A Public Improvement District Will Not Fix Niwot’s Roads

A feasibility study of the two district mechanisms available, and why neither solves the problem. A townwide district is available in law and has twice failed in practice. A neighborhood district works, and fixes one subdivision while leaving the network. This paper asks whether a district would actually fix the roads, and finds that it would not.

What This Paper Is

Niwot has 31 miles of road. Twenty-three of them have no funded repair plan, and the average pavement condition index across those 23 miles is 44 — well below any standard a Colorado municipality holds itself to.

That is the problem. This paper is about one family of answers to it.

A public improvement district is a county-created taxing unit, authorized by Part 5 of Article 20 of Title 30 of the Colorado Revised Statutes. It is a real mechanism, and Boulder County has one in Niwot already, at Burgundy Park. It comes in two forms — one district covering the whole community, or one covering a single subdivision — and it is the thing people mean when they ask “why not just do a PID?”

That is a fair question and it deserves an arithmetic answer. This paper is that answer. It asks whether a district would work here: not whether it is a good idea, but whether it is available, and whether it would actually fix the roads.

It would not, in either form, and for two different reasons.

A townwide district is available in law and has twice failed in practice. There is no proposal today and nobody assembling one, and its economics depend on the single participant with both the strongest reason and the clearest route to leave.

A neighborhood district is demonstrably formable — two exist in the county — but it costs the typical household several times what the alternatives do, and it fixes one subdivision’s streets while leaving the roads between neighborhoods unfunded permanently. It is a real mechanism that does not add up to a road network.

Incorporation appears in this paper only as a price comparison, at the end. The question here is not which camp is right. It is whether the district mechanism does the job.

We have tried to write down the places where our own case is weaker than we would like it to be. There are several, and they are marked — including the two neighborhoods where the comparison comes out even, which are the reason this study exists at all.

1. What a District Is, and What the Statute Actually Permits

An improvement district under CRS 30-20-503(3) is

a taxing unit that may be created by any county in this state for the purpose of constructing, installing, acquiring, operating, or maintaining any public improvement or for the purpose of providing any service so long as the county that forms the district is authorized to perform such service or provide such improvement.

CRS 30-20-504(1) grants the same authority in the same words.

A district is not a capital-only body. This is worth stating plainly because we have said otherwise. Earlier versions of our own material claimed a district could not lawfully fund snow plowing. That is not supported. Operation, maintenance and service provision appear twice, in the definition and in the grant of authority, and the test is whether the county may perform the service — which, for plowing roads, it plainly may. The only carve-outs in 503(3) are solid-waste disposal facilities under 30-20-101(8)–(9) and the waste-collection services in 30-15-401(4) to (7.7). Neither reaches snow.

What is true, and is a different claim, is that no district proposed for Niwot has ever included snow. Boulder County’s improvement-district practice has been capital work — Burgundy Park was road reconstruction repaid on a county loan. A district’s levy buys what its petition asked for, and none on offer here have asked for plowing.

This paper is careful statutory reading, not legal advice. The statutory text it relies on is quoted from the official Colorado Revised Statutes 2024 printout published by the Office of Legislative Legal Services.

2. How a District Is Formed, and Why the Petition Is Not the Hard Part

The common assumption is that forming a district is difficult. The statute says otherwise, and anyone arguing against a district should know this before an opponent tells them.

The petition threshold is very low. CRS 30-20-505(1):

The petition shall be signed by not less than thirty percent or two hundred of the electors of the proposed district, whichever is less. After the filing of a petition, no signer shall be permitted to withdraw his or her name therefrom.

Niwot has about 3,598 registered voters across its 34 neighborhoods. Thirty percent would be 1,080. Two hundred is less. So 200 signatures — about 5.6% of electors — is the binding threshold, and a signer cannot take it back. Note also that “elector” under 30-20-503(1)(a) includes non-resident property owners, so the eligible pool is wider than the registered-voter roll.

The boundary requirement is minimal. CRS 30-20-505(2)(d) asks a petition for

A general description of the boundaries of the district or the territory to be included therein, with such certainty as to enable a property owner to determine whether or not his or her property is within the district.

