The theory might be wrong, at least as far as welfare fraud is concerned. All of the evidence suggests that false claims on expenses by politicians and other forms of middle class fraud by government and employers is of higher incidence than welfare fraud but, even allowing the very high proportion of Australians receiving one form or another of welfare benefits, almost all of the concern and almost all of the detective energy is going on welfare fraud.

A recent British report on social security fraud might provide an interesting backdrop. The British government hands out about $300 billion in welfare benefits a year. By its best calculations, including close random audits, there are problems with about $9 billion of these payments, although mistakes by officials seem to be responsible for most of the problems. About $2.8 billion will have been underpaid to clients, because of mistaken assessments of eligibility by the British equivalents of Centrelink staff, other federal, state, federal, municipal or private sector contractors, or perhaps misunderstood information supplied by the claimant. About $6 billion will have been overpaid to claimants. The reason for the overpayments will involve mistakes made by officials ($1.2 billion), innocent mistakes made by "clients" ($2.8 billion) and deliberate fraud by claimants (about $2 billion).

Mistakes by officials are thought to be responsible for about 0.4 per cent of all welfare expenditure; innocent mistakes by "clients'' about 0.9 per cent and by deliberate fraud about 0.7 per cent of all expenditure. If a wrong payment is identified, in short, fraud will be involved in about one in every five cases.

These figures seem comparable with Australia's. Australia's social security and welfare bill is about half of Britain's and, if anything, the rate of mispayment could be expected to be a bit lower, given that Australia's present system of doling out money through Centrelink is (even if in a state of dismantlement) intrinsically more centralised, integrated with tax and other systems, and more focused than the British system. To that, no doubt, some might point to Australia's origins as a dumping ground of Britain's bludgers, skivers and crooks but, my own ancestors apart, the composition of the Australian population is now much changed and, in any event, it must be remembered that Great Britain is where all those awful convicts came from.

The average "fraud" on Centrelink seems to be about $1000. The war, and internal security, fraud control and quality assurance programs lead to about 4 million reviews of payments to individual clients every year. The cost of this fraud and quality control is about half of the money saved by reviews. By contrast, investigation of fraud by the Tax Office is sometimes said to produce a $7.50 return for every dollar spent on compliance review, although, no doubt, that is a figure that will soon be up in the air with the implementation of government directions to reduce taxation staff by more than 10 per cent. Even before those staff cuts, the Tax Office was reviewing only about 2 million files a year, although its "client" population is at least twice that of Centrelink and other welfare benefit services.

But increasing the resources going into the accurate reporting of income, or deductible expenditure, has political costs that hardly ever arise with welfare fraud, particularly if the sort of welfare fraud that is envisaged involves unemployed beneficiaries with undeclared sources of income, supposedly single mothers with undeclared partners, or invalid pensioners thought to be more fit and able than they claim to be. Dob-in rates from members of the public are high, as are investigations spurred by suspicions of Centrelink staff. In fact, about half of those "caught" catch themselves, because data holding on them elsewhere in the Commonwealth (including at the Tax Office) is at variance with information they have volunteered to Centrelink.

No one is suggesting that conscious and deliberate fraud on the welfare system should not be detected and punished. But the attention given to the problem, and the media and political spotlight on fraud, cheating, bludging and a bloated sense of entitlement, is out of all proportion to government resources focused on other types of fraud, or checks and balances against abuse of various systems.

A paper by Greg Marston, of the social policy unit at the University of Queensland, a few years ago pointed to a steadily increasing rate of prosecutions for social security fraud, while the number of prosecutions for fraudulent tax claims had remained steady over more than a decade at about 250 a year. In 2002, there were about 24 times as many Australians being charged with making a fraudulent claim on social security as every person charged with tax fraud. Two years later, 80 per cent of all criminal cases being prosecuted by the Commonwealth Director of Public Prosecutions involved social security fraud. Very few of the fraud prosecutions are over tax and other concessional payments going to ordinary working Australians, although there is little evidence that the fraud or mistake rate would be lower than with benefits dealing with unemployment, invalidity or disability.

Marston quoted, approvingly, a comment by ANU professor John Braithwaite that "the DPP is serving Australia very poorly with respect to serious white-collar crime ... it is tough on the fraud of the poor and very soft on the fraud of the rich''.

The tough talk about "cracking down" on welfare fraudsters, bludgers and scroungers, a constant dialogue coming from the tabloid and infotainment media that would seem to suggest that many folk on welfare are cheats, and talk by politicians of cutting relief for extended periods to young unemployed, and making others work for the dole, sometimes seems to divide Australians into whether they are "contributors" or whether they are dependents on others. No doubt this perception helps those such as Treasurer Joe Hockey in calling for an end to the culture of entitlement. Yet no one knows better than Hockey that the real problem with a culture of entitlement – if there is one – is not that too many of the underclass, the young and the aged are receiving welfare benefits but that too many of the middle classes are getting it. Perhaps a tiny percentage of the very poor are getting fractionally more than their legal entitlement but, with or without it, they are nearly all poor. By contrast, the charge about cutting back the reach and cost of welfare is about weaning working people, middle class people and others from reliance on government. And even assuming that such "salt of the earth'' people (once characterised as Howard's battlers) may be intrinsically more honest than the workshy scroungers (a proposition not supported by evidence of tax evasion, or expenses fiddling), the extra entitlement exceeds by a big amount all of the fraud at the bottom.

In much the same manner that some politicians are deterred from the more blatant fraud on their generous expense entitlements only by the very slight chance of exposure, Joe Hockey's campaign might be better served by demonising the welfare dependency of "ordinary'' and typical Australians. The problem is that all too often it is just the Commonwealth benefits that have made them dependent on the public purse that were, as often as not, given to them by previous Liberal governments.