Vol. 011: The Say-Do Gap

Current Conditions

Every municipality has a comprehensive plan. Most of them say the right things about housing, density, mixed-use development, and growth management. The question that matters for anyone putting capital at risk is whether the municipality does what it says.

We analyzed 16 municipalities in a major metro area in central North Carolina: 20 comprehensive plans, roughly 20,000 zoning permissions, and 195,000 political meeting records. The goal was to measure the gap between three things: what a municipality says it supports in its community plan, what it allows in its zoning code, and what it actually does at the hearing.

The results suggest that the standard methods for evaluating development climate (reading the comprehensive plan, checking the zoning code, asking a local broker) miss the variable that matters most. The behavioral record.

Signals Observed

Signal 1: Process burden does not predict approval outcomes.

Nine of 16 municipalities in the dataset show what we’re calling a “Practice Over Policy” pattern, they approve development despite restrictive codes. One municipality scored lowest in the dataset on by-right permitting (7.4 out of 100), meaning virtually nothing is permitted without a discretionary review. Its approval rate: 90.1%. Another municipality requires conditional approval for nearly every development type. Its denial rate: 0.8%.

The zoning code says no. The board says yes. If you’re evaluating these markets based on what the code allows by right, you’re reading the wrong document.

Signal 2: Fee variance within the same metro is staggering, and it’s not correlated with approval difficulty.

We ran a standardized 100-unit townhome project through the fee schedules of all 16 municipalities. The results:

• Lowest total fee burden: approximately $1,400
• Highest total fee burden: approximately $320,000
• That’s a 22x spread for functionally similar entitlement processes within the same metro area

Rezoning fees alone ranged from $250 in a rural county to over $5,100 in the urban core. Transportation impact fees ranged from $0 to $3,200 per unit. Two sites 20 miles apart can carry a $100,000+ fee differential on the same product type.

The assumption that low fees signal an easy approval is common and wrong. One of the lowest-fee municipalities in the dataset carries a 9.7% denial rate, the highest in the region. Another municipality with moderate fees has the highest approval rate at 93.8%.

Signal 3: The municipality most often cited as developer-friendly carries the highest delay rate in the dataset.

One large municipality in the metro, widely regarded as a good market for development, shows a 27% overall delay rate. Its conditional use permit delay rate hits 57%. Mixed-use applications face a 39% delay rate despite explicit community plan support for mixed-use development.

The approval eventually comes. But the time cost erases the fee advantage. A project delayed 6 months in a low-fee market costs more than a project approved on schedule in a high-fee market, once carry costs, engineering extensions, and resubmittal fees are factored in.

Signal 4: Only 2 of 16 municipalities are consistent across all three dimensions.

Two municipalities in the dataset say they support development, allow it in their code, and approve it at the hearing, consistently. They are the only markets where the community plan, the zoning code, and the behavioral record tell the same story. In the other 14, at least one dimension diverges from the others.

This is not a criticism. It’s a measurement. Most municipalities have gaps between policy and practice, and those gaps are where deals are mispriced.

Forecast

Developers selecting markets based on published zoning codes or reputation are working with incomplete data. Two municipalities with identical comprehensive plan language can produce 18-month outcome differences on the same product type. The actual approval climate is only visible in the behavioral record (votes, delays, conditions, denials) and that record is public.

The “Practice Over Policy” pattern is likely stable in the near term. Municipalities that have been approving development through conditional processes have institutional muscle memory around that approach. The risk is not that they’ll suddenly start denying. It’s that the time cost of the conditional process is invisible in the proforma until month 8.

Watch Items

• Fee schedule updates. Three municipalities in the dataset are under active fee revision. Impact fee increases of 40-90% are on adopted escalation schedules in several jurisdictions. The fee landscape is not static.
• Zoning code rewrites. Two municipalities are in active UDO rewrite processes. Transition periods between old and new codes create temporary windows where applicants may have flexibility to file under either framework, a timing advantage with a short shelf life.
• Election cycles. The behavioral record reflects the current board. Board composition changes, even a single seat, can shift the gap between policy and practice in either direction.

Confidence Level

High. N=195,000+ political meeting records across 16 jurisdictions. The patterns are structural, not anecdotal. Fee data is sourced from published municipal fee schedules. Approval and delay rates are calculated from classified political meeting transcripts with action-type confidence scoring.

The word “comprehensive” appears in 20 different plan documents across 16 municipalities in this dataset. It does not appear in the behavioral record. The record has other words: approved, denied, continued, tabled, withdrawn. Those are the ones that affect your timeline.

Why This Matters

Site selection models that weight zoning code restrictiveness as a proxy for approval difficulty are measuring the wrong variable. The municipality with the most restrictive code in the dataset has a 90% approval rate. The municipality with the best reputation has a 57% CUP delay rate. The municipality with the lowest fees has a 9.7% denial rate.

The gap between what a municipality says, what it allows, and what it does is where proformas break. Measuring that gap requires reading the behavioral record, not the brochure. The data is public. The analysis is straightforward. The cost of not doing it is discovering in month 12 that your “friendly” market has a time problem nobody mentioned.