5 Rolling Expenditure Plans: Australian Experience and Prognosis

A. Premchand
Published Date:
June 1990
  • ShareShare
Show Summary Details


As in other countries, Australian Governments in the 1970s and early 1980s found themselves presiding over apparently inexorable growth in the public sector. This growth was fueled by public expectations that had failed to adjust to slower economic growth, and that were arguably a cause of that slower economic growth. When a new government took office at the nadir of an economic recession in 1983, it was faced with a quite unacceptable prospective budget deficit. This largely reflected a blowout in outlays, and as there was deemed to be no possibility of increasing the burden of taxation, savings relative to prospective outlays were the first priority for fiscal policy.

This background is essential to the Australian Government’s reforms of public sector management and the budget. However, the reforms have not only been concerned to establish tighter expenditure control but also to establish a system of public administration that was both more responsive to government priorities and more effective in implementing those priorities. Indeed, in its first year in office, the Government released separate discussion papers on public sector reform and budget reform that were directed to these overlapping purposes. Thus, a principal aim of budget reform was “to develop a better means of identifying and setting budget priorities” (“Budget Reform” (April 1984), p. 1), which depended in significant measure on achieving the first objective of public sector reforms, namely, establishing “an administration that is more responsive to Ministers and Parliament” (“Reforming the Australian Public Service” (December 1983), p. 1). In addition, the budget reforms were intended to upgrade the financial management of programs, to focus attention more closely on the goals and objectives of particular programs, to review regularly their cost effectiveness and efficiency, and to ensure that the results of such reviews are taken into account in the ongoing evaluation of budgetary priorities.

What is interesting about the Australian experience, therefore, is how the Government has sought to combine fiscal restraint with the devolution necessary to achieve a more responsive administration, better resource allocation, and improved public management. Normally those two objectives of restraint and devolution are hard to combine. As a general rule, the greater certainty and stability necessary to underpin devolution to one level of management implies less flexibility to reallocate resources at a higher level of management. How, for example, does one ensure that the advantages of greater stability for program managers are not thwarted by the need for budget flexibility at the political level, including the need to be able to shift priorities from one budget to the next? Similarly, how useful to managers is their devolved responsibility for results in an environment of centrally imposed expenditure restraint?

The establishment of reliable and centrally determined forward estimates for all government programs, in conjunction with a system of program management and budgeting, has been essential to the resolution of these tensions. In particular, the present stability in program funding has depended upon the improved quality of the forward estimates and the avoidance of unexpected blowouts. It is within the tight aggregate controls based on the forward estimates that a system of program management and budgeting has been established that aims to maximize managerial flexibility and incentives.

It is against this background that this paper describes the development of, and experience with, rolling expenditure plans based on forward estimates of expected budget outlays in Australia.


Ever since 1973, the Australian Government has taken the view that preparation of the annual expenditure budget requires a perspective beyond the year in question. Very few decisions taken by the Government have an impact that does not transcend the present year—indeed, most are open-ended in time and can be expected to affect estimates beyond the time horizon of politicians and bureaucrats.

In Australia, as in most countries, the “budget” is a single-year document, in that the Parliament can only appropriate funds for general government operations for one year at a time, and, by convention, the event of presenting the budget to the Parliament constitutes an annual statement of the state of the economy and an occasion to promulgate the Government’s plan of action in all its areas of activity for the next year. This annual submission to Parliament is largely an illusion, however, as nearly three fourths of the outlays covered in the budget are authorized by standing legislation and are characterized more by their stubborn permanence than by their susceptibility to the changing priorities of governors and legislators. Furthermore, much of the remainder is also predetermined by implicit contracts of employment or expectations of future funding. The projections of expenditure over the coming years, therefore, are well determined in principle.

Prior to the incumbency of the present Australian Government, the projection of a spending “base” over the next three years1 formed the starting point for the budget. All budget-dependent agencies, which constitute departments of state and noncommercial statutory authorities, would be requested to submit estimates of the expenditure levels required over three years in order to maintain those programs and activities authorized by the Government in past budgets (or possibly since the last budget). By definition, these estimates were not to cover new programs or significant extensions of the scope or size of existing programs. These estimates were collected by the Department of Finance, which is the central agency with responsibility for public expenditure management and policy. The estimates were generated by individual agencies and expressed in then-prevailing prices, so that price adjustment was under central control. Any terminating program, for example, subject to sunset legislation or a capital program requiring ongoing authorization, was excised from the total estimates. Once submitted, the estimates were collated by the Department of Finance for government consideration under the cloak of secrecy that characterized budget preparation. Their purpose was to allow the Government to assess the scale of the budgetary task facing it and thus to settle on a broad strategy. These estimates were then promulgated to agencies as guidance in preparing their bids for budget funding.

The product of this process suffered serious deficiencies. First, the budget process encouraged agencies to raise their bid to create some padding for the inevitable cuts and to minimize real pain. The forward estimates constituted “wish-lists” and not the base estimates that were intended. That this was possible reflects the breadth of interpretation that can be attributed to government decisions that define existing policy, and consequently program costs, especially over the mists of time. Indeed, it is part of the political armory to leave open the prospect of better circumstances ahead, and agencies, as well as clients, do not need to be invited to take advantage of any uncertainty.

