) view is that the project is considered an overall success “if there is a high level
of satisfaction concerning the project outcome among key people in the parent organization”
(p. 165). Zwikael and Smyrk (2012) developed a conceptual triple-test performance
judgement framework including project management evaluation, project ownership
evaluation, and project investment evaluation.
The analysis of existing benefit-oriented project evaluation frameworks reveals the
dispersion and incoherence of these studies because a framework applied to one stage of
project evaluation is useless in the other stages. Although there are rare studies which deal
with two stages of project evaluation, such as Barclay and Osei-Bryson (2010) that developed
a framework for project monitoring and performance judgement, there is still a need to
develop an integrated framework to coherently combine all three stages of project evaluation.
Moreover, in such integrated framework, the combined analysis of projects’ returns and risks,
particularly in project appraisal and monitoring, is important because a project with a very
18
attractive expected return but high risk might expose the organization to a large loss, whereas
a low-risk project might secure the organization a lower but more certain return (Hillson,
2002; Sefair et al., 2017). Thus by and large, it can be asserted that, there is a need to develop
a coherent integrated project evaluation framework which takes into account projects’ returns
and risks, from the early phase of projects’ lives until after their completion.
To close the gap, because such evaluation framework is well explained by “utility
theory” (a powerful theory in the Finance discipline for asset evaluation which considers both
assets’ returns and risks), we employ the principles of this theory to develop an integrated
benefit-oriented project evaluation framework. Next, we discuss utility theory and explain
how its principles can be applied in developing the new project evaluation framework.
2.2. Utility theory
2.2.1. Principles
According to Bodie et al. (2014), it is assumed that each investor can assign a welfare,
or utility, score to competing assets on the basis of the expected returns and risks of those
assets. Higher utility values are assigned to assets with more attractive risk–return profiles.
Assets receive higher utility scores for higher expected returns and lower scores for higher
risks. Here, we pick up the quadratic von Neumann–Morgenstern utility function which
assigns an asset with expected return and risk , the following utility score, U:
19
21
2U A (1)
Where:
A is an index of the investor’s risk aversion, and
and are estimated by the mean and standard deviation of the asset’s historical returns.
Accordingly, investors choosing among competing investment assets will select the
one providing the highest utility level. Equation (1) is consistent with the notion that utility
is enhanced by higher expected return and diminished by higher risk. Notice that the return
of a risk-free asset (e.g., the return of placing money in the bank) receives a utility score equal
to its (known) rate of return, fr , because it has no risk. As can be seen from Equation (1),
the extent to which the variance of risky assets lowers utility depends on A, the investor’s
degree of risk aversion. Regarding the degree of risk aversion, investors can be divided into
three major categories as follows:
Risk-averse investors (for whom A>0) penalize the expected rate of return of a risky
asset by a certain percentage (or penalizes the expected profit by a dollar amount) to
account for the risk involved. The more risk averse the investor, the larger parameter
A is.
Risk-neutral investors (with A=0) judge risky assets solely by their expected rates
of return. The level of risk is irrelevant to the risk neutral investors, meaning that
20
there is no penalty for risk. For these investors, an asset’s certainty equivalent rate is
simply its expected rate of return.
Risk-lover investors (for whom A<0) are happy to engage in fair games and
gambles. These investors adjust the expected return upward to take into account the
“fun” of confronting the prospect’s risk. Risk lovers will typically take a fair game
because their upward adjustment of utility for risk gives the fair game a certainty
equivalent that exceeds the alternative of the risk-free investment.
In project evaluation, most organizations are risk-averse and thus they will not
undertake a high-risk project unless the project delivers a high return (Singh, 1986).
2.2.2. Indifference curves
Assume a risk-averse investor identifies all assets that are equally attractive as asset
P, depicted in Figure 1. Starting at P, an increase in standard deviation, which lowers utility,
must be compensated by an increase in expected return. Thus, point Q in Figure 1 is equally
desirable to this investor as P. In other words, this investor will be equally attracted to assets
with high risk and high expected returns compared with other assets with lower risk but lower
expected returns. These equally preferred assets will lie in the “return-risk plane” on a curve
called the indifference curve (Bodie et al., 2014), which connects all assets points with the
same utility values. As a result, a higher indifference curve (e.g. the curve point O lies on in
Figure 1) represents more attractive assets than a lower one (e.g. the curve point P and Q lie
21
on in Figure 1) to this investor. Figure 1 exemplifies indifference curves of risk-averse, risk-
neutral and risk-lover investors.
------------------------------------
Insert Figure 1 about here
------------------------------------
2.3. The application of utility theory in project evaluation
Researchers (e.g., Luo, 2012; Sefair et al., 2017) have identified important similarities
between asset and project evaluation that justify the use of utility theory in project evaluation.
These similarities include: (1) Both assets and projects have expected returns; (2) These
expected returns have uncertainties, i.e. risks; and (3) Both investors and organizations have
their own levels of risk aversion.
Despite these basic similarities between asset and project evaluation, some
differences also exist between these evaluations. First, some critics say that in practice, few
companies use utility theory to help in decision-making as within the same organization, one
manager may typically champion risky projects, while another in a similar position may be
more conservative (Bailey, Couët, Lamb, Simpson, & Rose, 2000). This issue can be
resolved because in project evaluation, there is only one decision maker, i.e. the funder. Thus,
indifference curves in organizations should be drawn regarding the risk aversion level of the
funder. Second, Perlitz, Peske, and Schrank (1999) assert that usually there are not sufficient
historical returns available to calculate projects’ returns and risks. This problem can be
22
addressed through replacing historical returns by the effects of triggering events, i.e. threats
and opportunities, around projects (Merikhi & Zwikael, 2017). Third, Casault, Groen, and
Linton (2013) argue that unlike assets which only have monetary benefits, projects have both
monetary and non-monetary benefits. This difference can be addressed through converting
non-monetary benefits and disbenefits to monetary ones (dolor values) by using some
techniques such as Delphi (Abbassi, Ashrafi, & Tashnizi, 2014) which its explanation is out
of the scope of this paper. Last but not least, unlike assets, some mitigation actions exist in
projects to reduce their risks. The objective of risk mitigation actions is to increase the
probabilities or impacts of opportunities and decrease the probabilities or impacts of threats
(Fang & Marle, 2015), which can in turn result in changing projects’ returns. These
mitigation actions cause shifting a project to a different indifference curve.
Altogether, Table 1 summaries the approaches used by utility theory in asset
evaluation and those should be used in project evaluation which introduces some implications
for the proposed integrated project evaluation framework.
------------------------------------
Insert Table 1 about here
------------------------------------
The analysis of Table 1 shows that although utility theory provides a good foundation
for project evaluation, two additional elements that are unique to project evaluation should
be integrated to the proposed framework. According to Table 1, these elements includes: (1)
The effects of triggering events that can have either damaging impacts as threats or assisting
23
impacts as opportunities on the expected return of a project through some chain of
consequences (Zwikael & Smyrk, 2011); and (2) the effects of risk mitigation actions. Next,
we propose a utility-based framework for three stages of project evaluation.
3. THE DEVELOPMENT OF PROJECT EVALUATION FRAMEWORK
Here, we first discuss how the return-risk plane should be drawn for project
evaluation. We call this plane “Project evaluation map”, which is the foundation of the
proposed project evaluation framework. Then we explain the usage of this map in each
project evaluation stage.
3.1. Project evaluation map
To develop the “project evaluation map”, we are inspired by the Project Investment
Evaluation (PIE) map (Zwikael & Smyrk, 2012), which uses project worth and its riskiness
to make decision regarding project appraisal. In the following sections we discuss the
dimensions of the project evaluation map and two core indifference curves required to make
evaluation decisions.
3.1.1. The dimensions of the project evaluation map
24
In order to estimate projects’ returns and risks, according to Table 1, we need first to
identify the effects of threats and opportunities around projects. The risk register is an
effective project management tool that includes this information used as part of a business
case which is updated through the project’s life (Project Management Institute, 2017).
Because most projects include a risk register as a core management tool, the proposed
framework uses data included in this document to estimate projects’ returns and risks. There
are different formats for risk registers proposed by various studies. One of the most
comprehensive formats is that proposed by Merikhi and Zwikael (2017) that is used in this
paper and exemplified in Table 2, in which a risk register is a table where rows are associated
with threats/opportunities and columns are relevant to their attributes including their
likelihoods and damaging/assisting impacts. The advantages of implementing this format
here is taking into account all threats and opportunities with all their possible damaging (i.e.
benefit reduced, benefit delayed, disbenefit increased, disbenefit advanced, cost increased,
and cost advanced) and assisting impacts (i.e. benefit increased, benefit advanced, disbenefit
decreased, disbenefit delayed, cost decreased, and cost delayed) on the project. Furthermore,
to develop the evaluation framework, we assume that the organization is going to implement
projects in a fixed planning period of time [0 T] assuming all project proposals, if get funded,
start at time 0.
------------------------------------
Insert Table 2 about here
------------------------------------
25
Employing the risk register introduced by Merikhi and Zwikael (2017), exemplified
in Table 2, the total number of potential situations can surround project i, im , is 2 in reached
from Equation (2) in which in is the number of threats/opportunities mentioned in the risk
register of project i.
... 20 1 2 3
iii i i i n
i
i
nn n n nm
n
(2)
Furthermore, we apply the “Modified Internal Rate of Return” (MIRR) as the
representative of a project’s rerun. MIRR (r in this paper) assumes that benefits/disbenefits
generated from a project are reinvested at the risk-free rate of return rather than at the
project’s internal rate of return (Lin, 1976). Equation (3) represents this assumption regarding
continuous compounding instead of discrete one.
rT
Futurevalue of benefits and disbenefitsPresent valueof costs
e (3)
Where: e is Euler's number which represents continuous compounding.
Thus, the return of project i in potential situation s, isr , is calculated as follows by
considering the fact that benefits and disbenefits have continuous flows, whereas costs are
discrete ones which only occur during the project’s life:
( )is is
is is isisis
FB FDLn
Ln FB FD Ln PCPCr
T T
(4)
26
( )
11
( )
11
1( )
(1 )
(1 )
(1 )
i
f
ik ikuB Bik iku
u s
f
i ik iku
B Bik ikuu s
f B Bik iku ik ikuu sf
KT r T t
is B BS S
k u s
r T tT
K B B
u s
S Sk f
r T S SB Br T Tu s
f
Where FB M M e dt
M M e
r
M M
e er
1
1
1
(1 )
1
i
i f B Bik iku ik ikuu s
K
k
K r T S SB B
u s
k f
M M
er
(5)
( )
11
( )
11
1( )
(1 )
(1 )
(1 )
i
f
il iluD Dil ilu
u s
f
i il ilu
D Dil iluu s
f D Dil ilu il iluu sf
LT r T t
is D DS S
l u s
r T tT
L D D
u s
S Sl f
r T S SD Dr T Tu s
f
FD M M e dt
M M e
r
M M
e er
1
1
1
(1 )
1
i
i f D Dil ilu il iluu s
L
l
L r T S SD D
u s
l f
M M
er
(6)
1
1
1i f C Cif ifu
u s
if ifu
F r S S
is C C
f u s
PC M M e
(7)
Where:
27
isFB is the total future value of benefits relevant to project i at time T if potential situation s
materializes,
isFD is the total future value of disbenefits relevant to project i if potential situation s
materializes,
isPC is the total present value of costs relevant to project i if potential situation s materializes,
iK is the total number of benefits relevant to project i,
iL is the total number of disbenefits relevant to project i,
iF is the total number of costs relevant to project i,
ikBM is the estimated/realized magnitude of kth benefit relevant to project i,
ikuBM is the estimated percentage of changes in the magnitude of kth benefit relevant to
project i if threat/opportunity u materializes,
ikBS is the estimated/realized scheduling for the realization of kth benefit relevant to project i,
ikuBS is the estimated percentage of changes in the scheduling for the realization of kth benefit
relevant to project i if threat/opportunity u materializes,
ilDM is the estimated/realized magnitude of lth disbenefit relevant to project i,
iluDM is the estimated percentage of changes in the magnitude of lth disbenefit relevant to
project i if threat/opportunity u materializes,
28
ilDS is the estimated/realized scheduling for the realization of lth disbenefit relevant to project
i,
iluDS is the estimated percentage of changes in the scheduling for the realization of lth
disbenefit relevant to project i if threat/opportunity u materializes,
ifCM is the estimated/spent magnitude of fth cost relevant to project i,
ifuCM is the estimated percentage of changes in the magnitude of fth cost relevant to project i
if threat/opportunity u materializes,
ifCS is the estimated/realized scheduling for the spending of fth cost relevant to project i, and
ifuCS is the estimated percentage of changes in the scheduling for the spending of fth cost
relevant to project i if threat/opportunity u materializes.
On the other hand, the likelihood of situation s materializing in project i, isp , is
calculated as the multiplication of the likelihoods of the threats/opportunities included in
situation s materializing and likelihoods of the others not materializing, demonstrated as
follows:
(1 )is u u
u s u s
p l l
(8)
Where: ul is the likelihood of threat/opportunity u materializing.
Thus, the estimation for return (i.e. mean, ˆi ), risk (i.e. standard deviation, ˆ
i ), and
cost ( ˆi ) of project i can be reached as follows:
29
1
ˆ (1 )im
is is is
i u u
s u s u s
Ln FB FD Ln PCl l
T
(9)
2
1
ˆ ˆ(1 )im
is is is
i u u i
s u s u s
Ln FB FD Ln PCl l
T
(10)
1
ˆ (1 )im
i u u is
s u s u s
l l PC
(11)
Since surplus budget is rarely left idle, we assume that the investor can always invest
unallocated budget in a risk-free asset with return fr (Findlay, McBride, Yormark, &
Messner, 1981). Thus, the estimated “overall expected return” and “overall risk” resulted
from funding project i (i.e. ˆi and ˆ
i respectively, which constitute the project evaluation
map’s dimensions, i.e. and ), and its estimated attractiveness (i.e. utility, ˆiU ) can be
reached as follows:
ˆ ˆˆ ˆ( ) ( )i i
i i fr
(12)
ˆˆ ˆ( )i
i i
(13)
21ˆ ˆ ˆ2
i i iU A (14)
Where: is the available budget to fund project i.
30
3.1.2. Indifference curves
We employ two types of indifference curves in project evaluation: (1) Funder
investment frontier; and (2) Attractiveness contours (Zwikael & Smyrk, 2011) discussed
next.
“Funder investment frontier” is an indifference curve with the utility equal to risk-
free rate of return, which represents all combinations of “overall returns” and “overall risks”
that a funder would be prepared to accept as a worst case. In the Funder’s point of view, the
risk-free rate of return can be for example placing money in the bank, the return of a well-
established production line or the like. Building on utility theory, this frontier is unique for
each funder according to how much risk averse he is and how much rate of return he considers
risk-free. Funders’ levels of risk aversion, “A”, can be extracted from analyzing their previous
decisions (see Bailey et al., 2000 for details). Having considered Equation (1), the function
of “Funder investment frontier” is reached as follows, depicted in Figure 2.
21
2fA r (15)
The “Funder investment frontier”, i.e. Equation (15), is applied in project appraisal
and project performance judgement explained later.
