Sensitivity analysis in CRE is the step that turns a single projected return into a defensible range. Without it, a pro forma states one internal rate of return with false precision and hides every assumption that produced it. A model that outputs 8 percent IRR and nothing else is a guess wearing the costume of precision. A model that shows how that 8 percent moves as the exit cap rate, rent growth, hold period, and leverage change is an argument you can defend to an investment committee. The difference is not the base case. It is whether you tested it.
Key Takeaways
Sensitivity analysis varies one input at a time to measure its effect on the output. Scenario analysis moves several inputs together into a coherent case. The CFA Institute curriculum treats scenario analysis as a specific form of sensitivity testing, not a substitute for it.
The exit cap rate moves IRR more than any other single variable in a stabilized deal, because the reversion can account for more than half of a ten-year discounted cash flow, per RICS discounted cash flow valuation guidance.
A two-way data table crossing exit cap rate against rent growth is the canonical CRE sensitivity grid. It shows the output surface, not one point on it.
In a worked five-year hold, a 150 basis point move in the exit cap rate swings unlevered IRR by 5.3 points, while a 300 basis point move in annual rent growth swings it only 3.0 points. The exit assumption is doing most of the work.
Sensitivity analysis is how you locate the margin of safety, Benjamin Graham's central concept: the distance between your base case and the point where the deal stops working.
What is sensitivity analysis in CRE, and how does it differ from scenario analysis?
Sensitivity analysis measures how much a single output, usually IRR or net present value, moves when one input changes and everything else holds constant. Scenario analysis moves several inputs together into a coherent story, such as a recession case where rent growth falls, exit caps widen, and vacancy rises at once. Both test the same model. They answer different questions.
Corporate Finance Institute defines sensitivity analysis as the study of how one independent variable affects a dependent variable under a set of assumptions. The value of the one-at-a-time approach is attribution. When you change only the exit cap rate, any movement in IRR belongs to the exit cap rate, which is what lets you rank variables by influence.
Scenario analysis trades that isolation for realism. The CFA Institute curriculum frames scenario analysis as examining best-case, worst-case, and most-likely combinations of the critical variables, which captures how risks correlate in the real world. The two are complements. Sensitivity analysis tells you which levers matter. Scenario analysis tells you what happens when the levers move together. An underwriting that runs one but not the other is incomplete. See pro forma assumptions buyers should challenge for where these inputs most often go soft.
Which variables move IRR the most in a real estate model?
In a stabilized acquisition, the exit cap rate typically moves IRR the most, followed by leverage, then rent growth, then hold period. The ranking shifts by deal type: heavy value-add and development deals lean more on rent growth and lease-up speed, because more of their value is created rather than inherited. But for a core or core-plus hold, the reversion dominates.
The reason is structural. RICS discounted cash flow valuation guidance notes that the reversion, the modeled sale at the end of the hold, can represent more than half of a property's total present value in a ten-year DCF. That number is set by the exit cap rate, so a small change there rescales the largest cash flow in the model and moves IRR more than inputs that touch only annual income.
The tool for ranking this is the tornado chart. You take each input, flip it between a low and a high value while holding the rest at base case, and record the swing in IRR. Sort the swings from largest to smallest and the bars form the funnel shape that names the chart. The variable at the top is the one your diligence should attack first. A tornado chart does not tell you the answer. It tells you which question is worth the most.
Input, ranked by influence | Why it moves IRR |
|---|---|
Exit cap rate | Rescales the reversion, often the largest single cash flow |
Leverage | Amplifies both gains and losses on the equity |
Rent growth | Compounds through NOI and into the exit value |
Hold period | Shifts the timing and weight of the reversion |
Going-in vacancy | Sets the base the whole projection builds from |
How do you read a two-way sensitivity table for exit cap rate and rent growth?
A two-way sensitivity table crosses two inputs and reports one output in every cell, so you read the deal as a surface instead of a point. The canonical CRE version crosses exit cap rate against rent growth and prints IRR. You find the base case in the interior, then read how far IRR falls toward the pessimistic corner of a higher cap and lower growth.
Consider a worked example. A property is acquired for 10,000,000 dollars all cash, producing 500,000 dollars of first-year NOI, a 5.0 percent going-in cap rate. NOI grows at a rate g each year. The hold is five years. The sale price equals the sixth-year NOI divided by the exit cap rate. Solving for the unlevered IRR across a range of exit caps and growth rates produces this grid.