Disagreements between the central and operating departments about what were properly authorized expenditure levels were expected to be resolved by officials, but many disputes had to be brought to the Cabinet to resolve. This process of negotiations and “disagreed bids” was time-consuming and inefficient, and it had developed an exaggerated role in settling the budget estimates.

A second deficiency was the conservative bias built into the estimates by the strict interpretation of continuing authority. Understandably, perhaps, there was an attitude within central agencies that inclusion of an item in the forward estimates constituted de facto approval. But, with the exclusion of major items, such as revenue-sharing agreements with other levels of government, because the term of the agreement expired during the forward period, the estimates became poor forecasts of pressures for expenditure.

As a result of these deficiencies, but also because the prevailing administrative procedures made forward estimates of little relevance to managing the present year’s finances, scant attention tended to be given to the second and third years’ estimates. In their old guise, therefore, the forward estimates represented little more than inflated bids for the budget year. There was no attempt to use them for medium-term expenditure planning, or even to take account of the medium-term implications of individual expenditure decisions.


When the present Australian Government assumed office in 1983, it was faced with estimates of outlays at their cyclical peak and heavily locked in, a fiscal situation that placed severe constraints on its ability to carry out its mandate. The Government therefore embarked on a program of public awareness of the need for restraint and better targeting of public expenditure. Indeed, its first day in office was dramatized by an announcement of the prospective budget deficit implied by the forward estimates. The publicity campaign was developed at a National Economic Summit conference held one month into its term of office, at which leaders of industry, trade unions, and community groups debated the economic agenda publicly with the Government. One of the information documents released to participants by the Government to enhance discussion was the (previously secret) “Report of the Forward Estimates of Budget Outlays” for the ensuing three years. Publication has since been a continuing feature and in the recently released budget (August 1989), the forward estimates were fully integrated into the budget presentation, setting the outlays estimates and their analysis into a proper medium-term perspective.

Two important disciplines accompanied this more open approach to economic policymaking. First, the economic parameters underlying the projections were revealed and the uncertainty surrounding them made clear. A sensitivity analysis of the outlays estimates to the key economic parameters was also included so that the effect of the uncertainty in the economic forecasts could be assessed by readers.

The second discipline was the publication in the budget documents of a reconciliation of the budget and forward estimates showing the influence of policy decisions, both positive and negative, and parameter and other nonpolicy changes by functional classifications. This reconciliation analysis has been refined over the years to take its place now as a centerpiece of the presentation of outlays information (an example from the 1989 budget is presented in Table 1). In an environment in which a virtue has been made out of budgetary savings, this table clearly exposes the degree of fiscal discipline to which the Government has subjected itself, and consequently greatly encourages that discipline. More generally, heightened public awareness accompanying better disclosure has been a crucial factor in providing acceptance of the extended period of restraint experienced in Australia in recent years. Moreover, the information now available on the factors underpinning changes in program outlays has enabled such nongovernmental bodies as the Business Council of Australia to produce well thought out and influential proposals for expenditure restraint. To cite just one recent example, the information published in association with the forward estimates has played an important role in underlining the need for government action to control the burgeoning costs of pharmaceutical benefits, including such controversial initiatives as reducing the rate of remuneration of pharmacists.

Table 1.Australia: Reconciliation of Forward and Budget Estimates, by Function, 1989/90(In millions of Australian dollars)
Function1988/89 Actuals1989/90Difference Between Budget and Forward EstimatesParameter and Other Estimates Variations1990/91 Effect of Policy Decisions
Forward estimatesBudget estimatesPolicy Decisions
Social security and welfare23,80224,77226,3341,5629828502705801,386-455931
Housing and community

Culture and recreation1,0781,0211,081604813-2112431053
Economic services
Transport and communication1,7231,3011,554253126130-312759-257
Industry assistance and development2,0311,9442,11417084129-24105142-14128
Labor and employment1,0801,1131,23111810742-321054-486
Other economic services32615421763402323100100
Total Economic Services5,1604,5125,115603338323-58285354-63291
General public service
Legislative services41830438682774444
Law, order, and public safety69860683022316360604747
Foreign affairs and overseas aid1,5511,4931,564716513-862424
General and scientific research523530582511537-1364343
Administrative services2,5742,9382,485-454-511105-4857231-14217
Total General Public Services5,7855,8725,84626-190220-5716434914335
Not allocated to function
Assistance to other governments13,60614,28813,978-31128225-618-59326-623-597
Public debt interest7,4037,3457,28035127-92-92-500-500
Contingency reserve700110-590-590-60-60
Total Not Allocated to Function21,00922,23321,368-865-18125-710-68526-1,183-1,157
Outlays Excluding Asset Sales82,66685,99587,9531,9581,6381,621-1,2993232,435-1,940495
Asset sales-538-1,100-1,200-100-100-100-800-800

The decision to publish the forward estimates and the increasing reliance upon them has greatly encouraged the improvement in their quality. Indeed, with the Government called upon to account for the extent of fiscal restraint against its previous projections of forward budget outlays, any blowout in those outlays would itself be very embarrassing. Nevertheless, the experience was that public exposure by itself did not lead to sufficient improvement in the quality of the estimates. While they remained departmental estimates, the dynamics of the budget process meant that they continued to reflect “bids” or even “wish lists.” In particular, the first year of each new set of estimates tended to display strong real growth and bore little relation to the previous forward estimates for that year even though all estimates were meant to represent a minimum cost of existing policies.

