Risk Simulator

Sample Models
2005 Copyright. Dr. Johnathan Mun.
All Rights Reserved.

Tornado, Spider and Sensitivity Charts

This sample model illustrates how to use Risk Simulator for:
1. Running a pre-simulation sensitivity analysis (Tornado and Spider Charts)
2. Running a post-simulation sensitivity analysis (Sensitivity Charts)
Model Background
File Name: Tornado and Sensitivity Charts (Linear).xls
This example illustrates a simple discounted cash flow model and shows how sensitivity analysis can be performed prior to running a simulation
and after a simulation is run. Tornado and Spider charts are static sensitivity analysis tools useful for determining which variables impact the key
results the most. That is, each precedent variable is perturbed a set amount and the key result is analyzed to determine which input variables are
the critical success factors with the most impact. In contrast, Sensitivity charts are dynamic, in that all precedent variables are perturbed together
in a simultaneous fashion (the effects of autocorrelations, cross-correlations, and interactions are all captured in the resulting Sensitivity chart).
Creating a Tornado and Sensitivity Chart
To run this model, simply:
1. Go to the DCF Model worksheet and select the NPV result (cell G6).
2. Select Simulation l Tools l Tornado Analysis (or click on the Tornado Chart icon).
3. Check that the software's intellegent naming is correct for the precedent values and click OK.
Interpreting the Results
The report generated illustrates the sensitivity table (starting base value of the key variable as well as the perturbed values and the precedents).
The precedent with the highest impact (range of output) is listed first. The Tornado chart illustrates this analysis graphically. The Spider chart
performs the same analysis but also accounts for nonlinear effects. That is, if the input variables have a nonlinear effect on the output variable, the
lines on the Spider chart will be curved. Rerun the analysis on the Black-Scholes Model sheet in the Tornado and Sensitivity Charts (Nonlinear)
example file.
Creating a Sensitivity Chart
To run this model, simply:
1. Create a new simulation profile called (Simulation l New Profile)
2. Set the relevant assumptions and forecasts on the DCF Model worksheet
3. Run the simulation (Simulation l Run Simulation)
4. Select Simulation l Tools l Sensitivity Analysis
Interpreting the Results
Notice that if correlations are turned off, the results of the Sensitivity chart are similar to the Tornado chart. Now, reset the simulation, and
turn on correlations (select Simulation l Reset Simulation then select Simulation l Edit Profile and check Apply Correlations,and then
Simulation l Run Simulation), and repeat the steps above for creating a Sensitivity chart. Notice that when correlations are applied, the resulting
analysis may be different due to the interactions among variables. Of course you will need to set the relevant correlations among the assumptions.
Tip
Sometimes, the chart axis variable names can be too long. If that happens, simply rerun the Tornado but this time, truncate or rename some of
the long variable names to something more concise and the charts will be more visually appealing.
Disclaimer
DEVELOPER SPECIFICALLY DISCLAIMS ALL OTHER WARRANTIES, EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. As standard practice for software development and end-user applications, it is important for
the Licensee to note that the valuation results attached herein are accurate to the software Developers best knowledge and are solely based on the information furnished
by the Licensee or end-user. While the software Developer has used his best efforts in preparing this report, he makes no representations or warranties with respect to
the accuracy or completeness of the contents of this model and specifically disclaims any implied warranties of merchantability of fitness for a particular purpose. The
Licensee hereby agrees that the Developer is not held liable for any loss of profit or any other commercial damages, including, but not limited to, special, incidental,
consequential or other damages. This model is only an illustration of using the software and in no way represents the correct and complete picture of an investor's
investment, risk, and return profile. The user is advised to take great care in using and interpreting the model and its results.

Copyright 2005. Real Options Valuation, Inc.

Discounted Cash Flow Model

Base Year
Market Risk-Adjusted Discount Rate
Private-Risk Discount Rate
Annualized Sales Growth Rate
Price Erosion Rate
Effective Tax Rate
Set Assumptions Here...

