Sensitivity Analysis in RIM
Sensitivity analysis is the bridge between a point-estimate valuation and a defensible investment thesis. In RIM, it reveals which assumptions drive valuation the most, which inputs deserve the deepest scrutiny, and how robust your valuation is to realistic assumption changes. A rigorous sensitivity analysis doesn't eliminate uncertainty—it maps it, showing how value swings across ranges of inputs. This clarity lets you distinguish between assumptions you're confident about and those where you're guessing, informing both your target price and your position sizing.
Quick definition
Sensitivity analysis in RIM tests how intrinsic value changes when one or more key inputs shift. One-way tables isolate individual variables (terminal ROE, cost of equity, payout ratio); two-way tables show interaction effects. Tornado diagrams rank which assumptions have the greatest valuation impact, while scenario modeling tests combinations of assumptions (bull case, base case, bear case).
Key takeaways
- Terminal ROE is typically the highest-impact variable in RIM; 1% changes can swing valuation 15–30%
- Cost of equity impacts both residual income spread and discounting, making it doubly sensitive
- Payout ratio sensitivity depends on ROE levels; high ROE makes reinvestment valuable, so lower payouts increase intrinsic value
- Two-way sensitivity tables reveal whether assumptions are correlated (e.g., higher growth typically pairs with lower terminal ROE as competitive advantages erode)
- Sensitivity tables should include realistic ranges, not extreme outliers; a 15% change in assumptions should feel feasible
- Graphical presentations (tornado diagrams, waterfalls) communicate sensitivity more effectively than tables to non-analysts
One-way sensitivity tables
The simplest and most common approach: vary one input while holding all others constant, and recalculate intrinsic value.
Example: Terminal ROE sensitivity
Assume a company has $500 million book value, cost of equity 9%, current ROE 12%, payout ratio 50%. Terminal ROE is the key assumption.
Terminal ROE | Intrinsic Value ($ billions)
8% | $3.20
9% | $4.10
10% | $5.05
11% | $6.05
12% | $7.10
13% | $8.25
14% | $9.50
Reading across: as terminal ROE rises 1%, intrinsic value increases roughly 10–15%. This signals that terminal ROE is a critical driver. A 2% shift (from 11% to 13%) changes valuation by $2.2 billion—material and worth defending explicitly.
Example: Cost of Equity sensitivity
Same company, now varying cost of equity while holding terminal ROE at 11%:
Cost of Equity | Intrinsic Value ($ billions)
7.5% | $8.65
8.0% | $7.50
8.5% | $6.60
9.0% | $5.80
9.5% | $5.10
10.0% | $4.50
Cost of equity sensitivity is typically more severe than terminal ROE sensitivity because it affects both the residual income spread (ROE - COE) and the discounting denominator. A 100 basis point rise in cost of equity cuts intrinsic value by roughly 15–20%.
Example: Payout Ratio sensitivity
Same company, varying payout ratio (which affects equity growth and thus scales residual income):
Payout Ratio | Sustainable Growth | Intrinsic Value ($ billions)
30% | 8.4% | $6.50
40% | 7.2% | $6.25
50% | 6.0% | $6.00
60% | 4.8% | $5.75
70% | 3.6% | $5.50
Lower payout (higher retention) accelerates equity growth, which scales future residual income. The impact is less dramatic than terminal ROE or cost of equity changes but still material (15% spread from 30% to 70% payout).
The intuition: with 12% ROE and 9% cost of equity, retention is accretive (spreads profits over a growing capital base). Lower payouts increase intrinsic value.
Two-way sensitivity tables
Two-way tables show how intrinsic value changes when two variables shift simultaneously. They reveal correlation and interaction effects.
