Sensitivity Analysis: How to Stop Trusting Your Base Case and Start Stress Testing Every Assumption

Every assumption in your underwriting will be wrong. The question is not whether your model is accurate. It is whether your deal survives the range of outcomes that reality is likely to produce. Most operators build a single scenario, call it the base case, and make their decision based on that one number. That process is not underwriting. It is storytelling with a spreadsheet. Sensitivity analysis replaces that false precision with an honest range of outcomes, and it is the single most important analytical discipline in real estate underwriting that almost nobody does well. In this episode, Louis walks through how to build real sensitivity analysis into your underwriting process, using a 30 unit multifamily acquisition as a working example. Covered in this episode: 

1. Why your base case is fiction and why the probability of all your assumptions landing exactly where you predicted is effectively zero
2. How to build one variable sensitivity tables and what they reveal about where your return is actually coming from
3. The two variable matrix: exit cap rate versus rent growth, and how to read the range of outcomes it produces
4. Three specific downside scenarios every operator should run: flat revenue with rising expenses, occupancy shock, and exit cap rate expansion
5. The tornado chart: how to rank your assumptions by impact and know exactly where to focus your due diligence and negotiation energy
6. How to use sensitivity analysis as a negotiation tool, not just an analytical exercise
7. Applying sensitivity analysis to hold decisions and existing portfolio positions
8. The three most common mistakes operators make: too many variables at once, unrealistic ranges, and ignoring correlation between inputs

 Plus a historical note on how sensitivity analysis was originally developed for military logistics at the RAND Corporation during World War II, and why real estate, with its long hold periods and high leverage, may be the asset class where it matters most. This episode is for operators who want to stop relying on a single point estimate and start understanding the full range of outcomes their capital is exposed to before they commit it.

BHPA - https://bhpropertyadvisors.com/

Transcript

SPEAKER_00 0:00

Welcome to Beyond IRR. This podcast examines real estate investments through the lens of structure, risk, and capital durability, not just headline returns. I'm your host, Louis Heiza. This podcast is sponsored by Beacon Hill Property Advisors. Today let's talk about the single most important analytical discipline in real estate underwriting that almost nobody does well if they do it at all. And this doesn't apply to the big boys, the institutional guys. Of course, they're doing this. I'm talking about uh the hobbyists, if if I can call them that. Uh these are people with other careers who are investing in real estate on the side. Uh they go to underwrite a deal, or maybe they're looking at their own portfolios and they download a spreadsheet off of Google and uh start plugging in numbers to underwrite future deals. Now, these people uh generally aren't going to have issue building a pro forma. Most operators and investors can build a pro forma. Uh, it's not calculating IRR or cash-on-cash return. Those are outputs. Um, what we're talking about here is sensitivity analysis. It's the practice of systematically asking what happens to my deal when this assumption is wrong. Because every assumption in your underwriting will be wrong, make no mistake about it. Um it's not that they might be wrong, it's that they will be, every single one. The question is not whether your model is accurate. It's not. It's whether your deal survives the range of outcomes that reality is likely to produce. And this is the question sensitivity analysis is designed to answer. Most operators build a single scenario. They call it the base case. They put their best estimate into a spreadsheet, press calculate, and get a return number. And if that number clears the hurdle, they proceed. If it doesn't, then they might adjust the input until it does or they pass. Now that process I would not consider underwriting. That is storytelling with a spreadsheet or something along those lines. That's guessing. Could be an educated guess, sure, but I think that that's guessing. And today I want to show you what real sensitivity analysis looks like, how to build it into your process, and why the operator should do it consistently, make better decisions than the ones who trust their base case. So let's start with discussing why your base case is fiction. And I don't mean to be disparaging here. Uh, the base case is a necessary starting point. You need a central estimate to anchor the analysis. The problem is that most operators stop there. So consider what goes into a typical multifamily acquisition pro forma. You're making assumptions about rent growth, you're assuming a trajectory for occupancy, you're projecting operating expenses often as a percentage of revenue or as a flat growth rate. You are assuming an interest rate on your financing, you're assuming a cap rate at the exit, you're assuming a hold period, you're assuming capital expenditure, timing and amounts. Uh notice I said the word assumption a lot, and that's on purpose. Every single one of those inputs is uncertain. And they do not move independently. Rent growth and occupancy are correlated, sure. Expense and revenue are loosely correlated, sure. Interest rates affect cap rates, which affect exit value, which affects your total return. So a lot of these are uh um correlated in some return. Often they can be uh either uncorrelated or causation uh based. Um, but by changing one input, often you're changing more than