This led the Government to decide in 1986 that in future the Department of Finance would have responsibility for setting the estimates. This key change has resulted in a fundamental improvement in the quality of the forward estimates and their present status as a realistic and consistent statement of the baseline for the preparation of the budget. By no longer asking agencies for their own generous assessment of what they “need” to implement government policy, most of the creeping incrementalism that used to characterize successive budgets has disappeared. Indeed, as the New Zealand Treasury recently recognized, “The effect of (the bidding) system … was to create incentives for managers to demonstrate a need for more resources, rather than to manage those they have more effectively. This has contributed to the great increase, in the size of the public sector, simply by accretion.”2


Incorporation of the forward estimates as an integral component of the Australian budget has meant that they now do reflect the essential attributes of a rolling expenditure plan. The initial strategic planning, collation of spending plans and savings options from departments, and preparation of the now entrenched autumn (April or May) budget statement take place on the basis of three years of estimates. In May, when the most up-to-date and accurate figures are presented for the major budget deliberations by the Expenditure Review Committee of Cabinet, the cycle is advanced and a new third year is added. (During this phase, four years of estimates exist: the base estimates for the emerging budget and three forward years.) As the budget is produced in August of each year, the baseline for the next year is already established, publicized, and debated as part of the Government’s fiscal stance.

Estimates for all years undergo major reviews in January, when the midyear review of the current budget is carried out, and in May when the cycle is updated. Between these formal reviews, the estimates are kept up to date with amendments as necessary to take account of government decisions or major external shocks such as payouts for natural disasters or revaluations of capital works. The updating becomes intense as budget time approaches. Each forward estimate is therefore reviewed formally six times (there being no review prior to the first publication) and published four times up to and including its release as the budget estimate. After that again, the estimates are reviewed twice more during the budget year and finally, more than four years after their initial collection, they are reported and archived as the budget outcome, in which form they are reconciled to the budget estimates. A depiction of the estimates cycle is presented in Chart 1.

Chart 1.The Estimates Cycle

Initial generation of the estimates and their progressive updates are carried out by the Department of Finance in consultation with operational agencies. Estimates are generated on the basis of a standard set of economy-wide parameters, in the form of quarterly projections of such variables as activity and employment levels, wage and price movements, and exchange and interest rates. The economic parameters are estimated using an official short-term forecasting model operated by the Treasury Department working within the constraints of policy targets. The considerable judgmental input contained in the forecasts is subject to the consensus of a committee of central agencies.

Economic parameters for the forward years are simple projections grafted onto the forecast period. Because of that and the uncertainties attached to noneconomic program variables, the third-year outlay estimates can be quite uncertain and more in the nature of a consistent projection than a forecast.

With economic parameters stipulated and any program-specific parameters generally mutually agreed with the operating department that is likely to have relevant expert knowledge, the influence exercised by the Department of Finance is in the sphere of interpreting the intent of Cabinet decisions and settling the estimates for administrative and other discretionary components of resource requirements. Dispute rarely arises at the point of inclusion of the new out-year estimate. This is partly because of the historical unimportance of the forward estimates in influencing government decisions, but increasingly it is because the rules for updating the estimates are clear (even if not always appreciated), and where there might be some room for discretion, ministers are now pressed to determine future funding for the out-years when they establish or significantly change a program.

The significance of central control of the forward estimates is apparent from Chart 2, which shows the strong growth built into the first forward year in the 1983 and 1984 reports, with lower, but significant, positive growth recorded in the subsequent three reports. It is also apparent that the second and third forward years generally contained only moderate growth. The first-year figure was seen by departments as staking an ambitious claim at the beginning of the budget negotiations. As such, the usefulness of the forward estimate as a baseline for ministerial deliberations, or as a publishable medium-term view of the Government’s spending intentions, was greatly diminished. Since the change in responsibility for establishing the estimates, however, the trend in the estimates has become a faithful reflection of the fruits of expenditure restraint. Not only has this had desirable publicity effects for the Government, but it has also stretched the outlook of ministers beyond the budget year as the effects of savings that are really only deferrals, or of back-end loaded new policies, become only too apparent.

Chart 2.Budget and Forward Estimates of Outlays: Cumulative Real Growth (Excluding Asset Sales)

Source: Based on estimates published in the Forward Estimates Reports up TO 1988 and the 1989/90 Budget Statement 3.

