B.

Quantitative Risk Analysis

Quantitative risk

analysis is a way of numerically estimating the probability that a project will

meet its cost and time objectives. Quantitative analysis is based on a

simultaneous evaluation of the impact of all identified and quantified risks.

The result is a probability distribution of the project’s cost and completion date based on the risks in the project.

The quantitative

methods rely on probability distribution of risks and may give more objective

results than the qualitative methods, if sufficient current data is available.

On the other hand, qualitative methods depend on the personal judgment and past

experiences of the analyst and the results may vary from person to person.

Hence the quantitative methods are preferred by most analysts.

Quantitative

risk analysis considers the range of possible values for key variables, and the

probability with which they may occur. Simultaneous and random variation within

these ranges leads to a combined probability that the project will be

unacceptable Quantitative risk analysis involves statistical techniques that

are most easily used with specialized software. Quantitative risk analysis is

to assign probabilities or likelihood to the various factors and a value for

the impact then identify severity for each factor.

When thorough

quantitative risk analysis is necessary it can take two alternative approaches

(Kuismanen, 2001):

1. Risks can be

quantified as individual entities while looking at the big picture. This way

can include the cumulative effects (to certain accuracy) into each individual

risk and thus make more accurate estimations of the net value of the risks.

2.

Alternatively modeling the mathematical properties of

the interrelations from the bottom up can be started and then calculate the net

impact of each risk including the effects of interrelations.

In Figure 2.4 the basic steps of a

quantitative risk analysis and a simplified relationship between risk analysis,

risk assessment and risk management is presented.

· Basic Steps of quantitative risk analysis

As discussed previously, the aim of

risk analysis is to determine how likely an adverse event is to occur and the

consequences if it does occur. When quantitative risk analysis is to be done,

it is attempted to describe risk in numerical terms. To do this, it should go

through a number of steps:

1. Define the

consequence; define the required numerical estimate of risk.

2. Construct a

pathway; consider of all sequential events that must occur for the adverse

event to occur.

3. Build a model –

Collect data; consider each step on the pathway and the corresponding variables

for those steps.

4. Estimate the

risk; once the model has been constructed and the data collected the risk can

be estimated. Included in this estimation will be an analysis of the effects of

changing model variables to reflect potential risk management strategies.

5. Undertake a

sensitivity and scenario analysis; Undertaking a risk analysis requires more

information than for sensitivity analysis.

· Methods of Quantitative Risk Analysis

Any specific risk analysis technique

is going to require a strategy. It is best to begin by providing a way of

thinking about risk analysis that is applicable to any specific tool might be

used.

· Probability Analysis is a tool in investigating problems which do not have a

single value solution, Monte Carlo

Simulation is the most easily used form of probability analysis.

· Monte Carlo Simulation is presented as the technique of

primary interest because it is the

tool that is used most often.

· Sensitivity Analysis is a tool that has been used to great extent by most

risk analysts at one time to another.

· Breakeven Analysis is an application of a sensitivity analysis. It can be

used to measure the key variables

which show a project to be attractive or unattractive.

· Scenario Analysis is a rather grand name for another derivative of

sensitivity analysis technique which

tests alternative scenarios; the aim is to consider various scenarios as

options.

Sensitivity Analysis and Monte Carlo

Simulation are discussed briefly:

· Sensitivity Analysis

Sensitivity analysis is a

deterministic modeling technique which is used to test the impact of a change

in the value of an independent variable on the dependent variable. Sensitivity

analysis identifies the point at which a given variation in the expected value

of a cost parameter changes a decision. Sensitivity analysis is performed by

changing the values of independent risk variables to predict the economic

criteria of the project. Sensitivity analysis is an interactive process which

tells you what effects changes in a cost will have on the life cycle cost.

Sensitivity Analysis is the calculating procedure used for

prediction of effect of changes of input data on output results of one model. It

does not aim to quantify risk but rather to identify factors that are risk

sensitive.

Sensitivity analysis enables the

analyst to test which components of the project have the greatest impact upon

the results, thus narrowing down the main simplicity and ability to focus on

particular estimates. The advantage of sensitivity analysis is that it can

always be done to some extent. Specific scenarios of interest can be reasonably

well described. Extreme outcomes, like the maximum or minimum possible costs,

can often be estimated.

Figure 2.4 Simplified relationship

between risk analysis, risk assessment and risk management.

The major disadvantage of

sensitivity analysis is that the analyst usually has no idea how likely these

various scenarios are. Many people equate possible with probable, which is not

the case with sensitivity analysis.