one input, whether you know it or not, uh, or at least you should be. So when you build a single scenario, you are implicitly assuming um that all of these uncertain inputs land exactly where you predicted. And the probability of that it happening is, I mean, it's effectively zero. It's not low. It's it's well, let's call it zero. Sensitivity analysis replaces that false precision with an honest range of outcomes. It tells you not what the deal returns in your best guess scenario, but what it returns across the range of scenarios that are plausibly going to happen. Uh, anyone who took any sort of math or statistics or physics in, let's say, high school or college, you're gonna remember the Gaussi curve, the bell curve. Uh, that's exactly, you know, without diving into the details of that, effectively why that was created is because often in nature and often in natural processes, uh, we see that there are no single numbers. Uh, there are no integers, often is the case. What there are are probability ranges. And the Gaussian curve uh stands out because it's the most common probability range. It's the one we see crop up all across the spectrum and all sorts of disciplines. But that's what we're talking about here. We need to look at that full range uh because the zero standard deviation case, the center, the tallest part of that bell curve, uh, that's all it is. It's just the center of a curve. It's not, that's not the answer. That's not what's going to happen with your property. That's just probably the most likely scenario. And your base case, again, needs to be that most likely scenario. That's the whole point of creating a base case. You are creating that zero standard deviation um uh uh outcome. And then you're gonna iterate off of that in both directions, both the upside and the downside. So let's start with a very simple um sensitivity table. This is gonna be a one-variable table. Uh it's the simplest form. And you basically take a simple uh single input, vary it across a range, and observe what happens to your key output metric. So let's build one here so we can use a live example. Let's use a 30-unit multifamily acquisition as your working example. So let's say purchase price is $3.2 million. Uh let's do a 75% LTV at 7% interest, 25 year amp. That's uh that's about what you're looking at right now if you're to buy something at this moment. Might be able to get a little below seven, but plus or minus, that's about what we're at with uh commercial rates. So that equity is $800,000. Let's project in year one NOI of $224,000. Um, let's project annual rent growth at 3%. Let's project our exit cap rate at 6.5% at the end of the five year. That's pretty good. That probably means this is a BB minus asset at a 6.5 cap. Um, and so in the space case, your deal produces roughly a 16% IRR internal rate of return and a 2.1% uh 2.1x equity multiple uh over that five years. So that looks pretty solid. So now let's sensitize on the rent growth. That's gonna be our one variable that we're changing. So let's say instead of assuming 3%, let's see what happens at 0%, 1%, 2%, 3%, 4, and 5%. At 0% rent growth, meaning rents stay completely flat for five years, IRR drops about 9%. So that's down to uh, let's see, that's 5% now. Because we are excuse me, 7%. We started at 16. Uh so the deal could still work, uh, but the return is dramatically different from the base case. At 1% rent growth, IRR is 11%. At 2% rent growth, we're at an IRR of 14. All the way up to a 5% rent growth. This is our optimistic case, we're at it looking at 20% IRR. So that's a big swing. We went from 7 to 20 just on that sensitivity analysis. And what this table now tells you immediately is how sensitive the deal is to rent growth. The spread between zero, uh, the zero percent scenario and the base case is seven percentage points of IRR. That's a massive range, just from zero to three percent rent growth. It tells you that a significant portion of your return is coming from the rent growth assumption, not from the property's current operating expense. And understanding that from the beginning means if you were to go forward with this deal, you know in your head rent growth is very important. I'm going to assume you're going to hire a property management company, uh, in which case, now you know when leases start getting renewed and new leases are signed, that rent growth is very important. And that's a conversation you'd have to have with your property manager that we need to be sticking uh to my, let's say, 3% assumption. And any more than that is excellent. So this is very valuable information. It does not tell you uh what rent growth will actually be. That's not the point of this, but it does tell you how much your return depends on getting that assumption right. And if most of your return is concentrated in an assumption you have limited control over, you should want to know that before you close. So let's up the empty here and go with a slightly more uh complicated system here. Let's go with a two-variable sensitivity table. Um, one variable tables are useful, but they're limited, of course. Real deals have multiple uncertainties interacting simultaneously. So the two variable table, sometimes called a data table or matrix, this lets you see the interaction between two inputs at the same time. The most common and most useful two-variable table in real estate is exit cap rate versus rent growth. Uh so whenever you're underwriting deals, maybe we're using a software, uh, BHBA has a native um uh deal analyzer um software. And we do uh allow you to change both uh both drivers. Uh so in the case of a table, this is would be both the um uh the column uh columns and the rows. You can change those across several different inputs. Uh, but at the end of the day, you're comparing two different