Title: Forecasting Free Cash Flows
1Forecasting Free Cash Flows
Where We Are Going
We will discuss forecasting free cash flow and
using that forecast to estimate the value of the
firm
Chapter 10
2 The Free Cash Flow Model
Value of core operations
3A Forecast Valuation Model
- Is a set of equations that describe how each
element of free cash flow behaves
4Forecasting and the Valuation
Has four stages
- Model the free cash flows
- Set the model assumptions and compute the results
- The major step in valuation
- Refine the model
- Ex, Break forecast down by produce line, forecast
sales growth by price and volume changes - Sensitivity analysis
5Business Analysis
Forecasting and the Valuation
Financial Statement Analysis
Forecast Assumptions
Historical Ratios
Forecast Ratios
GAAP Financial Statements
Historical Free Cash Flow Statements
Free Cash Flow Forecast
Valuation
Time
Historical Periods
Valuation Date
Forecast Periods
61. Model the Free Cash Flows
- Specify cause-and-effect relationships among FCF
components - For example, the equation defining sales might be
prior years sales increased by a growth rate
7Model the Free Cash Flows Continued
Each component of model is estimated from
- Already computed component of model of historical
value - One or more assumptions, often in form of a ratio
8Model the Free Cash Flows Continued
A "good" forecast is
- Reasonable
- Internally consistent
9Model the Free Cash Flows Continued
Reasonableness
- Model is a decent representation of
cause-and-effect relationships among FCF
components - Assumptions are realistic forecasts of actual
values that will be realized
10Model the Free Cash Flows Continued
Internal Consistency
- The assumptions are logical in relation to each
other - If project selling expense as sales times an
assumed ratio, then these expenses should
actually vary with sales levels, perhaps due to
commissions. In contrast, forecasting deprecation
in the same way would not make sense
11Example of a Free Cash Flow Model (Exhibit 10.2)
- The model allows for a different growth rate in
every year
12Example of a Free Cash Flow Model (Exhibit 10.2)
- For the REV and COS to be internally consistent,
COS forecast must depend on REV. The gross margin
percentage defines that dependency
13Example of a Free Cash Flow Model (Exhibit 10.2)
- Example of internal inconsistent forecast on page
215 - If gst is increased without changing ht, all the
incremental revenue will be profit, meaning the
margin assumed on an additional dollar of sales
was 100! - COS do not simply grow at some rate independent
of sales growth rather sales growth directly
affects COS
14Example of a Free Cash Flow Model Continued
- Operating expenses include all expenses that
occur in the normal course of operations except
for those associated with preparing the product
for sale, tax expense, and financing expenses - Eg. Selling, marketing, rent, administrative
expenses..
15Example of a Free Cash Flow Model Continued
- Whenever we apply a tax rate, we use the tax rate
that pertains to the associated income - In this case, we use the effective tax rate on
core operations
16Example of a Free Cash Flow Model Continued
- Now turn to the components of free cash flow that
come from the cash flow statement rather than the
income statement - The reversals of items included in NOPAT that are
not cash flows, such as deprecation and
amortization expense - Cash flows related to core operations that are
not included in NOPAT, such as capital
expenditures and changes in working capital
17Example of a Free Cash Flow Model Continued
- Examine historical changes in depreciation levels
from year to year and forecasting analogous
changes - Assume depreciation is unlikely to change much
form one year to the next
- The proportion of the income tax provision that
is not a current obligation to the government - Arise primarily due to book/tax differences in
depreciation
18Example of a Free Cash Flow Model Continued
- Represents additional investment
- Generally related to revenue increases
- Changes to the revenue forecast will
automatically alter the forecast of change in
working capital
19Example of a Free Cash Flow Model Continued
- Maintenance capital expenditure, ctDEPRt those
required to maintain the existing level of
productive capacity - Expansion capital expenditure, ft(REV-REVt-1)
capital expenditures beyond those required to
maintain existing productive capacity
202. Set the Model Assumptions and Compute the
Results
To Set Assumptions
- Use historical values as guide where reasonable
- Distinguish between forecast horizon and
perpetuity period
21Forecast and Valuation
22Setting Assumptions in the
Perpetuity Period Continued
How many years in the forecast horizon?
Go out far enough so that
- Regular FCF pattern is feasible
- Competitive advantage likely to have dissipated
23Setting Assumptions in the
Perpetuity Period
Assumptions in Perpetuity Period
- No more comparative advantage
- Implies any incremental investment is zero-NPV
24Set the Model Assumptions and Compute the Results
Using Historical Relationships
- Some historical relation will remain stable,
others will not - It is important to adjust the historical relation
based on business analysis, whenever necessary
25Starbucks Valuation Example
- Ex. 10.3
- Forecast ratios based on 5 years of historical
data - Use the forecasted ratios and the relation among
the components of free cash flow statement to
forecast free cash flow from core operation
26Starbucks in the Perpetuity Period
Assumptions in Perpetuity Period
- No more comparative advantage
- Implies any incremental investment is zero-NPV
- Okay to ignore incremental investment
- Nominal growth
- CAPEX maintenance only
- No differed taxes
273. Refine the Model
After building the model, the analyst often
decides to refine it
For example
Separate by business units
Separate a ratio into two or more parts
Link model to external economic forecasts
28Refine the Model Continued
Detailed Cost of Sales Forecast Submodel
Detailed Revenue Forecast Submodel
- Free Cash Flow Forecast
- Revenue forecast
- Cost of sales forecast
- ?
- ?
- ?
29Separate by Business Unit
- Separate forecasts down to some point in FCF
forecast - Aggregate separate forecasts
- Most useful when economics of business units
differ
30Separate by Business Unit Cont.
Unit A
Consolidated Forecast
Unit B
Unit C
31Separating Ratios
Break relationship into two or more
assumptions For example (see equations on page
231)
- Sales growth into volume growth and price changes
- A cost into fixed and variable components
32Link to External Forecasts
- Sales into forecasted market demand and market
share - Example first equation on page 232
- Sales based on an indicator of economic activity,
like housing starts or GDP
334. Sensitivity Analysis
Analysts commonly use four types of sensitivity
analysis
- Single-assumption
- Combined
- Scenario-based
- Reverse valuation
34Single-Assumption Sensitivity Analysis
- Calculate a range of values given optimistic and
pessimistic values for the assumption (ex. 10.9) - What would happen if revenues were higher or
lower than what we projected in the base case
forecast
35Combined Sensitivity Analysis
- Vary two or more assumptions simultaneously (ex.
10.10) - Variables do not usually change one at a time
- Changes in advertising affect sales volume and
operating expenses
36Scenario-based Sensitivity Analysis
- Recalculate value given some event, such as a new
entrant in the market - For example, assume a new competitor enter the
market - Revenue growth and operating margin percentage
might decline as the competitor takes market
shares and forces prices lower - Other operating expense might increase as the
firm responds by increasing its advertising and
promotion
37Reverse Valuation Sensitivity Analysis
- Determine the assumptions that are consistent
with the observed market price and assess whether
these assumptions are reasonable
38Summary
We have learned
- A forecast model is not just numbers it is a set
of equations describing cause and effect of free
cash flow components - Assumptions must be reasonable and internally
consistent
39Summary Continued
- To get basic model in place first then refine
- To understand potential errorssensitivity