# Showing a council what a budget buys, and what the next one down costs

*"We need more for roads" is a position. "At $2 million the network holds, at
$1.4 million it drops four points and we lose these roads" is an argument. The
second one needs a scenario, and a scenario needs to be checkable.*

The hardest conversation in local road management is not technical. It is the one
where an engineer has to explain to people with other legitimate priorities why
the road line should not be the one that gives.

Condition charts do not win it. A line sloping gently downward over ten years
invites the reply that it is sloping gently, and gently is survivable. What
changes the conversation is a comparison: the same ten years, under two or three
spending levels, with the roads named.

## The size of the question

The Grattan Institute's *Potholes and pitfalls: How to fix local roads* (2023)
estimated that Australian councils are collectively underspending on roads by
**at least $1 billion a year — almost a quarter of current maintenance
expenditure**. The councils that are actually underspending are underspending by
**$1.3 billion** a year; the lower figure nets off councils, mostly in major
cities, that are spending more than maintenance requires.

Two things about that number make it usable rather than rhetorical. Its method is
published — digital mapping data to model the sources of variation in maintenance
cost, such as urban density and rainfall, applied to standardised prices and
compared with what each council currently spends. And it is deliberately modest:
Grattan is clear that $1 billion is only what is needed to **hold roads at their
current condition**, which for many councils is already below what their
communities expect.

So a scenario that shows a network merely holding steady is not a conservative
scenario. It is the floor.

## What a scenario has to contain

**A budget, a horizon, and a rule for spending it.** The rule is the part usually
left unstated and it decides everything. Worst-first, lowest-cost-per-point,
preserve-the-good-first and a benefit–cost ranking will each produce a different
programme from the same money, and each is defensible for different reasons. A
scenario that does not say which rule it used is not reproducible.

**The roads, not just the curve.** A network average is the output. The thing a
councillor can act on — and the thing that makes the trade-off real — is which
segments are in and which are out.

**What it leaves behind.** The roads that did not make the cut at this budget are
the cost of that budget. Presenting only what gets done presents half of the
decision.

**The unsurveyed length, stated separately.** A programme built on the roads that
were surveyed is not a whole-network programme. Condition that has not been
measured cannot be ranked, treated or priced, so it should be reported alongside
the programme rather than silently excluded from it.

## The comparison that does the work

Run the same budget two ways: once spending it on the worst roads, once spending
it on keeping sound roads sound, and report where the network average lands after
ten years in each.

That comparison is the preventive maintenance argument made as two columns rather
than as a principle. Austroads defines preventative periodic maintenance as work
"intended to reduce future deterioration by timely surface interventions to limit
the need for expensive rehabilitation" (*Guide to Pavement Technology Part 7*,
2009) — and a scenario pair is how that definition becomes a number somebody can
approve.

Be careful about the magnitudes. The widely quoted multipliers for preservation
versus rehabilitation come from US research — the Federal Highway
Administration's 1999 overview reports a dollar of preservation deferring four to
five dollars of rehabilitation, and Michigan rehabilitation at roughly fourteen
times preventive work per lane-mile. The logic transfers. The numbers are from a
different country, a different climate and a different funding structure, and
quoting them as though they were local is exactly the thing a sceptical reader
will catch.

Your own network will produce better numbers, if the pre-treatment condition and
cost of each completed job are recorded. Most councils already record the cost.

## Beyond condition: the question a condition chart cannot answer

A network can hold a steady condition score for years while the budget stops
covering what the network is losing in value. Condition is a *lagging* measure —
it reports where the pavement is, not whether this year's spending kept up with
this year's wear.

The measures that answer that are ratios of money to money: renewal spending
against depreciation, budget against need, unfunded backlog against replacement
value. They are arithmetic on figures a council already produces for its
financial statements, and they tend to land with finance officers and general
managers in a way a condition curve does not.

They are also the point at which a scenario stops being an engineering artefact
and becomes a financial one — which is the room where road budgets are actually
decided.

## Three habits that make a scenario survive questioning

1. **Name the rule.** Whatever ranked the programme, say it in a sentence on the
   page the councillor reads.
2. **Show the inputs.** Unit rates and treatment thresholds are what convert
   condition into dollars. If they are not inspectable, the dollar figure is not
   checkable.
3. **Report the gaps in the same table.** Unsurveyed length belongs beside the
   programme, not in an appendix.

## How PaveKeep handles this

PaveKeep takes a budget and a number of years and shows what it buys: which roads
get done, what condition the network ends up at, and what a different budget
would have bought instead. Each recommended road carries the reason it is on the
list. The works programme also reports the unsurveyed length alongside it, so a
programme is never presented as covering more of the network than it does.

What it does not yet do: it does not report asset sustainability, asset
consumption or service-level ratios as named measures. The inputs those need —
the network's value, and the cost to reach a condition target — are in its
outputs and can be traced to their sources, so a council can calculate them. That
gap is recorded rather than left to be discovered.

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

## Sources

- Grattan Institute (2023), *Potholes and pitfalls: How to fix local roads*,
  Report No. 2023-11, Terrill, Bradshaw and Jones — the $1 billion and $1.3
  billion annual underspend figures and the method behind them.
- Victorian Auditor-General's Office (9 September 2026), *Maintaining State
  Roads*, Report 2026–27 — insufficient funding shifting effort from preventative
  to reactive and safety-critical work.
- Austroads (2007), *Guide to Asset Management Part 5B: Roughness*, AGAM05B-07 —
  intervention levels, and using a roughness time series to test whether the ones
  in force are holding a network stable.
- Austroads (2009), *Guide to Pavement Technology Part 7: Pavement Maintenance* —
  the definition of preventative periodic maintenance quoted above.
- Austroads (2025), *Guide to Pavement Technology Part 5: Pavement Evaluation and
  Treatment Design* — network-level versus project-level analysis.
- Federal Highway Administration / National Highway Institute (1999), *Pavement
  Preservation: The Preventive Maintenance Concept — Executive Overview*,
  FHWA-HI-00-006 — the preservation-to-rehabilitation ratios and the New York,
  Michigan and Wisconsin network simulations. **US evidence.**

---

Source: https://pavekeep.com.au/knowledge/what-a-budget-buys · PaveKeep · bound content