drivers uh in this matrix. And so this is the most common uh sized um sensitivity analysis table. And the two most common drivers, your row and your column uh headers, is going to be exit cap rate versus rent growth. That's the most common. Uh and they tend to move in related but unpredictable ways, which is why it's very interesting to run this analysis and see what comes out. So let's use that same deal we started with, that 30-unit multifamily, and build this matrix. So on the one axis, let's go with rent growth uh from zero to five percent in one percent increments. Um on the other axis, let's do exit cap rate. So let's start with a 5.5%, which is, you know, let's if we assume this was a B minus asset in a tertiary market. Uh 5.5 is definitely optimistic. Um and let's go up to 8%, which is probably um that's conservative. You know, you're probably a decent asset with good cash flow. Uh, it's probably in that kind of six and a half, seven percent cap. Um, cap rates are compressing right now. That just came out of CBRE's um uh most recent analysis, Q uh quarter one, 2026. I think national average was just above 5.5% for multifamily, but I tend to stay in the sixes, even up to seven percent because I underwrite conservatively. So let's do on the cap rate driver side, let's ink do increments of a half a percent uh because we've got a smaller range here. So in the base case, right in the middle, we got a 3% rent growth and a 6.5 uh percent exit cap rate. That IRR uh and and the the metric we're gonna be looking at uh that we're gonna be producing at the end of this analysis is internal rate of return. So that IRR in that base case was 16%. Let's look at the corners. If you can picture, we've just made a table uh with rows and columns. The upper left-hand corner and the lower right hand corner are effectively your extremes. So at 0% rent growth and an 8% exit cap, this is your worst case scenario. Um, you are looking at uh the realistic downside scenario in a, let's call it a softening, uh IRR drops to approximately 2%. You are barely getting your money back. And your equity multiple, therefore, is pretty close to 1x over your five years. So that's your worst case scenario. It's good to know that. Let's go to the other side of the spectrum at 5% rent growth and a 5.5 exit cap, uh, which is the optimistic case, the most optimistic case. We've got an IRR of approximately 26%, which is a home run. Um, the spread between the worst case and the best case in this matrix is 24 percentage points of IRR. That range is the actual risk profile of the deal, given the range that we put in. Of course, if you increase that range or tighten the range of both of your drivers, uh, that is going to change. And so this comes back to it's all about your inputs. Whatever you put in is the data you're gonna get out. Now, I think that our exit cap and uh in this particular case, the exit cap and the rent growth, um, the the range we chose is very realistic. Though those really are optimistic numbers on the one hand, and those really are conservative numbers on the other hand of the spectrum. And so you're looking at a pretty big swing here. And that really is the risk profile of this deal. So this matrix changes how you evaluate the deal. Instead of asking, does this deal return 16%? you are asking across the realistic range of outcomes, how often does this deal produce an acceptable return? If 70% of the sales in your matrix are above your hurdle, let's say your hurdle's a 10, 11, 12% IRR, if 70% of the sales in the matrix are above that, you probably have a deal. And if only 30% uh, on the other hand, if only like let's say 30% are above your hurdle rate, your base case return is misleading you uh and or this uh deal is not worth your time. Most likely. So let's talk about sensitivity and analysis on the downside now. Most operators, when they run sensitivity analysis, if they do it all, they run it symmetrically. They look at upside and downside equally. But the purpose of sensitivity analysis is not to see how good things could get. As fun as that is, it's to understand how bad things could get and whether the deal survives. This means the downside scenarios deserve more weight and more attention than the upside scenarios, because in real estate, the downside is where capital is destroyed. The upside that takes care of itself. The three downside scenarios I recommend running on every deal are scenario one, flat revenue, rising expenses, no, so therefore, no rent growth and let's say expense growth. Let's make it high. Let's make it four to five percent annually, which reflects um, you know, that really actually does reflect where insurance, property taxes, and like labor and even materials are right now. Um, that's kind of the growth we're looking at in this inflationary environment. This scenario compresses NOI over time. And so we want to calculate what happens to DSCR cash flow and your refinance coverage at year three and year five. If DSCR drops below 1.15 in this scenario, the deal has limited margin for a revenue miss. Now, this is gonna be a tight deal, and uh things would have to go well. So now that's scenario one. Scenario two, uh, let's look at an occupancy shock. So let's say current occupancy drops by five to seven percentage points and stays there for a 12-month period before recovering. That could be for a number of reasons. It could be related to the property, maybe a major capital or maybe a major system failure that requires a big capex. A bunch of people didn't renew their leases because of it. You know, maybe that's a leaking roof, maybe that's issues with an elevator, whatever it is. This could also be uh a for reasons outside of your control, such as a brand new complex just got built next door to you and you experience an occupant occupancy