Routine, preventive, rehabilitation: three kinds of road work, and what each is for

"Fix the worst first" is not a strategy, it is the absence of one. Austroads already names the three kinds of maintenance and says plainly what the middle one is for — and the middle one is the one councils under pressure stop doing first.

Every council does road maintenance. Far fewer can say, in a sentence a councillor will follow, which kind of maintenance a given job was and why that was the right kind.

Austroads does say it. Guide to Pavement Technology Part 7 divides maintenance into three, and the definitions are worth reading in the original because the middle one carries the whole argument.

The three kinds

Routine maintenance addresses minor defects on the carriageway, off- carriageway work like drain clearing and grass cutting, and the essential activities that keep a base level of road safety — removing obstacles, making a hazard safe. Austroads notes these works are "normally unplanned or planned with short lead time", using light equipment and small quantities of material.

Pothole patching lives here. So does most of what a council does in a wet winter.

Preventative periodic maintenance is the one to read twice. Austroads defines it as work "intended to reduce future deterioration by timely surface interventions to limit the need for expensive rehabilitation, and to ensure minimum skid resistance and general safety levels do not fall below minimum accepted levels". It is planned, with lead time generally greater than a month.

Three words in that definition do the work: reduce future deterioration, timely, and limit the need for expensive rehabilitation. A reseal on a sound pavement is not a small repair. It is the thing that stops a large one.

Rehabilitation targets roads whose ride quality has deteriorated significantly, or which no longer have the structural capacity for the traffic they carry. Planned with months of lead time, usually inside an annual or three- to five-year programme.

Why the middle one goes first

Routine maintenance is forced on you: a pothole opens and somebody rings. The road is already failing, and the work is already overdue.

Rehabilitation is visible: it is a capital line item, it appears in a budget, and somebody has to approve it.

Preventive work is neither. Nobody rings about a road that is still sound. It does not look urgent in a photograph. And the year you skip it, nothing happens — the consequence arrives three to seven years later, by which time it looks like general deterioration rather than a decision anyone made.

So it is the line that gets deferred, and deferring it is the most expensive thing a road budget can do slowly.

It is happening now, and an auditor-general has written it down

This is not a theoretical risk. In September 2026 the Victorian Auditor-General published Maintaining State Roads, and its findings describe the mechanism above almost exactly.

Around 48% of Victoria's state roads were in poor or very poor condition in 2024, up from around 39% in 2021. The gap between the number of defects recorded and the number repaired more than doubled between January 2025 and June 2026 — the network deteriorating faster than the pace of repairs.

And the cause, in VAGO's words: the department "has advised the government that funding is insufficient to maintain or improve the condition of state roads. This has meant it focuses more on reactive and safety-critical repairs rather than preventative works."

That is the middle category being squeezed out, observed and audited.

One important qualification. VAGO was auditing state roads, which are around 15% of Victoria's network by length and are managed by the department, not by councils. Councils manage the rest. The findings are not a statement about any council's roads — they are evidence that the funding-to-reactive-work mechanism is real at scale, in an Australian jurisdiction, now.

The evidence, and where it comes from

The strongest numbers on this question are American and old, and they should be labelled as both.

The US Federal Highway Administration's Pavement Preservation: The Preventive Maintenance Concept (FHWA-HI-00-006, 1999) is the standard reference. Its findings: a dollar spent on preservation defers four to five dollars of rehabilitation; in Michigan, rehabilitation cost roughly fourteen times preventive work per lane-mile; network simulations in New York, Michigan and Wisconsin found that "preservation first, then worst roads" outperformed "worst first" at the same budget.

Two cautions a careful reader will raise, and should:

• It is US evidence from a different funding structure and a different climate, now twenty-seven years old. The logic — right treatment, right pavement, right time — is current practice. The multipliers are not Australian and should not be presented as though they were.
• The ratios are idealised. They come from modelled comparisons, not from a controlled trial on your network.

Which is the honest reason to test it on your own roads rather than to quote it: a before-and-after record of your own treatments, with the condition before the work and the cost, is worth more to your council than any number from 1999.

What this looks like in a works programme

If the three kinds are real, the programme should be able to say which kind each job is, and the answer should not be "rehabilitation" for everything.

Four questions that make the distinction concrete:

1. What was the condition before the work? Preventive work by definition happens on a structurally sound pavement. If the pre-treatment condition is not recorded, you cannot later tell whether a treatment worked, or even whether it was the right kind.
2. What would this road have cost if left another three years? That is the comparison preventive work wins, and it is the only comparison in which it does.
3. At this budget, what gets deferred? Deferral is a decision. It is worth making it visibly.
4. How much of the programme is reactive? A programme that is nearly all routine maintenance is a network being managed by its own failures.

How PaveKeep handles this

PaveKeep produces a ranked, costed works programme with the reason each road is on the list, and shows what a given budget buys over five or ten years — including what a different budget would have bought instead. That comparison is how the preventive case gets made to a council: not as a principle, but as two columns.

What it does not yet do: it does not learn what a treatment actually achieved from before-and-after condition, which is the one thing that would turn the US evidence above into your own. The feature exists, is modelled from a small number of observations rather than measured, is off unless a council turns it on, and is not a basis for a spending decision yet.

PaveKeep is built for small and regional councils and is not yet in use at any council.

Sources

• Austroads (2009), Guide to Pavement Technology Part 7: Pavement Maintenance, §1 — the routine / preventative periodic / rehabilitation division and the definitions quoted above.
• Austroads (2025), Guide to Pavement Technology Part 5: Pavement Evaluation and Treatment Design — where periodic maintenance and rehabilitation design are treated.
• Victorian Auditor-General's Office (9 September 2026), Maintaining State Roads, Report 2026–27 — the 48%/39% condition figures, the defect gap, the 15% and 82% network shares, and the funding finding quoted above.
• Federal Highway Administration / National Highway Institute (1999), Pavement Preservation: The Preventive Maintenance Concept — Executive Overview, FHWA-HI-00-006. US evidence.