The higher profile of the forward estimates has required a greater discipline on the Department of Finance also. Being a key player in the budgetary debate, it must be conceded that the department used also to take a strict minimal position in respect of departmental claims for resources and preferred to let ministers—hard-pressed to find savings—decide on so-called disagreed bids. The success of the present system depends to some extent on departments’ perceiving that the baseline estimates are fair, if firm. This involves such inclusions as technical new policy additions for minor public works, funds for reasonable replacement of minor capital equipment or renewal of leases, increases in rent, etc. It has also required the Department of Finance to establish clear rules and better records of past agreements and decisions to support the estimation of general operational costs, an area where dispute can take place out of proportion to the funding involved.


The level of staff that can be employed and the associated salaries and other administrative costs will always be close to the hearts of decision makers, both ministerial and bureaucratic. In conjunction with the absence of bottom-line pressures or other equally effective accountability mechanisms, this has led successive governments in Australia, as elsewhere, to embark on a variety of methods to contain, or to reduce, staff numbers and operational expenditures. Consequently, the end piece of hard-fought budget rounds has sometimes been arbitrary cuts, selective or across the board, in staff numbers and administrative costs.

The Australian Government in recent years has tackled this problem progressively, moving in 1987 to a comprehensive running costs system. This system is characterized by a number of radical concepts:

  • amalgamation of all running costs, salaries, and nonsalaries, into a single controlled amount with a single appropriation item for each agency;

  • virtual elimination of controls on staff numbers, except at the senior executive level where departmental structures are determined;

  • considerable flexibility for managers to allocate aggregate running costs, without reference to the Department of Finance, across programs, between salaries and other costs, and (within limits) between adjacent years;

  • strict cash limiting within the budget year;

  • determination of the forward estimates of running costs by a formula that essentially assumes their maintenance in real terms less an efficiency dividend to the budget of 11.25 percent per annum (in recognition of the scope for productivity gains), and adjustments are also usually allowed for changes in workload or new policies (above a predetermined threshold level);

  • allowance within net appropriations for retention of some or all revenues through charging for services or disposing of underutilized minor assets;

  • a system of loans from the budget to finance worthwhile new investments and other initiatives, with the loans plus a real return to be repaid by reducing the forward estimates of running costs for later years.

This running costs system represents a major reform in favor of devolved responsibility and managing for results. It has greatly diminished the need for detailed examination of the estimates, intrusion by the Department of Finance, and the scope for contention with departments in setting forward estimates of their running costs. Although on occasions challenges to the application of the operating rules of the system still occur, they rarely proceed to ministerial level or require Cabinet to adjudicate.

The system, however, no longer encourages managers to waste their time pursuing unwarranted increases in their estimates. Instead the Australian running costs system now provides managers with clear incentives to pursue improved efficiency and better management. First, all savings from better efficiency over and above the efficiency dividend are retained by departments’ management. Second, the introduction of user charging and the system of loans means that services to agencies and new investments are not rationed but are decided on their merits in terms of their ability to pay for themselves. Moreover, when a service is no longer provided free or a loan with interest has to be repaid, then there are clear pressures on agencies to reconsider how much they really want those services.

In addition to these incentives for better management, the flexibilities in the system also allow managers to manage better. The carryover facility means that there is no longer the same pressure for an end-of-the-year spending spree to avoid being left with unused funds, and that could then result in a lower base for next year’s appropriation. Similarly, the former system of expenditure control based on adherence to detailed appropriations for individual inputs (such as travel, postage, telephones, etc.) provided no flexibility to switch resources in response to changing priorities and circumstances. Without that flexibility, cash limiting was impossible. Instead there was frequent resort to additional estimates to provide the necessary flexibility, but this practice made Department of Finance officers, rather than line managers, responsible for financial management and further reduced the scope for forward expenditure planning as well as good management.

The significance of the running costs system to the topic at hand is that none of these improvements would be available without the forward estimates. For example, the return of savings from efficiency gains to agencies’ managers is meaningless unless their base estimates have been decided in advance and are reasonably firm. Similarly, the repayment of loans can only be effected by adjusting the forward estimates and that also provides a clear signal to the agencies’ managers of what is expected from them.

For the system to work as intended, however, the forward estimates must be stable and consequently of high quality. The necessary improvements in their integrity have been achieved by setting the forward estimates for running costs essentially by formula. Moreover, this improved quality and stability has greatly reduced the risk of ad hoc budget cuts to running costs and staff numbers, which has encouraged greater managerial responsibility and thus, in turn, further enhanced the integrity of the estimates.


