Should the financial model decide what a community project builds?

Seventy percent of young children in remote rural America live in a licensed child care desert — more than three children under six for every licensed slot. Nationally the figure is 46%, and across all rural areas it is nearly 66% (Center for American Progress, April 2026). Over roughly the same period, 197 rural hospitals have closed or converted away from inpatient care since 2005 (UNC Sheps Center).

Those two numbers describe the same thing from two directions: in a lot of rural communities, the buildings where daily life used to happen are gone, and nothing replaced them.

I am an architect and a development underwriter. I build the financial models that decide whether projects like these get financed. So I want to be precise about what I am arguing, because it cuts against my own instrument: the financial model should test the program. It should not choose it.

Key takeaways

• The most common failure in community development is not a project that doesn't pencil. It is a project that pencils and serves the wrong need.
• Underwriting is a filter, not a generator. It tells you whether a program is viable — it cannot tell you what a community actually needs.
• Programs that emerge from real community process are harder to finance and more durable once built, because the demand is real.
• At the West Marion Resilience Hub in North Carolina, the program came out of the community first, and the capital stack was built around it.

What happens when the model picks the program

I have watched this sequence more times than I can count.

An organization has a site and an aspiration. Someone — often well-meaning, often a consultant — runs early numbers. Certain uses model well: market-rate units, self-supporting commercial space, program with a clear revenue line. Other uses model badly. Childcare, which is chronically underpriced relative to its cost. Behavioral health, which depends on payor mix. Community gathering space, which generates almost nothing.

The uses that model badly are quietly dropped. Not by anyone's decision, exactly. They just don't survive the spreadsheet.

What's left is a building that works financially and serves a need the community didn't have. The organization then spends three years trying to raise philanthropic money for it, and cannot understand why funders aren't moved. The funders aren't moved because the project is no longer about anything.

The model did its job. The job was just the wrong one. A financial model answers whether a given program can be sustained. It has nothing to say about whether that program should exist. When we let it decide, we have outsourced a moral question to an arithmetic tool.

What the alternative actually looks like

At the West Marion Resilience Hub, in Marion, North Carolina, the building came second.

The site is the former Hudgins High School, later Mountain View School — the only high school for Black students in McDowell County under segregation. It closed in the mid-1960s when the county's schools integrated. It sat. In 2024 the property was donated to West Marion Inc., a community organization founded in 2016 that was the county's first Black-led nonprofit, and which has grown its operating budget from roughly $100,000 in 2017 to $1.5 million today under executive director Paula Swepson Avery. Several of its board members attended the school.

The program was set by what the community said it needed, and the list is not one a spreadsheet would have generated: a primary care and behavioral health clinic, licensed childcare, a commercial kitchen, coworking and small-business space, an event hall, and a gallery of local Black history that also serves as the building's lobby. The building is designed with the county's emergency services to shelter up to 100 people overnight, with its own solar array, generator, energy control center, and a 2,500-gallon potable water reserve.

Some of those uses carry themselves. Several do not, and never will. The childcare center will not be a profit center. The gallery generates no revenue at all.

My job was not to remove them. It was to build a capital structure in which they could exist — one where the uses that generate income carry the ones that don't, where grant and tax-credit capital covers what debt cannot, and where the whole thing can still make its payments in year ten. That is a harder problem than optimizing a rent roll. It is also the actual problem.

The largest single funder of the predevelopment work was the Kate B. Reynolds Charitable Trust, whose grant made it possible to do the technical work — the environmental assessments, the design, the underwriting — before anyone knew whether the project would be financeable. That is the least glamorous money in community development and the most decisive.

Why does the mission-first order make the financing easier, not harder?

This is the part that surprises people, and it is the strongest argument I can make to a board that thinks community process is a delay.

Grant and philanthropic capital is not underwritten like debt. A lender asks whether you can pay. A foundation asks whether the outcome is worth buying. Those are different questions and they reward different projects. A program that came out of genuine community process has something a financially-optimized program structurally cannot have: evidence that the demand is real, and a constituency that will show up for it.

The projects I have seen raise money fastest are the ones where the answer to "why this program?" is a person in the room rather than a market study. The market study matters — I produce them — but it corroborates the answer. It isn't the answer.

There is also a durability argument. A building programmed around what financed easily has no defenders when conditions change. A building programmed around what a community asked for has a constituency that fights for it. Over a thirty-year asset life, that is worth more than a favorable year-one pro forma.

What this requires of people like me

If the community sets the program, then the technical work has to get harder in order to make it possible. Three things follow.

Underwrite the whole building, not each use. Cross-subsidy is the mechanism. The coworking space and the event hall exist partly so that the childcare center can.

Stop treating the exit as the measure. On projects with no sale planned, a ten-year IRR that assumes a hypothetical disposition is a fiction. Stabilized cash-on-cash return — can this building pay for itself indefinitely — is the honest metric, and it is the one a community board should be shown.

Build the capital stack around the program rather than the reverse. New Markets Tax Credits, program-related investments, state appropriations, in-kind contributions from public agencies, CDFI debt. Each one has its own logic, and assembling them is genuinely difficult work. It is also the work. Telling a community that its clinic doesn't pencil is not analysis. It is a failure to do the harder version of my job.

The argument

I am not arguing that numbers don't matter. I run them for a living, and I have told clients to stop projects that could not work. Enthusiasm is not a capital stack, and a project that cannot pay its operating costs will fail the community more completely than never building it.

What I am arguing is about order. Ask the community what it needs. Then find out whether it can be built and paid for. If it can't, say so plainly and go back — but go back to the community, not to the spreadsheet, because the spreadsheet will always hand you back the same answer, and it is always the safe one.

The former Hudgins High School has been empty for sixty years. A financial model, asked what to do with it, would have produced something. It would not have produced a building with a clinic, a daycare, a commercial kitchen, and a room full of the community's own history, designed to shelter a hundred people when the power goes out.

The community produced that. My job was to make the numbers work around it.

Sources

• Gibbs, Peeks, Lee, Pepin, Sadowski & Sojourner, America's Licensed Child Care Deserts, Center for American Progress (April 2026) — 46% of US children under 6 in licensed child care deserts; 70% in remote rural areas; 65.6% across all rural areas.
• UNC Cecil G. Sheps Center for Health Services Research, Rural Hospital Closures tracker — 197 rural hospital closures and conversions since January 2005 (accessed September 2026).
• McDowell News, "Reunion set for Hudgins, Mountain View schools" — Hudgins as the only high school in McDowell County for African American students.
• Appalachian Voices, "West Marion Resilience Hub" (July 2025) — West Marion Inc. as the county's first Black-led nonprofit; building donated in 2024; board members attended the school.