Senior Housing Deep Dive: Levels of Care, Financing & Underwriting Metrics

The property-type overview establishes senior housing as a single category defined by housing-plus-care and operator dependency. That framing is a starting point, not an underwriting model. Level of care — independent living, assisted living, memory care, skilled nursing, and where the resident's entrance fee sits inside a CCRC — is the variable that actually determines which lender will touch the deal, what the operator's Regulatory Agreement obligates them to do, how labor cost behaves as a share of revenue, and whether the community is carrying an actuarial liability that a purely rental asset never would. Two 120-unit senior housing communities half a mile apart can be underwritten in almost unrelated ways once one turns out to be a private-pay rental AL/MC campus and the other a Type A CCRC with a 90%-refundable entrance fee — same property type on a stacking plan, entirely different credit.

This topic assumes the level-of-care taxonomy and RIDEA's basic vocabulary from the overview are already familiar, and goes directly into HUD Section 232's program mechanics, how a RIDEA lease actually separates ownership from operations, why operator quality is treated as a primary credit factor rather than a qualitative afterthought, and the three metrics — level-of-care occupancy and rate, the labor cost ratio, and CCRC entrance-fee refund liability — that a senior housing underwrite cannot skip.

HUD Section 232 insures long-term, fixed-rate, non-recourse mortgages on licensed residential care facilities — nursing homes, board-and-care homes, and assisted living, including memory care — processed almost entirely through the Section 232 LEAN program administered by HUD's Office of Residential Care Facilities. LEAN standardizes underwriting through HUD-approved lenders and a common electronic submission process, but 'streamlined' is relative: initial application through firm commitment routinely runs six to twelve months, driven by state licensure review, an independent physical-needs assessment, and HUD's own credit review — which is why 232 fits a stabilized permanent-financing or planned-refinance timeline far better than a competitive acquisition.

Two variants cover the deal lifecycle: 232 new construction/substantial rehabilitation insures ground-up development or major renovation, while 232/223(f) insures acquisition or refinance of an existing, stabilized facility — the more common execution for underwriting a going-concern purchase. Both carry an upfront and annual mortgage insurance premium (MIP), require funding a reserve-for-replacement escrow sized to the physical-needs assessment, and cap distributions to surplus cash under the HUD Regulatory Agreement rather than allowing an unrestricted cash sweep to ownership.

The mechanic that most distinguishes 232 from a conventional or agency execution is who signs the Regulatory Agreement: HUD requires both the mortgagor (the borrower/owner entity) and the licensed Operator to execute it, and it holds the Operator itself to HUD's net-worth, liquidity, and experience requirements — not just the borrower. That structure assumes the borrower and the operator are either the same party or closely aligned; changing the operator mid-term requires HUD approval of the incoming operator and an amended Regulatory Agreement, adding a layer of lender-controlled friction to any operator transition that a Fannie/Freddie seniors housing loan or a conventional bridge loan doesn't impose to the same degree. Sponsors who need faster, deeper flexibility to swap operators — a common feature of the RIDEA structures below — often find 232 a worse fit than agency or life-company debt, even though 232 offers the longest amortization and lowest fixed-rate coupon available for licensed care real estate.

Before 2008, a REIT that wanted to own senior housing had almost no choice but to lease the property to an operator-tenant under a triple-net structure, because REIT tax rules require the bulk of a REIT's income to come from real property rents, not from operating an active trade or business — and running a licensed care community is unambiguously an operating business. That forced every REIT-owned senior housing deal into a bond-like credit lease: the REIT collected fixed or CPI-escalated rent from the operator-tenant and had no direct exposure to occupancy, rate, or labor-cost swings in either direction.

The REIT Investment Diversification and Empowerment Act (RIDEA), enacted as part of the Housing and Economic Recovery Act of 2008, created a specific carve-out for qualified health care and senior living properties: a REIT can lease the property to its own taxable REIT subsidiary (TRS), and that TRS can in turn engage a third-party 'eligible independent contractor' to manage day-to-day operations and hold the operating licenses, without the REIT losing its tax status. The economics that flow up to the REIT are no longer a fixed rent check — they are the TRS's operating results, net of the management fee paid to the independent operator, so occupancy, rate, and labor cost all flow through to the ownership entity's income almost as directly as they would for a wholly owned operating business.

That distinction changes how the deal gets underwritten, not just how it gets taxed. A triple-net lease to an operator-tenant is underwritten like a credit lease — the analysis centers on the tenant's lease guarantee and a coverage ratio, rent measured against facility EBITDARM (earnings before interest, taxes, depreciation, amortization, rent, and management fee) — because the landlord's cash flow is contractually insulated from day-to-day operating performance as long as the tenant keeps paying rent. A RIDEA structure has no such insulation: there is no intervening fixed-rent tenant to look through, so the underwriting has to be built the way an operating business would be underwritten — facility-level occupancy and rate by level of care, the labor cost ratio, payor mix — because the ownership entity now owns that performance directly, upside and downside both. RIDEA assets accordingly gravitate toward agency seniors housing programs, life-company debt, or CMBS rather than HUD 232, since 232's Regulatory Agreement anticipates a borrower and operator that are closely aligned or identical, which sits awkwardly against a REIT/TRS/independent-operator stack built specifically to keep the operator at arm's length from the ownership entity.