2005
15.00%
5.00%
2.00%
5.00%
40.00%

Sum PV Net Benefits

Sum PV Investments
Net Present Value
Internal Rate of Return
Return on Investment

$1,896.63
$1,800.00
$96.63
18.80%
5.37%

2005

2006

2007

2008

2009

Change in Net Working Capital

Capital Expenditures
Free Cash Flow

$10.00
$12.25
$15.15
50.00
35.00
20.00
$1,231.75
$184.76
$1,046.99
$157.50
$15.75
$873.74
$10.00
$3.00
$860.74
$2.00
$858.74
$343.50
$515.24
$13.00
$0.00
$0.00
$528.24

$9.50
$11.64
$14.39
51.00
35.70
20.40
$1,193.57
$179.03
$1,014.53
$160.65
$16.07
$837.82
$10.00
$3.00
$824.82
$2.00
$822.82
$329.13
$493.69
$13.00
$0.00
$0.00
$506.69

$9.03
$11.06
$13.67
52.02
36.41
20.81
$1,156.57
$173.48
$983.08
$163.86
$16.39
$802.83
$10.00
$3.00
$789.83
$2.00
$787.83
$315.13
$472.70
$13.00
$0.00
$0.00
$485.70

$8.57
$10.50
$12.99
53.06
37.14
21.22
$1,120.71
$168.11
$952.60
$167.14
$16.71
$768.75
$10.00
$3.00
$755.75
$2.00
$753.75
$301.50
$452.25
$13.00
$0.00
$0.00
$465.25

$8.15
$9.98
$12.34
54.12
37.89
21.65
$1,085.97
$162.90
$923.07
$170.48
$17.05
$735.54
$10.00
$3.00
$722.54
$2.00
$720.54
$288.22
$432.33
$13.00
$0.00
$0.00
$445.33

Investments

$1,800.00

$440.60
$0.00
$506.69

$367.26
$0.00
$485.70

$305.91
$0.00
$465.25

$254.62
$0.00
$445.33

Prod A Avg Price

Prod B Avg Price
Prod C Avg Price
Prod A Quantity
Prod B Quantity
Prod C Quantity

Total Revenues
Cost of Goods Sold

Gross Profit
Operating Expenses
SG&A Costs

Operating Income (EBITDA)

Depreciation
Amortization

EBIT
Interest Payments

EBT
Taxes

Net Income
Depreciation

Financial Analysis
Present Value of Free Cash Flow
Present Value of Investment Outlay
Net Cash Flows

$528.24
$1,800.00
($1,271.76)

Copyright 2005. Real Options Valuation, Inc.

Tornado and Spider Charts

Statistical Summary

One of the powerful simulation tools is the tornado chartit captures the static impacts of each variable on the outcome of the model. That is, the tool automatically perturbs each precedent
variable in the model a user-specified preset amount, captures the fluctuation on the models forecast or final result, and lists the resulting perturbations ranked from the most significant to
the least. Precedents are all the input and intermediate variables that affect the outcome of the model. For instance, if the model consists of A = B + C, where C = D + E, then B, D, and E
are the precedents for A (C is not a precedent as it is only an intermediate calculated value). The range and number of values perturbed is user-specified and can be set to test extreme
values rather than smaller perturbations around the expected values. In certain circumstances, extreme values may have a larger, smaller, or unbalanced impact (e.g., nonlinearities may
occur where increasing or decreasing economies of scale and scope creep occurs for larger or smaller values of a variable) and only a wider range will capture this nonlinear impact.

A tornado chart lists all the inputs that drive the model, starting from the input variable that has the most effect on the results. The chart is obtained by perturbing each precedent input at
some consistent range (e.g., 10% from the base case) one at a time, and comparing their results to the base case. A spider chart looks like a spider with a central body and its many legs
protruding. The positively sloped lines indicate a positive relationship, while a negatively sloped line indicates a negative relationship. Further, spider charts can be used to visualize linear
and nonlinear relationships. The tornado and spider charts help identify the critical success factors of an output cell in order to identify the inputs to simulate. The identified critical variables
that are uncertain are the ones that should be simulated. Do not waste time simulating variables that are neither uncertain nor have little impact on the results.

