Australian Cost-Risk & Contingency Frameworks Compared

The one-line story

Three Frameworks, One Method

The TMR Project Risk Management & Contingency Development Process Manual (1st Ed, March 2023), the RES Contingency Guideline (3rd Ed, 2025) and the Commonwealth's DITRDCA Cost Estimation Guidance (Guidance Notes, v2.0, Nov 2023) read like three separate rulebooks. In substance they prescribe the same discipline:

A first-principles base estimate; risk modelled probabilistically with Monte Carlo simulation; contingency reported at P50 and P90; escalation kept separate; and flat-percentage contingency discouraged. A firm that genuinely speaks all three doesn't switch methods per client — it applies one rigorous method and reports it in each framework's language.

The three core frameworks

Side by Side, Dimension by Dimension

Fourteen dimensions across the three frameworks a Queensland or federally funded project is most likely to be assessed against. The differences are in vocabulary and emphasis — the underlying discipline is shared.

Dimension	TMR (Queensland)	RES (national best practice)	DITRDCA (Commonwealth)
Full title	Project Risk Management & Contingency Development Process Manual, 1st Ed (Mar 2023)	RES Contingency Guideline, 3rd Ed (2025)	Cost Estimation Guidance (Guidance Notes, v2.0 Nov 2023)
Status	Mandatory for QTRIP / state-funded transport projects	Guidance, not a mandatory standard	Mandatory for Infrastructure Investment Program funding
Estimate structure	Base + contingency + escalation	Base + contingency + Management Reserve (escalation out of scope)	Base + contingency + escalation
Base method	First-principles	First-principles assumed	First-principles preferred (client + construction costs; PLMS build-ups)
Probabilistic trigger	State > $10M; federal > $25M	Simulation recommended > $10M	Out-turn incl. contingency > $25M
Preferred QRA method	Monte Carlo (@RISK named); inherent risk on line items + discrete contingent risks	Monte Carlo / FPRA; bare 3-point cautioned at FID	Monte Carlo; risk-factor method preferred (line-item ranging cautioned against)
Iterations	~10,000 (rule of thumb)	— (method-dependent)	5,000–10,000 recommended
Distributions	Uniform, Triangular, PERT, Discrete	Triangular, BetaPert (prefers "Alt" forms), Lognormal	Beta PERT, Triangular, Trigen
Correlation	Must be modelled	Strong emphasis (functional + applied)	Must be defined ("cardinal rule")
Confidence levels	P50 budget · P90 mandated beyond business case · P75 APDV at tender award	P50 → PMB, P90 → Management Reserve (per org risk appetite); report P10/P50/P90	P50 = approval basis · P90 notionally held, released on demonstrated need
Flat-% deterministic	"Not recommended"; < $25M only	Not recommended (early/small only)	"Simplistic… not recommended"
Optimism bias / RCF	Cites Flyvbjerg in the project-delay category	RCF = benchmark/governance check, not a method	"Reference Class" dropdown in the PCB template
Reporting outputs	S-curve, histogram, tornado, P5–P95 at 5% intervals, re-runnable @RISK files	S-curve at P10/P50/P90; FPRA report	S-curve / CDF, histogram, tornado, P50/P90
Escalation	In PCEM; → P90 out-turn	Excluded from the guideline's scope	Note 4: RCOCI (road) / Rail-COCI (rail), zero-floor; → out-turn
Distinctive device	APDV at P75 + portfolio drawdown / savings	Management Reserve outside the PMB; Contingency X-Factor	PCB template (auto-escalation); held-P90 funding

Reading the table

Where one framework names a tool (TMR's APDV, RES's Management Reserve, DITRDCA's PCB template) the other two have a functional equivalent. The estimate that satisfies one — bottom-up base, modelled correlation, Monte Carlo S-curve, P50/P90 — is structurally the estimate that satisfies all three.

The thresholds reconciled

$10M and $25M Point the Same Way

The triggers look like they conflict. They don't — each one says "model it probabilistically", just at a different cost band and funding source.

> $10M — TMR requires probabilistic modelling for state-funded projects; RES recommends simulation above this value too.

> $25M — both the Commonwealth (DITRDCA) and TMR (for federally funded work) require Monte Carlo for total out-turn cost including contingency above this figure.