“Attractiveness contours” are indifference curves representing the attractiveness of
project proposals, depicted in Figure 2. The “Attractiveness contour” of project i with
attractiveness ˆiU is as follows:
31
21 ˆ2
iA U (16)
Where ˆiU is drawn from Equation (14).
Taken together, Figure 2 represents the “Project evaluation map” including “Funder
investment frontier” and “Attractiveness contours”.
------------------------------------
Insert Figure 2 about here
------------------------------------
3.2. Project appraisal
A project’s life consists of four phases: initiation, planning, execution, and benefit
realization (Zwikael & Smyrk, 2011). The decision whether a project gets funded is made at
the first phase of a project’s life, i.e. initiation. Accordingly, project appraisal is done in
initiation phase to prioritize project proposals before one or several projects are funded
(Laursen et al., 2017) based on their “overall returns” and “overall risks”, through applying
Equations (2), and (4) to (14).
According to Equation (15), all project proposals which their overall returns and
overall risks applies to this function will lie on the “Funder investment frontier” and their
attractiveness is equal to the known risk-free rate of return fr . As a result, project proposals
which lie bellow (and to the right of) “funder investment frontier” are unacceptable for
funding, while projects that lie above (and to the left of) this frontier are suitable for funding.
32
Furthermore, according to Equation (16), all project proposals lying on an
attractiveness contour represent the same attractiveness to the funder. The higher the location
of an attractiveness contour, the more attractive its corresponding project proposal is.
3.2.1. The effect of risk mitigation actions on attractiveness contours
Different risk mitigation actions, either to enhance the probabilities and impacts of
opportunities or decrease the probabilities and impacts of threats, can result in shifting a
project proposal from one attractiveness contour to one another, either higher or lower. A
risk mitigation action decreases the project’s risk and depending on its cost, increases or
decreases the project’s overall return and overall risk. Different combinations of candidate
risk mitigation actions construct candidate risk mitigation programs. Accordingly, here,
Equations (4) to (14) as well as Equation (16) should be again calculated based on the updated
risk register resulting from each candidate risk mitigation program and its corresponding cost.
Through this calculation, the cost of risk mitigation program should be considered as a new
project’s cost in Equation (7).
In Figure 2, if we assume “I0” as the location of project proposal i before applying
any mitigation programs, i.e. pre-risk mitigation program scenario, three possible post-risk
mitigation program scenarios are shown as “I1”, “I2” and “I3” where:
I1 is more attractive than I0 (because the cost of risk mitigation program is more than
adequately compensated for by the reduction in overall risk, I1 lies on a higher
33
attractiveness contour than I0). Such risk mitigation programme is cost effective and
can be considered for adoption.
I2 is equally attractive as I0 (because the cost of risk mitigation program is
compensated for by the reduction in overall risk, I2 lies on the same attractiveness
contour as I0). Such risk mitigation program makes no difference if it is adopted.
I3 is of less attractiveness than I0 (because the reduction in overall risk is only small
while the cost of the associated risk mitigation program is large, I3 lies on a lower
attractiveness contour than I0). Such risk mitigation programme is not cost effective
and so should not be adopted.
Regarding this argument, the best risk mitigation program for a specific project
proposal would be the one that moves the project proposal to the highest possible
“Attractiveness contour” given limited resources, as each risk mitigation program would
bring cost to the funder. Such highest attractiveness contour is called “Approved business
case contour”, depicted in Figure 2, which is the foundation of project monitoring and project
performance judgement explained in the following.
3.3. Project monitoring
Project monitoring, measures the ongoing performance of a project during planning,
execution and benefit realization phases, to understand whether the project is on the track of
its initial goals in order to improve the strategy or the processes (Crawford & Bryce, 2003).
34
During these phases: (1) Some threats/opportunities materialize, some new ones appear,
some become irrelevant, and some remain either with or without changes in their likelihoods
or impacts; and (2) Some benefits/disbenefits are realized and some costs are spent. All of
these changes result in an updated risk register for each project’s milestone. Thus, at different
project’s milestones, first, the project’ overall return and overall risk should be updated
according to its corresponding updated risk register and applying Equations (2), and (4) to
(14). Then, the corresponding point (i.e. the coordinate of updated “overall risk” and “overall
return”) should be located on the project evaluation map to see whether the project has been
shifted to a different attractive contour than the “approved business case contour”. If project
monitoring indicates project shifting to a higher attractiveness contour than or remaining on
the “approved business case contour”, demonstrated as scenarios X and Y respectively in
Figure 2, the project progresses better than or as expected. Otherwise, represented as scenario
Z in Figure 2, the appropriate corrective actions should be implemented to bring back the
project to the “approved business case contour”.
3.4. Project performance Judgement
Project performance judgement evaluates the realized project performance at the end
of the project’s life, through applying two tests: (1) Project business case judgement:
represented by the project’s performance in realizing the initial goals, i.e. “approved business
case contour”; and (2) Project investment judgement: represented by the project’s
35
performance in meeting the funder’s minimum preferences, i.e. “Funder investment frontier”
(Zwikael and Smyrk, 2012).
After all benefits and disbenefits of a project are realized, all threats and opportunities
become irrelevant for the project and so we remain with the realized overall return, but no
risk (Zwikael and Smyrk, 2011). Regarding Figure 2, having located the corresponding point
(i.e. the coordinate of zero “overall risk” and realized “overall return”), achieved through
applying Equations (2), and (4) to (14), on the project evaluation map, three possible
scenarios are shown as “F”, “G”, and “H” where:
F lies on the overall return dimension bellow fr . In this situation, because F’s
corresponding attractiveness contour is bellow both the “approved business case
contour” and “funder investment frontier”, both project business case and investment
are judged as failures.
G lies on the overall return dimension between fr , and 1
ˆIU (the attractiveness of the
“approved business case contour”). Here, project business case is judged as failure,
while project investment is judged as success.
H lies on the overall return dimension above 1
ˆIU . In this situation, both project
business case and investment are judged as successes.
36
4. A NUMERICAL EXAMPLE
The purpose of this section is to employ a numerical example in order to illustrate
how the proposed project evaluation framework can be applied. Let us assume that a funder
with the risk aversion of 4 faces a hypothetic project proposal “E” for a period of 15 months,
the risk-free rate of return of 0.2% per month and a limited budget of $20,000. Table 2 depicts
the risk register developed for this project, which shows three threats, their pre-likelihoods
and -damaging impacts.
Regarding Equation (15), the “funder investment frontier” is drawn as follows, which
is depicted in Figure (3)
22 0.002
------------------------------------
Insert Figure 3 about here
------------------------------------
4.1. Project appraisal
Regarding Equation (2), as this project has three threats, there are eight potential
situations to materialize. Table 3 shows these potential situations, the likelihood of each
situation materializing ( Esp ) drawn from Equation (8), the return of project “E” in each
situation ( Esr ) calculated by Equations (4) to (7), and the total present value of the project’s
costs in each situation ( EsPC ) calculated by Equation (7).
37
------------------------------------
Insert Table 3 about here
------------------------------------
Applying Equations (9) to (11), the estimated return, risk and cost of project “E”
before applying any risk mitigation programs are calculated as follows:
00.21 0.14
0.09
ˆ ( 0.0446) (0.09 0.0033) (0.21 0.0072) ( 0.0425)
( 0.057 0.06 0.141) ( 0.0072) ( 0.0 0.034) ( 0.0661)
0.0074 0.
06
74%
E
0
2 2 2
2 2
2 2
2 2
ˆ ( ) 0.09 ( 0.0033 )
0.21 (0.0072 ) ( )
(
0.21 0.0446 0.0074 0.0074
0.0074 0.14 0.0425 0.0074
0.09 0.0571 0.0074 0.06 0.0072 0.0074
0.14 0.0034 0.0074 0.06 0.0661 0.0074 0.0011867
) (
) ( ) 2
)
(
ˆE
E
00.0344 3.44%
00.21 0.14
0
ˆ ( 14,841) (0.09 16,818) (0.21 14,841) ( 14,841)
( 16,818) ( 16,818.09 0.06 0.1) ( 14,841)4 0.06
15,43
( 16,818)
$ 4
E
Furthermore, according to Equations (12), (13) and (14), the overall return and overall
risk resulted from funding project “E” and its attractiveness are reached as follows:
0
15,434 20,000 15,434ˆ 0.0074 0.002 0.0062 0.62%
20,000 20,000E
0
15,434ˆ ( ) 0.0344 0.0266 2.66%
20,000E
38
0
21ˆ 0.0062 (4)(0.0266 ) 0.00482
EU
Thus, according to Equation (16), the attractiveness contour of project “E” is reached
as follows, depicted in Figure 3 as the curve point E0 (2.66%, 0.62%) lies on:
22 0.0048
In order to decrease the risk of project “E”, the funder faces three candidate risk
mitigation programs demonstrated in Table 4 as M1, M2 and M3.
------------------------------------
Insert Table 4 about here
------------------------------------
Having applied Equations (4) to (14) on all three risk registers resulted from risk
mitigation programs M1, M2 and M3 and their corresponding costs, the estimated return, risk,
cost, overall return, overall risk and attractiveness of project “E” are calculated as: (1) For
M1- 0.57%, 3.16%, $18,034, 0.54%, 2.85%, and 0.0037; (2) For M2- 0.78%, 3.02%, $16,737,
0.68%, 2.53%, and 0.0055; and (3) For M3- 0.48%, 2.89%, $18,737, 0.46%, 2.71%, and
0.0031 respectively. Figure 3 represents the corresponding attractiveness contours of risk
mitigation programs M1, M2 and M3, achieved by Equation (16), as the curves points E1
(2.85%, 0.54%), E2 (2.53%, 0.68%), and E3 (2.71%, 0.46%) lie on respectively. As can been
seen the best risk mitigation program is M2 because E2 sits on the highest attractiveness
contour, which results in the “approved business case contour” of project “E”, depicted in
Figure 3. Furthermore, the cost of this mitigation action would be added to the risk register
39
as C4 with the magnitude of $1500 and scheduling for spending at time zero. Altogether, it
can be concluded that if the funder approves project “E” to be funded, he should apply the
risk mitigation program M2 and use the “approved business case contour” for the project
monitoring and performance judgement.
4.2. Project monitoring
The funder considers 14 milestones to monitor project “E” during its life with respects
to the “approved business case contour” (the 15th milestone coincides with project
performance judgement). As an example, we demonstrate the monitoring of project “E” in
Milestone t=9. The corresponding updated risk register is demonstrated in Table 5, in which
C4 has been added to the risk register as the cost of approved risk mitigation action M2.
------------------------------------
Insert Table 5 about here
------------------------------------
As can be seen form Table 5, the likelihood of threat T02 has changed from 50% to
40%, threat T03 has become irrelevant for the project, and a new threat T04 has been added
to the risk register. Furthermore, cost C1 has been spent as expected (i.e. with the magnitude
of $5,000 at time 1, whereas cost C2 has been spent differently to what expected with the
magnitudes of $5,200 at times 4.5. Moreover, the project’s disbenefit, i.e. D1, has been
realized as expected at time 5 with the magnitude of $1,000. Having employed Equations (2),
and (4) to (14) on the updated risk register demonstrated in Table 5, the updated estimated
40
return, risk, cost, overall return, overall risk and attractiveness of project “E” at time t=9 are
calculated as 1.05%, 2.97%, $16,887, 0.92%, 2.51%, and 0.0079 respectively, which its
corresponding point is depicted in Figure 3 as V (2.51%, 0.92%). As can be seen this point is
above the “approved business case contour” and thus the project progresses better than
expected. Also, the ongoing attractiveness of the project has increased from 0.0055 to 0.0079.
4.3. Project performance judgement
To judge the project performance, assume that benefits B1, B2 and B3 have been
realized at times 11, 12 and 12 with the magnitudes of $2,900, $3,000 and $3,000
respectively. Also, the remaining cost, i.e. C3, has been spent at time 10 with the magnitude
of $5,300. Having employed Equation (2) and (4) to (14) on the realized benefits, disbenefit
and spent costs, the realized return, risk, cost, overall return, overall risk and attractiveness
of project “E” are achieved as 1.01%, zero, $16,838, 0.88%, zero, and 0.88% respectively,
which its corresponding point is depicted in Figure 3 as W (0, 0.88%). As this point is above
the “approved business case contour”, both project business case and investment are judged
as successes. Furthermore, the realized attractiveness, i.e. 0.0088 is higher than expected, i.e.
0.0055.
41
5. CONCLUSIONS
Project evaluation is a crucial topic in project management to appraise project
proposals, monitor projects during their lives and judge projects’ performances upon their
completions. In project evaluation, not only delivering the outputs is important, but also
realizing the projects’ benefits is crucial. Furthermore, as the intended benefits are affected
by some threats and opportunities around projects, it is crucial to incorporate such effects
into the project evaluation framework. Accordingly, we proposed a new project evaluation
framework to such incorporation inspired by utility theory, an effective framework for asset
evaluation in the Finance discipline.
Theoretically, this paper builds synthesized coherence (Locke & Golden-Biddle,
1997) across project evaluation and utility theory, which are not typically cited together, to
suggest an integrated benefit-oriented project evaluation framework. It also contributes to the
project evaluation literature by developing an integrated framework for all three stages of
project evaluation regarding projects’ returns and risks. In practice, having used risk registers
which are crucial, well established and practiced documents exist in most projects, the
proposed evaluation framework is asserted to be practical and effective in order to help
organizations ensuring the realization of projects’ intended benefits. To apply the proposed
integrated framework: (1) When developing a business case, one should ensure it includes a
risk register with different candidate risk mitigation programs; (2) When a funder approves
42
a project proposal, one should keep updating risk register in determined milestones; and (3)
When a project is completed, one should calculate the realized overall return of the project.
Our paper also has some limitations that need to be acknowledged. First, as we apply
projects’ risk registers as the source of information to develop the project evaluation
framework, adopting the limitation of this document is undeniable. As a result, the proposed
framework does not consider unknowns-unknowns, i.e. uncertainties which are not known at
the beginning of projects (Lechler, Edington, & Gao, 2012). Future research that takes into
account unknown-unknowns in the project evaluation framework can improve its quality.
Second, we use some sample data to demonstrate how our project evaluation framework
should be applied. Future studies can test the framework using data from real projects in
various industries, countries and cultures. Third, future research can investigate different
ways to incorporate non-monetary benefits and disbenefits into the project evaluation
framework. Finally, future research may explore the application of some other theories such
as “prospect theory” in the developing of the project evaluation framework.
43
REFERENCES
Abbassi, M., Ashrafi, M., & Tashnizi, E. S. 2014. Selecting balanced portfolios of R&D
projects with interdependencies: A Cross-Entropy based methodology. Technovation,
34(1): 54-63.
Angus, G. Y., Flett, P. D., & Bowers, J. A. 2005. Developing a value-centred proposal for
assessing project success. International Journal of Project Management, 23(6): 428-
436.
Andersen, E. S. 2010. Rethinking project management–An organisational
perspective. Strategic Direction, 26(3).
Ashley, D. B., Lurie, C. S., & Jaselskis, E. J. 1987. Determinants of construction project
success, Project Management Institute.
