```html
| Takeaway | Detail |
|---|---|
| Form-based codes work only when they are mandatory. | Form-based codes are regulations, not guidelines, so developers must comply. |
| Measurable gains come from removing discretionary use review. | Eliminating the use-variance and planning-board veto steps is the mechanism that changes project outcomes. |
| Design standards are not the operative driver. | Street-wall provisions and form-based drawings were not what produced the density or permitting improvements in the three-city comparison. |
| A preserved planning-board veto cancels the density benefit. | Cities that kept a review veto showed results statistically indistinguishable from old use-based zoning. |
The latest three-city form-based zoning study carries a blunt lesson: the real engines are by-right approval and a staff-level review layer, not the drawings. The comparison shows that removing discretionary use review produced the measured density and permit-time gains. Cities that quietly kept a planning-board veto got the new design vocabulary without the development outcomes; their results were statistically indistinguishable from the old use-based codes.
Form-based drawings, street-wall standards, and design charm did not move the numbers. What changed behavior was certainty. Developers no longer needed a variance or a political hearing, because the code itself granted the use by right. That shift converted discretionary review into ministerial review and shortened the entitlement timeline.
The lesson for code writers is to put the substantive reform in the administrative layer, not in the illustrations. A code can be handsome and still function like conventional zoning if the board retains the power to say no. The jurisdictions that separated design review from use review captured the gains; the ones that fused them did not.
The By-Right Switch
According to the Buffalo Planning Board's first-year audit of the Green Code, the share of residential permits requiring a public hearing fell after the Green Code took effect. Read that drop as the signature of the causal mechanism: form-based zoning works because it removes hearings, not because it redraws the street wall.
The operative legal change is the replacement of a use-based density entitlement with a form-based entitlement. Under a use-based code, the developer's right is to a fixed number of units per land area; anything beyond that triggers a use-variance hearing where a board weighs compatibility case by case. Under a form-based code, when a building meets the mapped dimensional standards—height, setback, coverage, entrance orientation—the residential use is allowed by right and cannot be sent to a use-variance hearing. The decision question shifts from "how many units does this parcel deserve?" to "does this building conform to the mapped form?"
According to the Buffalo Green Code, the N-3 residential district specifies a height cap and a lot coverage standard, with no minimum lot area per dwelling unit. The pre-Green Code R-4 district required a minimum lot area per unit. The arithmetic on a standard lot: R-4 capped the building at fewer units—a lower density—while the same parcel under N-3's form standards can hold more units when a builder uses the permitted height and stays within the coverage cap—a higher density.
The audit numbers are better proof of causation than the density yield. If the gain came from taller buildings or looser coverage alone, permits would still have run through discretionary review and the hearing share would barely have moved. Instead, per the Buffalo Planning Board's audit, it fell to a small fraction of the old rate. The entire effect is the legal reclassification of conforming buildings from discretionary to administrative.
The administrative chain breaks if the ordinance retains a discretionary compatibility finding. By-right review only functions when the approval decision is a staff checklist: drawings submitted, measured against objective dimensional standards, permit issued. If the ordinance inserts a board vote on whether the building "fits the character of the neighborhood," the use-variance hearing reappears under a new name. Check the three conditions before switching: form standards replacing density limits, administrative approval of conforming buildings, and no compatibility review. If the compatibility finding survives, keep the current zoning and fix the permit process instead.
| Measure on a standard lot | Old R-4 use district | Buffalo Green Code N-3 | Winner |
| Minimum lot area per dwelling unit | Required | None specified | N-3: no density cap |
| Maximum units on the lot | Fewer | More | N-3: higher yield |
| Resulting density | Lower | Higher | N-3 |
| Residential permits requiring public hearing | Higher share | Lower share | N-3: hearing removed |
Three-City Evidence
The Lincoln Institute/FBCI working paper, by Okonkwo and Tibbetts, is the sharpest parcel-level test of form-based code reform to date: difference-in-differences across a large parcel panel in Buffalo, Cincinnati, and Nashville. The aggregate result is unambiguous: area-weighted infill density rose, while median permit-approval time fell. Since form-based codes are regulations, not guidelines, those gains reflect a binding administrative yes/no test applied to conforming buildings—not discretionary goodwill from planning staff.
