Risk Management
Vignesh Sivadass
2022MANG6143slidesWeek10p.pptMANG 6143
Project Risk Management
Mario Brito
Recommended reading for this week
- Chapman(2019), Chapter 7, page 347-351
- Chapman and Ward (2011), Chapter 11 page 289-324
*
Learning outcomes
- Having learnt how to build sensitivity and decision diagrams, this week you will learn:
- How to add cumulative curves of the project cost to estimate the overall cost profile of a project
- How to use decision diagrams to make decisions about the best of two different choices using the risk efficiency concept
*
Flowchart
for the
basic SPP
identify all relevant sources,
responses & conditions
structure
all uncertainty
clarify
ownership
quantify
some uncertainty
evaluate
all the relevant implications
to the
appropriate
gateway
stage
capability-culture
assets
capability-culture
liabilities
create & enhance plans
for all relevant concerns
from
project
initiation or a
gateway
stage
select & focus the process
for appropriate clarity
capture the context
with appropriate clarity
shape base plans using
models of some key issues
Your task this week:
- Imagine that you are the manager of a project in the Execution and Delivery stage. The aim of the project is to design an eco friendly car rental app to manage car bookings. Now consider that you are in the Evaluate phase of the SPP. Can you give one example showing how sensitivity diagrams or decision diagrams can be used in this phase of the SPP for this project? Write a comment of approximately 150 words in the discussion board.
- Write a reply to a comment written by one of your colleagues.
*
Evaluate phase specific tasks
deliverables
fit for purpose?
from the
quantify
phase
diagnose
the implications
interim
or gateway
report
portray the effect
to the
capture
phase
no – restructure
or extend
global yes
combine the
subset of sources
select an appropriate
subset of sources
specify dependence
local yes
- Assumptions
- Independence
- Correlation
- Positive or negative
- Causation
- Cost of item A is dependent on variable X.
*
Adding sensitivity curves
*
Example: Discrete probability arithmetic
Assuming that we are combining the costs of item A and item B, each with the same probability distribution of costs represented by three values as shown in Table, defining Ca and Cb.
| Cost (k£), Ca and Cb | Probability |
| 8 | 0.2 |
| 10 | 0.5 |
| 12 | 0.3 |
*
0.40
0.45
0.50
Discrete density probability format
most likely value
(most probable
class mark)
Cost
Probability
6 7 8 9 10 11 12 13
0.05
0.15
0.25
0.35
0
0.10
0.20
0.30
expected value (point of balance)
*
1.0
Cost
0
6 7 8 9 10 11 12 13
0.2
0.4
0.6
median (50%) value
Cumulative
probability
0.8
Cumulative probability format
Example: Discrete probability arithmetic
The basic discrete probability arithmetic to calculate the distribution of Ci = Ca + Cb assuming the costs of A and B are independent)
*
| Cost (k£) | Probability computation | Probability |
| 16 | 0.2 x 0.2 | 0.04 |
| 18 | 0.2 x 0.5 + 0.5 x 0.2 | 0.20 |
| 20 | 0.2 x 0.3 + 0.5 x 0.5 + 0.3 x 0.2 | 0.37 |
| 22 | 0.5 x 0.3 + 0.3 x 0.5 | 0.30 |
| 24 | 0.3 x 0.3 | 0.09 |
*
Example: Casual model of dependence
*
The reason the costs of items A and B are correlated is that they will be the
responsibility of either contractor X or contractor Y, contractor X being very efficient and Y being very inefficient, but political pressures determining the outcome, it may be useful to model this choice directly.
Example: casual model of dependence
*
The basic discrete probability arithmetic to calculate the distribution of Cm = Ca + Cb assuming the costs of A and B are independent)
Sensitivity diagrams as a key tool: activity level example
Sensitivity diagrams as a key evaluate phase tool:
offshore project intermediate level output example
5th year
2nd year
3rd year
4th year
Award of design
contract (1)
1
2
3
4
5
6
Fabrication
complete (6)
Probability
of
achievement
by
dates
indicated
0.2
0.4
0.6
0.8
1.0
0
0.3
0.5
0.7
0.9
0.1
Award of fabrication contract (2)
and major material orders (3)
Steel (4) and other (5)
deliveries complete
Offshore installation
complete (7)
7
Sensitivity diagrams as a key evaluate phase tool:
addition of items A, B and C as in the BCS example
Cost
Cumulative
probability
0.2
0.4
0.6
0.8
1.0
0
A
A + B
A + B + C
P90
P10
Decision Theory – Portfolio theory – Risk frontier
*
Markowitz, H. (1952) Portfolio Selection. The Journal of Finance, 7, 77-91.