That is the whole requirement. There is no parcel-level benefit test anywhere in Part 5.

The only benefit test is district-wide. Under 30-20-508(2) the county commissioners must dismiss a petition if the improvement “will not confer a general benefit on the district” or if its cost “would be excessive as compared with the value of the property in the district.” Both are measured on the district as a whole, and no appeal lies from a dismissal.

We note, and do not rely on, the second ground. Niwot’s road bill runs between 0.54% and 1.68% of roughly $2.10 billion of residential actual value. A commissioner could call that excessive; nothing in the record suggests one would, and we have no precedent on what the word has meant.

Findings are conclusive. Under 30-20-508(3) the commissioners’ findings on signatures and related facts are “final and conclusive on all parties in interest, whether appearing or not.” Once the organizing resolution passes, 30-20-508(5) gives thirty days to attack the organization in court, after which it stands.

So the paperwork is not the obstacle. What a petition triggers is an election, and the election is where both Niwot attempts have ended.

3. How You Get Out, and Why Mostly You Do Not

There is no neighborhood opt-out. This matters because the campaign — ours included — has said otherwise, and it is not true. Nothing in Part 5 lets a subdivision decline to be in a district that includes it.

What exists is CRS 30-20-520, and it is much narrower:

The boundaries of any district organized under the provisions of this part 5 may be changed… The owners of property proposed to be included or excluded may file with the board a petition… accompanied by a deposit of money sufficient to pay all costs of the inclusion or exclusion proceedings… If the petition is granted, the board shall adopt a resolution…

Four things follow. Exclusion is available only after the district exists. The owner seeking it pays for the proceeding. The board decides. And the statute specifies no standard at all for granting or refusing.

And under 30-20-521, leaving does not clear the debt:

All property included within or excluded from a district shall thereafter be subject to the levy of taxes for the payment of its proportionate share of any indebtedness of the district outstanding at the time of inclusion or exclusion.

A district is easier to enter than to leave. It can be voted in by a majority of those who turn out; there is no corresponding vote to take it out.

One exception, and it points somewhere interesting. CRS 30-20-504(2):

If a municipality annexes or incorporates any territory within an established district, such territory shall remain in the district unless the municipality notifies the district’s board of the municipality’s intent to exclude the territory annexed or incorporated from the district.

So incorporation does not dissolve a district automatically — but a town government may vote territory out of one, effective the following 1 January, where an individual owner may only petition and pay. Under 30-20-504(3), once all of a district’s territory is incorporated and it carries no outstanding debt, the municipality simply takes over the district’s duties.

That asymmetry is the practical point. Inside a district you have a petition and a deposit. Inside a town you have a vote.

4. The Townwide Record: Twice Tried, Twice Not Delivered

2013 — defeated at the ballot. Boulder County itself placed a county-wide subdivision paving district on the November 2013 ballot as Issue 5C: up to 7.15 mills, up to $3.3 million a year, debt authorization up to $30 million. It failed.

2025 — withdrawn before the ballot. A citizen committee assembled a townwide proposal at about 12 mills. After resident input, surveys and attorney letters, the organizers withdrew it. It never reached a vote. Several of those organizers are part of the incorporation effort today.

It is worth being precise about what happened, because we have described it carelessly before. The 2025 proposal did not “fail” — it never reached a ballot to fail at. And neighborhoods did not ask to be excluded, because there is no such mechanism. What happened is that the people organizing it concluded it would not pass and stopped.

2026 — nothing on the table. There is no proposal today and nobody assembling one. This is the honest gap, and it is a stronger point than the often-repeated claim that “no organization exists to run one.” Running a district is the county’s job. The missing thing is upstream: there is no proposal for a voter to weigh against incorporation at all.

What the record supports, and what it does not. Two attempts in thirteen years, one defeated by voters and one abandoned by its own organizers, is evidence that a townwide district is hard to pass in this community. It is not evidence that it is impossible, and this paper does not claim that. The defensible statement is that nobody has managed it, twice, and nobody is trying now — and that a low petition threshold has never been the binding constraint.