Whether your project crosses TMR's $10M state trigger or the Commonwealth's $25M out-turn trigger, the answer is the same — model it probabilistically. Cenex builds the estimate to satisfy both.

From RES Appendix G

Every Australian Jurisdiction Is Converging

RES's Appendix G compiles each jurisdiction's own requirement. The pattern is unmistakable — probabilistic, P-value-based estimation is now the norm nationwide.

Jurisdiction	Confidence levels	Threshold / trigger	Mandated method
Commonwealth — DITRDCA (IIP)	Out-turn P50 / P90	Probabilistic > $25M out-turn	Monte Carlo; PCB template
Commonwealth — Dept of Finance	P50 (Stage 1) / P80 (Stage 2)	Two-Stage Capital Works	Monte Carlo S-curve
QLD — TMR	P50 & P90 (P75 at award)	Probabilistic > $25M (fed) / > $10M (state)	Monte Carlo / risk workshop
NSW — INSW / TfNSW	P50 & P90	Contingency ID > $100M	Monte Carlo (min. for High-Profile / High-Risk)
VIC — DTF / OPV	P90 approval; P60 at contract award	HVHR framework	Probabilistic, staged release
WA — Treasury SAMF	P90 or P50 (HVHR)	High-value / high-risk	Probabilistic (buildings use layered %)
SA — TI17 / DIT EST600	P50 & P90	Full compliance ≥ $11M	Probabilistic risk methods
TAS — SIIRP	Not specified	—	Documented in business case
NT	None identified	—	—
ACT — Capital Framework	Tier-based	—	Tier 1 stochastic (Monte Carlo)

Takeaway: seven of eight states and territories, plus both Commonwealth pathways, now expect probabilistic, P-value contingency. The method Cenex uses is the method the country has standardised on.

Speak all three

Shared Glossary

The core terms mean the same thing in every framework. Framework-specific terms are defined on each framework's own page.

Base estimate — the expected cost of the defined scope, excluding contingency and escalation.

Contingency — the provision above the base estimate for residual (inherent + contingent) risk, sized to a chosen confidence level; expected to be expended.

Inherent risk — variability in items already measured in the base estimate (quantities, rates).

Contingent risk — discrete risk events not in the base estimate (probability × impact).

Escalation — cost growth over time (inflation / market); converts real cost to out-turn (nominal) cost; kept separate from contingency.

Out-turn cost — base + contingency + escalation; the nominal total.

P50 / P90 — cost with a 50% / 90% probability of not being exceeded (read off the S-curve).

Monte Carlo simulation — sampling the modelled distributions over thousands of iterations to produce the cost probability distribution (S-curve).

Framework-specific terms

Defined on the relevant page only: APDV, EFCT, RCP (TMR); PMB, Management Reserve, FPRA, QSRA, iQSCRA, Contingency X-Factor (RES); PCB, RCOCI, Rail-COCI, NoA (DITRDCA).

One evidence pack

What Cenex Delivers

Every one of these frameworks expects the same evidence pack. Cenex produces it as standard — and reports it in whichever framework's language your project is funded under.

First-principles base estimate (PLMS build-ups; direct / indirect / margin; auditable to a 4-level WBS).
Facilitated risk workshop and a dollarised risk register.
Monte Carlo model (@RISK, 5,000–10,000 iterations) with inherent risk on line items, discrete contingent risks, and correlation modelled.
S-curve, histogram and tornado outputs; contingency reported at P50 and P90 (plus P75 for TMR award / P60 for VIC award where relevant).
Escalation to out-turn cost handled separately (framework-appropriate indices).
Reference class forecasting as an independent benchmark cross-check on the modelled result.
Chartered-Engineer sign-off, with re-runnable native model files for the funder's review.

Most estimators can produce a contingency number. Far fewer can produce one that satisfies the specific framework a project is funded under, defend it against an independent Commonwealth or TMR review, and report it in P50/P90 with the model evidence behind it. Cenex's independence — no downstream delivery interest — means the contingency is challenged, not inflated, so the P90 holds and savings are realised at the P75 / P60 award reset.

Which Framework Is Your Project Funded Under?

Tell us the framework — TMR, Commonwealth / DITRDCA, or a state mandate — and we'll build the probabilistic estimate and contingency it requires.

Contact Us Today Risk Modelling Service