How a consultant converts age-cohort demand, mandated ratios, and a realistic staffing model into a lender-grade projection for SBA, USDA, and conventional child care financing — in a sector where 59% of centers have closed classrooms while holding waiting lists.
The demand is real, and it is not the question
Child care is one of the few asset classes where unmet demand is documented to a level most sectors never achieve. Roughly 46% of American children under six live in a licensed child care desert — defined as a community with at least 50 children under five and more than three young children for every licensed slot. Child care is the second-largest household expense after housing. Infant and toddler care in family-based settings now averages around $18,500 a year, an increase of roughly 80% since before the pandemic.
Any competent analyst can establish that a given trade area is underserved. Most feasibility studies for this sector do exactly that, at length, and then conclude that demand supports the project.
They have proven the wrong thing.
Approximately 76% of centers report staffing shortages. Worker turnover runs around 26% annually against a median wage near $14.60 an hour. And the consequence is the single most important statistic in this sector: 59% of centers have reduced enrollment or closed classrooms despite having families on waiting lists.
Demand is not the binding constraint. Staffed capacity is.
A projection built on licensed capacity in a documented desert will not be met, because the centre will not be able to hire the teachers required to open the rooms. The consultant's central task is to model the capacity the operator can actually staff, not the capacity the licence permits.
Why the economics resist scale
Child care has a cost structure almost uniquely resistant to operating leverage, and understanding why is prerequisite to modelling it.
Ratios are mandated, not managed. State licensing sets maximum children per caregiver by age group. Infant ratios commonly run 1:3 to 1:4; preschool 1:7 to 1:12. Those ratios are the dominant cost driver, and unlike almost every other business, an operator cannot improve margin by serving more customers per employee. The ratio is the regulation.
Infant care loses money nearly everywhere. The tightest ratio applies exactly where parental need is most acute and willingness to pay is most constrained. Centres routinely run infant rooms at a loss and cross-subsidise from preschool rooms where ratios are looser. A project proposing to serve infants — which is what most communities need — is proposing the least profitable part of the sector, and the model has to show where the offsetting margin comes from.
Wages are simultaneously too low to retain and too high to be covered. At a $14.60 median, early educators leave for retail and hospitality work. Replacing them costs recruitment, training and lost enrolment when a room cannot open. A projection carrying low wage assumptions and low turnover assumptions is internally inconsistent, and a lender's analyst will notice.
Revenue is capped by household income, not by value delivered. National average full-time centre-based care runs roughly $1,200 to $2,400 per month depending on region, with weekly averages near $321 overall and $376 for infants. Families cannot pay beyond their capacity regardless of how badly they need the place.
Utilisation is seasonal. Centres experience an average 23% enrollment decline over summer months as school-age children leave and preschool families take extended absences. An annual revenue model conceals this entirely.
Break-even sits high. Most centres require 60% to 70% enrollment to cover overhead, and in higher-cost markets 70% to 80%. That is a narrow operating band, and it means the ramp period is genuinely dangerous rather than merely slow.
Three different capacity numbers
The most valuable discipline a consultant brings to this asset class is separating three figures that sponsors routinely conflate.
Licensed capacity. What the state permits given square footage, egress and facility standards. This is a property fact.
Staffed capacity. What the centre can actually operate given the teachers it can recruit and retain at the wages the revenue supports. Almost always lower than licensed capacity, and the gap widens in tight labour markets.
Enrolled capacity. What is actually filled, by age group, net of attrition and seasonality.
Revenue is a function of the third. A pro forma computing revenue as licensed capacity multiplied by tuition multiplied by twelve months is describing a building, not a business.
What the consultant establishes: the staffing plan room by room at licensed ratios, the local wage required to fill those positions, whether that wage is supportable by the tuition the market will pay, and how long it realistically takes to hire a full complement. Where the answer is that the wages the revenue supports will not attract staff in that labour market, that is the finding — and it is more useful to the sponsor before the loan than after.
The revenue stack
Very few financeable centres run on private tuition alone. The consultant models each source separately, with its own eligibility rules, rates and timing.
Private-pay tuition. The base, evidenced against what comparable local providers actually charge and collect — not against state or national averages, which conceal enormous regional variation.
Child Care and Development Fund subsidy. Federal funding administered through state lead agencies for eligible low-income families. The state's payment rate relative to private tuition is one of the most consequential inputs in the entire model. In some states subsidy approaches market rate; in others it sits well below, and a subsidy-heavy enrollment mix then reduces revenue per child rather than stabilising it. Payment timing also matters, since subsidy reimbursement cycles affect working capital.
Child and Adult Care Food Program reimbursement. Modest per child but reliable, and it materially changes the food cost line.
State pre-kindergarten contracts. Where a state operates a pre-K programme, a centre may contract to deliver it. This frequently pays better than private tuition and provides enrollment stability, at the cost of additional quality and staffing requirements. It is competitive and capacity-limited.
Head Start partnership or delegate arrangements, in some circumstances, with their own service-area and designation rules.
Employer partnerships. In markets with a dominant employer — hospital, manufacturer, processor, university — an employer-supported arrangement can underwrite meaningful capacity. Where contracted, this materially strengthens the credit.
State facility and quality grants. These affect the capital stack rather than operating revenue, and therefore the loan sizing.
The analytical requirement is an enrollment mix by revenue source, evidence of eligibility for each, and a sensitivity showing what happens when the mix shifts — because it will.
What the consultant actually does
Defines the catchment by drive time
Child care is a twice-daily trip and tolerance for distance is low. A catchment drawn at fifteen minutes describes a different market from one drawn at thirty. For centres dependent on employment-centre proximity rather than residential proximity, the catchment follows the commute, not the neighbourhood.
