Last week, the UK Treasury sparked controversy when it published examples of customised 'annual statements' being sent to taxpayers to show them how their income tax and National Insurance Contributions are spent.
Such statements could be useful in giving taxpayers a clearer view of which parts of public spending are the big ticket items, and which relative small change.
But in this case the statements are designed to draw the eye to a large orange quadrant labelled 'welfare'. As state pensions had been separated out, the impression left was that quarter of our taxes go on what is now the deeply stigmatised 'welfare' in the US sense - cash hand-outs to people who are of working age but not in work.
That picture would ring bells with many members of the British public. When the social security budget was described to them - covering state pensions, child benefits, tax credits for those in work, benefits for unemployed and disabled people - half of people said they thought that 40% or more of spending went on the unemployed.
In fact, it's a tenth of that - four per cent. Even if all benefits and tax credits going to working age people who are not working - sick and disabled and lone parents as well as unemployed - it's under a fifth of all tax credits and social security.
And social security and tax credits are less than half of all spending on the welfare state including the NHS, education, social care and housing. Most of the debate on 'welfare' is around what is, in fact, less than £40billion out of the £489billion we spend on the welfare state - £1 in every £12.50 we spend.
With people believing that unemployment benefits are such a large part of what we spend on social security, their other beliefs about them cloud attitudes to the welfare state as a whole. These include the widespread belief that a substantial minority of claimants are doing so fraudulently, including presumably actually working for 'cash in hand', but still claiming.
The Department of Work and Pensions' own estimate across the whole of the benefit system (including pensions), based on its detailed examinations of a random sample, is that a total of £3.5billion was overpaid in 2012-13 as a result of combined fraud and error by administrators or by claimants. These are large amounts, but they are from budgets that are very large indeed. And two-thirds of this total for 'fraud and error' - often quoted as if it measured fraud alone - was said by the DWP to be due to error. Just 0.7% of all benefits was over-paid as the result of fraud, less than the amount underpaid as a result of official error. For the main benefit for unemployed people, Jobseeker's Allowance, estimated fraud was 2.9%, or an annual total of £150million.
But when people were asked what proportion of total benefits and tax credits were claimed fraudulently, half thought that this was 20% or more, and the average amount was 27%, equating to £58billion. This is nearly 50 times the grossed-up result from the DWP's probes into random cases.
A quarter thought that 40% or more of benefits were being claimed fraudulently. This would require virtually all non-pensioner claims to be the result of fraud (including there being no children at all in the country for whom Child Benefit is paid), or most of them plus a good proportion of pensioners to be fraudsters too, perhaps having cunningly lied about their age.
But there's a more pervasive myth around the 'welfare system'. That is that its beneficiaries are largely unchanging and are different from the rest of 'us' who pay for them through our taxes. We are always in work, pay our taxes and get nothing from the state. They are a welfare-dependent underclass, pay nothing to the taxman, and get everything from the state. If only we could get them to work through ever-more stringent conditions on getting benefits and through cutting back the value of what people who get them are allowed, we'd fix the public finances and get the economy moving.
As I explore in detail in a book published this week, Good Times, Bad Times: The Welfare Myth of Them and Us, this ignores both the evidence from large scale surveys and what we know from our own lives. Most of what costs the bulk of our spending on the welfare state - and the part whose cost is rising as the population ages - are the things that nearly everyone benefits from as they move through the life cycle - schools, the NHS and pensions, on top of child benefits and tax credits for families when they have children.
Instead of most spending even on out-of-work benefits going on the same people year after year, there is huge churn. A million people lose their jobs and a million gain work every three months. Fewer than half of new Jobseeker's Allowance claims last more than two months, even now. Going through the 'low pay, no pay' cycle may not mean that people are escaping far from poverty, but they are not standing still. Researchers spending months in some of the bleakest labour markets in the country describe looking fruitlessly for the famous families where 'three generations have never worked' as like 'hunting the Yeti'.
Those with an interest in keeping down the contribution from taxes on higher incomes or greater accumulations of wealth may continue to feed these misperceptions, so this will remain an up-hill battle, but how perceptions could be better brought into line with the reality of what is going on is now one of the central challenges facing those making and debating social policies and their future.
Where there are rising demands for services and the legacy of crisis and slow growth in living standards mean tight constraints on resources, we cannot afford to make choices and decisions by myth, rather than in the light of reality.
John Hills is Professor of Social Policy and Director of the Centre for Analysis of Social Exclusion (CASE) at the London School of Economics
Good Times, Bad Times: The Welfare Myth of Them and Us is published by Policy Press. For further information, click here