Ashurst, C., Doherty, N. F., & Peppard, J. 2008. Improving the impact of IT development
projects: the benefits realization capability model. European Journal of Information
Systems, 17(4): 352-370.
Atkinson, R. 1999. Project management: cost, time and quality, two best guesses and a
phenomenon, its time to accept other success criteria. International journal of project
management, 17(6): 337-342.
Bailey, W., Couët, B., Lamb, F., Simpson, G., & Rose, P. 2000. Taking a calculated
risk. Oilfield Review, 12(3): 20-35.
44
Barclay, C., & Osei-Bryson, K. M. 2010. Project performance development framework: An
approach for developing performance criteria & measures for information systems (IS)
projects. International Journal of Production Economics, 124(1): 272-292.
Bell, D. E. 1995. Risk, return, and utility. Management science, 41(1): 23-30.
Bodie, Z., Kane, A., & Marcus, A. J. 2014. Investments (10th ed.). New York: McGraw-Hill
Education.
Casault, S., Groen, A. J., & Linton, J. D. 2013. Selection of a portfolio of R&D projects. In
A. N. Link & N. S. Vonortas (Eds.), Handbook on the theory and practice of program
evaluation: 89-112. Massachusetts: Edward Elgar publishing.
Crawford, P., & Bryce, P. 2003. Project monitoring and evaluation: a method for enhancing
the efficiency and effectiveness of aid project implementation. International journal of
project management, 21(5): 363-373.
De Wit, A. 1988. Measurement of project success. International journal of project
management, 6(3): 164-170.
Dvir, D., & Lechler, T. 2004. Plans are nothing, changing plans is everything: the impact of
changes on project success. Research policy, 33(1): 1-15.
Fang, C., & Marle, F. 2015. A Framework for the modeling and management of project risks
and risk Interactions. In Handbook on Project Management and Scheduling Vol.
2 (pp. 1105-1117). Springer International Publishing.
45
Findlay, M. C., McBride, R. D., Yormark, J. S., & Messner, S. D. 1981. Mean-variance
analysis for indivisible assets. Omega, 9(1): 77-88.
Frechtling, J. 2002. The 2002 User-Friendly Handbook for Project Evaluation.
Hillson, D. 2002. Extending the risk process to manage opportunities. International Journal
of project management, 20(3): 235-240.
Laursen, M., & Svejvig, P. 2016. Taking stock of project value creation: A structured
literature review with future directions for research and practice. International Journal
of Project Management, 34(4): 736-747.
Laursen, M., Svejvig, P., & Rode, A. L. G. 2017. Four Approaches to Project Evaluation.
In The 24th Nordic Academy of Management Conference.
Lechler, T. G., Edington, B. H., & Gao, T. 2012. Challenging classic project management:
Turning project uncertainties into business opportunities. Project Management
Journal, 43(6): 59-69.
Lee, M., Son, B., & Om, K. 1996. Evaluation of national R&D projects in Korea. Research
Policy, 25(5): 805-818.
Lewis, M. W., Welsh, M. A., Dehler, G. E., & Green, S. G. 2002. Product development
tensions: Exploring contrasting styles of project management. Academy of
Management Journal, 45(3): 546-564.
Lientz, B., & Rea, K. 2016. Breakthrough technology project management. Routledge.
46
Lin, S. A. 1976. The modified internal rate of return and investment criterion. The
Engineering Economist, 21(4): 237-247.
Locke, K., & Golden-Biddle, K. 1997. Constructing opportunities for contribution:
Structuring intertextual coherence and “problematizing” in organizational
studies. Academy of Management journal, 40(5): 1023-1062.
Luce, R. D. 1956. Semiorders and a theory of utility discrimination. Econometrica, Journal
of the Econometric Society, 178-191.
Luo, L.-M. 2012. Optimal diversification for R&D project portfolios. Scientometrics, 91(1):
219–229.
Merikhi, E., & Zwikael, O. 2017, January. Customizing Modern Portfolio Theory for the
Project Portfolio Selection Problem. In Academy of Management Proceedings (Vol.
2017, No. 1, p. 13178). Academy of Management.
Morris, P. W. 1997. The management of projects, Thomas Telford.
Müller, R., & Turner, R. 2007. The influence of project managers on project success criteria
and project success by type of project. European management journal, 25(4): 298-309.
Ong, H. Y., Wang, C., & Zainon, N. 2016. Integrated Earned Value Gantt Chart (EV-Gantt)
Tool for Project Portfolio Planning and Monitoring Optimization. Engineering
Management Journal, 28(1): 39-53.
Perlitz, M., Peske, T., & Schrank, R. 1999. Real options valuation: the new frontier in R&D
project evaluation? R&d Management, 29(3): 255-270.
47
Project Management Institute. 2017. A Guide to the project management body of knowledge
(6th ed.). Pennsylvania: Project Management Institute.
Remer, D. S., & Nieto, A. P. 1995. A compendium and comparison of 25 project evaluation
techniques. Part 1: Net present value and rate of return methods. International Journal
of Production Economics, 42(1): 79-96.
Sapountzis, S., Yates, K., Kagioglou, M., & Aouad, G. 2009. Realising benefits in primary
healthcare infrastructures. Facilities, 27(3/4): 74-87.
Sefair, J. A., Méndez, C. Y., Babat, O., Medaglia, A. L., & Zuluaga, L. F. 2017. Linear
solution schemes for Mean-SemiVariance Project portfolio selection problems: An
application in the oil and gas industry. Omega, 68: 39-48.
Singh, J. V. 1986. Performance, slack, and risk taking in organizational decision making.
Academy of management Journal, 29(3): 562-585.
Winter, M., Andersen, E. S., Elvin, R., & Levene, R. 2006. Focusing on business projects as
an area for future research: An exploratory discussion of four different perspectives.
International Journal of Project Management, 24(8): 699-709.
Winter, M., & Szczepanek, T. 2008. Projects and programmes as value creation processes:
A new perspective and some practical implications. International Journal of Project
Management, 26(1): 95-103.
Zwikael, O., & Meredith, J. 2018. Who’s who in the project zoo? The ten core project roles.
International Journal of Operations & Production Management (forthcoming).
48
Zwikael, O., & Smyrk, J. 2011. Project management for the creation of organisational
value, London: Springer Science & Business Media.
Zwikael, O., & Smyrk, J. 2012. A general framework for gauging the performance of
initiatives to enhance organizational value. British Journal of Management, 23(S1).
49
Table 1
The differences between the approaches used by utility theory in asset evaluation with
those should be used in project evaluation
# Parameter Asset evaluation Project evaluation
Implication for
the proposed
framework
1 The source
of risk
aversion
The risk aversion of
the individual investor
is considered.
The risk aversion of the funder
should be considered.
The investor is
replaced by the
funder.
2 Required
data
An appropriate amount
of reliable historical
returns is available.
Often there are not enough
historical returns or the accessible
historical returns are not reliable
enough. This is because the
triggering events, i.e. threats and
opportunities, faced by a project
can be unique and not possible to
generalize to other projects. The historical
returns are
replaced by the
effects of
triggering
events, i.e.
threats and
opportunities,
around projects.
3
Mean as
the measure
of expected
return
( )
Assets’ expected
returns are estimated
by using a “backward
approach”. In other
words, the mean of
historical returns
relevant to asset i are
used to estimate its
expected return.
Projects’ expected returns should
be estimated by using a “forward
approach”. In other words, some
forecasting techniques specific to
project management, i.e. taking
into account the effects of threats
and opportunities, should be
considered to estimate projects’
expected returns.
4
Standard
deviation as
the measure
of risk
( )
Assets’ risks are
estimated by the
standard deviations of
their historical returns.
Projects’ risks should be
estimated by using some
forecasting techniques specific to
project management, i.e. taking
into account the effects of threats
and opportunities.
5 Benefit /
Disbenefit
Only monetary
benefits and
disbenefits, measured
in private terms, are
considered in assets’
expected returns.
Both monetary and non-monetary
benefits and disbenefits, measured
in organizational terms, should be
considered in projects’ expected
returns.
Out of the scope
of this paper
6 Mitigation
action
No risk mitigation
actions can be done to
decrease an asset’s
risk.
Various mitigation actions can be
applied on threats and
opportunities to decrease projects’
risks, which can cause shifting a
project to a different indifference
curve.
The effects of
risk mitigation
actions are
added.
50
Table 2
Project E’s risk register in appraisal stage
ID Threat / Opportunity
Pre
Lik
elih
ood
Pre damaging/assisting impact
Benefit Disbenefit Cost
B1 B2 B3 D1 C1 C2 C3
M
($) S
(Mo)*
M
(I)**
S
(Mo)
M
(I)
S
(Mo)
M
(I)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
3100 10 3600 11 3000 12 1000 5 5000 1 5000 5 5000 10
∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S
T01 Exchange rate increases 30% -20% +20% -30% -10% +10% -20% +30%
T02 Environmental
organization protests 50% -20% +30%
T03 City requirements
estimate deviates 40% -10%
Where:
“M” is the “estimated/realized magnitude” of benefits, disbenefits or costs in the absence of any threats and opportunities,
“S” is the “estimated/realized scheduling” for the realization of benefits or disbenefits or the spending of costs in the absence of any threats and opportunities,
“∆M” is the “estimated percentage of changes in the magnitude” of benefits, disbenefits or costs (in two directions: increased or decreased, represented by “+” and “–
” respectively) resulted from corresponding threat or opportunity materializing, and
“∆S” is the “estimated percentage of changes in the scheduling” for the realization of benefits or disbenefits or spending of costs (in two directions: delayed or
advanced, represented by “+” and “–” respectively) resulted from corresponding threat or opportunity materializing.
* Mo: Month
** I: Index (the unit of non-monetary benefits/disbenefits can be converted to monetary values by Delphi approach)
51
Table 3
The return and cost of project “E” in different potential situations in appraisal stage
Situation
(s)
Threat/opportunity
included in the
situation s
likelihood of situation s
materializing
(Esp )
Return
(Esr )
Total present
value of Costs
(EsPC )
1 N/A 21% 4.46% $14,841
2 T01 9%* -0.33%** $16,818***
3 T02 21% 0.72% $14,841
4 T03 14% 4.25% $14,841
5 T01 & T02 9% -5.71% $16,818
6 T01 & T03 6% -0.72% $16,818
7 T02 & T03 14% 0.34% $14,841
8 T01 & T02 & T03 6% -6.61% $16,818
* According to Equation (8): 0.3×(1-0.5)×(1-0.4) = 0.09 = 9%
** According to Equations (4) to (7):
0.002(15 10(1 0.20)) 0.002(15 11(1 0)) 0.002(15 12(1 0))
2
0.002(15 5(1 0.10))
2
2
3100(1 0.2)( 1) 3600(1 0.30)( 1) 3000(1 0)( 1)
0.002 0.002 0.002$26,610
1000(1 0)( 1)
0.002$10,611
$16,818***
E
E
E
E
e e eFB
eFD
PC
r
2
(26,610 10,611) (16,818)0.0033 0.33%
15
Ln Ln
*** According to Equation (7):
0.002 1(1 0) 0.002 5(1 0.2) 0.002 10(1 0)
2 5000(1 0) 5000(1 0.1) 5000(1 0.3)$16,818
EPC e e e
52
Table 4
Candidate risk mitigation programs for project “E” in appraisal stage
Ris
k m
itig
ati
on
pro
gra
m I
D
Th
reat
/ O
pp
ort
un
ity
ID
Cost of
mitigation
Post
Lik
elih
ood
Post damaging/assisting impact
Benefit Disbenefit Cost
B1 B2 B3 D1 C1 C2 C3
M
($) S
(Mo)
M
(I)
S
(Mo)
M
(I)
S
(Mo)
M
(I)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
3100 10 3600 11 3000 12 1000 5 5000 1 5000 5 5000 10
∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S
M1
T01 - - 30% -20% +20% -30% -10% +10% -20% +30%
T02 2600 0 30% -20% +30%
T03 - - 40% -10%
M2
T01 1500 0 20% -20% +20% -20% -10% +10% -20% +30%
T02 - - 50% -20% +30%
T03 - - 40% -10%
M3
T01 1500 0 20% -20% +20% -20% -10% +10% -20% +30%
T02 1200 0 40% -20% +30%
T03 800 0 30% -10%
53
Table 5
The updated risk register of project “E” at t=9 in monitoring stage
ID Threat /
Opportunity P
ost
Lik
elih
ood
Post damaging/assisting impact
Benefit Disbenefit Cost
B1 B2 B3 D1 C1 C2 C3 C4
M
($) S
(Mo)
M
(I)
S
(Mo)
M
(I)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
3100 10 3600 11 3000 12 1000 5 5000 1 5200 4.5 5000 10 1500 0
∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆M ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S
T01 Exchange rate
increases 20% -20% +20% -20% +30%
T02 Environmental
organization protests
parliament
40% -20% +30%
T03 City requirements
estimate deviates 0
T04 Project manager
leaves 10% +10% +10% +10%
54
Figure 1
The indifference curves in the return-risk plane
55
Figure 2
The project evaluation map
56
Figure 3
The project evaluation map of the numerical example
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
0 1 2 3 4 5
E0
E1
E2
E3
V W
Overall Return %
Overall
Risk %
Funder
Investment
Frontier
Approved
Business case
contour
57
CHAPTER 3: PAPER 2
CUSTOMIZING MODERN PORTFOLIO THEORY FOR THE PROJECT
PORTFOLIO SELECTION PROBLEM
59
ABSTRACT
Because an organization’s performance depends on the projects it implements, selecting
the most appropriate portfolio of projects given limited resources is a crucial decision.
Modern portfolio theory in (financial) portfolio selection problem suggests that in
addition to the expected return, risk should also be considered. Because projects are
unique and lack of historical data, Modern portfolio theory needs to be tailored to the
“risk-return optimization” of project portfolio selection problem. A core element in risk-
return optimization is “risk interdependencies among projects” leading to achieving a
project portfolio with lowest level of risk for the same level of return. However, the
literature is underdeveloped in providing a clear approach to estimate risk
interdependencies among projects. In order to fill this gap, we substitute the historical
data in MPT with projects’ risk registers and propose a new risk-return optimization
model for. To demonstrate how to apply the proposed new model, we employ a numerical
example and report the results.
Keywords:
Project portfolio selection problem; risk-return optimization; Modern portfolio theory;
risk register; risk interdependency
60
1. INTRODUCTION
Senior executives approve project proposals that have the potential to enhance
organizational performance (Lewis, Welsh, Dehler, & Green, 2002; Zwikael & Smyrk,
2012). However, because resources are limited, not all project proposals can be funded
(Carazo, 2015). Selecting the most appropriate set of projects is a crucial organizational
decision (Ghasemzadeh, Archer, & Iyogun, 1999) because a wrong decision can result in
two destructive consequences: (1) resources are wasted on inappropriate projects that
have been funded (type II error), and (2) the benefits that could have been realized from
allocating such resources to better projects are lost (type I error) (Christensen & Knudsen,
2010; Martino, 1995). The selection of subset of projects to be funded to optimize
organizational performance given limited resources is known in the literature as the
“project portfolio selection problem” (PPSP) (Li, Fang, Tian, & Guo, 2015; Shou, Xiang,
Li, & Yao, 2014). The combined analysis of portfolio’s return and risk is important in
this problem since a portfolio with a very attractive expected return but high risk might
expose the organization to a large loss, whereas a low-risk portfolio might secure the
organization a lower but more certain return (Hillson, 2002; Sefair, Méndez, Babat,
Medaglia, & Zuluaga, 2017).