Terminal ROE vs. Cost of Equity
COE 8.0% | 8.5% | 9.0% | 9.5% | 10.0%
T-ROE 10% | $4.50 | $3.95 | $3.50 | $3.10 | $2.75
11% | $5.80 | $4.95 | $4.25 | $3.65 | $3.15
12% | $7.50 | $6.20 | $5.20 | $4.35 | $3.60
13% | $9.80 | $7.95 | $6.50 | $5.30 | $4.35
Reading the table:
- Best case (top-left): 13% terminal ROE, 8% cost of equity → $9.80B intrinsic value
- Base case (middle): 12% terminal ROE, 9% cost of equity → $5.20B intrinsic value
- Worst case (bottom-right): 10% terminal ROE, 10% cost of equity → $2.75B intrinsic value
The spread from worst to best case is 3.5x valuation—indicating high sensitivity to joint assumptions and the importance of getting both right.
Terminal ROE vs. Payout Ratio
Payout 30% | 40% | 50% | 60% | 70%
T-ROE 10% | $5.10 | $4.85 | $4.65 | $4.50 | $4.40
11% | $6.85 | $6.40 | $6.00 | $5.65 | $5.35
12% | $8.95 | $8.15 | $7.45 | $6.85 | $6.35
13% | $11.60 | $10.45| $9.45 | $8.60 | $7.85
Higher terminal ROE (more profitable reinvestment) pairs with lower payout (more retention) to maximize value. The spread between 30% and 70% payout is largest at high terminal ROE (nearly 50% difference) and smaller at low terminal ROE (roughly 5%).
This reveals an important insight: if you're uncertain about terminal ROE, payout policy becomes less important. But if terminal ROE is high (strong competitive advantages), retention policy matters significantly.
Tornado diagrams (sensitivity ranking)
A tornado diagram ranks variables by their impact on valuation, with the highest-impact variable at the top and lowest-impact at the bottom. The visual is helpful for communicating which assumptions deserve the most scrutiny.
Example tornado diagram for RIM valuation:
Terminal ROE |==========================════| (30% impact)
Cost of Equity |===================════| (25% impact)
Terminal Growth |==========| (12% impact)
Payout Ratio |======| (8% impact)
Current Book Value |===| (5% impact)
Each bar shows the valuation range for ±1 standard deviation from the base assumption. The ranking reveals that terminal ROE and cost of equity dwarf other inputs in impact. This guides where to invest analytical effort: calibrating cost of equity precisely (via CAPM) and grounding terminal ROE in competitive analysis matter far more than perfecting year-by-year revenue forecasts.
Scenario analysis
Rather than isolated sensitivity, scenario analysis combines assumptions into coherent narratives: bull case, base case, bear case.
Bull Case: Strong Competitive Advantages
Assumptions:
- Terminal ROE: 14% (sustained premium to peers due to network effects, brand, scale)
- Cost of Equity: 8% (lower risk given stable cash flows)
- Payout Ratio: 40% (reinvest aggressively to maintain competitive position)
- Terminal Growth: 3.5% (slightly above GDP, justified by market share gains)
Calculation: Year 5 Book Value: $650M (grown from $500M at 6% annually) Year 5 Residual Income: $650M × (14% - 8%) = $39M Terminal Value: $39M × (1.035) / (0.08 - 0.035) = $898M PV of Terminal Value: $898M / (1.08^5) = $611M
Explicit Period RI: $85M (sum of discounted residual income Years 1–5) Intrinsic Value: Book Value + PV(Explicit RI) + PV(Terminal) = $500M + $85M + $611M = $1.196B per share (example)
Base Case: Moderate Competitive Position
Assumptions:
- Terminal ROE: 11% (revert toward historical average and peer median)
- Cost of Equity: 9% (market-rate cost of capital)
- Payout Ratio: 50% (balanced capital allocation)
- Terminal Growth: 2.5% (GDP-aligned)
Calculation: Year 5 Book Value: $600M (grown at 5.5% annually) Year 5 Residual Income: $600M × (11% - 9%) = $12M Terminal Value: $12M × (1.025) / (0.09 - 0.025) = $185M PV of Terminal Value: $185M / (1.09^5) = $120M
Explicit Period RI: $65M Intrinsic Value: $500M + $65M + $120M = $685B
Bear Case: Competitive Erosion
Assumptions:
- Terminal ROE: 9% (converges to cost of equity; no economic moat)
- Cost of Equity: 10% (higher risk due to deteriorating position)
- Payout Ratio: 70% (limited reinvestment opportunity)
- Terminal Growth: 1.5% (below GDP, market share loss)
Calculation: Year 5 Book Value: $525M (grown at 2.1% annually) Year 5 Residual Income: $525M × (9% - 10%) = -$5.25M
Negative residual income signals value destruction. Intrinsic value converges toward book value or below. Intrinsic Value: $525M (book value as conservative bound; terminal value adds negative residual income)
Summary Table
Scenario | Intrinsic Value | Implied Multiple | Probability
Bull | $1.196B | 2.4x book | 25%
Base | $685M | 1.4x book | 50%
Bear | $525M | 1.0x book | 25%
Expected | $739M | 1.48x book | —
This framework shows that intrinsic value ranges from $525M to $1.196B depending on assumption realism. The base case probability-weighted value of $739M suggests the stock is fairly valued if it trades near that price, with meaningful upside in the bull case and downside in bear.
Stress testing critical assumptions
Beyond ranges and scenarios, stress testing isolates single assumptions at extreme levels to test model robustness.
Stress Test: ROE Never Exceeds COE
What if the company never builds a sustainable competitive advantage and terminal ROE always equals cost of equity (9%)?
- Residual income = 0 after the explicit forecast period
- Intrinsic value = Book Value + PV(Explicit Period RI) only
- No terminal value component
This is a useful sanity check: if the model implies intrinsic value relies heavily on terminal residual income, but the business operates in a competitive market with no moat, the valuation is fragile.
Stress Test: Terminal Growth = 0
If sustainable growth collapses (company shrinks):
- Terminal Value = Residual Income / COE (no growth term)
- Value declines because equity doesn't expand to amplify residual income
- Applicable to mature or declining businesses (e.g., utilities, legacy manufacturing)
Stress Test: Cost of Equity increases 200 basis points
Represents a shift in market risk premium or company-specific risk:
- Residual income spread (ROE - COE) shrinks
- Discounting denominator increases
- Intrinsic value typically falls 25–40% depending on scenario
This stress test is critical if your valuation assumes a low cost of equity. If a 200 basis point increase collapses the thesis, your model has hidden risk sensitivity.
Building sensitivity tables in spreadsheets
Create a two-way table using Excel or Google Sheets formulas for transparency:
Column headers: Terminal ROE (9% to 14% in 0.5% increments)
Row headers: Cost of Equity (7.5% to 10.5% in 0.5% increments)
Cell formula: =(Book_Value + PV_Explicit_RI + Terminal_RI_Perpetuity) / Shares_Outstanding
Where Terminal_RI_Perpetuity = (Terminal_Book_Value × (Terminal_ROE - COE)) × (1 + Growth) / (COE - Growth)
Conditional formatting (color scale from red to green) highlights value ranges visually, making it obvious which assumptions matter most.
Real-world examples
Apple sensitivity analysis: Given Apple's high ROE (80%+ in some periods) but cyclical gross margins, sensitivity to terminal ROE is critical. A 2% swing in terminal ROE assumptions (from 20% to 22%, not unrealistic) can change valuation by 10–15%.
Utility sensitivity analysis: Utilities operate in regulated environments with visible cost of equity (based on allowed returns) but stable ROE. Sensitivity here centers on terminal growth (tied to inflation expectations) and cost of equity, not terminal ROE, which is pinned to regulatory requirements.
Financial services sensitivity: Banks' ROE depends heavily on interest rates and credit cycles. Sensitivity analysis must isolate cost of equity (linked to risk premiums) from terminal ROE (linked to profitability margins), which are independent drivers.
