The Pro Forma Is Where Multifamily Development Actually Happens

One thing Kelly Moncrief made clear early: the four stages of a development project — feasibility, pre-development, construction, stabilization — don't carry equal weight.

They look symmetrical on a planning diagram.
But feasibility is where you still have all your choices.

By the time construction starts, the land is closed, the financing is committed, and the contractor is on contract. You can still have bad luck — supply chain disruptions, weather delays, a subcontractor going under — but the project's fundamental economics were decided earlier, in the phase nobody can see from the street.

Here's what's puzzling: every developer knows feasibility is the critical phase.
It's in every textbook, every panel discussion, every war story about deals that went wrong.
And yet the same inputs keep getting modeled loosely — impact fees estimated from a prior project, absorption taken from the optimistic end of the market study, a developer's cap with no real buffer.
Why does that keep happening?

The feasibility phase isn't just the beginning of a project. It's where the project earns the right to move forward — and the pro forma is the instrument of that test. Most developers who blow up on a project didn't lose it in construction. They lost it in the assumptions made in feasibility that couldn't be backed out of later.

After working through this material, three questions now anchor how I evaluate any feasibility analysis: Where did the impact fee number come from — a site-specific lookup, or an estimate from a prior project?
What does the conservative end of absorption imply in monthly construction loan interest?
And what's the current market cap rate for this submarket — and how much spread does this model actually carry above it?
Those three questions separate a real feasibility from a feasibility exercise.

So in this post I want to work through what actually happens in feasibility: how a site gets evaluated, how the financial model gets assembled, and why certain inputs — density calculations, impact fees, absorption timelines — end up mattering far more than most people outside development expect.

The four stages are a sequence, but not an equal one

Multifamily development moves through four stages: feasibility, pre-development, construction, and stabilization.

Feasibility is when you've identified a promising area, maybe have a site under contract, and are doing initial underwriting — testing whether the numbers can possibly work before committing to the site.

Pre-development starts after you've decided to proceed. The site is under control, design begins, consultants get hired, entitlements get processed if needed, and financing starts to come together.

Construction is construction.

Stabilization is after the building is complete and you're leasing it up to the occupancy threshold — typically 70 to 90 percent — required to refinance from a construction loan into permanent debt.

Here's why feasibility is disproportionately important: it's the only stage where you can walk away before spending real money. Feasibility costs — a market study, an architect's concept plan, some early environmental work — are modest compared to what pre-development and construction require. Once you're past feasibility and into design and permitting, the path forward has inertia. Once you've closed on the land and drawn the construction loan, you're committed.

The decision gate at the end of feasibility is the most consequential decision a developer makes. Everything downstream is an execution of that decision.

Site selection is the most irreversible early decision

Before any financial model gets built, the developer has to evaluate whether the site itself makes sense.

The core principle: site selection is the most critical influencer of a project's success.

For market-rate projects, the question is whether the location supports rents that work — and whether those rents will grow over time. Compound annual rent growth matters because it means revenue rises faster than expenses, which improves the long-term economics of holding the asset.

For affordable projects, location is evaluated on a completely different axis: can the site score enough points in subsidy program applications to win the funding it needs? Proximity to transit, grocery stores, clinics, and high-performing schools determines whether a subsidized project gets financed — not market rents. I'll get into affordable housing separately in the next post; for now, the point is that these two product types look for different things in a site.

Once location passes the initial screen, the developer has to answer: how many units can actually be built here?

Zoning governs this through two common measures.

FAR — floor area ratio — is a multiplier on lot size. A 20,000 square foot site with an FAR of 5 yields 100,000 buildable square feet.

DU — dwelling units — is a per-area cap on unit count. At one unit per 1,000 square feet, that same site has a base density of 20 units.

Neither number is final, because California layers a density bonus program on top of base zoning. Include affordable units in a market-rate project and you can add market-rate units beyond the base density. And in Transit Priority Areas, newer programs can push density dramatically higher — enough to change whether a project is worth pursuing at all.

The early yield calculation is also why developers hire an architect for a feasibility study before committing.
The architect doesn't just sketch a building.
They produce a concept plan that tests whether the zoning yield is real — accounting for egress requirements, setbacks, parking access, and the actual shape of the site.
The unit count that fits "on paper" often shrinks once someone starts drawing the building.

Getting that concept plan done in feasibility costs relatively little.
Discovering the yield problem two years in does not.

The pro forma converts assumptions into a decision

Once the site passes initial screening and the unit count looks realistic, the conceptual underwriting begins.

The pro forma is sometimes called the financial model, the analysis, the underwriting — terminology varies, but the structure is the same: inputs go in, returns come out, and the question is whether those returns justify the risk.

Market-rate pro forma inputs divide into revenues and costs.

On the revenue side, the key input is rent — specifically, market rents for that site, supported by a market study and rent comparables from nearby properties. This is not a number you can estimate loosely. If comparable apartments nearby are leasing at $2,800 per month for a one-bedroom, that's a data point. If comparable apartments are sitting 10% vacant at $3,100, that's a different data point. The rent assumption in the pro forma has to be justifiable to an outside reviewer — a lender, an equity partner, an appraiser — not just to the developer who wants the deal to work.

On the cost side, the standard line items are land (plus carry costs for the years between acquisition and construction), soft costs (consultants, entitlements, legal), hard costs (construction), and financing costs (construction loan interest).