A major limitation with most of the existing PPSP models is disregarding risk
interdependencies among projects. Risk interdependencies occur when two or more
projects have greater or lower risks if carried out simultaneously than if they were
accomplished at different times (Mutavdzic & Maybee, 2015). Such interdependencies
arise over time from overall social and economic changes and can affect multiple projects
61
(Gear & Cowie, 1980; Guo, Liang, Zhu, & Hu, 2008). For example, a potential increase
in the exchange rate can increase the expected return of a project which its final product
would be exported (an opportunity), and simultaneously reduce the expected return of a
project that its raw material should be imported (a threat). Considering such
interdependency in the optimization model of PPSP to reach the optimal project portfolio
is crucial as it provides the condition to reach a project portfolio with lowest level of risk
for the same level of return (Mutavdzic & Maybee, 2015). Disregarding such
interdependencies, i.e. assuming that projects are independent, can leads to funding an
inappropriate project portfolio as the result of underestimating or neglecting an
appropriate project portfolio as the result of overestimating the risks of project portfolios.
Accordingly, literature (e.g. Gear & Cowie, 1980; Guo et al., 2008) calls to add risk
interdependencies among projects into the PPSP optimization model through applying
financial principles of (financial) portfolio selection problem (PSP) in which the concept
of such interdependency is well analyzed.
PSP is one of the most studied topics in finance (Chien, 2002), and is concerned
with the allocation of limited capital to a number of potential assets for a profitable
investment strategy (Lwin & Qu, 2013). An effective optimization model to PSP is
“Modern portfolio theory” (MPT) - for which its pioneer, Harry Markowitz, was awarded
a Nobel Prize (Varian, 1993). MPT views PSP as a mean-variance optimization problem
with regard to two criteria: to maximize the portfolio’s return and to minimize its risk,
measured by the mean and standard deviation of its return respectively (Markowitz,
1952). The underlying principle of MPT is that assets should not be selected solely based
62
on their individual returns and risks, so it is concerned with quantifying how each asset’
return changes in relation to other assets in the portfolio, i.e. assets’ correlations, in the
overall market fluctuations (Casault, Groen, & Linton, 2013). This crucial principle also
applies to the projects in a portfolio as their expected returns are affected by some threats
and opportunities such as overall social and economic changes, which in turn can result
in risk interdependencies among projects.
Having considered the similar concepts between PSP and PPSP, i.e. return, risk,
correlation/risk interdependency and the available budget, as well as the importance of
analyzing the risk interdependencies among projects, researchers tried to apply MPT into
PPSP (e.g., Esfahani, Sobhiyah, & Yousefi, 2016; Luo, 2012; Sefair et al., 2017).
However, there is a research gap in these studies as they disregard some fundamental
differences exist between the characteristics of financial portfolios and those of project
portfolios (Casault et al., 2013). The major difference is based on the fact that unlike
financial assets, because each project is unique (Westerveld, 2003), often there is not
enough projects’ historical data, or the accessible historical data is not reliable enough to
estimate projects’ returns and risks as well as risk interdependencies among them.
Thus, although MPT provides a good foundation for considering risk
interdependencies among projects in the project portfolio selection model, some other
customizations in the model of MPT are required to make it suitable to be used in PPSP.
Such research gap will be addressed by this paper in which first the essential differences
between PSP and PPSP are distinguished and then some customizations in the MPT
model are proposed in order to make it suitable to be used in PPSP. Accordingly, the
63
paper aims to answer the following research question: “How can MPT principles be
applied in the project portfolio context to improve the optimization model of PPSP?” Our
paper contributes to the literature in two ways: (1) the proposed optimization model
expands PPSP models by incorporating the effects of threats and opportunities around
projects into the optimization model, and (2) it expands MPT to make it suitable to
investment decisions that may also include projects with no historical data to allow a
reliable estimation of their return and risk levels. Furthermore, in practice, this paper may
improve the quality of portfolio selection decision in organizations.
The rest of this paper is organized as follows. Section 2 reviews different
definitions and models have been used by the researchers into PPSP, describes MPT,
exemplifies studies that have applied MPT in PPSP, and discusses their limitations. The
principles of designing the optimization model are clarified in Section 3. Section 4
proposes the optimization model of the problem which customizes MPT to PPSP. A
numerical example is presented in Section 5 to demonstrate how the proposed model
should be applied and finally, the conclusions are drawn in Section 6.
2. LITERATURE REVIEW
2.1. The definition of “project portfolio selection problem” (PPSP)
The literature has suggested multiple definitions for PPSP. Table 1 compares the
leading definitions in their objectives and constraints. As can be seen from this table,
there is an agreement that the objective of PPSP is to optimize organizational
performance and that there are limited resources available to the decision maker. After
64
considering the similarities and the consensus in the literature, the following definition
for PPSP will be used in this paper: “The selection of subset of project proposals to be
funded to optimize organizational performance given limited resources”.
------------------------------------
Insert Tables 1 about here
------------------------------------
2.2. PPSP models in the literature
Since the mid-1950s researchers have developed various models to formulate and
solve PPSP (Lorie & Savage, 1955). These models can be categorized into four distinct
groups. The first group of models proposes one-criterion optimization models in which
only one criterion is considered to assign a score to each project. Thus, projects are
selected from the highest to the lowest score of this criterion until the budget available is
spent. For instance, Lorie and Savage (1955), Myers (1972) and Weingartner (1962)
apply economic assessment measurements (e.g., net present value and internal rate of
return) to calculate the project score.
Multi-criteria models form the second group of models. As the sphere in which
decisions are taken in any organization is usually characterized by a set of competing
criteria (Carazo, 2015), this category of methods presents different multi-criteria models
in order to incorporate the decision maker’s preferences into the process. For example,
Gear, Lockett, and Muhlemann (1982), Melachrinoudis and Rice (1991) as well as
Vepsalainen and Lauro (1988) apply “comparative models” such as analytical hierarchy
process (AHP) to combine different criteria to a single objective criterion and then
65
compare one project to either another project or some subsets of alterative projects. Albala
(1975), Cooper (1978) and Krawiec (1984) use “Scoring models” to combine the merit
of each project with respect to a small number of decision criteria to specify its desirability
score and then rank projects according to their scores. Benjamin (1985), Golabi,
Kirkwood, and Sicherman (1981) as well as Neely, North, and Fortson (1977) apply
“mathematical programing models” to maximize objectives such as profit, revenue and
utility or minimize others like resource use, cost and runtime simultaneously.
The third group of models includes projects’ interdependency models. As there
may be technical (e.g. complementarities), cost and benefit interdependencies (i.e.
synergies produced by sharing resources and benefits respectively) among projects
derived from conducting more than one project at the same time (Czajkowski & Jones,
1986; Fox, Baker, & Bryant, 1984), this group of solutions incorporates one or more
categories of these interdependencies into the optimization model. For example,
Carraway and Schmidt (1991) propose a model by formulizing the benefit and cost
interdependencies among pairs of projects quantitatively. Klapka and Piños (2002), Lee
and Kim (2000), Santhanam and Kyparisis (1995) as well as Schmidt (1993) suggest
different models which reflect benefit, cost and technical interdependencies among the
sets of two or three projects. Doerner, Gutjahr, Hartl, Strauss, and Stummer (2006),
Santhanam and Kyparisis (1996) as well as Yu, Wang, Wen, and Lai (2012) developed
various models that allowed the technical, benefit and cost interdependencies among any
numbers of projects.
66
A major limitation of the above-mentioned models is disregarding the risk
interdependencies among projects that arise over time from overall social and economic
changes (Gear & Cowie, 1980; Guo et al., 2008). Such interdependencies can affect a
portfolio’s risk in a way that it is inferior or superior to the weighted average sum of the
risks of all its projects (Mutavdzic & Maybee, 2015). Disregarding such
interdependencies, i.e. assuming that projects are independent, can be catastrophic and
leads to funding an inappropriate project portfolio as the result of underestimating or
neglecting an appropriate project portfolio as the result of overestimating the risks of
project portfolios.
The resolution of this limitation resulted in the advent of the fourth group of
models, which tries to consider the risk interdependencies among projects in addition to
the projects’ returns and risks in their optimization models. To reach this objective,
models in this group employ MPT. Next, we discuss the suitability and structure of MPT
and summarize some of its implementations in PPSP, i.e. PPSP-MPT models.
2.3. Modern Portfolio Theory (MPT)
One of modern finance theory’s main tenets is MPT which led to Markowitz’s
Nobel Prize in Economics in 1990. Markowitz (1952) originally formulated the
fundamental theorem of mean-variance portfolio model for risk-return optimization in
PSP, which trades off between expected return and risk of a portfolio (represented by
mean and standard deviation of that portfolio’s return respectively) to reach the optimal
portfolio of various assets. The underlying principle of MPT is that assets should not be
67
selected solely based on their individual returns and risks, so it quantifies assets’
correlations which represent how each asset’ return changes in relation to other assets in
the portfolio in overall market fluctuations (Casault et al., 2013). Such correlation is
called “risk interdependency” in project portfolio. Having considered the correlation
among assets, MPT argues that by the same level of expected return, portfolio’s risk can
be reduced by creating a diversified portfolio of unrelated assets, i.e. with lowest
correlation coefficients (Moriarty, 2001). Then, a portfolio is considered “efficient” if for
a given level of expected return there are no portfolios with a lower risk, or conversely
for a given level of risk there are no portfolios with a higher expected return. The complete
set of these efficient portfolios forms the efficient frontier that represents the best trade-
offs between return and risk (Markowitz, 1952, 1959; Markowitz, Todd, & Sharpe, 2000).
Accordingly, the mean-variance optimization model is as follows (Bodie, Kane, &
Marcus, 2014):
P f
P
P
rMax S
(1)
1
N
p i i
i
Where X
(2)
2
1 1
N N
p i j i j ij
i j
X X
(3)
1
1N
i
i
Subject to X
(4)
0 1 1,2,...,iX i N (5)
Where:
pS is the slope of capital allocation line (CAL), called “Sharpe ratio” (Sharpe, 1994),
68
p is the expected return of portfolio p,
fr is the risk-free rate of return (e.g., the return of placing money in the bank),
p is the standard deviation of return relevant to portfolio p,
i is the expected return of asset i,
i is the standard deviation of return relevant to asset i,
ij is “Pearson correlation coefficient” between assets i and j,
N is the number of available assets, and
iX is the decision variable of the budget proportion invested in asset i.
To estimate two parameters i and i , MPT uses the historical data relevant to
asset i, e.g. as follows:
1
ˆiM
i is is
s
p r
(6)
22
1
ˆ ˆiM
i is is i
s
p r
(7)
Where:
ˆi and ˆ
i are the estimates of i and i respectively,
iM is the number of historical situations available for asset i,
isr is the return of asset i in historical situation s, and
isp is the proportion of happening historical situation s.
Figure 1 depicts the “optimal capital allocation line”, CAL (P), which is tangent
to the “efficient frontier” at the “optimal portfolio” (P).
69
------------------------------------
Insert Figure 1 about here
------------------------------------
Having summarized the structure of MPT, that is both well-constructed in its
theoretical foundations and successful in its applications in PSP (Chien, 2002), we can
now turn to PPSP-MPT models.
2.4. PPSP-MPT models
A few researchers have identified important similarities between PSP and PPSP
(e.g., Boasson, Cheng, & Boasson, 2012; Luo, 2012; Sefair et al., 2017). Table 2
summarizes these similarities in relation to four concepts, i.e. return, risk, correlation
among assets/risk interdependencies among projects, and available budget. This analysis
suggests strong similarities between PSP and PPSP and encourages the implementation
of a strong financial theory, i.e. MPT, to deal with the research limitation of the first three
groups of PPSP models discussed earlier.
------------------------------------
Insert Table 2 about here
------------------------------------
Building on the similarities, PPSP-MPT models applies the mean-variance
foundation of MPT in forming optimal risk-return project portfolio. These models can
further be divided into two distinct subgroups.
The first subgroup belongs to those models that use projects’ historical data or not
propose an approach how to estimates different MPT’s parameters, i.e. projects’ returns
70
and risks as well as risk interdependencies among them. For instance, Boasson et al.
(2012) apply MPT to municipal financial and capital budgeting decisions by considering
the historical data relevant to benefits and cost of the similar projects. Esfahani et al.
(2016) apply MPT to PPSP considering the historical returns of similar projects and
proposed the harmony search algorithm to solve it. Findlay, McBride, Yormark, and
Messner (1981) apply MPT to indivisible assets/projects to develop a quadratic integer
programming model without explaining how the projects’ return, risk as well as risk
interdependencies can be estimated. Sefair et al. (2017) developed a linear solution for
the Mean-SemiVariance PPSP in the oil and gas industry by considering the historical
data relevant to the net present value of the similar projects.
The second subgroup of models implementing MPT for PPSP addresses the lack
of projects’ historical data through considering a few common elements among projects.
For example, Luo (2012) concentrated on the risk-side control for Research and
Development (R&D) project portfolio and developed a method for optimal diversification
of R&D project portfolio incorporating market and technology risk. Ball and Savage
(1999) explain five sources of risk interdependencies among projects including prices,
places, profiles, politics and procedures and declare that simulation or decision tree model
can be used to calculate return, risk and risk interdependencies among production and
exploration projects in oil industry. Mutavdzic and Maybee (2015) consider the same five
sources of correlation and propose a two-way data table to estimate the risk
interdependencies and constructed a preferred portfolio of petroleum assets.
71
2.5. Limitations of the existing PPSP-MPT models
An analysis of models applying MPT in PPSP suggests that they ignore some core
differences that exist between PSP and PPSP, discussed below. Boasson et al. (2012) and
Casault et al. (2013) argue that unlike financial assets which only have monetary benefits,
projects have both monetary and non-monetary benefits. Mutavdzic and Maybee (2015)
assert that usually there is not sufficient historical data available to calculate the risk
interdependencies among projects. This argument highlights the main difference between
financial portfolios and project portfolios. In projects, given the lack of historical data,
the effects of triggering events on the return of multiple projects determine risk
interdependencies among them. A triggering events’ effect can be either a damaging
impact as a threat or an assisting impact as an opportunity through some chain of
consequences (Zwikael & Smyrk, 2011). As the threats and opportunities that surround a
project can be unique, one cannot use projects’ historical data or limit the source of risk
interdependencies to some certain elements such as prices, places, profiles, politics and
procedures to estimate MPT’s parameters in PPSP. This means that historical data
employed by MPT in PSP should be substituted by the effects of threats and opportunities
in PPSP. However, the existing PPSP-MPT models are underdeveloped in estimating
MPT’s parameters appropriately, i.e. taking into consideration the effects of threats and
opportunities.
Table 3 compares the approaches used in PSP with those should be used in PPSP,
which introduces some initial guidelines for the proposed design principles.