Unlike some other budgeting systems, such as Canada’s Policy and Expenditure Management System (PEMS), the Australian budget estimates do not make allowance for prospective variations to budget figuring for decisions or other external influences. This apparent omission is part of the philosophy of budgeting adopted for many years: namely, that reviews of priorities, and spending and taxation decisions, are taken at one intensive annual review session and, barring unforeseen and unavoidable circumstances, the budget is set for the year. There is a process of additional, or supplementary, estimates late in the financial year to provide necessary additional appropriations generally for unavoidable “blowout” in the estimates, balanced to an extent by withdrawal of authority for known underspent appropriations. But this time is not meant for changing budget strategy. Of course, there is always leakage from this discipline, but the tight adherence to estimates in recent years (see Table 2) is testimony to the broad success of the approach. (In practice, the extra spending that does arise from intra-budget decisions is largely offset by the inevitable underspending from annual appropriations brought about by prudential management. In other words, the lack of a contingency reserve for new spending decisions during the course of the budget year counters the lack of an allowance for underspending utilized in some other countries.)

Table 2.Budget Estimates and Outcomes for Outlays, 1979/80–1988/89
Budget YearOutlaysChange
(in millions of Australian dollars)(percent)
Note: As the outcomes shown for each year reflect the classification treatment adopted for the following year’s budget, the outcome figures in this table should not be compared between years.
Note: As the outcomes shown for each year reflect the classification treatment adopted for the following year’s budget, the outcome figures in this table should not be compared between years.

The forward estimates similarly do not contain the reserves that characterize the public expenditure management system. As such, the estimates are not forecasts of expected levels of expenditure, but rather baseline estimates to or from which all future decisions will be added or subtracted.

But the forward estimates are also intended to provide a direct measure of the scale of the budgetary task and a baseline against which the public can assess the Government’s fiscal performance. It is therefore important that the growth in total outlays revealed by the estimates is not subject to any conservative bias. Understatement of that growth would not only mislead the Government in its future forward planning but its subsequent correction could also lead the community to conclude that the Government had exercised less fiscal restraint than had in fact happened. On the other hand, correction of the bias could encourage expectations of future expenditure that is not presently authorized. This is most likely when the estimates for individual programs are maintained beyond the period for which they are presently authorized. But even a nonattributed general contingency reserve can encourage bids by spending agencies in the expectation that the Government has made sufficient provision to finance those bids.

In seeking to balance these considerations the Australian forward estimates do contain a modest contingency reserve. This reserve includes provision for new technical policy additions referred to earlier, such as minor public works and the replacement of minor capital equipment. A notional allowance is also incorporated for the extension of some programs. However, these notional allowances are not attributed to individual programs, and the contingency reserve is tightly defined to exclude any allowances whereby the Government could be expected at least to consider terminating or reducing the program.

The contingency reserve therefore tends to increase progressively, being largest for the third out-year, and diminishes as the contingencies become actual and are incorporated into the appropriate estimates. In aggregate the present forward estimates show a contingency reserve of about 0.1 percent of total outlays in 1989/90 increasing to about 1.7 percent of total outlays in 1992/93.

The remaining influences on the forward estimates are changes in economic parameters (the impact of which can be broadly estimated using the published sensitivity factors), and government decisions to vary programs; no contingency reserve is provided for these.


Developments in the forward estimates process in Australia have underpinned reforms in resource management and have proved as important to program managers and planners as to budgeters.

As recently as 1983, a survey of senior managers found that 94 percent considered good financial management consisted of spending no more and no less than their budget allocation. Only 5 percent of them recognized their responsibility to consider improved resource allocation and to pursue more cost-effective results. The program management and budgeting system that has been introduced in conjunction with forward expenditure planning has revolutionized managers’ thinking about what constitutes good financial management.

Because the program and running costs estimates on which managers base their planning are established quite firmly in advance, there is a reasonable expectation that they will not be varied up or down unless the Government makes an explicit decision to change its policy. Managers now waste much less time haggling over the details of the estimates with the Department of Finance, but instead get on with their real job of improving performance. When the Government does decide to change policy—possibly in response to budgetary exigencies, but also to improve value for money—the program parameters will be specifically adjusted and there is no arbitrary squeezing of the resources available to achieve the expected objectives. Thus, the implementation of programs and the delivery of services is greatly improved, as managers are able to plan activities and to focus on achieving planned objectives—an essential requirement when managers are expected to manage for results. Moreover, the flexibilities that allow managers to manage by allocating resources across inputs and programs, within years and between years, depend on their knowing in advance, and with surety, to what base level of resources these flexibilities are being grafted. When program funds are cut, it is generally in response to an explicit change in the program’s objectives, so that the performance of managers is not assessed against unrealistic targets.

Devolution of managerial responsibility to program managers puts an onus on agencies to replicate within their own structures the relative independence that the forward estimates system has given to agencies. But that devolution requires, in turn, granting reliable forward estimates to agencies. Such devolution, in practice, is limited by the size of the units involved, since splintering managerial autonomy can cause the loss of administrative efficiencies and reduce the scope for an agency to reallocate priorities at the margin.

Greater reliability and integration of forward estimates into the budget have led to more accurate and considered costing of effects of future years by the proponents of spending or saving options. Ministers have therefore been less inclined to resort to short-term expedients, such as arbitrary cuts in agencies’ allocations or deferrals, to meet their savings requirements. Instead, the requirement to consider the next three years’ estimates encourages them to achieve expenditure reductions through considered changes to policy.