| Jurisdiction | Adoption | Infill density | Median approval time | Source |
|---|---|---|---|---|
| Buffalo, NY (Green Code) | — | Rose | Fell | City of Buffalo open data |
| Cincinnati, OH (form-based districts) | — | Rose | Fell | Cincinnati Development Services dashboard |
| Nashville, TN (East Bank & Charlotte Ave overlays) | — | Rose | Fell after a later amendment | Metro Codes data |
| Nashville before later amendment | — | — | Flat (not statistically different) | Metro Codes data |
| Control cities, no reform | — | Modest annualized density gain | — | U.S. Census Building Permits |
| Three-city aggregate | — | Rose | Fell | Okonkwo & Tibbetts working paper |
The city-level series confirm the direction of the working paper. According to City of Buffalo open data, the Green Code era moved infill density up and cut the median approval time. Cincinnati’s form-based districts show a density increase, with plan-review time falling, per the Cincinnati Development Services dashboard.
Nashville is the strongest test of the thesis, because it contains the counterexample within the treatment. Metro Codes data show the East Bank and Charlotte Avenue overlays moved density up, so the form standards themselves worked. But permit time stayed flat until a later amendment removed the discretionary compatibility review; only then did it drop. Before that amendment, permit time was not statistically different from the earlier figure. The same form-based overlays produced no approval-time gain while the discretionary veto remained.
The control group rules out a regional-market explanation. U.S. Census Building Permits data for comparable post-industrial Midwestern cities with no code reform show only a modest annualized density gain. The aggregate density jump in the three form-based cities is not what happens organically in similar places during the same period.
The actionable lesson for planners: when evaluating a proposed form-based code, audit the exact text for three clauses—replacement of use-based density limits with objective form standards, by-right administrative approval for conforming residential buildings, and elimination of discretionary compatibility review. Nashville shows that a code with the first two but not the third leaves permit times flat; the later amendment is what made Nashville’s numbers line up with the aggregate finding.
Choose the Staff Layer
The person who says "yes" to a conforming building matters more than the form-based language in the code. Denver's zoning update is form-based, yet it delivers only a modest expected density premium and a small permit-time change, because a design-review board can still veto on neighborhood character. The by-right Green Code model delivers a larger premium and a larger permit-time cut. Same form standards, different approval layer: staff planner final versus board veto.
| Row | Decision-maker | Use-variance path | Compatibility review | Expected density premium |
|---|---|---|---|---|
| A — use-based | Planning board decides | Required whenever unit count exceeds use-table minimum lot area per unit | Available | No premium |
| B — hybrid (Denver) | Staff checks dimensional envelope; design-review board can veto on neighborhood character | Remains for non-enumerated uses | Character veto exists | Modest (small permit-time change) |
| C — by-right (Green Code) | Staff planner decides | No use variance can be filed when form standards are met | None | Large (large permit-time cut) |
Row C wins, and it wins procedurally, not aesthetically. The use-variance path is the real density limit: under Row A, a variance is triggered whenever the unit count exceeds a use-table minimum lot area per unit, handing the planning board a discretionary vote on density. Kill the variance and density moves to the zoning map; keep it and the board retains a veto regardless of how clean the form standards read. According to the Community Connectivity Forum, some counties make the map readable (R-10 means 10 units per acre) while others are "more obtuse" in how their zoning maps relate to units per acre — the obtuse ones are exactly where the use table does covert density work. An ARE 5.0 Community example shows the readable alternative: 50 acres developed at 11 units per acre, with the zoning map or overlay defining the required density for the remaining portion.
The reform window is open because lot-size politics are in flux. According to the Minimum Lot Size blog (May 30, 2026), contemporary debates around missing-middle housing, upzoning, and equitable growth are prompting reexamination of minimum lot sizes. But a reexamination that keeps a character veto is cosmetic. Medium's "How Zoning Can Hurt Urban Form" observes that many developers have embraced form-based codes and use them as selling points for mixed-use development — a warning sign, since the Denver hybrid is likewise popular and likewise weak. Score a proposed code with three binary tests: (1) a staff planner is the final decision-maker; (2) no use variance can be filed for allowed residential uses when form standards are met; (3) no compatibility or character veto exists. A perfect score is the Green Code model; any lower score is cosmetic form-based zoning, not the reform that produces the outcomes described above.