Payor source is the variable that decides how much of that operating performance is even within the operator's control. Private-pay independent living, assisted living, and memory care communities set their own rates and can reprice against local competitive supply and wage inflation with a rate letter. A Medicaid-dependent skilled nursing facility cannot: reimbursement is set externally by state legislatures and CMS formulas, is frequently rebased on a lag behind actual cost inflation, and can be frozen or cut in a state budget cycle regardless of the facility's occupancy or operating quality. Medicaid-heavy long-stay facilities often show higher, stickier occupancy than private-pay competitors — custodial residents don't discharge quickly, and there is no affordability ceiling limiting demand — but that occupancy strength coexists with thinner, policy-exposed margins, which is why lenders underwrite Medicaid-dependent assets to lower leverage and a higher minimum DSCR than a comparably occupied private-pay community.

Operator quality sits on top of payor mix as a second, independent credit variable — and in senior housing it functions as a primary underwriting factor in its own right, not a qualitative footnote the way property-management quality is treated in conventional multifamily. A struggling or thinly capitalized operator can impair a well-located, well-built community's performance regardless of payor mix, because staffing decisions, survey and licensure compliance, and day-to-day care quality drive both revenue (through occupancy and rate) and expense (through the labor cost ratio) more directly here than in almost any other property type. Institutional underwriting reflects that by pulling the operator's own financial statements and portfolio-wide occupancy/staffing trends alongside the subject property's, sizing debt against the operator's EBITDARM coverage rather than the real estate's NOI alone, and negotiating loan-document rights — a replacement-operator provision, minimum liquidity or net-worth covenants on the operator, and lender consent rights over any change to the management or lease agreement — so operator distress triggers a contractual remedy rather than a surprise. Lenders with several loans to the same regional operator also track operator concentration risk, since that single management team's execution is exposed across every one of those communities at once.

A blended occupancy or average-rate figure for a multi-level senior housing campus averages together levels of care with different rate structures, different staffing intensity, and different marginal value per vacant unit — which means the blended number can look healthy while the highest-revenue level of care is quietly underperforming underneath it. The correct sequence is always to compute occupancy and average rate separately at each level of care before ever looking at the blended figure.

Take a 220-unit rental senior housing campus with three levels of care. Independent living has 100 units, 90.0% occupied (90 occupied units) at an average rate of $3,000/unit/month. Assisted living has 80 units, 82.5% occupied (66 occupied units) at $5,500/unit/month. Memory care has 40 units, 90.0% occupied (36 occupied units) at $7,000/unit/month.

Occupied revenue by level: IL is 90 × $3,000 = $270,000/month; AL is 66 × $5,500 = $363,000/month; MC is 36 × $7,000 = $252,000/month. Total monthly revenue is $270,000 + $363,000 + $252,000 = $885,000. Total occupied units across the campus are 90 + 66 + 36 = 192, against 220 total units, for a blended occupancy of 192 ÷ 220 = 87.3%. Revenue per occupied unit (a blended rate) is $885,000 ÷ 192 ≈ $4,609/month, and revenue per available unit — RevPAU, the occupancy-adjusted figure — is $885,000 ÷ 220 ≈ $4,023/month.

An analyst who stops at 87.3% blended occupancy and roughly $4,600 blended rate would call the campus healthy. Decomposed by level, assisted living — the largest single revenue-per-unit segment on the campus after memory care — is running the weakest occupancy of the three at 82.5%, ten points below both IL and MC. Its 14 vacant units (80 − 66) represent 14 × $5,500 = $77,000/month, or $924,000/year, of unrealized revenue at the current rate — a concentrated gap the blended 87.3% figure never surfaces, because IL and MC's stronger occupancy does the averaging.

In conventional multifamily or triple-net retail, payroll is a minor line against contract rent. In senior housing it is typically the single largest operating expense, commonly running 45%–55% of effective gross revenue (EGR) in private-pay assisted living, 55%–60%+ in memory care given its round-the-clock, dementia-trained staffing model, and higher still in skilled nursing under state-mandated licensed-nurse ratios. Because the labor cost ratio also varies structurally by level of care — independent living runs the lowest ratio of the continuum, since it needs hospitality staff rather than caregivers, while memory care runs the highest — a campus's blended labor cost ratio moves not only with wage inflation but with the campus's level-of-care mix itself, a variable most underwriting models never isolate.

Continue the 220-unit campus above and add a labor cost ratio to each level: IL at 32% of its EGR, AL at 52%, and MC at 60%. Labor cost by level is IL: $270,000 × 32% = $86,400; AL: $363,000 × 52% = $188,760; MC: $252,000 × 60% = $151,200. Total labor cost is $86,400 + $188,760 + $151,200 = $426,360 against total revenue of $885,000, for a blended labor cost ratio of $426,360 ÷ $885,000 = 48.2%.

