Two years ago, in February 1993, Michael Portillo, then Chief Secretary to the Treasury, announced that the Government was out to do just that. He announced a series of "fundamental expenditure reviews" which would take a full parliamentary term and more to complete and which would examine every spending programme to determine "whether its purpose remains right for the 1990s".
"We shall," he said, "be seeking to identify areas where better targeting can be achieved, or from which the public sector can withdraw altogether" (my italics).
The first departments to be reviewed were, in the main, the big spenders - health, education, social security and the Home Office: the core of the welfare state. For a government wanting to cut spending, this was the right thing to do. Withdrawing altogether means real cuts, not cuts on plans. It means taking things away from people or making them spend more on services themselves. The political cycle dictates this can be done only early in a parliament. Doing it late carries a huge political price - as Margaret Thatcher discovered with the poll tax.
Two years on, what has this unprecedented review of government expenditure produced? As the graphics show, not a lot.
The biggest savings have in fact come from defence - not formally part of the reviews but already subject to "Options for Change" and the Defence Costs Study when Portillo's review was getting under way. Cash provision for defence is falling from £26bn in 1991/92 to an estimated £20bn by 1997/98. With the end of the Cold War, there have been parts of defence from which the state has withdrawn.
In health, education and the Home Office, no such withdrawal has taken place. Only in social security have marked changes been made. And huge though the figures are - totalling towards £4bn over the next three years - they still amount to only around 1.5 per cent of total social security spending during that period.
In addition, the effect of these changes has yet to be felt. Between now and the next election, MPs' postbags will fill with howls of protest from those the Government has excluded, and lower unemployment figures, about which the Government has been boasting, will be inflated by more than 200,000 as a result of these changes.
This does not mean the reviews have been a waste of time. They have heightened the search for economy in government and given a new edge to measures already in motion - the drive for market testing, for example, and the search for private capital in place of public capital for roads, hospitals, universities and much else. Most spectacularly they have accelerated the Government's determination to cut its own administrative costs. A thousand jobs are going from the Department of Health, 600 from the Treasury, many hundreds more elsewhere. A billion pounds is set to be sliced off central government's administrative costs over the next three years.
But the central conundrum remains. The reviews may have constrained government spending, but they have not fundamentally changed it. And the smaller departments now being put through the review mill cannot, by definition, produce the big savings that have failed to come from the big spenders.
The state - the perpetual search for efficiency aside - appears to have reached its irreducible minimum unless something really dramatic and new happens. The point is underlined by looking at the pattern of government spending since 1979.
Since 1979, the government has successfully shrunk spending on trade and industry by privatising the nationalised industries. After a sharp rise to 11.4 per cent of all government expenditure in 1984, defence has been curtailed. Government has also - despite the current size of the deficit - reduced interest charges on debt repayment. The only other sector to have been reduced significantly is housing - and there the figures hide a transfer of housing benefit to the social security budget which means the saving is appreciably smaller than it appears.
What has gone up? Education has stayed broadly level. Health has risen, chiefly under demographic pressures, and social security has soared - the result of the same demographic pressures that have affected health (more elderly), plus more benefit-dependent lone parents, the housing benefit transfer, and much more unemployment. In 1979, health, education, social security and housing took just under 53 per cent of all central and local government expenditure. In the year just ended, they took a fraction under 60 per cent.
Spending on law and order has also risen, and some would argue that the record numbers of police and prisoners are a function at least in part of higher unemployment. Add in training, which has stayed static, and the welfare state now takes two-thirds of all government spending against rather more than half in 1979.
In other words - and despite all perceptions to the contrary - government spending has increasingly concentrated on the welfare state since 1979. Which means that if ministers are really serious about reducing spending, they will have to cut, and not just constrain, the welfare state.
This is, of course, the agenda of Michael Portillo, and, less noisily, of Michael Howard, John Redwood and Peter Lilley, though the latter is the only cabinet right-winger so far to deliver.
How might it be done? The means are already well known. Vouchers for all forms of education which, for the slightly better off, would cover less than the whole cost so that parents have to top up. Loans for student tuition as well as maintenance. Widespread charges for health care. A substitution of private insurance for tax spending in health - the route the Government considered during the 1988 NHS review and rejected on the grounds it would push up costs and save little. Privatisation of the basic state pension, rather than merely encouraging people to opt out of Serps, the state-run second pension. Enforcement of private cover for much more than mortgage interest payments - for industrial injuries, unemployment and invalidity benefit, for example.
There are just two problems. Virtually everything on that list was considered in the fundamental expenditure reviews - and rejected on grounds that ranged from the practical to the political. And those who believe that such solutions should have been adopted face, in Kenneth Clarke, a Chancellor who has dismissed hospital charges as "corny old rubbish" and who has declared of the basic state pension, "I do not see that as being opted out of". In his first speech at the Treasury, he announced that he did not come into the job "in order to dismantle the welfare state".
In addition, they face an Education Secretary who described as "rubbish" any suggestion that the Government would adopt a recent proposal for vouchers to be used to end free compulsory schooling, and various other cabinet ministers who are no keener on such ideas.
These figures and those positions, however, define the divide that the Conservative Party will face after the next election, whether John Major wins or loses. And they define, too, the problem Labour faces if it fails to crack high unemployment and still wants to be a tax-cutting party. Sustained employment-rich growth would heavily reduce the problem, scaling down the £13bn that social security currently spends on the jobless, and raising revenues that would in turn allow individual tax cuts. But without that, those who really want dramatically to cut taxes will have to cut the welfare state and shift the burden on to individuals. Any votes in that?
What have the reviews
No withdrawal by the Government. Twenty programmes went through the mill. Action already in the programme, such as moves to cut NHS drugs bill, were re-badged as part of the review. The signal that no drastic changes were planned - such as charges for hospital stays or GP visits - came with prescriptions rising as usual by 50p last year, with no change in exemptions, for example by charging better-off pensioners.
Next to no withdrawal. Only change Department for Education can point to is acceleration of student loans so that they match student grants 50/50 from 1997/8 rather than the year 2000.
No withdrawal. Review "did not result in any major savings or in any decision to drop or significantly reduce the functions being carried out," Sir Clive Whitmore, the permanent secretary, told the home affairs select committee.
Real changes. Commencing last week, new incapacity benefit will withdraw invalidity payments to 200,000 claimants by end of 1996, cutting out an extra 55,000 claimants a year. Saving £410m this year, estimated £2.3bn a year by 2000. Next year, Jobseeker's Allowance will in effect cut entitlement to non-means tested unemployment benefit from 12 months to six, removing 100,000 claimants from books. Saving £270m first year, and £410m in second. Both remove, for the first time in 45 years, entitlement to national insurance benefits to which people have contributed. In addition, pension age for women to be equalised upwards to 65, saving £3bn a year sometime next century, and new home buyers will from October have to rely on private insurance to cover first nine months of mortgage repayments.