268  Forecasting Performance
likely to change as you learn about the company, so at this point, a work-
ing model should be your priority. Once the entire model is complete, 
return to the forecast page and enter your best estimates.
3. Multiply the forecast ratio by an estimate of its driver. Since most line items 
are driven by revenues, most forecast ratios, such as cost of goods sold 
(COGS) to revenues, should be applied to estimates of future revenues. 
This is why a good revenue forecast is critical. Any error in the revenue 
forecast will be carried through the entire model. Ratios dependent on 
other drivers should be multiplied by their respective drivers.
Exhibit 13.4 presents the historical income statement and partially com-
pleted forecast for a hypothetical company. To demonstrate the three-step 
process, we forecast cost of goods sold. In the first step, calculate historical 
COGS as a function of revenues, which equals 37.5 percent. To start the model, 
initially set next year’s ratio equal to 37.5 percent as well. Finally, multiply the 
forecast ratio by an estimate of next year’s revenues: 37.5 percent × $288 mil-
lion = $108 million.
Note that we did not forecast COGS by increasing the previous year’s costs 
by 20 percent (the same growth rate as revenues). Although this process leads 
to the same initial answer, it reduces flexibility. By using a forecast ratio rather 
than a growth rate, we can either vary estimates of revenues (and COGS will 
change in step) or vary the forecast ratio (for instance, to value a potential im-
provement). If we had increased the COGS directly, however, we could only 
vary the COGS growth rate.
EXHIBIT 13.4  Partial Forecast of the Income Statement
Forecast worksheet
Income statement
%
2019
Forecast 
2020
$ million
2019 
Forecast 
2020 
Revenue growth
20.0
20.0
Revenues
240.0
288.0
Cost of goods sold/revenues
37.5
37.5
Cost of goods sold
(90.0)
(108.0)
Selling and general expenses/revenues
18.8
Selling and general expenses
(45.0)
Depreciationt /net PP&Et–11
9.5
Depreciation
(19.0)
EBITA
86.0
Step 1: Choose a forecast driver, 
and compute historic ratios.
Interest expense
(15.0)
Interest income
2.0
Step 2: Estimate 
the forecast ratio.
Nonoperating income
4.0
Earnings before taxes (EBT)
77.0
Provision for income taxes
(18.0)
Net income
59.0
Step 3: Multiply the forecast ratio 
by next year’s estimate of revenues 
(or appropriate forecast driver).
1 Net PP&E = net property, plant, and equipment.

Mechanics of Forecasting  269
Exhibit 13.5 presents typical forecast drivers and forecast ratios for the most 
common line items on financial statements. The appropriate choice for a forecast 
driver, however, depends on the company and the industry in which it competes.
Most valuation models, especially those of public companies, rely on ratios cre-
ated directly from the company’s financial statements. If you have access to other 
data that improves your forecast, incorporate it. For instance, the external valua-
tion of a delivery company such as UPS will tie fuel costs directly to revenue. A 
more sophisticated model might tie fuel costs to the price of fuel and the number 
of packages delivered. Be mindful about incorporating new data, however. While 
additional data often improves the realism of your model, it will also increase its 
complexity. A talented modeler carefully balances realism with simplicity.
Operating Expenses  For each operating expense on the income statement—
such as cost of goods sold; selling, general, and administrative expenses; and 
research and development—we recommend generating forecasts based on 
revenues. In most cases, the process for operating expenses is straightforward. 
However, as outlined in Chapter 11, the income statement sometimes embeds 
certain nonoperating items in operating expenses. Before you begin the fore-
casting process, reformat the income statement to properly separate ongoing 
expenses from one-time charges.
Depreciation  To forecast depreciation, you have three options. You can 
forecast depreciation as either a percentage of revenues or a percentage of 
property, plant, and equipment (PP&E), or—if you are working inside the 
company—you can also generate depreciation forecasts based on specific 
equipment purchases and depreciation schedules.
Although one can link depreciation to revenue, you will get better forecasts 
if you use PP&E as the forecast driver. To illustrate this, consider a company 
that makes a large capital expenditure every few years. Since depreciation is di-
rectly tied to a particular asset, it should increase only following an expenditure. 
EXHIBIT 13.5  Typical Forecast Drivers for the Income Statement
Line item
Typical forecast driver
Typical forecast ratio
Operating
Cost of goods sold (COGS)
Revenue
COGS/revenue
Selling, general, and administrative 
(SG&A)
Revenue
SG&A/revenue
Depreciation
Prior-year net PP&E
Depreciationt  / net PP&Et –1
Nonoperating
Nonoperating income
Appropriate nonoperating asset, 
if any
Nonoperating income/nonoperating 
asset or growth in nonoperating 
income
Interest expense
Prior-year total debt
Interest expenset  / total debtt–1
Interest income
Prior-year excess cash
Interest incomet  / excess casht–1