72
------------------------------------
Insert Table 3 about here
------------------------------------
According to substantial differences between PSP and PPSP highlighted in Table
3, in particular the lack of reliable projects’ historical data rooted in different threats and
opportunities exist around projects, as well as benefit, cost and technical
interdependencies among projects, it can be concluded that despite the similar concepts
between two problems mentioned in Table 2, direct application of MPT to PPSP is
problematic and thus some other customizations are also required to capture such
characteristics in the evaluation of project portfolio. To do such customization, in this
paper, we substitute the historical data in MPT with the effects of threats and opportunities
and propose a new risk-return optimization model for PPSP.
3. DESIGN PRINCIPLES
This section develops the core design principles of the model. Design principles
are prescriptive statements that show how to do something in order to achieve a goal
(Gregor & Jones, 2007). As the basis of any selection problem is the evaluation of
different alternatives (Dey, 2006; Oral, Kettani, & Lang, 1991), here, design principles
elaborate “how to evaluate the attractiveness of project portfolios” that play the crucial
role in developing the PPSP’s optimization model. These “know how” are developed in
the rest of this Section.
The evaluation foundation of a portfolio is well-developed in MPT, which tells a
single investor how to combine individual assets into a portfolio in order to accomplish
73
the goal of maximizing portfolio attractiveness, i.e. Sharpe ratio (Smith & Smith, 2005).
MPT uses assets’ historical returns to investigate their behaviors in different market
fluctuation to estimate assets’ individual returns and risks as well as correlation
coefficients among them (Bodie et al., 2014). However, as the threats and opportunities
exist around a project can be unique and not possible to generalize to other projects, one
cannot use projects’ historical returns to estimate the individual projects’ returns and risks
as well as risk interdependencies among them. In other words, the effects of threats and
opportunities on projects determine the individual projects’ returns and risks as well as
risk interdependencies among them. Thus, we propose the following design principles:
Design principle 1 (P1). The attractiveness of a project portfolio is evaluated
based on its return and risk.
Design principle 2 (P2). Project portfolio’s return is evaluated based on the
individual returns of its projects.
Design principle 3 (P3). The individual returns of projects are evaluated based
on the effects of threats and opportunities on those projects.
Design principle 4 (P4). Project portfolio’s risk is evaluated based on the
individual risks of its projects and the risk interdependencies exist among those
projects.
Design principle 5 (P5). The individual risks of projects and risk
interdependencies among them are evaluated based on the effects of threats and
opportunities on those projects.
74
While MPT is significant to provide the context of how the individual returns and
risks of projects and risk interdependencies among them should be evaluated based on
the effect of different threats and opportunities exist around those projects, it is not
sufficient to explain some other interdependencies among projects including benefit, cost
and technical interdependencies. This type of evaluation is well developed in the third
group of PPSP models, i.e. projects’ interdependency models. According to these models,
a portfolio’s return is calculated by considering the individual returns of its projects as
well as three more elements, i.e. projects’ benefit, cost and technical interdependencies,
which influence this calculation in a way that the portfolio’s return is inferior or superior
to the sum of the individual returns of all its projects (Spradlin & Kutoloski, 1999; Carazo,
2015). Benefit interdependencies derive from two or more projects producing greater or
less benefits when carried out simultaneously than if they were accomplished at different
times (Carazo, 2015). For example, suppose that the objective of one project is to develop
a new product feature that would make an existing product more attractive to a certain
demographic group and would have the effect of increasing sales by some amount as the
benefit. Suppose the objective of another project is to develop a second feature that would
make the same product more attractive to a different demographic group and would also
increase sales as the benefit. If both projects were funded and successful, sales may show
a net increase that is less than the sum of the two increases. In other words, it is possible
that the introduction of both new features would make the product less attractive to the
demographic groups than expected when considering the two features independently (Fox
et al., 1984). Cost interdependencies originate when the simultaneous implementation of
75
two or more projects require less or more resources than if they were carried out
separately. It implies that the required cost of a project portfolio is inferior or superior to
the sum of the costs of all its projects (Carazo, 2015). For example, two products might
share development resources, so that the cost of developing both might be less than the
sum of the individual projects’ costs (Spradlin & Kutoloski, 1999). Technical
interdependencies take place when the accomplishment of a determined project
necessarily involves the total or partial accomplishment or non-accomplishment of
another project or projects (Chien, 2002). For example, a project may be technologically
infeasible unless several enabling projects are undertaken to close existing technology
gaps (Czajkowski & Jones, 1986). Accordingly, it can be asserted that:
Design principle 6 (P6). Project portfolio’s return is also evaluated based on the
benefit, cost and technical interdependencies exist among the projects of that
portfolio.
Figure 2 demonstrates the above-mentioned six design principles of project
portfolio’s attractiveness evaluation, upon which an optimization model will be
developed in the next Section.
------------------------------------
Insert Figure 2 about here
------------------------------------
76
4. MODEL DEVELOPMENT
Here, we first introduce our model’s assumptions and the threat and opportunities
identification, then propose a customized approach to estimate MPT’s parameters as well
as benefit, cost and technical interdependencies in PPSP and finally develop the
optimization model.
4.1. Model assumptions
To develop the optimization model, we consider the following model assumptions
which are common in most PPSP models (Sefair et al., 2017; Ball & Savage, 1999;
Mutavdzic & Maybee, 2015):
Funding period is fixed. The output of the model would be the best subset
of project proposals to be funded in a planning period of time [0 T]
assuming all projects start at time 0.
Projects are not divisible. The decision variables are binary, representing
the selection, or not, of each project proposal. In other words, one cannot
fund a proportion of a project (e.g. building half a bridge).
Projects are considered before mitigation actions. The returns and risks
of projects before any risk mitigation actions are implemented to reflect the
available information during the project selection decision point in time.
Only new project proposals are considered. The selection decision takes
into account new project proposals, but does not deal with project
adjustments and terminations.
77
4.2. Threats and Opportunities Identification
According to design principles depicted in Figure 2, to develop our model, we
first need to identify threats and opportunities. The risk register is an effective project
management tool that includes this information used as part of a business case (Project
Management Institute, 2017). Because most projects include a risk register as a core
management tool, this paper uses data included in this document to estimate MPT’s
parameters in order to customize MPT for PPSP. There are different formats proposed by
various studies and standards for developing a risk register. One of the most
comprehensive formats is that proposed by Zwikael and Smyrk (2011), in which a risk
register is a table where rows are associated with threats and columns are relevant to their
attributes. Through applying the third model assumption, we use the first four columns of
their proposed format as follows:
1) Threat ID
2) Threat title: description of the triggering event
3) Likelihood of the threat in the absence of the proposed mitigation action
4) The damaging impacts of the threat on the project’ return in the absence of the
proposed mitigation action as at least one of the six potential effects: benefit
reduced, benefit delayed, disbenefit increased, disbenefit advanced, cost
increased, and cost advanced.
Furthermore, in order to generalize our formulation, we apply some modifications
in the above-mentioned risk register’s format. We consider both threats and opportunities
78
to cover both negatives and positive triggering events. Thus, the third column
demonstrates the likelihood of corresponding threat or opportunity materializing.
Moreover, there are six additional potential assisting impacts of opportunities: benefit
increased, benefit advanced, disbenefit decreased, disbenefit delayed, cost decreased, and
cost delayed, in addition to the above-mentioned potential damaging impacts of threats
in the fourth column.
Accordingly, Table 4 exemplifies the required part of a risk register that is used
in estimating MPT’s parameters explained in the rest of this section.
------------------------------------
Insert Table 4 about here
------------------------------------
4.3. The customized approach to estimate MPT’s parameters in PPSP
Here, we elaborate how projects’ risk registers can be applied to estimate MPT’s
parameter, i.e. projects’ returns and risks as well as risk interdependencies among them.
In the course of a project’s life, any number of the threats and opportunities mentioned in
its risk register can materialize. Thus, the total number of potential situations can surround
project i, im , is 2 inreached from Equation (8) in which in is the number of threats and
opportunities mentioned in the risk register of project i.
... 20 1 2 3
iii i i i n
i
i
nn n n nm
n
(8)
79
On the other hand, we employ “Modified Internal Rate of Return” (MIRR) to
extract a proper representative for projects’ returns, in which benefits and disbenefits can
be both monetary and non-monetary. MIRR, assumes that benefits/disbenefits generated
from a project are reinvested at the risk-free rate of return rather than at the project’s
internal rate of return (Lin, 1976). Equation (9) represents this assumption regarding
continuous compounding instead of discrete one, in which non-monetary
benefits/disbenefits can be converted to monetary ones (dolor values) by using some
techniques such as Delphi (Abbassi, Ashrafi, & Tashnizi, 2014) which its explanation is
out of the scope of this paper.
.MIRR T
Futurevalue of benefits and disbenefitsPresent valueof costs
e (9)
Where: e is Euler's number which represents continuous compounding.
Regarding the fact that the return of a project portfolio should be obtained by
proportionate combination of projects’ return (Findlay et al., 1981), here we consider
.MIRR Te as the representative of a project’s return. Furthermore, according to Table 4, each
threat and opportunity can affect the “magnitude” or “scheduling for realization” of
different benefits, disbenefits or costs. Thus, the return of project i in potential situation
s, isr , is calculated as follows by considering the fact that benefits and disbenefits are
continuous flows, whereas costs are discrete ones which only occur during the project’s
life:
is isis
is
FB FDr
PC
(10)
80
( )
11
( )
11
1( )
(1 )
(1 )
(1 )
i
f
ik ikuB Bik iku
u s
f
i ik iku
B Bik ikuu s
f B Bik iku ik ikuu sf
KT r T t
is B BS S
k u s
r T tT
K B B
u s
S Sk f
r T S SB Br T Tu s
f
Where FB M M e dt
M M e
r
M M
e er
1
1
1
(1 )
1
i
i f B Bik iku ik ikuu s
K
k
K r T S SB B
u s
k f
M M
er
(11)
( )
11
( )
11
1( )
(1 )
(1 )
(1 )
i
f
il iluD Dil ilu
u s
f
i il ilu
D Dil iluu s
f D Dil ilu il iluu sf
LT r T t
is D DS S
l u s
r T tT
L D D
u s
S Sl f
r T S SD Dr T Tu s
f
FD M M e dt
M M e
r
M M
e er
1
1
1
(1 )
1
i
i f D Dil ilu il iluu s
L
l
L r T S SD D
u s
l f
M M
er
(12)
1
1
1i f C Cif ifu
u s
if ifu
F r S S
is C C
f u s
PC M M e
(13)
Where:
isFB is the total future value of benefits relevant to project i at time T if potential situation
s materializes,
81
isFD is the total future value of disbenefits relevant to project i if potential situation s
materializes,
isPC is the total present value of costs relevant to project i if potential situation s
materializes,
iK is the total number of benefits relevant to project i,
iL is the total number of disbenefits relevant to project i,
iF is the total number of costs relevant to project i,
ikBM is the estimated magnitude of kth benefit relevant to project i,
ikuBM is the estimated percentage of changes in the magnitude of kth benefit relevant to
project i if threat/opportunity u materializes,
ikBS is the estimated scheduling for the realization of kth benefit relevant to project i,
ikuBS is the estimated percentage of changes in the scheduling for the realization of kth
benefit relevant to project i if threat/opportunity u materializes,
ilDM is the estimated magnitude of lth disbenefit relevant to project i,
iluDM is the estimated percentage of changes in the magnitude of lth disbenefit relevant
to project i if threat/opportunity u materializes,
ilDS is the estimated scheduling for the realization of lth disbenefit relevant to project i,
iluDS is the estimated percentage of changes in the scheduling for the realization of lth
disbenefit relevant to project i if threat/opportunity u materializes,
ifCM is the estimated magnitude of fth cost relevant to project i,
82
ifuCM is the estimated percentage of changes in the magnitude of fth cost relevant to
project i if threat/opportunity u materializes,
ifCS is the estimated scheduling for the spending of fth cost relevant to project i, and
ifuCS is the estimated percentage of changes in the scheduling for the spending of fth cost
relevant to project i if threat/opportunity u materializes.
On the other hand, the likelihood of situation s materializing in project i, isp , is
calculated as the multiplication of the likelihoods of the threats/opportunities included in
situation s materializing and likelihoods of the others not materializing, demonstrated as
follows:
(1 )is u u
u s u s
p l l
(14)
Where: ul is the likelihood of threat/opportunity u materializing.
After placing Equations (8), (10) and (14) in Equations (6) and (7), the final
customized estimation for return ( ˆi ) and risk (standard deviation, ˆ
i ) relevant to project
i can be reached as follows:
2
1
ˆ (1 )
ni
is isi u u
s u s u s is
FB FDl l
PC
(15)
22
2
1
ˆ ˆ(1 )
ni
is isi u u i
s u s u s is
FB FDl l
PC
(16)
Where isFB , isFD and isPC are calculated by Equations (11), (12) and (13) respectively.
Furthermore, the estimated cost of project i, ˆi , can be calculated as follows:
83
2
1
ˆ (1 )
ni
i u u is
s u s u s
l l PC
(17)
Having calculated returns and risks of projects, in the following, we propose a
methodology to calculate risk interdependencies among projects in a portfolio. Each one
of the projects in the pairs of projects has two categories of threats/opportunities:
particular threats/opportunities (which are specific to each one) and common ones (which
are common in the two projects). It should also be mentioned that a threat for a project
can be a threat or opportunity to another project. To estimate the risk interdependencies,
i.e. correlation coefficient, between projects i and j, ˆij , we first estimate their covariance,
ˆij , as follows:
2 2
1 1
2 2 2 2 2 2
1 1 1 1
2 2
1 1
ˆ ˆ ˆ( , )
,
,
,
n nji
n n n n nnij ij ji j jii
ij ij ij ij ji ji ji ji
ij ij ji ji
n ni ji
ij ij ji ji
ij ji
ij i j
is is js js
s s
ic ic ia ia jc jc ja ja
c a c a
ic ic jc jc
c c
COV
COV p r p r
COV p r p r p r p r
COV p r p r
2 2 2
1 1
2 2 2
1 1
2 2 2 2
1 1
,
,
,
j
n n nij j ji
ij ij ji ji
ij ji
n nn ij jii
ij ij ji ji
ij ji
n n nn ij j jii
ij ij ji ji
ij ji
ic ic ja ja
c a
ia ia jc jc
a c
ia ia ja ja
a a
COV p r p r
COV p r p r
COV p r p r
(18)
Where:
84
ijn is equal to jin and is the number of common threats and opportunities in projects i
and j,
ijc and jic are the cth common situations in projects i and j respectively,
ijicp and jijcp are the likelihoods of common situations
ijc and jic materializing in projects
i and j respectively,
ijicr and jijcr are the returns of projects i and j in common situations
ijc and jic
respectively,
ija is the ath particular situation in project i when it is compared to project j,
ijiap is the likelihood of particular situation ija materializing in project i, and
ijiar is the return of project i in particular situation ija .
As we can see in Equation (18), there are four terms which should be calculated
to reach the total covariance between projects i and j. By considering some assumptions
as there is no covariance among particular threats/opportunities of project i and those of
project j, and there is no covariance among common threats/opportunities in project i and
particular ones in project j and vice versa, the second, third and fourth terms of Equation
(18) are equal to zero. These assumptions of independence are justified because no
explicit relationships exist among these combinations of threats/opportunities. In other
words, if one occurs in project i, it does not give us any new information on occurrence
of the other one in Project j (Bakhshi & Touran, 2012). for the same reason, in the first
term of Equation (18), only the likelihoods of common threats and opportunities
materializing in projects i and j should be considered in calculating ijicp and
jijcp , which
85
makes these two likelihoods equal to each other. On the other hand, if there are no
common threats/opportunities in projects i and j, their covariance would be equal to zero.
