Session 3.3 Sensitivity and Risk Analysis, Study notes of Risk Analysis

Download Session 3.3 Sensitivity and Risk Analysis and more Study notes Risk Analysis in PDF only on Docsity!

Session 3.

Sensitivity and Risk Analysis Introductory Course on Economic Analysis of

Investment Projects5-9 July 2010

Why Sensitivity and Risk Analysis

^ Project returns are spread over time

^ costs and

benefits are subject to uncertainty and may vary fromthe base case  Sensitivity and risk analysis is an analytical framework fordealing with uncertainty.

The objective is to reduce the

likelihood of undertaking bad projects while not failing toaccept good projects  References: Guidelines (1997), Handbook (2002),Technical Note 2 (2002)

Sensitivity Analysis: Procedure  Determine key variables and their possiblechanges  Re-calculate EIRR and NPV  Calculate Switching Value

Switching Value

^ Switching value is the value of the variable at which theproject investment decision is changed. Usually isdefined as percentage change from the base case. ^ Percentage change in a variable required to reduce thenet present value (NPV) to ‘zero’ at the chosen discountrate:Set NPV = 0

^ v*

^ SV = 100% x (v* - v

0 0 )/v

OR ^ Percentage change in a variable required to make theproject EIRR equal a chosen discount rate (e.g., 12%):Set EIRR = r

^ v*

^ SV = 100% x (v* - v

0 0 )/v

Risk Analysis

^ A method to estimate the probability distribution ofproject outcome: F(EIRR), F(NPV) ^ Determine a probability distribution of values of keyvariables: F(v

), F(v 1

), … , F(v 2

)n

^ Sample these values to determine a probabilitydistribution of outcome ^ Process greatly simplified by use of PCs andstandard software packages

Distribution for Economic Rate of Return

Mean = 0.

EIRR<=0.2995%

EIRR<=0.126.9189%

(^1) 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.^

0.^ 0.^ 0.^

Probability

EIRR

ERD’S Retro 2005 Findings

^ Sensitivity and risk analyses were inadequate orpoor in about half of the RRPs reviewed. ^ Most projects applied sensitivity test but theanalysis is generally limited to a mechanistic “plus10% (20%) project costs”, “minus 10% (20%) ofproject benefits”, or “1-year delay in projectimplementation” (same as in Retros 2002 and2003).

Note that the focus of sensitivity test is on specific variables to inform project design.  Only a small number of projects conductedquantitative risk analysis.

Final Notes

^ There may be other risks very important to assessbut cannot quantify (e.g., political and institutionalrisks). ^ They are really a part of sensitivity and riskanalysis and generally treated under qualitativerisk analysis (e.g., risk matrix of typically low,medium, high impact). ^ When risk is high (e.g., very small switching value,high probability that EIRR<EOCC), need to workout mitigating measures or change project design