shock because of that. Um so we got to calculate the cash flow impact and whether you can service the debt through that trough. And and and uh sensitivity analysis can allow you to do that and see what the full spectrum is. Scenario three that I always run uh exit cap rate expansion. Uh your exit cap rate is uh 100 to 150 basis points wider than your base case assumption. So that's a that's a very good case to run because that is a real risk. If your IRR is affect is you know being driven largely by the exit of the deal and you miss on your cap rate assumption, your IRR when the deal is done is gonna plummet. So you got to calculate whether the deal still produces a positive return in this scenario. If you miss the cap rate and it's higher than you expected, um, if a hundred basis point cap rate move uh turns your deal from 16% IRR to 5%, the deal's return is almost entirely dependent on the exit insumption. And you've got to know that, which because you got to nail that exit assumption. Um, next I want to turn to what's called a tornado chart. Uh, once you've run sensitivity on multiple variables individually, the next step is to rank them by impact. This produces what we call a tornado chart, named for its shape when you plot the results. So for each variable, you show the range of the output metric. Typically IRR or equity multiple, I tend to look at IRR. And even that, um, and when that variable swings from its pessimistic, pessimistic value to its optimistic value while holding everything else uh at the base case. And so that's what you are, that's what you're going to be charting here. So for a 30-unit example, the ranking might look something like this. Exit cap rate. Uh, when we when we varied that, the IRR ranges from 8% to 24% as the cap rate moves from 8% to 5.5. So the total swing there, 16 points. The next biggest swing was rent growth. We had a range from 9% to 20% as rent growth moves from 0 to 5%. So that total swing is 11 points. And then uh we'll work down a little bit. Um, here's another variable you could run as a sensitivity. That's the purchase price. Um, so in this case, if we move the purchase price from 3.5 million to 2.9, you're gonna see an IRR range from 12% to 20%. That's an eight-point swing. So notice we're working down on how big of a swing um each of these variables produces. Exit cap rate at the top, 16. We're now down to purchase price, which only swung eight points. Let's add a couple more in here, uh, just so you get the picture. Interest rate, we ranged from 13 to 19% as we moved rates from 8% interest to six. That's a six-point swing on IRR. Vacancy rate uh was a five-point swing. We moved it from a 10% vacancy down to a 3% vacancy. And finally, operating expenses. Uh, that was a four-point range in the IRR, 14 to 18%, as we uh moved expense growth from 5% down to 1%. Um, so this ranking tells you exactly where to focus your due diligence and your negotiation energy. Exit cap rate and rent growth dominate the return. Purchase price is the third most important variable, and that's one you have uh the most direct control over, because that's you know, that's you're either gonna do the deal or not. So you have more or less complete control over it. And of course, you don't have control over the price, but you have control over whether you're gonna buy the asset or not. And so now you can picture that tornado shape. Uh at the top, you've got your biggest range, all the way down to the bottom, your smallest range. And this is, again, gonna order in importance the variables that have the biggest return, uh, big biggest swings on your return, which is not only gonna help your underwriting, but if you were to end up closing this deal, those are going to be the variables you're gonna want to watch the closest. And maybe if you have, if you're tracking your portfolio performance or this property's performance on a dashboard, you might want to put those variables at the top of your dashboard to ensure that you're hitting uh your base case or above your base case uh for these input assumptions. So the, you know, in this particular case, the interest rate mattered, but it mattered less than most people think. And in this case, it mattered it was, you know, towards the very bottom of our tornado chart, which is counterintuitive. Um, vacancy and operating expenses matter the least in terms of total return impact, although they matter enormously for cash flow and debt coverage on a month-to-month basis. But in this case, we are looking at IR. So the tornado chart, the whole point is to replace intuition with data. So instead of vaguely knowing that cap rates matter, you now know exactly how much they matter relative to everything else. And that precision changes how you allocate your time and attention during underwriting. So let's let's shift the focus here to sensitivity analysis and negotiation. So this is where sensitivity analysis becomes a tool and not just some analytical exercise that us data nerds love to run. When you know that your deal's return is most sensitive to exit cap rate, you know that the purchase price negotiation is really a cap rate negotiation. If you can buy 25 basis points tighter on entry by negotiating a lower price, the impact on your five-year return is quantifiable. When you know that rent growth is the second most important variable, you can evaluate whether the seller's rent growth assumption is defensible by looking at submarket data, competing supply and demand trends. And if you determine that a 3% growth rate is aggressive and 1.5% is more realistically more realistic, you know exactly how much that adjustment costs you in return. And you can adjust your price accordingly, pointing to real numbers to the seller to justify why you need this at a lower price. When you know that