Result
Base Value: 96.6261638553219
Precedent Cell
Investment
Tax Rate
A Price
B Price
A Quantity
B Quantity
C Price
C Quantity
Discount Rate
Price Erosion
Sales Growth
Depreciation
Interest
Amortization
CapEx
Working Capital

Output
Downside
$276.63
$219.73
$3.43
$16.71
$23.18
$30.53
$40.15
$48.05
$138.24
$116.80
$90.59
$95.08
$97.09
$96.16
$96.63
$96.63

Output
Upside
($83.37)
($26.47)
$189.83
$176.55
$170.07
$162.72
$153.11
$145.20
$57.03
$76.64
$102.69
$98.17
$96.16
$97.09
$96.63
$96.63

Effective
Input
Range
Downside
360.00 $1,620.00
246.20
36.00%
186.40
$9.00
159.84
$11.03
146.90
45.00
132.19
31.50
112.96
$13.64
97.16
18.00
81.21
13.50%
40.16
4.50%
12.10
1.80%
3.08
$9.00
0.93
$1.80
0.93
$2.70
0.00
$0.00
0.00
$0.00

Input Changes
Input
Base Case
Upside
Value
$1,980.00
$1,800.00
44.00%
40.00%
$11.00
$10.00
$13.48
$12.25
55.00
50.00
38.50
35.00
$16.67
$15.15
22.00
20.00
16.50%
15.00%
5.50%
5.00%
2.20%
2.00%
$11.00
$10.00
$2.20
$2.00
$3.30
$3.00
$0.00
$0.00
$0.00
$0.00

Sensitivity Analysis
Statistical Summary
Sensitivity charts are dynamic perturbations created after the simulation run. Sensitivity charts are dynamic perturbations in the sense that multiple assumptions are perturbed
simultaneously and their interactions are captured in the fluctuations of the results. In contrast, Tornado charts are static perturbations, meaning that each precedent or
assumption variable is perturbed a preset amount and the fluctuations in the results are tabulated. Tornado charts therefore identify which variables drive the results the most
and hence are suitable for determining which variables to simulate (that is, they are used before a simulation), whereas sensitivity charts identify the impact to the results
when multiple interacting variables are simulated together in the model (that is, they are used after a simulation).
The Nonlinear Rank Correlation charts indicate the rank correlations between each assumption and the target forecast, and are depicted from the highest absolute value to
the lowest absolute value. Positive correlations are shown in green while negative correlations are shown in red. Rank correlation is used instead of a regular correlation
coefficient as it captures nonlinear effects between variables. In contrast, the Percent Variation Explained computes how much of the variation in the forecast variable can be
explained by the variations in each of the assumptions by itself in a dynamic simulated environment. These charts show the sensitivity of the target forecast to the simulated
assumptions.

Sensitivity Analysis
Statistical Summary
Sensitivity charts are dynamic perturbations created after the simulation run. Sensitivity charts are dynamic perturbations in the sense that multiple assumptions are perturbed
simultaneously and their interactions are captured in the fluctuations of the results. In contrast, Tornado charts are static perturbations, meaning that each precedent or
assumption variable is perturbed a preset amount and the fluctuations in the results are tabulated. Tornado charts therefore identify which variables drive the results the most
and hence are suitable for determining which variables to simulate (that is, they are used before a simulation), whereas sensitivity charts identify the impact to the results
when multiple interacting variables are simulated together in the model (that is, they are used after a simulation).
The Nonlinear Rank Correlation charts indicate the rank correlations between each assumption and the target forecast, and are depicted from the highest absolute value to
the lowest absolute value. Positive correlations are shown in green while negative correlations are shown in red. Rank correlation is used instead of a regular correlation
coefficient as it captures nonlinear effects between variables. In contrast, the Percent Variation Explained computes how much of the variation in the forecast variable can be
explained by the variations in each of the assumptions by itself in a dynamic simulated environment. These charts show the sensitivity of the target forecast to the simulated
assumptions.