The present Australian Government has been greatly assisted in its pursuit of expenditure restraint by its readiness to submit itself to the discipline of an overall fiscal target or targets. But over the years the exact expression of these targets has varied. In its early years the Government committed itself to a so-called trilogy which set, as an upper limit, that outlays, revenues, and the budget deficit should not increase over a three-year horizon as a proportion of gross domestic product (GDP). More recently, this has been replaced by a much tighter target involving real reductions in outlays and target levels for total public sector debt repayment (a negative public sector borrowing requirement). The gap between the forward estimates and the target level provides a ready measure of the Government’s task in its expenditure deliberations.

The Government’s performance in closing this budgetary gap has been routinely revealed through the publication of the reconciliation table referred to earlier that compares the forward and budget estimates. That table (reproduced here as Table 1), by identifying the role of parameter movements and other external influences separately from the impact of government decisions, shows the extent to which the Government has held new spending in check, or alternatively, has aggravated the task for the present and future years by agreeing to new expenditures. Moreover, the impact of these decisions on each functional area is presented explicitly so that the winners and losers from budget decisions to reallocate resources can also be clearly identified.

The strategic or “top-down” approach to setting an overall fiscal target or targets has been complemented by a refinement of the more traditional “bottom-up” consideration of individual claims. Targets for each portfolio are also set on the basis of some consideration of overall priorities, new policy bids, and the scope for savings, and these targets are constrained to be consistent with the overall fiscal target. Portfolios are then asked to prepare options to meet their specific targets. This might involve net reductions or net increases to their forward estimates, but in either case portfolio ministers are able to increase their scope for new initiatives by finding more offsetting savings. Ministers and their senior executives are thus constrained to consider priorities and opportunity costs closely, and overall resource allocation is improved.

Since running costs are restrained to their forward estimate in aggregate terms only, there is also incentive to find savings within a portfolio’s running costs, through efficiency measures, charging users for services provided, or retaining the proceeds from disposal of (minor) underutilized assets. The running costs and portfolio targets are interdependent, and it has become commonplace in some administratively complex programs to assign increased running costs, as new policy spending, to achieve greater program savings. Fraud and overpayment control in welfare benefits, or the collection of defaulted child maintenance payments are cases in point. This thrust has supported the Government’s objective of improving targeting of budgetary assistance to the genuinely needy. But it has also created a significant offsetting factor to the restraint imposed on running costs by the efficiency dividend and other efficiency initiatives.

As the forward estimates have become more soundly based, it has become possible to set portfolio targets over the forward period. This has been done by imposing explicit dollar limits on new policy or savings packages to apply to all years, or by placing qualitative or quantitative constraints on full-year effects of proposals. It would not have been possible under earlier arrangements, because portfolios could generate “false savings” and increase their access to new policy funds by trading off their inflated bids for increases in base estimates.


The forward estimates, as operating in the Australian environment, attempt to serve two purposes: to establish a baseline for departments to prepare their budgetary proposals and to plan program implementation; and to present a realistic forecast of expenditure levels under existing policies. This dual purpose can lead to conflicting requirements.

A major difficulty in preparing budget estimates in the past has been the lack of precision in policy pronouncements. Governments have been exhorted to specify the implications of their policy intentions for the forward years to facilitate planning, but that tends to increase the expectations of client groups and departments to a committed level of expenditure.

Similarly, the desirable control practice of “sunsetting” programs and making their continuation conditional on satisfactory evaluation, leaves a “hole” in the estimates beyond the terminating year, which is a major source of conservative bias. To the extent that “neutral” assumptions, such as continuation of the program at the same real or nominal level, are made to maintain integrity of the estimates, this will reduce fiscal flexibility at the macro level, and can even weaken fiscal discipline by raising the political cost of shifting priorities through any decision to terminate or to curtail the program. The same is true of other forms of new technical policy, such as funds to continue capital programs, or allowance for expected price and wage increases or currency devaluations. Certainty in funding conflicts with the pressure that constrained resources can impose to resist cost increases.

As noted earlier, however, in practice forward commitment of outlays was already extensive before comprehensive and controlled forward estimates were introduced. Some 70 percent of outlays is covered by standing legislation. Even for those programs that do not have continuing legislative authority, ministers tend to make political statements that commit the Government to maintain broad present levels of expenditure. Although the enhanced status of forward expenditure planning may have raised expectations about commitment to some programs, therefore, and thus resulted in some loss of fiscal flexibility, this has been more than balanced by the improved expenditure planning.

As long as the Government maintains tight control of its forward expenditure planning the experience has been that there is less blowout in the estimates from one year to the next, with less need for the major savings exercises that were a feature of past budgets, and the loss of fiscal flexibility is consequently less serious. Additionally, by encouraging ministers to consider the out-year consequences of each decision, the net reductions agreed for the budget estimates have been matched by net reductions in the forward estimates for subsequent years. As a result, recent experience has been that the forward estimates that have formed the starting point for a new budget have either been falling in real terms or held to almost no growth (see Chart 2), and that also has reduced the pressure to maintain fiscal flexibility to allow for further savings.