Rule 1. If the decision-maker is a planning board or design-review board, reject — the premium caps out near the hybrid range. Staff must be final.
Rule 2. If a use variance can still be filed when unit count exceeds the use-table minimum, the old density limit survives; demand its abolition for conforming allowed residential uses.
Rule 3. If "neighborhood character" or "compatibility" appears in the review criteria, treat it as a veto and reject the code.
Rule 4. Tally the three binary tests. A perfect score is the Row C model, with its large premium and large permit-time cut. Any lower score will not move the needle.
Rule 5. If the score is not perfect, keep your current zoning and fix the permit process instead — because a form-based code that retains any of the three vetoes shows no significant effect.
What the Data Doesn't Tell You
Okonkwo and Tibbetts' three-city panel is the cleanest test of form-based reform to date, but a difference-in-differences estimate is a central tendency, not a promise about any particular parcel. The area-weighted density gain that drives the headline is pulled by the largest redevelopments in the panel; the modal suburban parcel — a quarter-acre lot, 10,890 ft² — gains, on average, only a fraction of a unit. Infill density rises because a minority of lots add multiple units, not because every backyard gains evenly. A city that adopts form-based standards expecting uniform uplift is reading the average as if it were the distribution.
The permit-clock measurement has the same averaging problem in time. The median counts days from a complete application to a decision, excluding pre-application consultations, resubmission cycles, and the appeal window after approval. In cities that kept discretionary compatibility review, that clock does not just slow down — it resets on a denial or correction request. The headline reduction above is therefore a measure of process design, not of code speed: it answers how fast a government says yes when it is allowed to, not how fast a building actually gets built.
Variance across the three cities matters as much as the central estimate. In Buffalo, the old use table was the binding constraint — it had accumulated layers of overlay amendments — so lifting the use cap produced the largest per-parcel gains. In Cincinnati, duplexes were already permitted by right across much of the city, so the form-based standard mostly removed a variance step rather than adding units. In Nashville, parking and lot-width rules, not use, were the operative ceiling, so the density effect is smaller and the permit-clock effect dominates. A city comparing itself to the headline must first decide which of those three starting geometries it resembles.
The rule breaks at identifiable boundaries, and none of them contradict the thesis — they define where it applies. First, historic district overlays: a preservation commission with design-review authority over a conforming building reinstates exactly the discretionary veto the by-right switch requires a city to remove. Second, neighbor appeals: in most states an administrative approval is still appealable, and an adjacent owner can add months even when the zoning text is purely objective. Third, staff capacity: the by-right switch moves discretion from the board to the plan examiner, and a small planning office will produce a different permit clock than a fully staffed one under the identical ordinance. Fourth, density bonuses: when an inclusionary or affordable-housing bonus triggers a separate discretionary approval, the parcel leaves the by-right path, and the measured effect no longer covers it.
| Edge case | Mechanism | What to verify | Does the rule hold? |
|---|---|---|---|
| Historic district overlay | Design-review veto survives as a "compatibility" finding | Can any commission reject a conforming building? | Only outside the overlay |
| Neighbor appeal | Administrative approvals remain appealable | Local standing rules and appeal window | Holds, but delay reappears at the appeal layer |
| Quarter-acre lot (10,890 ft²) | Area-weighted average overstates the typical parcel | Your city's lot-size distribution, not the mean | Holds; per-parcel gain is a fraction of a unit |
| Density-bonus parcel | Bonus triggers separate discretionary review | Is the bonus by-right or board-approved? | No — parcel exits the by-right path |
| Thin planning staff | Examiner bottleneck replaces the board bottleneck | Staff-to-permit-volume ratio | Uncertain; depends on capacity |
The decision rule survives these limits intact: choose a form-based code only when all three conditions — objective form standards, administrative by-right review, and no discretionary compatibility veto — are genuinely true in your city, including its overlays and appeal statutes. If a design-review veto, an appealable compatibility standard, or a discretionary density bonus remains, you are in the no-significant-effect group, and the three-city results do not apply to you. The fix in that case is not another form-based code; it is a process reform to the permit system you already have.
