The Coalition aims to reduce the size of the federal bureaucracy by eliminating the need for federal public servants to oversee the implementation of federally funded programs administered by state and territory governments. As I discussed in the Informant last month, most federal funds allocated to the states come with an ever-increasing level of performance requirements, set either directly by the federal government or in conjunction with the Council of Australia Governments through the COAG reform council. Compliance with this accountability structure is not only increasingly burdensome to the states and territories but also generates extensive work for the federal government departments that oversee it.

Robb's proposal is to remove as many of the reporting requirements as possible. They have become an unnecessary paper trail, intended simply ''to cover backsides''. Instead, the Commonwealth will confine itself to setting standards and granting funds, trusting others to observe these standards in their administration of federally funded programs. In effect, Robb says, the provision of Commonwealth-funded services will be outsourced to state and territory bureaucracies, without any need for the federal bureaucracy to become involved.

The Coalition has already flagged one example of this approach: its policy towards the environmental approval of major projects, such as mines and ports. As part of its campaign against ''green tape'', it intends to replace the current, cumbersome two-stage approval process conducted by both state and federal governments. Instead, a single assessment process will be entrusted to state governments. The Coalition has stopped short of agreeing fully with Queensland Premier Campbell Newman, who wants his government to have sole responsibility for environmental decisions. The federal government will still set the required legal standards but the administration of these standards will be entrusted to the state bureaucracy.

Strictly speaking, the notion of outsourcing the administration of particular federal policies to state and territory bureaucracies is not practicable. State and territory bureaucrats serve their own state and territory ministers. Constitutionally, they cannot also answer directly to federal ministers, even if the policies they are following and the standards they are applying originate with the federal government. If state public servants are to implement Commonwealth policies, they will need to do so under instructions from their own ministers and parliaments. In the proposed model of environmental assessment, for example, the Commonwealth standards would need to be incorporated into state legislation, with the agreement of the individual state government. State public servants would then implement these standards under the direction of their own ministers.

In theory, such a division of responsibilities between the Commonwealth and states and territories could work well, yielding substantial savings by removing bureaucratic duplication. But it would depend on a rare degree of shared values and goodwill between the two levels of government. Over highly contentious issues, such as environmental assessment, one can foresee immediate problems. State and federal ministers could well disagree over how the legislated standards are to be interpreted and applied. Could state public servants be expected to take their lead from Canberra, against the explicit wishes of their own ministers? More probably, they would follow their own ministers, even at the cost of making federal ministers look powerless.

Again, who would take formal responsibility for decisions? When federal ministers came under attack for decisions effectively made by state governments, how would they respond? Should they try to defend a decision as their own, even if made against their own better judgment? Or should they admit that the state government is responsible and wash their hands of the matter? If the federal government wishes to retain overall responsibility for policy and ultimate oversight, as Robb insists, outsourcing implementation to state and territory governments is an extremely risky and flawed strategy.

Any form of outsourcing, if it is to work successfully, must provide reasonable guarantees that the agent to whom the task is outsourced will follow the wishes of the outsourcing principal. This congruence of aims can be secured in various ways, ranging from strictly specified contracts that spell out all the requirements in detail, to more open-ended relational contracts that rely more on shared values and trust within a formal agreement to cooperate on certain terms.

The first approach, detailed specification, is predominant in the current management of Commonwealth-funded programs administered by the states and territories. It leaves little room for the states and territories to diverge from federal policies, but at a very high compliance cost. The Coalition prefers the second approach, that of trusting the states and territories to make their own decisions on the basis of their sharing the Commonwealth's values and policy vision. However, history suggests that such trust will often be misplaced. State and territory governments are just as likely to reject Canberra's vision as to embrace it. When conflict arises, the federal government will find itself responsible for decisions over which it has little control. Will ministers be able to admit that they cannot guarantee implementation of their own policies? Or, more likely, will they be driven to re-impose detailed instructions on the states and territories, with all the inevitable compliance costs?

The main problem with the Coalition's policy of outsourcing programs to states and territories is not that it is too radical, as Robb seems to fear, but that it is not radical enough. If the Coalition really wishes to slash the federal bureaucracy's oversight of federally funded policies, it should abandon Commonwealth responsibility for overall policy direction.

In this case, it should avoid the language of ''outsourcing'', which implies retention of overall responsibility and control. Instead, it should talk more of ''devolution'': that is, a genuine transfer of responsibility. In areas such as schools and hospitals, the Commonwealth could restrict itself to allocating funds but leave all other decisions to state and territory governments. This would mean abandoning any ambitious national policies and targets, other than those that may arise through mutual cooperation. All questions and debate on outcomes, other than those relating to overall funding allocations, would be referred to state and territory ministers.

The Coalition has an historic opportunity to rewrite the federal compact between the Commonwealth and the states and territories. In support, it can point to Labor's spectacular failure to impose a coherent national health policy and its overambitious schools policy, which is also set to founder on state intransigence. Federal Labor, which sees health and education as its strongest suits electorally, has little choice but to plough on with undeliverable national policies in these areas. But the Coalition could rediscover its federal principles and advocate a major shift in power, abandoning ultimate responsibility for the direction of policy in health, education and elsewhere. Unless it does so, any major, long-term reductions in the relevant federal bureaucracies will be hard to sustain against the relentless demands of ministerial responsibility and electoral competition.

State and territory governments are not the only planned beneficiaries of the Coalition's attack on bureaucratic compliance. Frontline managers are also to be given relief from their onerous reporting burden. In Robb's words, we must ''restore a culture of personal responsibility'' and ''put the onus on management''. Presumably, he has in mind managers such as school principals, vice-chancellors, hospital administrators and nursing home managers. In place of laborious reporting requirements, managers should be subject to periodic inspections from expert assessors, who will be able to weed out the odd ''bad apple'', while letting the conscientious majority get on with their work.

Seasoned public servants will be suspicious of giving frontline managers too much discretion. They know how easily well-intentioned funding schemes can be rorted by unscrupulous people. They also know how unforgiving the public can be of the occasional misuse of taxpayer funds, particularly when egged on by the government's political opponents and the media.

On the other hand, there is always room to reassess the risks associated with administrative discretion. Many honest professionals employed from public funds would welcome relief from excessive and time-consuming reporting. The balance between accountability and trust is never easy to set. But, as with the federalism balance, the key questions relate to public expectations and political culture.

Would a future Coalition government be willing to re-educate the public into accepting occasional lapses as a price worth paying for increased trust? Or would it allow itself to be hounded into re-imposing controls in the face of a public opinion intolerant of any failure?

Richard Mulgan is an emeritus professor with the Crawford School of Public Policy at the Australian National University.

richard.mulgan@anu.edu.au