R= expected return on investment for a single stock investment, pj = probability of a given return E(Rj); E(Rj) a possible return for the stock investment.
For a portfolio with say two stocks, where you invest w1 on stock 1 and w2 on stock 2. The expected return is:
Prof. Harry Markowitz
An example on return versus risk in investments
- Suppose you want to invest £100000 in one of two stocks: stock 1 or stock 2.
*
| Stock 1 | |
| Probability | Return |
| 0.05 | 50% |
| 0.25 | 30% |
| 0.4 | 10% |
| 0.25 | -10% |
| 0.05 | -30% |
| Stock 2 | |
| Probability | Return |
| 0.3 | 30% |
| 0.25 | 20% |
| 0.2 | 10% |
| 0.15 | 5% |
| 0.1 | -20% |
| E(R) | 0.1 |
| E(R)^2 | 0.01 |
| E(R^2) | 0.046 |
| standard deviation | 0.190 |
| E(R) | 0.148 |
| E(R)^2 | 0.0218 |
| E(R^2) | 0.0434 |
| standard deviation | 0.147 |
Efficient Frontier Obtainable from Risky Investments
*
1.0
Standard deviation of return
0
6 7 8 9 10 11 12 13
0.2
0.4
0.6
0.8
Expected
return
Hull, John C. 2018. Risk Management and Financial Institutions, John Wiley & Sons: New Jersey.
Efficient Frontier
Decision Theory – Portfolio theory – Risk frontier
*
Prof. Chris Chapman
Using decision diagrams to evaluate risk efficient
choices: one risk efficient choice example
Using decision diagrams to evaluate risk-reward
trade-offs: two risk efficient choices example
Using decision diagrams to evaluate other
criteria trade-offs: the photocopier example
Managing opportunity efficiency is the overall goal,
with risk efficiency in relation to each relevant attribute,
plus optimal trade-offs between all relevant attributes,
plus optimal trade-offs between expected outcomes and risk, whether or not objectives are measurable.
Optimal trade-offs between all objectives is the goal,
including all relevant clarity efficiency concerns,
and it is crucial to systematically seek this goal.
Clarity/opportunity efficiency interdependence is important – they are in effect different views of the same basic ‘best practice’ concern.
Build comprehensively from the bottom-up,
moving the bottom down selectively,
then present selectively top-down
- Understand specific sources of uncertainty, responses and conditions at the bottom level, moving the bottom down where this seems worth while.
- Understand the role of dependence as uncertainty accumulates and decomposition is refined.
- Understand composite uncertainty and general responses at higher levels.
- Manage the uncertainty plus any associated opportunity and risk as the build-up evolves.
- Present the ‘bottom line’ first.
- Explain where it came from working top-down.
- Make sure conditions and non-quantified objectives do not get overlooked, emphasising the key ones.
- Focus all presentations on what matters most.
Approaches to combining uncertainty
- Methods based on moments (mean-variance approaches in particular).
- Methods based on discrete probabilities (as used in decision analysis and other probability tree models).
- Methods based on simulation (Monte Carlo simulation in particular).
- Functional integration.
- Hybrids (HAT approaches in particular).
The current common practice norm is using simulation without a gradual build up of intermediate results – no ‘sensitivity diagrams’ are the obvious symptom. Simulation can be used for any approach to any decomposition structure, and simulation is a very reasonable default assumption, provided it is used effectively, which a HAT approach should facilitate.
*
Approaches to dependence to be aware of
- Assuming independence as a default position is the norm, often without effectively testing the independence assumptions.
- Assuming 100% dependence (perfect positive correlation) is a very useful alternative default position, because perfect positive correlation is a simple form of complete dependence, which can be a much better default position for a minimum clarity starting position in some circumstances.
- Intermediate correlation and percentage dependence can provide a very useful simple compromise between 0% and 100% dependence.
- Conditional specifications exploiting the tree structure aspect of the HAT approach start to provide the insight needed to understand simple and complex forms of statistical dependence.
- Causal structures exploiting decision tree aspects of a HAT approach provide a practical maximum insight approach sometimes, but they may not be clarity efficient in some contexts.
- Cascade and positive feedback loop effects can be more extreme than simple 100% positive dependence – 100% dependence can be optimistic.
Part two review, questions and discussion