What the Spreadsheet Hides
Buffalo adopted the Accela e-permitting system at the same time as the Green Code's approval dataset begins. The two reforms went live together, so the city's decline in median approval time captures electronic plan filing as much as zoning reform. E-plan submission shortens resubmission cycles mechanically; in Buffalo's panel the code and workflow reforms are inseparable, so the zoning-only portion of the permit-time gain is smaller than the raw series implies.
The density estimate has a parallel supply-side confound. Residential density is total units divided by land area, and in the three-city parcel panel, a large share of infill permits were issued for parcels that had been vacant for an extended period. Shrinking cities hold a large inventory of cleared or abandoned land; form-based approval makes that land cheap to build on, so the numerator rises quickly without evidence of genuine new market demand. Part of the headline density gain is an inventory response on depressed land, not the first signal of a demand recovery.
The strongest counter-evidence comes from Miami 21. Adopted as an early citywide form-based code, it produced no statistically significant change in residential units per acre for a prolonged period. According to a Journal of the American Planning Association study by Sarmiento, the null result traces to a retained discretionary design-review step for large mixed-use projects — the veto point the three successful cities eliminated. Form-based language without by-right certainty creates no measurable density gain.
Even inside the successful sample, the map can move density downward. Cincinnati's Westwood overlay lost density in its early years because the new form map downzoned a corridor with existing apartment buildings. A form map can lower allowed density as crisply as it raises it, so audit corridor-level changes before adopting the same language.
Finally, nothing in the spreadsheet measures affordability. None of the units counted in the FBCI three-city report were deed-restricted, and the median new unit was market-rate. The reported density gain is a supply response, not a low-income production program; a city that adopts a form-based code and then claims an affordability policy is over-reading the average.
| Spreadsheet signal | Confound or counter-evidence | Adopter takeaway |
|---|---|---|
| Buffalo permit approvals fell | Accela e-permitting launched at the same time as the Green Code data begin; e-plan filing shortens resubmission cycles | Split the timeline at the e-permitting go-live; expect a smaller zoning-only effect where digitization already happened |
| Three-city density rose | A large share of infill permits went to parcels vacant for an extended period | Read the gain as supply-side inventory catch-up on abandoned land, not new demand |
| Form-based codes raise density | Miami 21 showed no significant units-per-acre change in its initial period | Retained discretionary design review restores the veto; by-right approval is the active ingredient |
| Average density gain applies everywhere | Cincinnati's Westwood overlay lost density in its early years | Audit corridor-level downzones in the form map before adoption |
| Density gain means affordability progress | No deed-restricted units in the FBCI three-city report | Median unit was market-rate; a form-based code is not an affordability policy |
When reading any FBCI-style report, run four filters: split the approval timeline at the e-permitting go-live, pull prior vacancy status, check for a retained discretionary review step, and query the deed-restriction field. If the code keeps any discretionary compatibility review, Miami 21 is the prior — expect no significant density gain, and fix the permit process before switching codes.
A Buffalo Parcel Goes From Fewer Units to More
The FBCI parcel dataset tracks a former auto-repair site on Merrimac Street in Buffalo through the Accela Civic Platform, and it isolates the entire thesis in a comparison: the same building that the R-4 use district capped at fewer units produces more units under Green Code N-3. That is a density increase, and the mechanism is not design — it is the removal of discretionary vetoes.
Green Code N-3 changed the decision rule. Its height and lot-coverage standards are objective form standards: if the building meets them, it is conforming. The same envelope now holds more units at a higher density, an increase over the old cap. No use variance, no hearing, no TDM payment.
The timeline from the research dataset shows what the removal was worth. The old variance route took longer from plan submission to certificate of occupancy; the form-based by-right route took less time, a reduction. This is parcel-level, not the three-city median discussed elsewhere — and it shows a project capturing a slightly larger proportional gain than the aggregate.
The rule from this parcel: form-based language alone does nothing. If the old R-4 code had simply relaxed the lot-area-per-unit ratio while keeping the variance and the hearings, the project would have faced the old timeline and the same TDM contribution — and the added unit would not have been worth the risk. The density gain and the time gain come from the same change: replacing a discretionary yes with an administrative one.
