It is important to keep in mind that when a company analyzes a potential project, it is forecasting potential not actual cash flows for a project. As we all know, forecasts are based on assumptions that may be incorrect. It is therefore important for a company to perform a sensitivity analysis on its assumptions to get a better sense of the overall risk of the project the company is about to take.
There are three risk-analysis techniques that should be known for the exam:
1.Sensitivity analysis2.Scenario analysis3.Monte Carlo simulation
1.Sensitivity AnalysisSensitivity analysis is simply the method for determining how sensitive our NPV analysis is to changes in our variable assumptions. To begin a sensitivity analysis, we must first come up with a base-case scenario. This is typically the NPV using assumptions we believe are most accurate. From there, we can change various assumptions we had initially made based on other potential assumptions. NPV is then recalculated, and the sensitivity of the NPV based on the change in assumptions is determined. Depending on our confidence in our assumptions, we can determine how potentially risky a project can be.
2.Scenario AnalysisScenario analysis takes sensitivity analysis a step further. Rather than just looking at the sensitivity of our NPV analysis to changes in our variable assumptions, scenario analysis also looks at the probability distribution of the variables. Like sensitivity analysis, scenario analysis starts with the construction of a base case scenario. From there, other scenarios are considered, known as the "best-case scenario" and the "worst-case scenario". Probabilities are assigned to the scenarios and computed to arrive at an expected value. Given its simplicity, scenario analysis is one the most frequently used risk-analysis techniques.
3.Monte Carlo SimulationMonte Carlo simulationis considered to be the "best" method of sensitivity analysis. It comes up with infinite calculations (expected values) given a number of constraints. Constraints are added and the system generates random variables of inputs. From there, NPV is calculated. Rather than generating just a few iterations, the simulation repeats the process numerous times. From the numerous results, the expected value is then calculated.
