The Road Cost Model

How we priced fixing Niwot’s roads, segment by segment

Model version of August 31, 2026

We say repairing Niwot’s roads costs about $15 million if the work starts now, $22 million by 2034 and $33 million by 2040. This page is where those figures come from: all 179 road segments, the condition of each, the repair it needs, what a contractor bid to do it, and every assumption in between.

The result

The middle column fixes one error in the workbook, which applied 13 years of inflation to 2040 instead of 14. As built, the workbook shows $34.2M for 2040. See What we checked.

The published figures date from the June version of the model, which came to $15.73M for 2029; $15M is that figure rounded down. The August version repriced two segments that had been priced too low, which raised every year. We have kept the published figures. For 2029 and 2040 they are below the model; for 2034 the model rounds to the published figure.

How it works

1. Every road is measured

The model lists 179 segments: 31.70 miles and 5,102,459 square feet of pavement, an average of 30.5 feet wide. Each segment has its own length and width. Cul-de-sac bulbs are added separately, at 2,500 square feet each.

The model counts every road inside the proposed boundary, including the ones someone else maintains today:

2. Every road has a condition score

Pavement condition is scored from 0 to 100 on the Pavement Condition Index (PCI). The 23 miles that no one maintains were measured in 2024. Their simple average is 44. For the county arterials and the HOA roads, no survey exists, so the model assumes a score. See Assumptions.

3. The score decides the repair, and the repair decides the price

The bid prices are Son-Haul’s, from the bid reproduced in Appendix A of the road paper. The overheads cover what the bid excluded: the performance bond, mobilization, traffic control, erosion control, engineering and design, and inspection. Reconstruction also carries mix design and testing, water, and subgrade work. Each segment is priced whole, at the rate for its band.

4. Roads decay until someone fixes them

Each road loses 2.0 PCI points a year, in a straight line, and no one repairs anything in the meantime. That moves pavement out of the cheap bands and into the expensive ones:

By 2040, 90% of Niwot’s pavement sits in the two most expensive bands, and none is left in the cheapest. That shift alone takes the cost from $11.4M to $23.3M before a dollar of inflation.

5. Contingency and inflation

The total gets a 10% contingency for the roads and another 10% for ancillary work (×1.20). It is then escalated at 3% a year for construction inflation, from 2026 to the year the work is done.

Assumptions

Every estimate rests on assumptions. These are the model’s, with the direction each one pushes the result.

1. All 31.7 miles are priced, including roads others maintain today. The county’s 5.1 miles of arterials and the 3.0 miles of HOA roads are in the total. A town would own the public roads. Whether it takes on private roads is a decision for a future council and the associations. Raises the estimate.
2. The county arterials are assumed to be PCI 80, and HOA roads 50, 60 or 80. No survey of them exists. The model then lets them decay unrepaired with everything else. They are 3% of the cost today and 17% by 2040. If the county and the associations keep maintaining them, the 2040 figure is lower than shown. Raises the estimate, mostly in later years.
3. Roads lose 2.0 PCI points a year, in a straight line. Fourteen Niwot segments with known repair dates averaged 2.09 points a year. Pavement research describes decay as a curve that steepens once a road falls below roughly PCI 55 to 70, and Niwot’s average is already 44. Lowers the estimate.
4. Each segment gets one repair, priced at its band’s rate. The bands change at PCI 25, 50 and 70. A real repair program would price and phase work road by road, but these four bid prices come from that kind of program. Neutral.
5. Prices are Son-Haul’s bid plus overheads. The overheads add 40% to reconstruction, 22% to mill & overlay and 19% to seal treatments. Each line is listed in Appendix A. Neutral.
6. 20% contingency on top. 10% for the roads and 10% for ancillary work. Raises the estimate.
7. Construction inflation of 3% a year from 2026. Neutral.
8. Each year’s figure prices the whole network as if it were all repaired in that year. In practice the work would run over two or three construction seasons. Neutral at this level of estimate.
9. Cul-de-sac bulbs are 2,500 square feet each. The workbook’s note says an 80-foot diameter, which would be about 5,027 square feet. The smaller figure is the one used. At the note’s own size, the 2040 cost would be about $0.7M higher in 2026 dollars. Lowers the estimate.
10. Pavement only. No curbs, gutters, sidewalks or drainage. Niwot is almost entirely curbless, and the bid and its overheads cover the pavement work. Ongoing maintenance after the repair is budgeted separately in the pro forma. Leaves out work that may be needed.

Taken together, the assumptions that raise the estimate (the county and HOA roads, the contingency) are roughly offset by those that lower it (the slow straight-line decay, the small cul-de-sacs, and the errors below). On balance the published figures are a fair estimate that leans low.

What we checked

We rebuilt the model from its segment data, independently of the spreadsheet, and compared it with the workbook line by line. Every year’s total reproduces to the dollar. That check also found the following. Each one makes the estimate too low, not too high.

We have not raised the published figures. They are the ones the committee has used in print, and they are lower than the model, not higher.

Every road segment

Condition by year, under the model’s assumptions. Shading shows the repair band. Cost is the delivered price to repair the segment in 2029, in 2026 dollars, before contingency.

Reconstruction (PCI below 25)Mill & overlay (PCI 25 to 50)Seal (PCI 50 to 70)Crack seal (PCI above 70)

Neighborhood names are the workbook’s. “Assumed” marks a segment whose starting condition is assumed, not measured. Four Niwot Hills streets are new pavement and start at PCI 100. Segment costs sum to $12.2M; the 20% contingency brings that to $14.7M.

Test it yourself

The model is also published as a working spreadsheet. Change the decay rate, the unit prices, the contingency or the inflation rate, and every figure recalculates. It includes the corrections described above.

Download the road cost model (Excel)

Derived from the committee’s working file. Working notes on private-road ownership have been removed. They are our reading of other parties’ property records, not established fact. Nothing removed affects a cost figure.