Common mistakes
Using unrealistic ranges. A sensitivity table showing 5% to 25% terminal ROE is not useful if the company's ROE has ranged 10–15% for 20 years. Set realistic bounds: historical range ±2 percentage points, unless structural change justifies larger deviation.
Ignoring correlation between variables. If cost of equity rises, terminal ROE may also fall (more competitive risk). Presenting a scenario with 14% ROE and 11% cost of equity (positive spread) may be inconsistent—higher risk typically implies higher COE and lower expectations for ROE.
Presenting point estimates without sensitivity. A valuation memo that says "intrinsic value is $47 per share" without sensitivity analysis is incomplete. Readers can't assess the robustness of assumptions.
Sensitivity tables so large they're unreadable. A 10×10 or 15×15 table overwhelms. Focus on the 2–3 most critical variables. Use tornado diagrams and scenarios for clarity.
Not stress-testing the base case assumptions. If your base case assumes terminal ROE = historical average and cost of equity = CAPM estimate, separately test what happens if one or both are wrong. These are educated guesses, not certainties.
FAQ
Q: How many decimal places should I use in sensitivity tables?
A: One decimal place is standard (e.g., 8.5%, 9.0%, 9.5%). More precision is false accuracy; less precision (e.g., 8%, 9%, 10%) hides important gradations. Align precision to the confidence level of your base assumption.
Q: Should I always include a scenario with ROE = COE?
A: Yes. This is a useful reality check—the scenario where the company earns its cost of capital and has no economic moat. Compare this to your base case to see how much value depends on assumed competitive advantages.
Q: How do I explain sensitivity analysis to non-analysts?
A: Use ranges and scenarios, not tables. "Under conservative assumptions, the stock is worth $40. Under our base case, $55. Under aggressive assumptions, $75." Then explain which scenario you find most likely and why. Use a simple visual—tornado diagram or waterfall—to show which inputs matter most.
Q: Can I use historical volatility in earnings or ROE to set sensitivity ranges?
A: Partially. Historical volatility suggests realistic bounds but isn't binding. A company with volatile historical ROE might stabilize in the future, or might not. Use historical ranges as a starting point, then adjust based on your assessment of future competitive positioning.
Q: What if sensitivity tables show intrinsic value is negative under some scenarios?
A: That signals assumptions are incoherent or the company is a value trap. A negative intrinsic value means the company destroys shareholder capital. Either revisit your ROE and cost of equity estimates, or acknowledge that the bear case is truly catastrophic (and price risk accordingly).
Related concepts
- Terminal ROE and competitive advantage: The highest-impact assumption in most RIM models
- Cost of equity and CAPM: The denominator in residual income; small changes create large valuation swings
- Scenario modeling: Building coherent bull, base, and bear cases
- Tornado diagrams and visualization: Communicating sensitivity findings clearly to stakeholders
- Reverse-engineering market price: Testing whether market-implied assumptions are reasonable
Summary
Sensitivity analysis in RIM transforms uncertainty into insight. One-way tables isolate how intrinsic value changes with each key input; two-way tables reveal interaction effects and correlation risks. Tornado diagrams rank which assumptions matter most, guiding where analytical effort has the highest payoff. Scenario modeling (bull, base, bear) combines assumptions into coherent narratives, helping investors and boards assess probability-weighted intrinsic value. Terminal ROE and cost of equity are typically the highest-impact variables; terminal growth and payout ratio matter less but still warrant scrutiny. Set sensitivity ranges based on historical company performance and competitive positioning, not extreme outliers. Stress tests—asking what happens if critical assumptions are wrong—reveal hidden risks. Present sensitivity findings clearly using ranges, scenarios, and visuals rather than dense tables. Finally, sensitivity analysis isn't permission to avoid making assumptions; it's the discipline to recognize which assumptions are critical, which are secondary, and how robust your valuation is if assumptions prove wrong.