Budget reforecasting is the moment an Australian government IT project has to admit that its original numbers no longer reflect reality. It's uncomfortable, it's politically charged, and it happens on almost every significant project. The question is not whether a reforecast will be needed, but how early it gets done and how honestly it gets framed.
Most agencies treat reforecasting as a last resort. They hold to the original baseline until the gap becomes indefensible, then seek approval for a revised figure that is often still optimistic. The pattern repeats. A project that is approved at $40 million becomes a $60 million approval two years later, then $85 million at delivery. Each reforecast is presented as the definitive revision, and each one turns out not to be.
Why the original budget rarely holds
Government IT budgets are set early, under pressure, and with incomplete information. Agencies work backwards from funding available rather than forward from a scoped solution. Business cases are written to win approval, not to predict cost. That creates a structural gap between what gets approved and what delivery actually requires.
Three things reliably blow the gap wider. Scope changes that weren't anticipated at business case stage add cost without triggering a full reforecast, accumulating quietly until the total becomes undeniable. Integration complexity, particularly with legacy systems, is consistently underestimated at the front end of a project. And resourcing conflicts that push specialist work to higher-cost contractors extend both timeline and cost in ways that weren't modelled.
Procurement also plays a role. Fixed-price contracts that shift risk to vendors often see that risk repriced through variations and scope negotiations. Time-and-materials arrangements let costs drift without natural pressure points. Neither model makes reforecasting easier; they just change where the cost surprise lands.
The governance gap in most reforecasting processes
Most agencies lack a formal trigger for reforecasting. Reforecasts tend to happen when a senior officer raises the alarm or when quarterly reporting makes the variance impossible to ignore. There's rarely a predefined threshold, such as a 10 percent budget deviation, that automatically activates a structured reforecast process.
Steering committees are often the weakest link. They receive financial reports but don't have the technical depth to interrogate whether a projected underspend in one quarter signals good management or deferred work. Steering committees set up to hear good news tend to wave through financial reports that deserve harder questions.
Risk registers compound this. When risks are listed in a register but not linked to financial contingencies, a project can carry a high-probability risk for months without that risk appearing in any budget conversation. The risk and the budget live in separate documents, reviewed by separate people, on separate cycles.
What a defensible reforecast actually looks like
A good reforecast does three things: it accounts for completed work at actual cost, it re-estimates remaining work from current scope rather than original scope, and it applies an explicit contingency based on what is still uncertain. Most agency reforecasts do the first, partially do the second, and treat contingency as a number to minimise rather than an honest reflection of remaining risk.
The contingency problem is significant. Treasury guidance on contingency reserves exists, but project teams routinely argue for the lowest defensible number because higher contingency signals poor planning rather than good risk management. This incentive runs backwards. A project that carries realistic contingency is less likely to need a second reforecast than one that carries none.
Independent estimation is underused. Some agencies bring in external reviewers to validate a reforecast, but this is more common on very large projects than on mid-tier ones where the governance overhead is seen as disproportionate. For projects in the $20 million to $80 million range, which represent a large share of agency IT investment, independent cost validation is the exception rather than the norm.
How agencies present reforecasts to ministers and central agencies
Reforecasting a government IT project isn't only a technical exercise. It requires approval from Finance or Treasury, briefing of the responsible minister, and often notification to the portfolio department. The political dimension shapes how reforecasts are framed: agencies soften the language, emphasise what is being delivered for the additional investment, and avoid presenting a reforecast as a project failure.
This isn't dishonest, exactly. A reforecast that gets rejected by a minister gets replaced by a project that continues on an impossible budget, which is worse. But the instinct to package reforecasts favourably means the underlying causes, scope growth, poor original estimation, vendor disputes, don't get fully disclosed. Future project teams inherit the same structural problems without knowing they existed.
The Australian National Audit Office has repeatedly found that agencies don't adequately document the reasons for budget growth on major IT projects. That means the institutional knowledge from each reforecast, what drove the gap, what the options were, what was decided, largely disappears when the project closes.
What better practice looks like in Australian agencies
A small number of agencies have moved toward continuous reforecasting, where the financial model is updated monthly against actuals and remaining work is re-estimated at each reporting cycle. This reduces the shock of a formal reforecast because the numbers move in small steps rather than large jumps.
Linking the risk register directly to the financial model is another genuine improvement. When a risk is elevated from medium to high probability, the financial model should automatically update the contingency calculation. This connection is technically straightforward but organisationally difficult because it requires the project manager, financial manager, and risk manager to work from a single shared model rather than their own documents.
How lessons learned sessions handle budget reforecasting matters too. If a closeout review documents the full history of budget growth, including what triggered each reforecast and what the organisation knew at the time versus what it disclosed, that institutional knowledge can prevent the same pattern on the next project. Most don't go that deep.
The agencies that handle reforecasting best aren't the ones that never need to reforecast. They're the ones that reforecast early, disclose clearly, and adjust scope to match funding rather than repeatedly seeking more funding to match scope.