When the cash flow component values are not normally
distributed or when other features of the model are uncertain,
simple combinations of means and variances are not sufficient
to obtain valid statistical results. For these situations, we
use Monte-Carlo simulation.
The first step toward simulation is to change the point estimates
to simulated values. This is done by selecting *Change Project* from
the* Economics *menu and selecting *Simulate * from
the point estimate options. Column K in the worksheet below holds
simulated values of the cash flow components. Cell S2 contains
the *NPW *for the simulated values for a single replication.
Choose *Simulate_RV* from the *Random
Variables * menu to obtain the dialog below. Enter S2 as
the cell to simulate because it holds the simulated *NPW*.
The dialog specifies that 1000 observations are to be simulated
and the check boxes indicate the statistics to be gathered.
Part of the results of a simulation run are shown
below along with the results computed by the *Economics
*add-in. The mean and standard deviation obtained by the
simulation are close to those obtained by the previous section
where normality was assumed.
The frequency chart created by the simulation defines
a random variable called *SimS2*. The name comes from
the address of the simulated cell. Functions available from the* Random
Variables *add-in compute probabilities and percentile levels
based on the simulated distribution.
Probabilities
and percentiles obtained through simulation do not require
the assumption of normality. |