Scenario analysis in forecasting

Why build scenarios instead of one forecast?

Every forecast in this module rests on views about risks that may not play out. Key risk factors are the events that could make the future differ from the forecast. Four generic ones affect all companies to varying degrees, and company-specific ones add to them.

Key risk factorseach scenario takes a different view of them

• Generic risk factors Memorize affect all companies, to varying degrees
  • Business cycle Memorize volume, price and margin move together
  • Competition Memorize share and pricing
  • Inflation or deflation Memorize input costs and prices
  • Technological developments Memorize shift supply (cheaper production) or demand (a substitute)
• Company-specific risk factors Use it a product launch, a regulator's decision

Scenario analysis replaces a single point estimate with several complete forecasts, each taking a different view of the key risks, and a judgment of how likely each is. The usual set has a base case (the most central view), a bull case (favorable outcomes) and a bear case (unfavorable outcomes). The scenarios change linked assumptions together: a recession lowers volume, price and margin at once.

Sensitivity analysis is different. It changes a single input and holds everything else constant, to show how much the result depends on that input.

Handling forecast uncertainty

• Scenario analysis Memorize a few complete forecasts
  • Changes several linked assumptions at once
  • Output bear, base, bull; each with a probability
• Sensitivity analysis Memorize one lever at a time
  • Changes one input, all others held constant
  • Output how much the result moves per input (a tornado)

Technology deserves a special word. It can lower the cost of production, shifting the supply curve to the right, or create an attractive substitute, shifting demand away from the existing product. When it takes sales from a company’s own products or from a competitor’s, that effect is called cannibalization .

Exam tip Scenarios move assumptions together; sensitivities move one at a time. A “scenario” that changes one number is a sensitivity.

What do worked scenarios look like?

Lumen Ledger has 46,000 subscribers paying USD 1,260 each, so revenue this year is USD 57.96 million. The analyst builds three cases for next year, each changing both subscribers and price, and assigns probabilities.

Scenario	Probability	Assumptions	Revenue (USD million)
Bear	25%	subscribers −5%, price flat	46,000 × 0.95 × 1,260 = 55.06
Base	50%	subscribers +12%, price +4%	67.51
Bull	25%	subscribers +18%, price +6%	46,000 × 1.18 × 1,335.60 = 72.50

$$E = \sum_i p_i\,V_i \qquad \sum_i p_i = 1$$

What to compare scenarios with

Investors compare scenarios with the forecasts of other analysts, known as consensus forecasts , and with the forecasts implied by the current share price. If consensus sits above even your bull case, the market may be ignoring a risk. The spread of scenarios also tells you how big a position the uncertainty can justify: a narrow range supports a larger one.

Sensitivity around the base case

A tornado of Lumen’s base case moves one input at a time. Subscriber growth of 7% rather than 12% (a 5-point shift) gives 46,000 × 1.07 × 1,260 × 1.04 = 64.50, a swing of about 3.0 either way. Price growth of 1% rather than 4% (a 3-point shift) gives 65.56, a swing of about 1.9. Subscribers matter more than price here, which tells the analyst where to spend research effort. But it does not say what happens when both move together with the competitor’s bundle. That needs a scenario.

Go deeper Linking scenarios to valuation

Revenue is only the first line. A full scenario carries its assumptions through costs, capital spending and financing, and ends with a value for each case, such as a value per share. Fixed costs matter: a company with mostly fixed costs, like a software business, loses profit faster than revenue when sales fall. The probability-weighted value per share, not the weighted revenue, is what you compare with the share price in the next module on valuation.

Go deeper Why the base case is not automatically the most likely outcome

The base case is the analyst’s central view, but with three scenarios it is not guaranteed to hold the largest probability, and with skewed risks it is not the average either. In the Lumen example the weighted value (65.65) lies below the base case (67.51), because the downside is larger than the upside. Report the range and the weights, not just the base case.