Accordingly, the final formulation of covariance between projects i and j is as follows:
2 2
1 1
2 2 2
1 1 1
ˆ ,
0 /
n nij ji
ij ij ji ji
ij ji
n n nij ij ji
ij ij ij ij ji ji ji
ij ij ji
ij ic ic jc jc
c c
ic ic ic ic jc jc jc
c c c
COV p r p r
If nocommon threats pportunities exist between
projects i and j
p r p r r p r Oth
erwise
(19)
Having considered the relationship between correlation coefficient and covariance
as demonstrated in Equation (20), the final formulation of risk interdependency between
projects i and j is drawn from Equation (21).
ˆˆ
ˆ ˆ
ij
ij
i j
(20)
2 2 2
1 1 1
1
0 & /ˆ
ˆ ˆ
n n nij ij ji
ij ij ij ij ji ji ji
ij ij ji
ij
ic ic ic ic jc jc jc
c c c
i j
If i j
If i j Nocommon threats opportunities exist between
projects i and j
p r p r r p r
Otherwise
(21)
4.4. Benefit, cost and technical interdependencies among projects
86
Benefit interdependency between projects i and j, ijB , represents the difference
between the benefits of project i, when it is funded along with project j and when it is
funded in the absence of project j, calculated as follows:
( )
( )
( )
1
f
ijBij
f
ij
Bij
f Bij ijf
f Bij ij
T r T t
ij BS
r T tT
B
Sf
r T SB r T T
f
r T SB
f
B M e dt
M e
r
Me e
r
Me
r
(22)
Where:
ijBM is the estimated magnitude of benefit interdependency between projects i and j, and
ijBS is the estimated scheduling for the realization of benefit interdependency between
projects i and j.
Cost interdependency between projects i and j, ijC , represents the difference
between the costs of project i, when it is funded along with project j and when it is funded
in the absence of project j, calculated as follows:
f Cij
ij
r S
ij CC M e
(23)
Where:
ijCM is the estimated magnitude of cost interdependency between projects i and j, and
ijCS is the estimated scheduling for the realization of cost interdependency between
projects i and j.
87
Technical interdependencies among projects are added to the optimization model
as constraints (Carazo, 2015; Guo et al., 2008), which can be divided into three main
categories as follows:
Mutual exclusion technical interdependency: This type of technical
interdependency occurs when projects i and j cannot be funded simultaneously,
which is represented by Equation (24).
1i jX X (24)
Where:
iX and jX are binary decision variables, representing the selection, or not, of
projects i and j respectively.
Prerequisite technical interdependency: This type of technical interdependency
implies the situation in which project i cannot be funded unless project j also be
funded. In other words, if project j is rejected, project i have to also be rejected.
Equation (25) indicates prerequisite technical interdependency between projects i
and j.
0i jX X (25)
Compulsory technical interdependency: This kind of technical interdependency
occurs in three situations: (1) when project i is required to be funded, (2) when
project i is required to be abandoned, and (3) when project i is funded if and only
if project j is funded, that is, projects i and j should be both either funded or
abandoned simultaneously. These three types of Compulsory technical
interdependency are demonstrated in Equations (26), (27) and (28) respectively.
88
1iX (26)
0iX (27)
0i jX X (28)
4.5. The optimization model
Here, we develop an optimization model by summarizing the following model
parameters:
ˆpS : The estimated “attractiveness”, i.e. the Sharpe ratio, of project portfolio p
ˆp : The estimated expected return (mean) of project portfolio p
ˆp : The estimated risk (standard deviation) of project portfolio p
fr : The risk-free rate of return (e.g., the return of placing money in the bank or the return
of a well-established production line),
ˆi : The estimated expected return of project i drawn from Equation (15)
ˆi : The estimated risk (standard deviation) of project i achieved by Equation (16)
ijB : The benefit interdependency between projects i and j reached by Equation (22)
ijC : The cost interdependency between projects i and j achieved by Equation (23)
ˆij : The estimated risk interdependency between projects i and j drawn from Equation
(21)
M i : The set of projects with which project i has mutual exclusion technical
interdependencies
89
P i : The set of projects that are the prerequisites of project i
F : The set of projects required to be funded
A : The set of projects required to be abandoned
C i : The set of projects with which project i should be either funded or abandoned
simultaneously
: The total available budget
ˆi : The estimated cost of project i drawn from Equation (17)
N : The number of project proposals
iX : Binary decision variable, representing the selection, or not, of project i
Furthermore, since surplus budget is rarely left idle, we assume that the investor
can always invest unallocated budget in a risk-free asset with return fr T
e (Findlay et al.,
1981). Thus, we add the following decision variable to the optimization model.
oX : The decision variable of the budget proportion invested in a risk-free asset
Accordingly, the final non-linear optimization model of customizing MPT to
PPSP is developed as follows:
ˆˆ
ˆ
fr T
p
p
p
eMax S
(29)
0
1 1
1 ˆˆ ˆ f
N Nr T
p i i i j ij
i j
Where X X B X e
(30)
2
21 1
1 ˆ ˆ ˆˆ ˆ ˆN N
p i j i j i j ij
i j
X X
(31)
90
0
1 1
ˆN N
i i j ij
i j
Subject to X X C X
(32)
1 ; , 1,2,...,i jX X j M i i N (33)
1 ;iX i F (34)
0 ;iX i A (35)
0 ; , 1,2,...,i jX X j P i i N (36)
0 ; , 1,2,...,i jX X j C i i N (37)
1
1N
i
i
X
(38)
0,1 ; 1,2,...,iX i N (39)
0 0X (40)
Where: the objective function, i.e. Equation (29), maximizes the attractiveness of project
portfolio. Constraint (32) maintains the total available budget. The projects’ technical
interdependencies are guaranteed in Equations (33) to (37). Constraint (38) assures the
selection of at least one project. Equation (39) and (40) denotes the domains of the
variables.
5. A NUMERICAL EXAMPLE
The purpose of this section is to employ a numerical example in order to illustrate
how the optimization model proposed in Section 4 should be applied. Let us assume that
an organization considers 15 project proposals for a period of 24 months regarding the
91
risk-free rate of return of 0.5% per month. With a limited budget of $160,000, the
organization wants to select the best set of projects to fund. Each project proposal includes
a business case with a risk register. Tables 4 and 5 depict two relevant data of the risk
registers developed for projects 1 and 2 respectively. As can be seen, the projects have
two common triggering events, as P1.T01/P2.T01 (“Exchange rate increases”) with the
likelihood of 0.3, and P1.T02/P2.O03 (“Iron import law is passed by the government”)
with the likelihood of 0.2. Furthermore, it is derived that the former plays the role of threat
for both projects, while the latter has a role of threat for project 1 and that of opportunity
for project 2.
------------------------------------
Insert Table 5 about here
------------------------------------
We demonstrate how to calculate the return and risk of project 1. Regarding
Equation (8), as this project has three threats/opportunities, there are eight potential
situations to materialize. Table 6 shows these potential situations, the likelihood of each
situation materializing ( 1sp ) drawn from Equation (14), the return and cost of project 1
in each situation ( 1sr and 1sPC respectively) calculated by applying Equations (10) to (13).
------------------------------------
Insert Table 6 about here
------------------------------------
Applying Equations (15), (16) and (17), the estimated return, risk and cost of
project 1 are calculated as follows respectively:
92
1ˆ ( ) (0.216 11.82) (0.126 14.03) ( 14.28)
(
0.504 14.97 0.056
0.054 11.09 0.024 0.014 13.35 0.006 10.46 13.) ( 11.19) ( ) ( ) 79
1
2 2 2
1
2 2
2 2
2 2
ˆ ( ) 0.216 (11.82 )
0.126 (14.03 ) ( )
0.504 14.97 13.79 13.79
13.79 0.056 14.28 13.79
0.054 11.09 13.79 0.024 11.19 13.79
0.014 13.35 13.79 0.006 10.46 13.79 2.1876
ˆ 1.
( ) ( )
( )
8
)
4
(
1 0.504 0.056 14,608
0.054 15,86
ˆ ( 14,608) (0.216 15,867) (0.126 14,608) ( )
( ) ( ) ( ) ( )
$
7 0.024 15,867 0.014 14,608 0.006 15,867
14,986
Similarly, the estimated return, risk and cost of project 2 are 17.24, 1.71 and
$18,595 respectively. Table 7 shows the estimated returns, risks and costs of projects 1
and 2 plus the assumed those of the other 13 project proposals, which can be reached by
applying the same approach.
------------------------------------
Insert Table 7 about here
------------------------------------
To extract the risk interdependency between projects 1 and 2, we consider their
two above-mentioned common threats/opportunities. According to Equation (8) there are
four common situations in these two projects. Table 8 shows these common situations,
the likelihood of each situation materializing (121cp ) drawn from Equation (14), and the
returns of projects 1 and 2 in each common situation (as represented by 121cr and
212cr
respectively) have been calculated in the previous step.
------------------------------------
93
Insert Table 8 about here
------------------------------------
According to Equations (19), (20) and (21), the estimated covariance and risk
interdependency between projects 1 and 2 are calculated as follows:
12 12
12
21 21
21
4
1 1
1
4
2 2
1
( ) (0.24 11.82) (0.14 14.03) ( 11.09) 13.85
( ) (0.24 15.41) (0.14 19.00) ( 16.68) 17.2
0.56 14.97 0.06
0.56 17.73 0.06 8
c c
c
c c
c
p r
p r
12 21ˆ ˆ ( )( )
(11.82 13.85)(15.41 17.28)
(14.03 13.85)(19.0
0.56 14.97 13.85
0 17.28
17.73 17.28
0.24
0.14
0
)
(11.09 13.85)(16.68 17.. 28) 1.0 346
12 21
1.34ˆ ˆ 0.53
1.48 1.71
By the same way, assume that the risk interdependency matrix relevant to all 15
project proposals are as follows:
1 0.53 0 0.30 0.02 0 0.03 0 0 0.08 0 0 0.23 0 0.10
0.53 1 0 0.1 0 0 0.21 0 0.41 0.07 0 0.27 0 0.1 0.10
0 0 1 0 0 0.10 0 0.22 0 0.09 0 0 0.12 0 0.03
0.30 0.1 0 1 0 0.05 0 0.12 0 0.19 0 0 0 0.26 0
0.02 0 0 0 1 0 0.03 0.13 0.21 0 0.14 0 0.27 0 0
0 0 0.10 0.05 0 1 0 0.19
0.11 0.09 0.21 0.06 0.03 0.18 0
0.03 0.21 0 0 0.03 0 1 0.30 0.13 0 0 0.12 0 0.22 0.41
0 0 0.22 0.12 0.13 0.19 0.30 1 0 0.19 0.08 0.20 0 0 0.11
0 0.41 0 0 0.21 0.11 0.13 0 1 0.10 0.10 0 0.09 0.10 0.08
0.08 0.07 0.09 0.19 0 0.09 0 0.19 0.10 1 0.21 0.05 0
.05 0.12 0.14
0 0 0 0 0.14 0.21 0 0.08 0.10 0.21 1 0 0.11 0 0.13
0 0.27 0 0 0 0.06 0.12 0.20 0 0.05 0 11 0 0.04 0
0.23 0 0.12 0 0.27 0.03 0 0 0.09 0.05 0.11 0 1 0.14 0.09
0 0.1 0 0.26 0 0.18 0.22 0 0.10 0.12 0 0.04 0.14 1 0
0.10 0.10 0.03 0 0 0 0.41 0.11 0
.08 0.14 0.13 0 0.09 0 1
94
Furthermore, the organization develops the magnitude and realization scheduling
Matrixes of benefit interdependency relevant to 15 project proposals, i.e. BM and BS
respectively, as the following,
0 1000 0 0 0 0 3500 700 0 0 0 0 0 0 0
1000 0 0 200 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 100 0 0 0 0 0 0 0
0 200 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1000 0 0 0 0 0 0 0
0 0 0 0 0 0 2000 0 450 0 0 0 0 0 0
3500 0 0 0 0 2000 0 0 700 0 0 0 0 0 0
700 0 100 0 1000 0 0 0 0 0 0 0 500 0 0
0 0 0 0 0 450 700 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
BM
0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 500 0 0 0 0 0 0 3000
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 3000 0 0
0 12 0 0 0 0 12 20 0 0 0 0 0 0 0
12 0 0 10 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 18 0 0 0 0 0 0 0
0 10 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 15 0 0 0 0 0 0 0
0 0 0 0 0 0 18 0 13 0 0 0 0 0 0
12 0 0 0 0 18 0 0 10 0 0 0 0 0 0
20 0 18 0 15 0 0 0 0 0 0 0 11 0 0
0 0 0 0 0 13 10 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
BS
0 0 0
0 0 0 0 0 0 0 11 0 0 0 0 0 0 14
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 14 0 0
And the magnitude and realization scheduling Matrix of cost interdependency
relevant to 15 project proposals, i.e. CM and CS respectively, as follows:
95
0 500 0 0 0 0 1000 0 400 0 0 0 0 0 0
500 0 200 0 0 0 0 0 0 0 0 0 0 0 0
0 200 0 0 0 0 0 0 0 320 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 150 0 0 0 0 0 0 0
0 0 0 0 0 0 700 0 0 0 0 0 120 0 0
1000 0 0 0 0 700 0 0 0 0 0 0 1000 0 0
0 0 0 0 150 0 0 0 80 0 0 0 280 0 0
400 0 0 0 0 0 0 80 0 0 0 0 0 0 0
0 0 320 0 0 0 0 0 0
CM
0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 90 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 120 1000 280 0 0 0 0 0 0 500
0 0 0 0 0 0 0 0 0 0 90 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 500 0 0
0 5 0 0 0 0 6 0 10 0 0 0 0 0 0
5 0 7 0 0 0 0 0 0 0 0 0 0 0 0
0 7 0 0 0 0 0 0 0 5 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 11 0 0 0 0 0 0 0
0 0 0 0 0 0 8 0 0 0 0 0 8 0 0
6 0 0 0 0 8 0 0 0 0 0 0 5 0 0
0 0 0 0 11 0 0 0 7 0 0 0 11 0 0
10 0 0 0 0 0 0 7 0 0 0 0 0 0 0
0 0 5 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 9 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 8 5 11 0 0
CS
0 0 0 0 5
0 0 0 0 0 0 0 0 0 0 9 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 5 0 0
As an example, we calculate the benefit and cost interdependencies between
projects 1 and 2 through applying Equations (22) and (23) respectively as follows:
0.005(24 12)
12
0.005 5
12
1000( 1)$12,367
0.005
500 $ 488
eB
C e
Through using the similar approach, the benefit and cost interdependencies
matrixes relevant to all 15 projects are reached as follows:
96
0 12367 0 0 0 0 43286 2828 0 0 0 0 0 0 0
12367 0 0 2900 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 609 0 0 0 0 0 0 0
0 2900 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 9206 0 0 0 0 0 0 0
0 0 0 0 0 0 12182 0 5089 0 0 0 0 0 0
43286 0 0 0 0 12182 0 0 10151 0 0 0 0 0 0
2828 0 609 0 9206 0 0 0 0 0 0 0 6716 0 0
0 0 0 0 0 5089 10151 0 0 0 0
B
0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 6716 0 0 0 0 0 0 30763
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 30763 0 0
0 488 0 0 0 0 970 0 380 0 0 0 0 0 0
488 0 193 0 0 0 0 0 0 0 0 0 0 0 0
0 193 0 0 0 0 0 0 0 312 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 142 0 0 0 0 0 0 0
0 0 0 0 0 0 673 0 0 0 0 0 115 0 0
970 0 0 0 0 673 0 0 0 0 0 0 975 0 0
0 0 0 0 142 0 0 0 77 0 0 0 265 0 0
380 0 0 0 0 0 0 77 0 0 0 0 0 0 0
0 0 312 0 0 0 0 0 0 0 0 0 0
C
0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 86 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 115 975 265 0 0 0 0 0 0 488
0 0 0 0 0 0 0 0 0 0 86 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 488 0 0
Furthermore, the organization extracts technical interdependencies among the
project proposals as demonstrated in Table 9.