Counts children by age cohort, not in aggregate
Infant, toddler and preschool are three separate markets with three separate ratios, cost structures and price points. A single "children under five" figure cannot be used to size rooms, staff or revenue.
The consultant establishes children by single-year cohort within the catchment, the proportion with all available parents in the labour force, and household income distribution to establish ability to pay.
Establishes existing supply precisely, including its utilisation
Licensed centres, licensed family child care homes, Head Start, school-based pre-K — with capacity by age group and, critically, actual utilisation.
A community with nominal capacity that is fully enrolled and holding waitlists is a genuine opportunity. One with vacant capacity is not, regardless of what the desert ratio says. The desert metric counts licensed slots; it does not tell you whether those slots are staffed and operating.
Recent closures are also evidence, and in a sector losing providers, worth documenting.
Tests the local labour market
This is the section that distinguishes a serious study, and it is routinely omitted.
What do comparable centres in this market pay lead teachers and assistants? What early childhood credentials are required by state licensing, and how many credentialed candidates exist locally? Is there a community college or CDA programme producing them? What are competing employers — retail, hospitality, school districts — paying for comparable-skill labour?
If the answer is that the tuition the market supports cannot fund wages competitive with the local Target, the project has a staffing problem that no amount of demand analysis solves.
Builds enrollment as a ramp, room by room
Rooms open as staff are recruited, not on the day the building is finished. The model shows monthly enrollment build by room, reflecting realistic hiring timelines and the sequence in which rooms come online.
Preschool rooms typically fill faster than infant rooms and carry better economics, which creates a genuine sequencing question the model should address.
Models the cost structure honestly
Payroll dominates and is largely fixed. Staff must be present to meet ratios regardless of daily attendance. Payroll is calculated from the staffing schedule at local prevailing wages, plus payroll taxes, plus benefits where offered, plus the recurring cost of 26% turnover.
Occupancy at contracted rent with escalations.
Food, net of CACFP reimbursement.
Insurance, including general liability, professional liability and workers' compensation — the last of which is material in this sector.
Supplies, curriculum and classroom materials, which are recurring rather than one-time.
Marketing, which cannot be zero for a new centre.
Replacement reserves. Playground equipment, kitchen equipment, classroom furnishings and vehicles all have finite lives. Where financing is USDA-guaranteed this is a regulatory requirement, since Part 5001 defines coverage as EBITDA less reasonably expected replacement capital expenditures.
Models seasonality monthly
With a documented average 23% summer enrollment decline, an annual model materially overstates cash available for debt service in specific months. Monthly modelling exposes the trough; annual modelling hides it.
Accounts for the licensing timeline
A centre cannot enrol before it is licensed, and licensing offices in several states are backlogged. This sits directly on the revenue ramp and has to be funded rather than assumed away. For a construction project, the gap between certificate of occupancy and licence is a real financing period.
Sensitivity that matters here
- Staffing shortfall. Coverage if the centre operates at 80% of planned staffed capacity because two rooms cannot be filled.
- Wage inflation. What a $1.50 an hour increase across the staffing schedule does to coverage.
- Subsidy rate change. Where the enrollment mix is subsidy-heavy, what a rate reduction does.
- Ramp delay. Coverage if full enrollment is reached at month 30 rather than month 18.
- Licensing delay. What three additional months of fixed cost before first revenue does to working capital.
These are the variables that actually determine outcomes in this sector, and a lender who sees them modelled recognises an analyst who has done this before.
How the programmes differ
Child care sits unusually well across all the major programmes, and the pathway depends on sponsor type and location.
SBA 7(a). The dominant route for for-profit operators. Roughly 3,800 child care originations over the last five fiscal years, at an average approval around $845,000. SOP 50 10 8 sets when a third-party feasibility study is expected — including startups, businesses under two years old, and changes of ownership, which covers most of this sector.
SBA 504. Applies where the transaction includes real estate. Notably, roughly 17% of SBA-financed child care projects run through 504, well above the programme-wide average — this is a property-driven asset class, and purpose-built or converted facilities are common.
USDA B&I. Available for for-profit operators in communities of 50,000 or fewer, up to $25 million, with fiscal 2026 guarantees at 85% below $5 million and 80% at or above.
USDA Community Facilities. For public bodies, non-profits and tribes. Available as a guaranteed loan, a direct loan, and as grants — three different regulations with different processes. CF underwrites on essential service need alongside repayment capacity, and for child care in a documented desert the essentiality argument is close to self-proving.
The sponsor-structure question is worth raising early. A project that struggles as a for-profit credit may be viable as a non-profit able to access CF grant funding and philanthropic support alongside the loan. That decision is far cheaper to make before the entity is formed.
Across all programmes, the study must be prepared by an independent third party with no financial interest in the transaction.
What the lender is reading for
Can this centre be staffed? Not whether demand exists, but whether the teachers can be hired at wages the revenue supports.
What does it actually collect, and from whom? Enrollment mix by payer, at documented local rates, net of attrition.
How long until it covers, and what funds the gap? The ramp is long, break-even is high, and payroll starts before revenue does.
Does it survive the summer? A 23% seasonal decline against a fixed payroll is a real cash event.
A study that answers those four with local evidence will clear review. One that establishes the trade area is a child care desert and stops there has documented the opportunity without demonstrating the business — and in a sector where 59% of operators are turning families away for want of staff, the distinction is the whole analysis.
Prepared by feasibility-study-consultant.com. Industry data reflects published sources at the date of writing. Licensing requirements, ratios and subsidy rates vary by state and change; verify against current state and programme guidance. Last updated: July 30, 2026.