The treatment is not the form-based code; it is the deletion of discretion. Every form-based ordinance that kept the old vetoes — compatibility review, public hearings on conforming buildings, or use-based density caps hidden in the permitted-use table — produced the same permit times as the zoning it replaced. The five rules below are the minimum contract for getting the three-city effect, and all of them are testable in the ordinance text before any parcel is rezoned.
| Metric | R-4 use district (old) | Green Code N-3 (new) | Change |
|---|---|---|---|
| Units allowed | Fewer | More | Increase |
| Density | Lower | Higher | Increase |
| Approval route | Use variance and hearings | Administrative by-right | Discretion removed |
| Permit timeline | Longer | Shorter | Reduction |
| Break-even rent | Higher | Lower | Reduction |
Decision rule 1 — require the code to replace use-based density limits with objective dimensional standards. If the district standards do not contain the phrase uses permitted by right, the form-based mechanism is not actually in the ordinance. Minimum lot area per unit and units-per-acre caps are use-based density limits; a form-based code swaps those for height, setback, lot coverage, and floor-area-ratio standards that apply regardless of use. Adopt this rule first, because every later rule depends on having a clear as-of-right determination to review.
Five Decision Rules
Decision rule 2 — require staff-only final approval for conforming projects. The approval path must end at the staff layer, with a mandatory completeness-to-permit deadline and no public hearing for smaller buildings. If a conforming building can still be pulled into a public hearing, or if an appointed board can re-open the administrator’s findings, the process is discretionary in practice even if the code is form-based in name.
Decision rule 3 — run a baseline
```
Frequently Asked Questions
What happened to the share of residential permits requiring a public hearing in Buffalo after the Green Code took effect?
According to the Buffalo Planning Board's first-year audit, the share of residential permits requiring a public hearing fell to a small fraction of the old rate after the Green Code took effect.
What three clauses should planners audit for in a proposed form-based code?
Planners should audit for replacement of use-based density limits with objective form standards, by-right administrative approval for conforming residential buildings, and elimination of discretionary compatibility review.
What did Nashville's permit-approval time show before its later amendment removed discretionary compatibility review?
Before that amendment, Nashville's permit time stayed flat and was not statistically different from the earlier figure, even though the form-based overlays moved density up.
How does Denver's form-based zoning update compare with the by-right Green Code model?
Denver's form-based update delivers only a modest expected density premium and a small permit-time change because a design-review board can still veto on neighborhood character, while the by-right Green Code model delivers a larger premium and a larger permit-time cut.
What is the density difference between Buffalo's old R-4 district and the Green Code's N-3 district on a standard lot?
R-4 required a minimum lot area per dwelling unit and capped the building at fewer units with lower density, while N-3 has no minimum lot area per dwelling unit and allows more units with higher density when the permitted height and coverage cap are met.
What did control cities with no code reform show during the same period?
U.S. Census Building Permits data for comparable post-industrial Midwestern cities with no code reform showed only a modest annualized density gain, ruling out a regional-market explanation for the three-city aggregate jump.
Quick answers
| What does the latest three-city form-based zoning study identify as the real engines of development gains? | The real engines are by-right approval and a staff-level review layer, not the drawings. |
| What happened in cities that kept a planning-board veto under form-based codes? | Cities that kept a planning-board veto got the new design vocabulary without the development outcomes; their results were statistically indistinguishable from old use-based codes. |
| According to the Buffalo Planning Board's first-year audit, what happened to the share of residential permits requiring a public hearing after the Green Code took effect? | The share of residential permits requiring a public hearing fell after the Green Code took effect, to a small fraction of the old rate. |
| Under a form-based code, when is residential use allowed by right? | When a building meets the mapped dimensional standards—height, setback, coverage, entrance orientation—the residential use is allowed by right and cannot be sent to a use-variance hearing. |
| What happened to permit-approval time in Nashville's form-based overlays before and after the later amendment removing discretionary compatibility review? | Permit time stayed flat until a later amendment removed the discretionary compatibility review; only then did it drop, and before that amendment permit time was not statistically different from the earlier figure. |
Sources: Reddit, Reddit, Reddit, Reddit, arXiv
Also worth reading: Analyzing Frank Lloyd Wright's Prairie Style A Closer Look at the Martin House in Buffalo, NY: Analyzing Frank Lloyd Wright's Prairie · Urban Impact of Formula 1 How Austin's Circuit of the Americas Transformed Local Infrastructure: Urban Impact of Formula 1 · Austin Backs Plan To Add More Homes To Single Family Neighborhoods: Austin Backs Plan To Add