------------------------------------
Insert Table 9 about here
------------------------------------
Having considered the non-linear optimization model presented in Equations (29)
to (40), we employed the global solver of the optimization modeling software “Lingo 17”
97
to reach the optimal portfolio. Accordingly, the optimal project portfolio, P, consists
funding projects 1, 2, 5, 6, 7, 9, 11 and 13 that requires a budget of $145,930 out of
$160,000 total available budget. The rest of the budget, i.e. $14,070 is invested in the
risk-free asset. Accordingly, the estimated expected return, risk and attractiveness of
optimal portfolio P are reached as 13.77, 0.46 and 27.64 respectively. Figure 3 shows the
optimal project portfolio P and efficient frontier of the discussed numerical example.
In order to compare our model with those that do not consider either risk
interdependencies or benefit and cost interdependencies among projects, we solved the
numerical example disregarding the benefit, cost and risk interdependencies among
projects. Disregarding the risk interdependencies among projects, the optimal project
portfolio, W, consists funding projects 1, 2, 5, 6, 7, 11, 14 and 15 as well as investing $3,698
in the risk-free asset. Furthermore, the estimated expected return, risk and attractiveness
of optimal portfolio W are reached as 13.69, 0.51 and 24.82 respectively. Disregarding
the benefit and cost interdependencies among projects, the optimal project portfolio, Y,
consists funding projects 2, 5, 6, 9, 12, 13 and 15 as well as investing $12,961 in the risk-
free asset. Moreover, the estimated expected return, risk and attractiveness of optimal
portfolio Y are reached as 11.96, 0.44 and 24.54 respectively. Finally, Disregarding the
risk, benefit and cost interdependencies among projects, the optimal project portfolio, Z,
consists funding projects 1, 5, 6, 7, 9, 11, 13 and 15 as well as investing $5,898 in the risk-
free asset. Moreover, the estimated expected return, risk and attractiveness of optimal
portfolio Z are reached as 13.26, 0.51 and 23.64 respectively. Optimal project portfolios
W, Y and Z are depicted in Figure 3.
98
------------------------------------
Insert Figure 3 about here
------------------------------------
As can be seen from Figure 3, our model reaches better results than those models
which do not consider either the risk interdependencies or the benefit and cost
interdependencies among projects in the optimization model of PPSP.
6. CONCLUSIONS
PPSP is a key decision that allows organizations achieving their strategic goals.
As a project portfolio with a high expected return may also expose the organization to a
large loss, the combined analysis of return and risk (“risk-return optimization”) should be
considered in the optimization model of PPSP. A similar approach is applied in PSP, i.e.
the leading financial theory “MPT” (Markowitz, 1952, 1959). Furthermore, project
portfolio’s returns and risks are influenced by some threats and opportunities. These
threats and opportunities may affect the returns of one or more projects simultaneously
and cause risk interdependencies among projects. Thus, it is crucial to incorporate the
effects of threats and opportunities into the optimization model of PPSP. Accordingly,
we proposed a new model to such incorporation inspired by MPT and customized to PPSP
through considering critical information that is not discussed in other PPSP models, i.e.
particular and common threats/opportunities exist around projects.
Theoretically, this paper builds synthesized coherence (Locke & Golden-Biddle,
1997) across PPSP and MPT to take into account the risk interdependencies among
projects to the PPSP model. Moreover, the paper contributes to the PPSP literature by
99
incorporating the effects of threats and opportunities around projects into the optimization
model through customizing MPT to PPSP. It also expands MPT to portfolio decisions
where some projects are also considered and make it suitable in the project environment
by taking into account the effects of threats and opportunities. In practice, having used
risk registers which are crucial, well established and practiced documents exist in most
projects, the proposed solution is asserted to be practical and effective in order to improve
the quality of portfolio selection decision in organizations.
Our paper also has some limitations that need to be acknowledged. First, as we
apply projects’ risk register as the source of information to develop the optimization
model, adopting the limitation of this document is undeniable. As a result, this paper does
not consider unknowns-unknowns, i.e. uncertainties which are not known at the
beginning of the project (Lechler, Edington, & Gao, 2012), in the model. Finding a way
to take into account unknown-unknowns in the optimization model can improve its
quality. Second, we assume that all projects start at the same time. Expanding our study
to the scheduling of projects in the portfolio and also scheduling of activities in projects
can enrich the generalizability of the model. Third, this paper applies the characteristics
of projects before doing any mitigation actions, so other research can investigate the
effects of candidate mitigating actions on the proposed model. Fourth, we only consider
the selection decision of new project proposals in model development. Future studies can
incorporate project adjustment to the optimization model. Fifth, we use some sample data
to demonstrate how our model should be applied. Future studies can test the model and
design principles using data from real projects. Finally, future research can investigate
100
different ways to incorporate non-monetary benefits/disbenefits to the optimization
model.
101
REFERENCES
Abbassi, M., Ashrafi, M., & Tashnizi, E. S. 2014. Selecting balanced portfolios of R&D
projects with interdependencies: A Cross-Entropy based methodology.
Technovation, 34(1): 54-63.
Albala, A. 1975. Stage approach for the evaluation and selection of R&D projects. IEEE
Transactions on Engineering Management, (4): 153-164.
Bakhshi, P., & Touran, A. 2012. A method for calculating cost correlation among
construction projects in a portfolio. International Journal of Architecture,
Engineering and Construction, 1(3): 134-141.
Ball Jr, B. C., & Savage, S. L. 1999. Holistic vs. Hole-istic E&P strategies. Journal of
Petroleum Technology, 51(09): 74-84.
Benjamin, C.O. 1985. A linear goal-programming model for public-sector project
selection. Journal of the Operational Research Society, 36(1): 13-23.
Boasson, V., Cheng, J., & Boasson, E. 2012. Applying modern portfolio theory to
municipal financial and capital budgeting decisions. Public and Municipal
Finance, 1(1): 58-65
Bodie, Z., Kane, A., & Marcus, A. J. 2014. Investments (10th ed.). New York: McGraw-
Hill Education.
Carazo, A. F. 2015. Multi-criteria project portfolio selection. In J. Zimmermann & C.
Schwindt (Eds.), Handbook on project management and scheduling vol. 2: 709-
728. Switzerland: Springer International Publishing.
102
Carraway, R. L., & Schmidt, R. L. 1991. An improved discrete dynamic programming
algorithm for allocating resources among interdependent projects. Management
Science, 37(9): 1195-1200.
Casault, S., Groen, A. J., & Linton, J. D. 2013. Selection of a portfolio of R&D
projects. In A. N. Link & N. S. Vonortas (Eds.), Handbook on the theory and
practice of program evaluation: 89-112. Massachusetts: Edward Elgar publishing.
Chien, C. F. 2002. A portfolio-evaluation framework for selecting R&D projects. R&D
Management, 32(4): 359–68.
Christensen, M., & Knudsen, T. 2010. Design of decision making-organizations.
Management Science, 56(1): 71–89.
Cooper, M.J. 1978. An evaluation system for project selection. Research Management,
21(4): 29-33
Czajkowski, A. F., & Jones, S. 1986. Selecting interrelated R & D projects in space
technology planning. IEEE Transactions on Engineering management, (1): 17-24.
Dey, P. K. 2006. Integrated project evaluation and selection using multiple-attribute
decision-making technique. International Journal of Production
Economics, 103(1): 90-103.
Doerner, K. F., Gutjahr, W. J., Hartl, R. F., Strauss, C., & Stummer, C. 2006. Pareto ant
colony optimization with ILP preprocessing in multiobjective portfolio selection.
European Journal of Operational Research, 171(3): 830–841.
103
Esfahani, H. N., Sobhiyah, M. H., & Yousefi, V. R. 2016. Project portfolio selection via
harmony search algorithm and modern portfolio theory. Social and Behavioral
Sciences, 226: 51 – 58.
Findlay, M. C., McBride, R. D., Yormark, J. S., & Messner, S. D. (1981). Mean-variance
analysis for indivisible assets. Omega, 9(1): 77-88.
Fox, G. E., Baker, N. R., & Bryant, J. L. 1984. Economic models for R and D project
selection in the presence of project interactions. Management science, 30(7): 890-
902.
Gear, T. E., & Cowie, G. C. 1980. A note on modeling project interdependence in research
and development. Decision Sciences, 11(4): 738–748.
Gear, T.E., Lockett, A.G., & Muhlemann, A.P. 1982. A unified approach to the
acquisition of subjective data in R&D. IEEE Transactions on Engineering
Management, 29(1): 11-19.
Ghasemzadeh, F., Archer, N., & Iyogun, P. 1999. A zero-one model for project portfolio
selection and scheduling. Journal of the Operational Research Society, 50(7): 745–
755.
Ghasemzadeh, F., & Archer, N. P. 2000. Project portfolio selection through decision
support. Decision support systems, 29(1): 73-88.
Golabi, K., Kirkwood, C. W., & Sicherman, A. 1981. Selecting a portfolio of solar energy
projects using multiattribute preference theory. Management Science, 27(2): 174–
189.
104
Gregor, S., & Jones, D. 2007. The anatomy of a design theory. Journal of the Association
for Information Systems, 8(5): 312-335.
Guo, P., Liang, J. J., Zhu, Y. M., & Hu, J. F. 2008. R&D project portfolio selection model
analysis within project interdependencies context. In Industrial Engineering and
Engineering Management, 2008. IEEM 2008. IEEE International Conference
on (pp. 994-998). IEEE.
Hillson, D. 2002. Extending the risk process to manage opportunities. International
Journal of project management, 20(3): 235-240.
Klapka, J., & Piños, P. 2002. Decision support system for multicriterial R&D and
information systems projects selection. European Journal of Operational
Research, 140(2): 434–446
Krawiec, F. 1984. Evaluating and selecting research projects by scoring. Research
Management, 27(2): 21-25.
Lechler, T. G., Edington, B. H., & Gao, T. 2012. Challenging classic project management:
Turning project uncertainties into business opportunities. Project Management
Journal, 43(6): 59-69.
Lee, J. W., & Kim, S. H. 2000. Using analytic network process and goal programming
for interdependent information system project selection. Computers & Operations
Research, 27(4): 367–382.
Lewis, M. W., Welsh, M. A., Dehler, G. E., & Green, S. G. 2002. Product development
tensions: Exploring contrasting styles of project management. Academy of
Management Journal, 45(3): 546-564.
105
Li, X., Fang, S. C., Tian, Y., & Guo, X. 2015. Expanded model of the project portfolio
selection problem with divisibility, time profile factors and cardinality constraints.
Journal of the Operational Research, 66(7): 1132–1139.
Lin, S. A. 1976. The modified internal rate of return and investment criterion. The Engineering
Economist, 21(4), 237-247.
Locke, K., & Golden-Biddle, K. 1997. Constructing opportunities for contribution:
Structuring intertextual coherence and “problematizing” in organizational
studies. Academy of Management journal, 40(5): 1023-1062.
Lorie, J. H., & Savage, L. J. 1955. Three problems in rationing capital. The journal of
business, 28(4): 229–239.
Luo, L.-M. 2012. Optimal diversification for R&D project portfolios. Scientometrics,
91(1): 219–229.
Lwin, K., & Qu, R. 2013. A hybrid algorithm for constrained portfolio selection
problems. Applied intelligence, 39(2): 251-266.
Markowitz, H. 1952. Portfolio selection, The Journal of Finance, 7(1): 77–91.
Markowitz, H. M. 1959. Portfolio selection: efficient diversification of investments.
New York: Wiley.
Markowitz, H. M., Todd, G. P., & Sharpe, W. F. 2000. Mean-variance analysis in
portfolio choice and capital markets, vol. 66. New York: Wiley.
Martino, J. P. 1995. Research and development project selection, vol. 23. New York:
Wiley.
106
Melachrinoudis, E., & Rice, K. 1991. The prioritization of technologies in a research
laboratory. IEEE Transactions on Engineering Management, 38(3): 269-278.
Moriarty, N. 2001. Portfolio risk reduction: Optimising selection of resource projects by
application of financial industry techniques. Exploration Geophysics, 32(3/4): 352-
356.
Mutavdzic, M., & Maybee, B. 2015. An extension of portfolio theory in selecting projects
to construct a preferred portfolio of petroleum assets. Journal of Petroleum Science
and Engineering, 133: 518-528.
Myers, S. C. 1972. A note on linear programming and capital budgeting. The Journal of
Finance, 27(1): 89–92.
Neely, W.P., North, R.M., & Fortson, J.C. 1977. An operational approach to multiple
objective decision making for public water resources projects using integer goal
programming. American Journal of Agricultural Economics, 59(1): 198-203.
Oral, M., Kettani, O., & Lang, P. 1991. A methodology for collective evaluation and
selection of industrial R&D projects. Management Science, 37(7): 871-885.
Project Management Institute. 2017. A Guide to the project management body of
knowledge (6th ed.). Pennsylvania: Project Management Institute.
Santhanam, R., & Kyparisis, J. 1995. A multiple criteria decision model for information
system project selection. Computers & Operations Research, 22(8): 807-818.
Santhanam, R., & Kyparisis, G. J. 1996. A decision model for interdependent information
system project selection. European Journal of Operational Research, 89(2): 380–
399.
107
Schmidt, R. L. 1993. A model for R&D project selection with combined benefit, outcome
and resource interactions. IEEE Transactions on Engineering Management, 40(4):
403–410.
Sefair, J. A., Méndez, C. Y., Babat, O., Medaglia, A. L., & Zuluaga, L. F. 2017. Linear
solution schemes for Mean-SemiVariance Project portfolio selection problems: An
application in the oil and gas industry. Omega, 68: 39-48.
Sharpe, W. F. 1994. The sharpe ratio. The Journal of Portfolio Management, 21(1): 49-
58.
Shou, Y., Xiang, W., Li, Y., & Yao, W. 2014. A multiagent evolutionary algorithm for
the resource-constrained project portfolio selection and scheduling problem.
Mathematical Problems in Engineering, 2014: 1-9.