Rent growth \ Exit cap | 4.5% | 5.0% | 5.5% | 6.0% |
|---|---|---|---|---|
1% | 8.0% | 6.0% | 4.3% | 2.7% |
2% | 9.0% | 7.0% | 5.3% | 3.7% |
3% | 10.0% | 8.0% | 6.2% | 4.7% |
4% | 11.0% | 9.0% | 7.2% | 5.6% |
Read across the base-case row of 3 percent growth. Moving the exit cap from 4.5 to 6.0 percent, a 150 basis point widening, drops IRR from 10.0 to 4.7 percent, a swing of 5.3 points. Now read down the base-case column of a 5.0 percent exit cap. Moving rent growth from 1 to 4 percent, a 300 basis point range, lifts IRR from 6.0 to 9.0 percent, a swing of 3.0 points. The exit cap rate, moved half as far, moved IRR nearly twice as much.
The mechanism is visible in the reversion. At a 5.0 percent exit cap, the base-case sale value is 11,592,741 dollars. Widen the cap 100 basis points to 6.0 percent and the same NOI sells for 9,660,617 dollars, a 16.7 percent haircut on the largest cash flow in the deal, purely from an assumption. This is why underwriting that gets NOI right but the exit wrong still misprices the deal. The grid forces the exit assumption into the open where it can be argued.
What is the margin of safety, and how does sensitivity analysis reveal it?
The margin of safety is the distance between your base case and the point where the deal stops working. Benjamin Graham made it the central concept of investment, distilling sound investing into three words: margin of safety. Sensitivity analysis is how you measure that distance in a real estate model, because it shows exactly how much an assumption can move before the return falls below your threshold.
Read the grid again as a floor test rather than an upside case. If your equity requires a 6 percent unlevered IRR, the base-case deal at 3 percent growth and a 5.0 percent exit clears it at 8.0 percent, with room to spare. But the same deal breaks the 6 percent line the moment the exit cap reaches roughly 5.5 percent, holding growth flat. That single number, the exit cap you can tolerate before the deal fails your hurdle, is the margin of safety expressed in basis points.
A base-case IRR tells you what the deal returns if you are right. The margin of safety tells you how wrong you can be and still be fine. Graham's point, echoed by the CFA Institute, is that the function of the margin is to make an accurate forecast unnecessary. You do not need to predict the exit cap in 2031. You need to know that the deal survives a reasonable band of them. Sensitivity analysis draws that band. A guess does not. For the metric itself, see internal rate of return and the mechanics in the sensitivity analysis glossary entry.
Frequently Asked Questions
How many variables should a sensitivity analysis test?
Test the five to seven inputs that a tornado chart shows to be most influential, typically exit cap rate, leverage, rent growth, hold period, and going-in vacancy. Testing every input dilutes attention. The discipline is to rank first, then stress the variables that actually move the outcome.
Is a two-way data table better than a one-way table?
They serve different purposes. A one-way table isolates a single variable's effect and is cleaner for attribution. A two-way table crosses two inputs and reveals interaction effects, such as how a wider exit cap and slower rent growth compound in the pessimistic corner. Institutional underwriting usually shows both.
Does sensitivity analysis replace scenario analysis?
No. Sensitivity analysis moves one variable at a time and ranks influence. Scenario analysis moves correlated variables together into a coherent recession or upside case. The CFA Institute curriculum treats them as complements. Running one without the other leaves the underwriting exposed on the side it skipped.
Why does the exit cap rate dominate a stabilized deal?
Because the reversion, set by the exit cap rate, can exceed half of the total present value in a long-hold DCF, per RICS guidance. The cap rate rescales the single largest cash flow, so a small change there outweighs changes to inputs that touch only annual income.
Conclusion
A pro forma that outputs one number is a claim. A pro forma that outputs a surface is a case. Sensitivity analysis is the step in between, and it is not optional for anyone defending a deal after the fact. It ranks the assumptions by how much they move the return, exposes the exit cap rate as the variable doing most of the work in a stabilized hold, and measures the margin of safety as the distance between your base case and failure. Skip it and you have underwritten a guess. Run it and you have underwritten a model.