5. The Concentration Problem, and the Exit That One Taxpayer Has

This is the substantive feasibility argument, and it does not depend on any contested figure.

A townwide district needs about $3.05 million a year if the Boulder Tech Center is inside it, and about $2.85 million if it is not — a district pays for the roads it contains, and the Tech Center’s 1.4 miles of parkway leave with its parcels. The levy is 15 mills with the Tech Center in, 20 mills with it out. On the median Niwot home of $1,094,850 that is $1,045 to $1,394 a year.

Fifteen parcels. About a quarter of the town’s taxable value. Worth five of those mills.

So the affordability of a townwide district rests on the continued participation of a single commercial owner group — the participant with the most money at stake, the most reason to object, and the most resources to act on it. A homeowner facing 15 mills can vote. The Tech Center can do considerably more than vote.

The route out is real, it has local precedent, and the city limits are already at the fence line. Under Colorado annexation law a municipality may annex contiguous unincorporated territory. Boulder County’s own annexation records show the City of Boulder has done exactly this pattern here, twice, in the 1970s: the IBM campus, 487 acres, and Gunbarrel, 261 acres — the industrial and commercial land taken into the city while a majority of the residences around it stayed in the unincorporated county. The IBM campus is Kyndryl now, and it is City of Boulder.

This is not a distant precedent. On the county assessor’s own municipal-boundary layer, the City of Boulder’s limit runs up N 71st Street along the western frontage of the Boulder Tech Center, and the land on the far side of that line is city land — some of it owned outright by the City of Boulder. The Tech Center is not near Boulder. It is at the edge of it.

The incentive runs the right way for them. The City of Boulder’s municipal mill levy is about 11.65 mills, which buys a full municipality — police, fire, streets, utilities. A Niwot townwide improvement district would levy 15 to 20 mills and buy pavement. A commercial owner comparing those two numbers is not being asked a hard question.

What we are not saying. Nothing has ever proposed excluding the Tech Center, and all fifteen parcels were inside the 2025 boundary. We are not predicting that the Tech Center leaves, and anyone who reads that into this is reading something we did not write. We have established that a route exists and that someone has walked it before, in this exact place, with this exact kind of property. We have not established that anyone intends to walk it again, and Colorado’s one-sixth contiguity requirement would still have to be satisfied on a specific boundary that nobody has drawn.

What we are saying is narrower and harder to argue with: a financing mechanism whose affordability depends on its largest taxpayer, when that taxpayer has a documented and cheaper alternative available to it and the residents do not, is a fragile mechanism. A district takes years to reach a ballot, and a boundary can lose parcels in that time.

6. Neighborhood Districts: Formable, Expensive, and Partial

The county’s own guidance points a neighborhood that wants to act alone toward its own district, and Boulder County has set up the mechanism and will help with the process. This is the version of a PID that demonstrably works.

It is also the version almost nobody uses. In ten years, two have formed in the county: Homestead in Gunbarrel, and Burgundy Park here in Niwot. Forming one requires a sustained volunteer effort, a petition, a boundary, a cost estimate and an election, carried by residents who are not being paid to do it.

And it is much more expensive per household. Across 24 costed districts covering 1,114 homes:

The typical figure is weighted by homes rather than by district, because a 21-home district should not count as much as a 140-home one. The range is real and it is wide: what you pay depends almost entirely on how much pavement your neighborhood has per home.

The deeper problem is not the price. It is that it never adds up. A neighborhood district fixes the streets inside one subdivision. It does not touch the collector roads that carry everyone, and it does nothing at all for any neighborhood that did not organize. Run the neighborhood path to its conclusion and Niwot has a patchwork: some subdivisions repaired at two to eight times the per-household cost of incorporation, the connecting roads still unfunded, and no mechanism that ever covers the whole town.

That conclusion does not rest on any single figure, which is why we lead with it.

7. The Two Neighborhoods That Come Out Even

In two of Niwot’s neighborhoods — Cottonwood Park Condos and Cottonwood Park West — a townwide district and incorporation cost about the same. We publish that rather than leave it out.