Smith, K. V., & Smith, J. A. 2005. Strategies in Personal Finance: Basic Investment
Principles for Today and Tomorrow. Purdue University Press.
Spradlin, C. T., & Kutoloski, D. M. 1999. Action-oriented portfolio
management. Research-Technology Management, 42(2): 26-32.
Varian, H. 1993. A portfolio of Nobel laureates: Markowitz, Miller and Sharpe. The
Journal of Economic Perspectives, 7(1): 159-169.
Vepsalainen, A.P.J., & Lauro, G.L. 1988. Analysis of R&D portfolio strategies for
contract competition. IEEE Transactions on Engineering Management, 35(3):
181-186.
Weingartner, H. M. 1962. Mathematical programming and the analysis of capital
budgeting problems. New Jersey: Prentice-Hall.
108
Westerveld, E. 2003. The Project Excellence Model®: linking success criteria and critical
success factors. International Journal of project management, 21(6): 411-418.
Yu, L., Wang, S., Wen, F., & Lai, K. K. 2012. Genetic algorithm-based multi-criteria
project portfolio selection. Annals of Operations Research, 197(1): 71–86.
Zwikael, O., & Smyrk, J. 2011. Project management for the creation of organisational
value. London: Springer Science & Business Media.
Zwikael, O., & Smyrk, J. 2012. A general framework for gauging the performance of
initiatives to enhance organizational value. British Journal of
Management, 23(S1).
109
Table 1
A comparison of PPSP definitions in the literature
Source PPSP definition Objective of PPSP Constraints of
PPSP
Ghasemzadeh
& Archer
(2000)
Project portfolio selection is the
periodic activity involved in
selecting a portfolio of projects,
that meets an organization's
stated objectives without
exceeding available resources or
violating other constraints.
Selecting a portfolio
of projects, that meets
an organization's
stated objectives
Without
exceeding
available
resources or
violating other
constraints
Shou et al.
(2014)
Given a set of project proposals
and constraints, the traditional
PPSP is to select a subset of
project proposals to optimize the
organization’s performance
objective.
To select a subset of
project proposals to
optimize the
organization’s
performance
objective.
Given a set of
constraints
Carazo
(2015)
The problem of selecting a
project portfolio arises from the
everyday dilemma faced by
organizations in finding the best
possible way to distribute a
limited budget among candidate
projects to fulfil the needs of the
organization.
Finding the best
possible way to
distribute … budget
among candidate
projects to fulfil the
needs of the
organization.
distribute a
limited budget
Li et al.
(2015)
The problem of how to select a
portfolio that meets a
firm’s/organization’s objectives
without violating indispensable
constraints is called a project
portfolio selection problem
(PPSP).
To select a portfolio
that meets a
firm’s/organization’s
objectives
without
violating
indispensable
constraints
110
Table 2
The similarities between PSP and PPSP
# Similar concepts between
PSP and PPSP Explanation
1 Assets’/projects’ return Both assets and projects have their own expected returns.
2 Assets’/projects’ risk There are uncertainties in the expected returns of both
assets and projects.
3
Correlations among assets/
risk interdependencies
among projects
The expected returns of both assets and projects are affected
by some threats and opportunities like the exchange rate
changes, which in turn can result in correlations among
assets or risk interdependencies among projects. For
example, a change in the exchange rate can increase the
expected return of an asset or project and simultaneously
lower that of another asset or project.
4 Available budget Both investors and organizations have a limited budget to
distribute among a set of assets or projects.
111
Table 3
The differences between the approaches used in PSP with those should be used in PPSP
# Parameter PSP PPSP
1 Required data
An appropriate amount of
reliable historical data is
available.
Often there is not enough historical
data or the accessible historical data is
not reliable enough, as the threats and
opportunities faced by a project can be
unique and not possible to generalize
to other projects.
2
Mean as the
measure of
expected return
( i )
Assets’ expected returns are
estimated by using a
“backward approach”. In
other words, the mean of
historical returns relevant to
asset i is used to estimate its
expected return.
Projects’ expected returns should be
estimated by using a “forward
approach”. In other words, some
forecasting techniques specific to
project management should be used to
estimate the expected return relevant
to project i.
3
Standard deviation
as the measure of
risk ( i )
Assets’ risks are estimated
by the standard deviations
of their historical returns.
Projects’ risks should be estimated by
using some forecasting techniques
specific to project management.
4
Correlation
coefficient among
assets / risk
interdependencies
among projects
( ij )
The correlations among
assets are calculated by
using “Pearson correlation
coefficient formula” for
their historical returns.
The risk interdependencies among
projects should be calculated by
considering the common threats and
opportunities around them.
5 Benefit / Disbenefit
Only monetary benefits and
disbenefits, measured in
private terms, are
considered in assets’
expected returns.
Both monetary and non-monetary
benefits and disbenefits, measured in
organizational terms, should be
considered in projects’ expected
returns.
6 Benefit
interdependency
There are not any benefit
interdependencies among
assets.
Benefit interdependencies among the
projects of a portfolio should be
considered in that portfolio’s return
calculation. Such interdependencies
occur when two or more projects
producing greater or less benefits if
carried out simultaneously than if they
were accomplished at different times.
112
7 Cost
interdependency
There are not any cost
interdependencies among
assets.
Cost interdependencies among the
projects of a portfolio should be
considered in that portfolio’s return
calculation. Such interdependencies
originate when the simultaneous
implementation of two or more
projects require less or more resources
than if they were carried out
separately. It implies that the required
cost of a project portfolio is inferior or
superior to the sum of the costs of all
its projects.
8 Technical
interdependency
There are not any technical
interdependencies among
assets.
Technical interdependencies among
the projects of a portfolio should be
considered in that portfolio’s return
calculation. Such interdependencies
take place when the accomplishment
of a determined project necessarily
involves the total or partial
accomplishment or non-
accomplishment of another project or
projects.
113
Table 4
Project 1’s risk register
ID Threat / Opportunity
Lik
elih
ood
The damaging/assisting impacts of the threat/opportunity
Benefit Disbenefit Cost
B1 B2 B3 D1 C1 C2 C3
M
($) S
(Mo)*
M
(I)**
S
(Mo)
M
(I)
S
(Mo)
M
(I)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
6000 10 3000 11 9000 12 1000 5 5000 1 5000 5 5000 10
∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S
P1.T01 Exchange rate increases 30% -15% +10% -30% -10% +10% -8% +16%
P1.T02 Iron import law is
passed by the parliament 20% -20% +20%
P1.T03 Design changes 10% -9%
Where:
“M” is the “estimated/realized magnitude” of benefits, disbenefits or costs in the absence of any threats and opportunities,
“S” is the “estimated/realized scheduling” for the realization of benefits or disbenefits or the spending of costs in the absence of any threats and opportunities,
“∆M” is the “estimated percentage of changes in the magnitude” of benefits, disbenefits or costs (in two directions: increased or decreased, represented by “+” and
“–” respectively) resulted from corresponding threat or opportunity materializing, and
“∆S” is the “estimated percentage of changes in the scheduling” for the realization of benefits or disbenefits or spending of costs (in two directions: delayed or
advanced, represented by “+” and “–” respectively) resulted from corresponding threat or opportunity materializing.
* Mo: Month
** I: Index (the unit of non-monetary benefits/disbenefits can be converted to monetary values by Delphi approach)
114
Table 5
Project 2’s risk register
ID Threat / Opportunity
Lik
elih
ood
The damaging/assisting impacts of the threat/opportunity
Benefit Disbenefit Cost
B1 B2 D1 D2 C1 C2 C3
M
($)
S
(Mo)
M
(I)
S
(Mo)
M
(I)
S
(Mo)
M
(I)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
M
($)
S
(Mo)
9000 9 15000 10 800 7 500 5 15000 1 2000 4 2000 7
∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S ∆M ∆S
P2.T01 Exchange rate increases 30% -20% -20%
P2.O02 Project manager leaves 10% +7% -5% -12% -15%
P2.O03 Iron import law is
passed by the parliament 20% +10% -15%
P2.T04 Customer requirement
changes 20% +12% -10% +15%
115
Table 6
The return of project 1 in different potential situations
Situation
(s) Threat/opportunity
included in the situation s
likelihood of situation s
materializing
(1sp )
Return
(1sr )
Cost
( 1sPC )
1 N/A 50.4% 14.97 $14,608
2 P1.T01 21.6%* 11.82** $15,867***
3 P1.T02 12.6% 14.03 $14,608
4 P1.T03 5.6% 14.28 $14,608
5 P1.T01 & P1.T02 5.4% 11.09 $15,867
6 P1.T01 & P1.T03 2.4% 11.19 $15,867
7 P1.T02 & P1.T03 1.4% 13.35 $14,608
8 P1.T01 & P1.T02 & P1.T03 0.6% 10.46 $15,867
* According to Equation (14): 0.3×(1-0.2)×(1-0.1) = 0.216
** According to Equations (10) to (13):
0.005(24 10(1 0.10)) 0.005(24 11(1 0)) 0.005(24 12(1 0))
12
0.005(24 5(1 0.10))
12
12
6000(1 0.15) ( 1) 3000(1 0.30) ( 1) 9000(1 0) ( 1)
0.005 0.005 0.005$208,015
1000(1 0) ( 1)
0.005$20,482
$15,867***
e e eFB
eFD
PC
12
208,015 20,48211.82
15,867r
*** According to Equation (13):
0.005 1(1 0) 0.005 5(1 0.08) 0.005 10(1 0)
12 5000(1 0) 5000(1 0.1) 5000(1 0.16)$15,867
PC e e e
Table 7
116
The estimated returns, risks and costs relevant to 15 project proposals
Project number
( i )
Estimated Return
( ˆi )
Estimated Risk
( ˆi )
Estimated cost
( ˆi )
1 13.78 1.48 $14,986
2 17.24 1.71 $18,595
3 13.58 1.19 $21,644
4 17.34 2.70 $16,499
5 12.39 1.38 $18,925
6 11.36 1.16 $17,568
7 13.67 1.25 $22,297
8 12.96 1.64 $31,467
9 15.78 1.77 $20,478
10 17.32 2.95 $15,445
11 12.49 1.46 $16,404
12 14.15 1.87 $22,031
13 9.53 1.10 $23,880
14 9.21 1.00 $26,799
15 13.44 1.16 $24,817
117
Table 8
The returns of projects 1 and 2 in their common potential situations
Situation
( 12c )
Threat/opportunity
included in the
situation
likelihood of situation
s materializing
(121cp )
Return of
project 1
(121cr )
Return of
project 2
(212cr )
1 N/A 56% 14.97 17.73
2 P1.T01/P2.T01 24%* 11.82 15.41
3 P1.T02/P2.O03 14% 14.03 19.00
4
P1.T01/P2.T01
&
P1.T02/P2.O03 6% 11.09 16.68
* According to Equation (14): 0.3×(1-0.2) = 0.24
118
Table 9
Technical interdependencies among 15 project proposals
Type of technical interdependencies Sets
Mutual exclusion
(1) 3
(3) 1
(4) 8
(8) 4( ) 2,5,6,7,9,10,11,12,13,14,15
M
M
M
MM i i
Prerequisite
(1) 11
(3) 10
(7) 5( ) 2,4,5,6,8,9,10,11,12,13,14,15
P
P
PP i i
Compulsory
(9) 13
(13) 9( ) 1,2,3,4,5,6,7,8,10,11,12,14,15
FAC
CC i i
119
Figure 1
The efficient frontier of risky assets with the optimal capital allocation line, CAL (P), and P as
the optimal portfolio-Source: Bodie et al. (2014, p. 221)
120
Figure 2
The design principles of project portfolio’s attractiveness evaluation
P6
P5
P3
P2
P1
P4
Threats & Opportunities
Individual projects’ risks
Risk interdependencies
among projects
Project portfolio’s risk
Project portfolio’s
attractiveness
Project portfolio’s return
Individual projects’
returns
Benefit interdependencies
among projects
Cost interdependencies
among projects
Technical interdependencies
among projects
121
Figure 3
The efficient frontier of the numerical example
0
2
4
6
8
10
12
14
16
18
20
22
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8σ
μ
CAL (P)
Y
PZ
Efficient Frontier
W
122
CHAPTER 4: CONCLUSIONS
123
Because an organization’s performance depends on the projects it implements,
project evaluation and PPSP have become two crucial topics in project management to
appraise and select project proposals, monitor projects during their lives and judge projects’
performances upon their completions. In project appraisal, as the first stage of project
evaluation which is conducted in the initiation phase, those project proposals that are
unacceptable for funding are removed. In project selection (PPSP), which is conducted at the
end of the initiation phase, the best portfolio of project proposals is selected to optimize
organizational performance given limited resources. In project monitoring, as the second
stage of project evaluation which in conducted in the planning, execution and benefit
realization phases, the ongoing performance of a selected project is compared to its initial
goals, to understand what went right or wrong in order to improve the strategy or the
processes. In project performance judgement, as the third stage of project evaluation which
is conducted after benefit realization phase, the realized project performance of a selected
project is measured to judge whether the initial goals have been achieved and enhance
organizational learning in order to achieve successful projects in future.
In project evaluation and selection, not only delivering the outputs is important, but
also realizing the projects’ benefits is crucial. Furthermore, as the intended benefits are
affected by some threats and opportunities around projects, it is crucial to incorporate such
effects into the project evaluation framework and project portfolio selection model. To do
such incorporation, I proposed (1) a new project evaluation framework inspired by utility
124
theory, and (2) a new PPSP model inspired by MPT. Thus, the two research questions
developed in Chapter 1 can now be answered: (1) For paper 1 - the principles of utility theory
are applied to introduce a “project evaluation map”, as the foundation of the integrated
benefit-oriented project evaluation framework, including “Attractiveness contours”,
“Approved business case contour” and “Funder investment frontier”; and (2) For paper 2 -
the principles of MPT are applied to add risk interdependencies among projects into the PPSP
model through substituting the projects’ historical data with information from their risk
registers and introducing a new approach to estimate risk, return and risk interdependencies
among projects.
Theoretically, this thesis contributes to the project management literature by
developing an integrated framework for all three stages of project evaluation regarding
projects’ returns and risks. Furthermore, it contributes to the PPSP literature by incorporating
the effects of threats and opportunities around projects into the optimization model through
customizing MPT to PPSP. It also expands MPT to portfolio decisions where some projects
are also considered and make it suitable in the project environment by taking into account
the effects of threats and opportunities. In practice, the proposed evaluation framework and
PPSP model can improve the quality of portfolio selection decision in organizations and in
turn lower their resource wastes, as well as help them ensuring the realization of projects’
intended benefits.
125
As the project management discipline is now also taking into consideration projects’
benefits as the underlying rationale for all projects, this thesis opens a new horizon in benefit-
based project evaluation and selection for future studies. An extension of this thesis can also:
(1) take into account unknown-unknowns in the project evaluation framework and PPSP
model; (2) test the framework and model in various industries, countries and cultures; (3) add
project adjustment, the effects of risk mitigation actions, scheduling of projects in the
portfolio, and scheduling of activities in projects to the proposed PPSP model; and (4),
investigate different ways to incorporate non-monetary benefits and disbenefits into the
project evaluation framework and PPSP model.