Renewal probability is doing more work in most office models than the cap rate is

The rent-to-market spread as a proxy for renewal probability is the right frame, and the point about above-market leases being misclassified as credit risk is one that gets missed consistently in office underwriting. The tenant holds optionality. They pay until expiration, leave cleanly, and the model that carried an 85 percent renewal assumption suddenly needs to fund a TI package, absorb six to twelve months of downtime, and lease into a submarket that may have moved further against the building in the intervening years.

The assumption doing the most work in suburban office right now is not the renewal probability itself. It is the re-leasing spread underneath it. When a model assumes 85 percent renewal probability and the 15 percent vacancy scenario uses current asking rents, that is two compounding optimisms in adjacent cells. Suburban asking rents in secondary markets have been under real pressure since 2022, and effective rents, after concessions and TI, are lower still. A model that does not stress the re-leasing rate independently of the renewal probability is not actually stress-testing the vacancy scenario.

The TI funding issue you named is the downstream consequence nobody prices correctly. Above-market leases feel like a buffer. They generate strong near-term cash flow, which is real. The problem is that cash flow tends to fund distributions or debt service rather than a TI reserve, so when rollover hits, the capital requirement lands on the equity at exactly the moment when the asset is most impaired. A lender watching that maturity wall will not extend easily.

One thing worth separating in your analysis: is the 40 percent occupied suburban building vacant because tenants left, or because it was never leased up? The rollover risk profile on those two situations is different, even at the same occupancy number.

What does the submarket's net absorption look like over the last eight quarters for your size range, and does the renewal probability in the model have any empirical anchor to that data?

The point I want to add is about the TI budget problem you raised, because most proformas get the direction wrong and not just the magnitude. When an above-market tenant rolls, the landlord is negotiating from weakness: the space has a mark-to-market problem the incoming tenant can see, the submarket probably has alternatives, and the concession package required to backfill is not the same TI allowance used for the original deal. It is the TI required to compete with Class A product that leased in the last 24 months, which in most suburban office markets means you are looking at $60 to $80 per square foot before you get to free rent. A proforma that funds $35 per square foot in a reserves line and calls it conservative has simply not asked what the competing space is offering. The number to check is what the last three leases in that submarket actually cost the landlord all-in, including free rent months converted to present value, because that is what you are bidding against when the space goes dark.

The risk that goes unnamed in most of these threads is sequencing. A 40 percent occupied building with two tenants whose leases expire in the same 18-month window does not have rollover risk twice, it has rollover risk that forecloses the first renewal conversation before the second one starts. A prospective tenant doing diligence on that building will see a co-tenancy profile that is about to look materially worse, and that changes their negotiating posture before you even get to economics. The model treats each lease as an independent probability, but the tenant on the longer lease is watching what happens to the tenant on the shorter one. What does the lease expiration schedule actually look like across the whole rent roll, and are any of the shorter leases anchor tenants whose exit would change how the remaining space markets?

What submarket and vintage is the building, because those two things alone can move defensible renewal probability by thirty points.

The piece most models miss is that renewal probability should shift based on the tenant's own unit economics, not just rent-to-market spread, and those two measures diverge more than people expect in suburban office right now. A tenant paying 12 percent above market may still renew if their buildout is fully amortized, their staff is geographically sticky, and moving costs them more than the rent delta over the next term. Conversely a tenant at market or slightly below may be running a hybrid policy that quietly reduces their footprint at every opportunity, making 80 percent renewal probability look reasonable on the rent roll and completely wrong operationally. The TI exposure you raised is the sharper version of this: a tenant who shrinks by 30 percent at renewal and asks for a full re-fit on the retained space costs the hold more than a clean departure with a straightforward re-leasing assumption, and that partial-renewal scenario almost never gets its own line in a waterfall model.

What is the average remaining lease term across your rent roll, and do you know whether any of these tenants have exercised renewal options before on this same space or negotiated contractions at prior rollovers?

Who set the renewal assumption and what was their incentive when they set it?

Nobody I have ever seen defend an 85 percent renewal assumption on an above-market lease has had anything more than "the tenant has been there twelve years" as the citation, and long tenancy is actually a sign they negotiated hard last time and will do it again or walk.

What does the current lease rate per square foot look like relative to what is actually transacting in that submarket right now, not asking about the stated term, asking about the spread?

Four to five IRR points is probably conservative if the TI gap is real. On a deal I looked at in Roseville in 2021, the proforma had $28 per foot budgeted for TI on a re-leasing assumption and actual comps in that corridor were running $52 to $58. The renewal probability was 80 percent, which meant the model was only pricing that gap 20 percent of the time. We passed, building traded anyway, tenant left in year three, and the new owner is still sitting on dark space.

The above-market lease as credit risk versus rollover risk framing is the part worth pushing on. Credit risk gets a reserve. Rollover risk gets an optimistic renewal cell and a prayer.

What was the average lease size in the deal you're looking at, because tenant concentration changes how much that renewal number actually matters. One 40 percent tenant at 85 percent renewal probability is a completely different animal than five tenants at equal weight with the same blended assumption.

Whoever funded the TI reserve on that above-market tenant as a line item in the acquisition model and then buried it below the waterfall needs to answer for it, because I have seen that specific move eat a deal whole in year three.

What I have never seen work is using broker opinion of market rent as the comp basis for the spread calculation. Brokers quote asking rent, and in suburban Atlanta right now actual executed leases are closing 18 to 22 percent below that, so the tenant who looks 8 percent above market is actually 10 to 14 percent above, and the renewal probability cell just became fiction.

Where did the market rent figure in your model come from?

The TI funding gap compounds when the above-market tenant leaves mid-hold, because the replacement lease clock starts at zero while debt service does not pause.

What year does the above-market lease expire relative to your projected exit date?

The renewal probability is an output of the rent-to-market spread, not an independent input, and most models treat it as the latter.