· Monte Carlo Simulation

Simulation is a probability-based

technique where all uncertainties are assumed to follow the characteristics of

random uncertainty. A random process is where the outcomes of any particular

process are strictly a matter of chance. The Monte Carlo process is simply a

technique for generating random values and transforming them into values of

interest, the methods of generating random or pseudo random numbers are more

sophisticated now and the mathematics of other distributions is more complex.

Different values of risk variables are combined in a Monte Carlo simulation.

The frequency of occurrence of a particular value of any one of the variables

is determined by defining the probability distribution to be applied across the

given range of values. The results are shown as frequency and cumulative

frequency diagrams. The allocation of probabilities of occurrence to each risk

requires the definition of ranges for each risk. Following are the risk

analysis simulation steps:

1. Start with a

project estimate done for each cost account.

2. Decide on the

most likely cost, pessimistic costs, and optimistic costs.

3. Insert data

into simulation software, then run the model.

4. Determine

contingencies based on desired risk level.

5. Prioritize “risky” cost accounts for risk response planning.

This method of sampling (i.e. random

sampling) will, lead to over- and under-sampling from various parts of the

distribution. In practice, this means that in order to ensure that the input

distribution is well represented by the samples drawn from it, a very large

number of iterations must be made. In most risk analysis work, the main concern

is that the model or sampling scheme we use should reproduce the distributions

determined for the inputs.

2.6 Risk Response Practices

PMI (1996) suggested three ways of

responding to risk in projects, they are as follows:

· Avoidance:

eliminating a specific threat, usually by eliminating the cause. The project

management team can never eliminate all risks, but specific risk events can

often be eliminated.

· Mitigation:

reducing the expected monetary value at risk events by reducing the probability

of occurrence (e.g., using new technology), reducing the risk event value

(e.g., buying insurance), or both.

· Acceptance:

accepting the consequences. Acceptance can be active by developing a

contingency plan to execute should the risk event occur or passive by accepting

a lower profit if some activities overrun.

Abu Rizk (2003) suggested some

actions to be taken in response to residual risks. Actions can include:

· Reduce

uncertainty by obtaining more information, this leads to re-evaluation of the

likelihood and impact.

· Eliminate or

avoid the risk factor through means such as a partial or complete re-design, a

different strategy or method etc.

· Transfer the

risk element by contracting out affect work.

· Insure against

the occurrence of the factor.

· Abort the

project if the risk is intolerable and no other means can be undertaken to

mitigate its damages.

2.6.1 Risk Avoidance

Risk avoidance is sometimes referred

to as risk elimination. Risk avoidance in construction is not generally

recognized to be impractical as it may lead to projects not going ahead, a

contractor not placing a bid or the owner not proceeding with project funding

are two examples of totally eliminating the risks. There are

a number of ways through which risks can be avoided, e.g. tendering a very high

bid; placing conditions on the bid; pre-contract negotiations as to which party

takes certain risks; and not bidding on the high risk portion of the contract

(Flanagan & Norman, 1993).

2.6.2 Risk Transfer

This is essentially trying to

transfer the risk to another party. For a construction project, an insurance

premium would not relieve all risks, although it gives some benefits as a

potential loss is covered by fixed costs (Tummala & Burchett, 1999) Risk transfer

can take two basic forms:

· The property or

activity responsible for the risk may be transferred, i.e. hire a subcontractor

to work on a hazardous process;

· The property or

activity may be retained, but the financial risk transferred, i.e. by methods

such as insurance and surety.

2.6.3 Risk Retention

This is the method of reducing

controlling risks by internal management (Zhi, 1995); handling risks by the

company who is undertaking the project where risk avoidance is impossible,

possible financial loss is small, probability of occurrence is negligible and

transfer is uneconomic (Akintoyne & MacLeod,1997). The risks, foreseen or

unforeseen, are controlled and financed by the company or contractor. There are

two retention methods, active and passive;

a. Active retention (sometimes referred to as self-insurance) is a

deliberate management strategy after

a conscious evaluation of the possible losses and costs of alternative ways of

handling risks.

b. Passive retention (sometimes called non-insurance), however, occurs

through negligence, ignorance or absence

of decision, e.g. a risk has not been identified and handling the consequences

of that risk must be borne by the contractor performing the work.

2.6.4 Risk

Reduction

This is a general term for reducing

probability and/or consequences of an adverse risk event. In the extreme case,

this can lead to eliminate entirely, as seen in “risk avoidance”.

However, in reduction, it is not sufficient to consider only the resultant

expected value, because, if potential impact is above certain level, the risk

remains unacceptable.