Computers make it possible to simulate various economic and financial outcomes, using a large number of variables. Thus simulation is one way of dealing with the uncertainty involved in forecasting the outcomes of capital budgeting projects or other types of decisions. A Monte Carlo simulation model uses random variables for inputs. By programming the computer to randomly select inputs from probability distributions, the outcomes generated by a simulation are distributed about a mean, and instead of generating one return or net present value, a range of outcomes with standard deviations is provided. A simulation model relies on repetition of the same random process as many as several hundred times. Since the inputs are representative of what one might encounter in the real world, many possible combinations of returns are generated.
One of the benefits of simulation is its ability to test various possible combinations of events. This sensitivity testing allows the planner to ask “what if” questions: What will happen to the returns on this project if oil prices go up? Go down? What effect will a 2 percent increase in interest rates have on the net present value of this project? The analyst can use the simulation process to test possible changes in economic policy, sales levels, inflation, or any other variable included in the modeling process. Some simulation models are driven by sales forecasts with assumptions to derive income statements and balance sheets. Others generate probability acceptance curves for capital budgeting decisions by informing the analyst about the probabilities of having a positive net present value.
For example, each distribution in Figure 13-7 will have a value picked randomly and used for one simulation. The simulation will be run many times, each time selecting a new random variable to generate the final probability distribution for the net present value (at the bottom of Figure 13-7). For that probability distribution, the expected values are on the horizontal axis and the probability of occurrence is on the vertical axis. The outcomes also suggest something about the riskiness of the project, which is indicated by the overall dispersion.