Two of those cost inputs turn out to be far more variable than they look: impact fees, and the construction loan interest driven by how long it takes to lease up the building.

Development impact fees are the variable most people underestimate

When you build housing, you're adding households that will drive on local roads, send children to local schools, use local parks, and draw on local water and sewer infrastructure. Most jurisdictions charge development impact fees to help fund those services.

In San Diego, those fees vary dramatically by location. In neighborhoods where the existing park and service infrastructure can absorb new demand, fees are lower. In areas where the master plan identifies gaps — places where the nearest park doesn't count toward satisfying requirements — fees are much higher.

The range in San Diego runs from roughly $10,000 to over $60,000 per unit just for parks, fire service, and libraries.
That's before water and sewer capacity fees, regional transportation fees, and school fees.

At 90 units, a $50,000-per-unit swing in impact fees is $4.5 million.
On a $35 million total project, that's 13% of cost.
If the pro forma assumed the lower number and the actual fees are the higher number, the developer's cap rate drops, the return metric fails, and the project may not pencil.

The per-unit fee structure also creates some odd outcomes. A studio apartment and a four-bedroom apartment are charged the same impact fee, even though a four-bedroom can realistically house six or more people versus one or two in a studio. The city has been considering a per-bedroom model that would better align fees with actual infrastructure demand. Until that changes, developers building larger family-sized units are paying higher fees on a per-person basis than developers building studio-heavy projects.

The more relevant point for feasibility: impact fees have to be researched for the specific site, not estimated from a prior project in a different neighborhood. They're too variable and too significant to guess.

What makes them particularly dangerous is the combination: maximum variance, minimum benchmarking.
Construction cost per square foot is published in a dozen indices.
Impact fees by neighborhood aren't — only a site-specific lookup answers it.
And that lookup is exactly what gets skipped when a developer is in early exploration, still deciding whether to pursue the site.
The fee that kills the deal later was the assumption nobody checked early.

A few things I'm taking away

The early-feasibility phase of multifamily development is where the most consequential decisions happen — not because the most money is spent there, but because the most reversibility exists there.

• Feasibility is the only stage where you can walk away cheaply — which makes it the most important stage to get right.
• Site selection is the first and most permanent decision, because location determines rents, and rents anchor everything downstream.
• Density calculations translate Zoning into unit count, but the architect's feasibility study is what tests whether the paper yield and the real yield actually match. The unit count that fits "on paper" often shrinks once someone starts drawing the building.
• Pro forma rent assumptions need to be supported by market study data and comparables, not optimism or the number you need the project to work at.
• Development impact fees vary dramatically by location and can swing total cost per unit by $40,000 or more. The Terner Center's California-cities study has fees ranging from $21,000 per single-family unit in Sacramento to over $157,000 per unit in Fremont — that scale of variation is why impact fees require site-specific research, not estimates from a prior project.
• Impact fees have maximum variance combined with minimum benchmarking — construction cost has indices, impact fees have site-specific lookups. The fee that kills the deal later is the assumption nobody checked early.
• California's density bonus law (Government Code § 65915) and Transit Priority Area programs can change yield enough to make a project worth pursuing that wouldn't pencil under base zoning alone — but those bonuses are conditional on affordable allocations or specific siting, so they don't apply to every site.

What this gets you ready for

The early half of feasibility — site, density, fees — fixes what you can build and roughly what it costs. But it doesn't tell you whether the resulting deal should be built. That depends on two later inputs: how fast the building leases up after construction (absorption), and whether the developer's cap rate beats the market by enough to justify the construction risk. Those are covered in the companion post on absorption and developer's cap rate.

The stakes go beyond any single developer's return. Most housing undersupply isn't a developer motivation problem — there are developers willing to build. It's a math problem. Impact fee structures, absorption risk, and rising construction costs have together made feasibility harder to clear than it was a decade ago. More housing gets built when the math gets easier, not when developers try harder.

The pro forma isn't just a private financial document. It's a signal about where housing policy is enabling production — and where it's blocking it without the zoning map ever changing.

This post is the first of two on multifamily feasibility in my ongoing series — Real Estate Development. The companion post covers absorption pace and developer's cap rate, the two later-stage inputs that decide whether a project actually pencils. Earlier posts cover the development process and why entitlements decide what's possible.

Sources

• Kelly Moncrief, Create Development — primary source for the four development stages, site selection criteria, density calculations (FAR, DU, density bonus), and impact fee mechanics.
• Terner Center for Housing Innovation, UC Berkeley — It All Adds Up: The Cost of Housing Development Fees in Seven California Cities — used for the impact-fee variation data ($21K Sacramento → $157K+ Fremont per single-family unit). https://ternercenter.berkeley.edu/wp-content/uploads/pdfs/Development_Fees_Report_Final_2.pdf
• Terner Center — Residential Impact Fees in California (overview) — companion overview of how California impact fees compare to the national baseline. https://ternercenter.berkeley.edu/blog/residential-impact-fees/
• California Government Code § 65915 — Density Bonus Law — official statute for the density bonus program referenced in the post. https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?sectionNum=65915&lawCode=GOV
• Mike E. Miles, Laurence M. Netherton, Adrienne Schmitz — Real Estate Development: Principles and Process (5th ed., ULI, 2015) — the standard textbook on the development process; recommended deep reference for the four stages and site selection. https://www.amazon.com/Real-Estate-Development-Principles-Process/dp/0874203430