Now suppose ownership repositions the campus by converting 10 IL units to memory care — a common play chasing MC's higher rate — leaving IL at 90 units (81 occupied at the same 90.0% occupancy) and MC at 50 units (45 occupied at the same 90.0% occupancy), with AL unchanged at 66 occupied. Revenue is now IL: 81 × $3,000 = $243,000; AL: $363,000 (unchanged); MC: 45 × $7,000 = $315,000; total $243,000 + $363,000 + $315,000 = $921,000. Applying the same three level-specific labor cost ratios — nothing about staffing efficiency changed at any single level — labor cost is IL: $243,000 × 32% = $77,760; AL: $188,760 (unchanged); MC: $315,000 × 60% = $189,000; total $77,760 + $188,760 + $189,000 = $455,520. The blended labor cost ratio is now $455,520 ÷ $921,000 = 49.5% — a 1.3-point increase, and $455,520 − $426,360 = $29,160/month ($349,920/year) more labor cost, purely from shifting the unit mix toward the highest-labor-cost level of care. An underwriter tracking only the blended ratio would see labor cost creeping up year over year and might reasonably suspect wage inflation or staffing inefficiency; the actual driver here is a deliberate, otherwise value-accretive mix shift that a level-of-care-blind expense model cannot distinguish from a genuine cost-control problem.

A CCRC entrance fee is not senior housing's version of a security deposit, and it is not upfront revenue either — it is a payment that has to be split, on day one, into a deferred-revenue component the community earns over time and a liability component it may owe back. The contract type sets the split. A Type A (life care/extensive) contract bundles IL housing with unlimited or near-unlimited AL, MC, and SNF care for a fixed monthly fee regardless of the resident's actual acuity, so the community — not the resident — bears essentially all of the future cost-of-care risk embedded in that entrance fee. A Type B (modified) contract includes only a capped amount of higher-level care before the resident shifts to fee-for-service billing, sharing acuity risk between resident and community. A Type C (fee-for-service, sometimes called 'rental' CCRC) contract sells IL housing only; the resident pays market rate for AL, MC, or SNF care if and when needed, and the community carries little to none of the future-acuity risk that entrance fees on Type A and B campuses are pricing in.

Whatever the contract type, the entrance fee itself is typically split into a refundable portion, held as a liability because it is contractually owed back to the resident or their estate, and a non-refundable portion, recorded as deferred revenue and amortized into income over the resident's actuarially-adjusted remaining life expectancy rather than recognized as revenue when cash is received — the CCRC-specific guidance under U.S. GAAP (ASC 954, Health Care Entities) treats it this way because the community's obligation to house, and on Type A/B contracts to care for, that resident extends across their remaining stay, not just the year the fee was collected.

Work a simple cohort example. A CCRC signs 40 new Type B residents in a year at an average entrance fee of $320,000 each, with the contract 50% refundable to the resident's estate and 50% non-refundable, and an actuarially-adjusted average remaining life expectancy of 10 years for this cohort. Per resident: refundable portion = 50% × $320,000 = $160,000, booked as a refundable-fee liability, not revenue; non-refundable portion = 50% × $320,000 = $160,000, booked as deferred revenue and amortized straight-line over 10 years, or $160,000 ÷ 10 = $16,000/year recognized into income.

Across the 40-resident cohort: total cash collected is 40 × $320,000 = $12,800,000. Total refundable-fee liability booked is 40 × $160,000 = $6,400,000. Total non-refundable deferred revenue is also 40 × $160,000 = $6,400,000, of which only 40 × $16,000 = $640,000 is recognized as year-one income; the remaining $6,400,000 − $640,000 = $5,760,000 of deferred revenue amortizes into income over the cohort's remaining nine years. In other words, the community collected $12,800,000 of cash but recognized only $640,000 of it as current-year revenue — the other $12,160,000 sits on the balance sheet as a refundable liability and unearned deferred revenue, not as NOI available to service debt or distribute to ownership.

Type A and Type B communities carry one more obligation entirely absent from rental senior housing: because the entrance fee and future monthly fees are supposed to fund the resident's future care, GAAP requires the community to test whether the present value of the costs it expects to incur delivering that future care and any refunds due exceeds the present value of the future revenue — remaining unamortized entrance fees plus expected future monthly fees — it expects to collect from those same residents. Where expected future cost exceeds expected future revenue, the shortfall is booked as an additional liability, the Obligation to Provide Future Services and Use of Facilities, on top of the deferred-revenue and refundable-fee liabilities already on the balance sheet. That obligation is effectively an actuarial bet that acuity costs, medical inflation, and resident longevity will track the assumptions priced into the entrance-fee and monthly-fee schedule when the contract was signed — a real, quantifiable balance-sheet risk that a rental IL/AL/MC campus, or a Type C fee-for-service CCRC, simply does not carry.