interest rate is the fourth most important variable, you can make a more informed decision about whether to kit to pay points to buy down that rate with your lender up front. If a 50 basis point rate reduction adds two points to IRR, the cost of that reduction is one point up front. And let's assume that the cost was one point up front. The math tells you whether or not that's worth it. So this is the difference between negotiating on feel and negotiating on data. Sensitivity analysis gives you that data. Now, in the context of hold decisions, this framework applies equally well to assets you already own. In fact, it may be more valuable for hold decisions than for your acquisition decisions, because most operators never revisit the sensitivity of their existing portfolio. For each property you own, you can run forward-looking sensitivity on key variables, let's say rent growth, expense trajectory, cap rate at a hypothetical sale, refinance terms at maturity. The question you are asking is under what range of conditions does it make sense to continue holding this property versus selling selling or refinancing? If the sensitivity analysis shows that your current equity position produces an acceptable return across most scenarios, I would say holding is well supported. And if it shows that the return on your current equity is only acceptable in the optimistic case and the downside cases produce poor returns, you may be holding a position that is more speculative than you realize. And this connects directly to the equity yield analysis we've covered in earlier episodes. Sensitivity analysis adds dimensionality to that metric. It's not just what my equity yield is today, but what is my equity yield across the range of scenarios that might unfold over the next three to five years. Before we close here, I want to flag three of the most common mistakes I see operators make with sensitivity analysis. So mistake number one, uh running the sensitivity on too many variables at once. So when you vary everything simultaneously, the output really just becomes noise. You can't tell which variable is driving that change. So start with one variable at a time, build to two variable tables, and then use a tornado chart to identify the top three drivers. Then focus your analysis there. Mistake number two, using unrealistic ranges. If your rent growth sensitivity ranges from, let's say, negative 5% to positive 10%, you're modeling scenarios that frankly are never going to happen. And the point is not to capture every conceivable outcome. You've got to end your, your, you know, you've got to keep your tails fat, as they say, in a Gaussian curve. If you let those taper out to four, five, six uh standard deviations, again, you're introducing too many inputs and the output, um, the range, you know, the rate, the output range is going to be sure, I guess, you know, within the range of plausibility, uh, but you're going to introduce, again, more noise here. And so for rent growth between 0 and 5% is, I think, is almost always what I use. For exit cap rates, plus or minus 150 basis points from your base case is usually sufficient. Mistake number three here is ignoring the interaction effects. So variables in real estate are correlated. When interest rates rise, cap rates tend to widen. When occupancy drops, rent growth tends to stall. Running sensitivity on individual variables in isolation misses these correlations. And so the two variable table captures the most important interaction, but you should also think qualitatively about which scenarios tend to happen together and weight those combinations more heavily. So to wrap up here, here's my honest summary of sensitivity analysis is you gotta understand that your base case is a guess. It's a thoughtful, informed, well-researched guess, but a guess nonetheless. Sensitivity analysis does not eliminate uncertainty, it makes the uncertainty visible and quantifiable. The operators who do this consistently do not make better predictions about the future. That's not the point. They make better decisions in the face of uncertainty because they understand the range of outcomes their capital is exposed to before they commit it. So run the tables, build the matrix, rank the drivers. And when someone asks you what return a deal produces, give them a range, not a point estimate. The point estimate is the base case. The range is the actual truth. We'll close this episode today with a fun fact, which is that sensitivity analysis was originally developed not for finance, but for military logistics during World War II. Operations researchers at the RAN Corporation in the 40s and 50s needed to understand how changes and assumptions about enemy positions, supply line reliability, and weather conditions would affect mission outcomes. The mathematical framework they built called parametric analysis was later adopted by engineers and then by financial analysts. The real estate industry was among the last to adopt it systematically, and most independent operators still don't use it at all. And the irony is that in real estate, with its long hold periods, illiquid positions, and high leverage, it's honestly arguably the asset class where sensitivity analysis matters the most. Hope you got a lot out of this episode, and we will see you next week. This podcast is produced by Beacon Hill Property Advisors, where we focus on bringing clarity, structure, and rigor to real estate investment analysis. If you want to evaluate deals beyond headline metrics and better understand the mechanics driving performance, you can learn more about our tools and approach at bhpropertyadvisors.com. You can also connect with us directly for demonstrations, resources, and additional insights. Until next time, analyze deeply, allocate wisely, and always go beyond IRR.