Clearly, the success of the Australian system of forward expenditure planning is sensitively dependent upon the quality of the estimates. The trade-off between the sacrifice of fiscal flexibility for better expenditure planning will probably only remain acceptable while the integrity of the estimates is maintained. If, however, a major blowout occurred in those estimates, with consequential pressure for a major savings exercise to restore the estimates, the degree of “lock-in” implicit in the present system might well come into question.

Perhaps the major risk at present for the quality of the forward estimates is their dependence on forecasts and projections of economic parameters over a three-year horizon. Economic forecasting is always hazardous. In fact, official forecasts in Australia look only six quarters ahead, and the projections beyond that really only reflect a feasible and desirable economic scenario based on consistent assumptions.

Changes in the forecasts of economic parameters can however have perverse and offsetting effects on the forward estimates and it is these offsets that have probably helped to sustain the present approach to forward expenditure planning. For example, higher inflation tends to reduce the measured real growth in expenditure because the deflator rises immediately while the (adverse) impact on expenditure is lagged by indexation relationships.

Thus the typically upward corrections to the forecasts of inflation have not tended to pose real difficulties for the Government in meeting its expenditure targets, which are expressed in real growth terms, and indeed have even made their achievement easier. The essence of the problem in this case is, of course, not the setting of correct parameters, but an honest acknowledgment of the underlying problem for which the appropriate policy response could well involve tightening the original real growth target. Somewhat surprisingly, the converse is readily recognized, and an increase in the projected real growth in outlays in response to declining inflation is readily explained away.


Fundamentally, the real test of the success of Australia’s system of forward expenditure planning will be the results achieved. Judged by the imperative for fiscal restraint, the system has been associated with a remarkable turnaround in the budget over the last six years. And it has been achieved to a significant extent by correcting the structural imbalance. The budget deficit has been converted from a peak level of 4.2 percent of GDP in 1983/84 to a surplus estimated at 2.5 percent of GDP in 1989/90. This has been achieved while the proportion of revenue has risen by less than one percentage point. Thus almost all the burden of restraint has fallen on outlays that have been constrained to fall in real terms for four consecutive years and by 6.3 percentage points of GDP since 1985/86 (see Chart 3).

Chart 3.Outlays as a Proportion of GDP and Real Growth, 1979/80–1992/93

As the forward estimates are transformed into budget estimates during the budget process, the measure of real growth in the estimates is affected by decisions and other influences, the relative contributions of which for recent years are shown in Table 3. Since 1985/86 the relatively low growth in the forward estimates has been reduced further through the budget process. Ministers’ decisions provided the major contribution in the years from 1985/86 to 1987/88, but, in the last two years, there was a somewhat greater reliance on fortuitous improvements in the budgeting environment from other influences, such as movements in economic parameters. Pent-up pressures for the introduction of policies resulted in a significant shift in the balance between new policy spending and savings, as depicted in Chart 4. It shows that, after being held down to 0.8 percent of total outlays in 1987/88, new policy spending rose to 1.2 percent in 1988/89 and 1.9 percent in 1989/90.3 Similarly, the savings effort has declined from 5.3 percent of total outlays in 1987/88 to 1.9 percent in 1988/89 and 1.5 percent in 1989/90.

Chart 4.Savings and New Policy,1983/84–1989/90

Table 3.Real Growth in Budget and Forward Estimates of Outlays
Real Growth


Contribution from
(percent)(percentage points)

Forward planning for savings has arguably led to better decision making by providing more time and more information to identify future problems and pressures. The reductions made in outlays through ministerial decisions are therefore of higher quality, reflecting genuine shifts in policy priorities and improved targeting. This enhanced value for money is most clearly apparent in the area of social security and welfare. The growth rate in this function has been restrained to only 2.2 percent in real terms since 1985/86, compared with about 24.5 percent in the preceding four-year period. Simultaneously, however, funding has been redirected to high-priority programs such as support for low-income families. Major improvements in real per capita benefits have been achieved while the target population has been narrowed to the most needy.

These “quality cuts” have helped maintain the lower growth rates into the out-years, which is apparent in Chart 2. Since 1986, little or no real growth has been built into the forward estimates. The focus on the out-years through forward expenditure planning has been vital to achieving this low growth and thus easing future budgetary tasks. In addition, ministerial cooperation has been enhanced in what are inevitably difficult decisions by the move to top-down budgeting where ministers are asked to meet individual outlay targets in various forms. This at least gives them the incentive that genuine savings found within a portfolio can be allocated according to the portfolio’s priorities with relatively little central scrutiny of individual spending proposals. The establishment of reliable forward estimates has been a necessary precondition for meaningful outlay targets.


