Rent-to-market gaps at institutional scale rarely close the way the model says they will

Question

A 180-unit acquisition underwrites to 94 percent market rents on day one, with a 24-month glide path to get the in-place roll to market. That sounds orderly until you sit with what "market" actually means for a tenant who signed at $1,150 and is being told the renewal comes in at $1,420. That is a 23 percent increase in one letter. The model shows it as a line item. The tenant experiences it as a crisis, and the behavioral response does not follow a cap rate.

The part that institutional underwriting tends to handle badly is the segmentation problem. A 180-unit building is not one rent roll, it is three or four cohorts with completely different elasticity. The longest-tenured residents are often the furthest below market, which means the largest proposed increases land on the people with the strongest community ties and the least financial flexibility. Turnover cost at $4,500 to $6,000 a unit, plus 30 to 45 days of vacancy during repositioning, means a 15 percent resident exit rate on the roll-to-market cohort can eat the first year of gained revenue entirely. The model has to hold both the increase and the retention assumption simultaneously, and most decks I have reviewed treat them as independent variables when they are not.

The lever that actually changes the math is sequencing. Operators who phase increases by lease expiration date rather than by unit type tend to absorb the shock better, because natural turnover gives you renovated units at true market rate while the remaining in-place tenants see a smaller gap each renewal cycle. The blended yield gets there more slowly, but the downtime and concession costs are materially lower. A 36-month glide path with 8 to 12 percent annual increases on the below-market cohort almost always outperforms a 24-month push when you price in actual turn costs and lease-up concessions on the re-let units.

The assumption doing the most work in most models I have seen is that the replacement tenant signs at 100 percent of market with zero concessions. In submarkets with 12 to 15 weeks of supply at current absorption, that holds. In submarkets sitting at 8 or 9 percent vacancy with new deliveries still landing, the replacement tenant is signing at market minus one month free, and the effective rent is 7 to 8 percent below the face rate the model used. That gap compounds across 30 or 40 units and the IRR moves before anyone updates the rent roll.

What does your current deal look like on the below-market cohort specifically: how many units, how far below, and what turn cost did the model use?

brett_hollis21 · Answer

The sequencing point is right, but the cohort that breaks the model first is usually not the longest-tenured resident, it is the mid-tenure tenant who signed 18 to 24 months ago at a rent that already felt like a stretch. That cohort has no community-anchor loyalty to offset the payment shock, and they are mobile enough to leave quickly, which means your turn cost assumption crystallizes earlier in the glide path than the model expects. If 20 percent of a 180-unit roll falls into that band, you can lose 36 units of occupancy in months 6 to 12 before the below-market legacy cohort even receives their first repositioning notice.

The related strategy that changes the picture here is phased capital deployment tied to lease expiration sequencing. Operators who stage unit renovation spend to match the natural expiration calendar reduce carrying cost on vacant units because they are not holding a renovated unit dark waiting for a qualified applicant at a price point the submarket may not yet support. Say you renovate 15 units in month one and the absorption rate at true market is 60 days rather than the 30 the model assumes: you are now carrying $90,000 to $120,000 in unrealized revenue against renovation cost already spent, and that gap compounds if the submarket has any softness in months 8 through 14.

The 100 percent of market replacement assumption is where I would push back hardest in any deck. Concession-adjusted effective rent in a repositioning lease-up is almost never face rent, and the gap between the two is where IRR projections go quiet. What does the submarket absorption data actually show for units in that price band over the past four quarters?

Rena Expert · Answer

The sequencing argument is right, and the 36-month glide path consistently wins on the math once you build in realistic turn friction. The assumption doing the most work, as you note, is that the replacement lease comes in at face market with no concessions. The second assumption doing nearly as much damage is the turn cost itself.

$4,500 to $6,000 is a reasonable range in markets where you are doing light cosmetic work. In a genuine value-add play where the repositioning scope includes flooring, appliances, and fixtures, all-in turn cost on a below-market unit can run $8,000 to $14,000 before you count the vacancy drag. If the model used the low end of that range and the operator is actually executing a full renovation to justify the $1,420 ask, the cost basis per turned unit is wrong from the start. The cohort-level underwriting looks fine on paper until unit 22 turns and the equity burn is already off-plan.

The retention and increase assumptions being modeled as independent variables is the structural error you are naming. They are correlated, and the correlation is non-linear. Residents with three-plus years of tenure and large absolute-dollar gaps do not exit at the same rate as residents with 14-month tenure and a $90 gap. Any deck that uses a single blended retention assumption across the below-market cohort is smoothing over the actual distribution of risk, and the IRR sensitivity to that cohort is disproportionate because those are the units where the value-add thesis lives.

The submarket supply question compounds everything. You are right that 8 to 9 percent vacancy with active deliveries means the replacement tenant has negotiating leverage, and a free month on a $1,420 unit drops effective rent to roughly $1,302 annualized. Model that across the units that actually turn on your 24-month timeline and the blended rent roll in month 25 looks materially different from the pro forma.

On your specific deal: what tenure profile does the below-market cohort carry, and did the underwriting segment turn cost by renovation scope or apply a single per-unit assumption across all 180?

Rent-to-market gaps at institutional scale rarely close the way the model says they will

2 replies