“About the same” is the whole finding and it will not carry more weight than that. Priced like for like, the condominiums land on the identical figure, $327 a year either way, and Cottonwood Park West prices about $7 a year lower under a district, on a bill of $632. That is roughly one percent, on a comparison whose driving input — how many people a multifamily household holds — carries a 45% margin at 90% confidence. Our own model’s verdict on these two is that the question “cannot be decided with available evidence.” A one-percent gap is not a result. These are ties, and we are not going to describe them as anything else in either direction.

Like for like means counting snow. A roads-only district and a town are not buying the same thing. The town’s budget carries snow removal, street sweeping, signs and striping; a district’s levy buys what its petition asked for. Set a district that bought only pavement against a town that also plows and you are comparing different baskets — the smaller one rings up lower, which tells you nothing about value. The statute permits a district to fund those services, as section 1 sets out, so the honest comparison prices both sides for the same work. Priced that way, these two neighborhoods come out even.

This is the calculation that prompted the study. Two neighborhoods where the arithmetic ran close is exactly where we should be most careful, and the question was whether we owed those households the news. Checking it is how we came to the rest of this paper — and the answer turned out to be that there is nothing to tell them. A townwide district is not something either neighborhood can choose. No proposal exists, nobody is assembling one, and the two attempts ended at a ballot box in 2013 and in a withdrawal in 2025.

A tie with an option that is not on offer is not a choice. We would rather say so plainly than leave those two neighborhoods believing there is a cheaper road ahead of them that nobody is building.

8. Conclusion

No public improvement district on offer fixes Niwot’s roads. That is the finding, and the two halves of it fail for different reasons.

A townwide district is legally available and practically unproven here. It can be petitioned with 200 signatures. It has been put to this community twice — defeated at the ballot in 2013, withdrawn before the ballot in 2025 — and there is no proposal today and nobody building one. Its economics depend on a single commercial owner group continuing to participate, and that owner has a cheaper alternative with local precedent behind it while residents have none.

Neighborhood districts work and do not scale. Two have formed in the county in ten years. They cost the typical household about 33 mills where incorporation costs 4 mills plus a shared sales tax, they cover only the subdivision that organized, and they leave the roads between neighborhoods unfunded permanently.

Incorporation is not cheaper because the pavement is cheaper. It is cheaper per household because the base is broader, and it is more durable because the authority is general. A town can levy for roads and also plow them, write a land use code, hold standing in county decisions, and — under 30-20-504(2) — vote territory out of an improvement district that an individual owner could only petition to escape.

What Would Change This Conclusion

We would revise this paper if any of the following turned out to be true:

• A townwide proposal is actually assembled and reaches a ballot, which would move the record from two attempts to three.
• The Boulder Tech Center’s owners state publicly that they intend to remain inside a district boundary.
• The City of Boulder states that it has no interest in annexing further commercial land north of Gunbarrel, which would weaken section 5 considerably.
• A survey of private-road pavement lands, which would let us price the eight neighborhoods currently excluded from the comparison for want of data.

Sources

• Colorado Revised Statutes 2024, Title 30, Article 20, Part 5 — County Public Improvement District Act — and section 30-15-401. Official uncertified printout, Office of Legislative Legal Services.
• Boulder County Assessor — 2025 actual and assessed values, parcel by parcel.
• Boulder County engineering — pavement condition index survey, scored street by street.
• Boulder County GIS — annexation history and municipal boundaries.
• Boulder County — Subdivision Paving PID program.
• City of Boulder — published municipal mill levy.
• Son-Haul — base bid for the town’s road program.
• PID Cost Methodology — the full derivation of every levy quoted here.
• The PID Proposal — the 2025 citizen proposal and the 2013 ballot issue.
• What a Road District Would Cost — all 34 neighborhoods priced at both ends of the range.

Every figure in this paper is derived by script from the source data and indexed at working/FIGURES.md, and the figures that appear on the public pages are asserted against those models on every build. Where this page and the scripts disagree, the scripts are right and this page is stale.