When the present Australian Government took office in 1983, it was faced with the need for expenditure restraint. At the same time it wished to achieve a significant reallocation of priorities and to improve public management. It was recognized that this required reforms of the management environment, including the budget process, and that financial management and accountability had to be given a higher prominence. It is doubtful whether either could have occurred without the other. The changes affected relationships within agencies, between central and operating agencies, and between those agencies and ministers.

Publication of the forward estimates was a key early initiative. It both improved public scrutiny and consequently their status. This then led the Government to shift authority for the determination of the forward estimates solely to the Department of Finance, and thus to establish them as the base, and not the subject matter, of the bidding and priority-setting processes. Within this framework, which provided for medium-term planning and overall control of program expenditures, significant devolution was possible. The improved forward estimates have therefore been a major contributing factor to the progress that the Government has been able to make in its endeavors to improve both fiscal policy and financial management.

The forward estimates have become an important aid for ministers, bureaucrats, and the public to make decisions and judgments about appropriate levels and directions of public expenditure. Their success in achieving restraint depends most importantly on the political environment, but they are part of the process of establishing a conducive environment.

Looking to the future, experience in all countries suggests that the one certainty is that the processes for formulation of the budget can be expected to change. This change is necessary if these processes are to be responsive to changes in the internal dynamics of government, which in turn reflect changing political and economic circumstances. Nevertheless, it is difficult to imagine that future Australian Governments will draw back from the present reliance on forward expenditure planning. Fiscal credibility requires the continued publication of forward estimates and associated analysis and the opportunities that provide for the public to scrutinize the expenditure implications of all government policies and new decisions. Thus, even if political priorities do shift from expenditure restraint, the benefits of the reform of the forward estimates and associated changes in the budget and resource management processes should remain through improved decision making and result-oriented management.

How forward expenditure planning in Australia might be further enhanced is more difficult to foresee. No doubt at the level of some individual programs, greater use could be made of the framework which has now been put in place. Certainly there is room for further improvement in Australia in the area of program evaluation, and the forward estimates provide an important information base in this context. At the macro level the potential of medium-term expenditure planning for reassessing overall priorities has probably still not been fully realized, but this area is difficult. Practical constraints affecting individual programs are the reason why so much of government decision making is done from the bottom up and incrementally. Nevertheless, by planning expenditure over the medium term, there is somewhat greater flexibility and therefore more scope to review priorities. The use of forward estimates for developing a strategic overview of the key issues for resource allocation will therefore be a major challenge for Australian budgeters over the next few years.

History of Reports of Forward Estimates
Initial Forward Estimates
1965—Informal projections of current estimates collected from department by Treasury
October 1971—New guidelines for collection of forward estimates that were to include costing of any new proposals that ministers firmly intended to put forward in the next budget
TitleIssue DateForward Estimate YearsComment
Unpublished Reports (1973/74–1982/83)
Forward Estimates

of Australian


November 19731974/75–1975/76Cabinet Committee

established to consider

forward estimates. Met

only once.
June 1974

April 1975

Cabinet agreed to upgrade

the estimates and to

integrate them into the

annual budget decision-

making process.
Treasury Report on

Forward Estimates

of Budget Outlays
March 19761976/77Cabinet decision that

officials from the

Department of Prime

Minister and Cabinet

should carry out a review

of the forward estimates

for 1976/77 to identify

possible options for

reducing outlays.
Review of Forward

Estimates of Budget

Outlays for 1977/78
April 19771977/78Same as above.
Report on the

Forward Estimates

of Budget Outlays
February 19781978/79–1980/81
February 19791979/80–1981/82Forward estimates

confined to the costs of

maintaining approved

ongoing programs and

would exclude new

February 1980

February 1981

February 1982


Published Reports

Report on the

Forward Estimates

of Budget Outlays
March 19831983/84–1985/86
March 19841984/85–1986/87
Report on the

Forward Estimates

of Budget Outlays

and Staffing
May 19851985/86–1987/88
November 19861986/87–1987/88Previous reports produced

in autumn and served as a

scene-setting document

for the budget. The second

report in November 1986

showed the out-year

implications of policies

contained in the August budget. Only two forward

estimate years published

as the budget was

intended to be year one.
Report on the

Forward Estimates

of Budget Outlays
December 19861987/88–1989/90
December 19871988/89–1990/91
Report on the

Forward Estimates

of Budget Outlays
November 19881989/90–1991/92First out-year published in

Budget Paper 1

(Statement 3).
Budget Paper 1

(Statement 3)
August 19891990/91–1992/93Published as part of

Budget Statement 3.

In 1974 and 1975, the forward estimates spanned only two years and, in the subsequent two years, reverted to single-year estimates. Three-year estimates have been the norm since 1978. A tabular history of the collection of forward estimates of outlays by the Australian Government is presented in the appendix.

“Government Management, Brief to the Incoming Government, 1987,” New Zealand Treasury (1987), p. 59.

Some 0.6 percent of the 1.9 percent increase in spending on new policies in 1989/90 was for improvement in family assistance and was a preferred alternative to further tax cuts negotiated as part of a wage-tax deal with the unions.

    Other Resources Citing This Publication