Here at the Conservative Party Conference, delegates are fond of mocking the 'magic money tree' that will apparently fund Labour's spending plans. Few of them realise that the party's own flagship policy – the National Living Wage (NLW), supposedly designed to offset cuts to tax credits for low-paid workers – is a magic money tree policy itself.
The chancellor's logic is this: working tax credits "subsidise" firms by allowing them to pay their workers less. Reduce the subsidy, and raise the minimum wage, and firms will be forced to pay their workers more. The government spends less and firms spend more, and workers don't feel the difference.
Virtually none of this is true. Tax credits do not subsidise firms, raising the minimum wage does not magic more money from firms to workers and even if it did it would hardly offset the tax credit cuts at all. Osborne is cutting one of the best kinds of welfare we have, and adding insult to injury with a minimum wage hike that may do more harm than good.
In-work poverty is a significant and probably growing problem for developed countries. Although globalisation and automation raise living standards overall, unskilled workers in developed countries, who could previously rely on industrial jobs to provide a decent wage, are now finding these jobs going overseas. At least for the time being the jobs they are finding instead are not as good.
That means that the old welfare state, designed to act as a temporary safety net to support people in between jobs, is outdated. We now have a large and growing number of people who are not economically productive enough to earn the wages that most people think are necessary to have a decent life even though they are in full-time work.
Tax credits, which are really a form of direct cash transfer to low earners, are designed to remedy this problem. The principle here is that is that a job may not be enough, and giving poor people money to top up their incomes is the best way to improve their standard of living.
Tax credits as they exist in Britain are deeply flawed. Administratively, they are a complex and confusing disaster – up to £7bn ($11bn) in tax credits may go unclaimed. Administrators within the system report people removing themselves because of the complexity and potential for sanctions if you get something wrong – or if you are overpaid by accident and fail to report this.
Then there is the problem of withdrawal, where earning more from work means receiving less in tax credits, which can reduce the incentive to work more. This effect can be dampened by withdrawing more gradually, but of course that costs money.
But all means-tested welfare faces this, and tax credits actually help to mitigate withdrawal effects from things like jobseeker's allowance, the loss of which can make low-paid work substantially less attractive. Working tax credits attract people into the labour force because they help to make work pay.
The chancellor's proposals to cut working tax credits in real terms reduce this positive incentive effect, and of course reduces the amount that people in work have.
Do tax credits subsidise employers? Except in very inefficient labour markets or for very narrowly targeted forms of tax credits, which we don't have in the UK, the answer is: probably not.
The Resolution Foundation's assessment of the evidence around this suggests that there is little or no major wage slippage associated with the UK's old Working Family Tax Credit system.
Forthcoming National Institute of Economic and Social Research (NIESR) research suggests that of the 76p an hour spent in tax credits for full-time minimum wage workers, 72%-79% goes to the worker.
Where does the rest of that 21%-28% go? Possibly not to the firm – it may be that because tax credits draw people into work, it means that wages are pushed downwards slightly for current low-paid workers. Overall, because it means more people are working, that's still a good thing.
Raising the minimum wage to a 'national living wage' level probably won't do much good either. Most economic evidence suggests that minimum wage increases lead to unemployment – a 2007 review of 103 papers found that a majority of studies showed that minimum wage increases destroyed jobs, an effect found in two-thirds of the most methodologically robust papers.
The Office for Budget Responsibility estimates that 60,000 people will be put out of work by the NLW, and total working hours reduced by 0.4% – a small fraction of the labour market, perhaps, but hardly "marginal" as the chancellor has claimed. And remember that this hike is well above any level recommended by the Low Pay Commission, whose mandate is to recommend minimum wage increases that will cause minimal unemployment.
The reason for this is that firms pay workers according to how much those workers are worth to the firm. This is true across the board – firms simply will not employ workers at loss-making wages for very long, and will increase their investment in labour-saving machines (like automatic checkouts) or cut back on their operations and staffing.
Unless we boost workers' productivity we can't boost their wages. Last year Andy Haldane, the Bank of England's chief economist, confirmed that since the 2008 crisis (as usual) wages and productivity have tracked each other "in lockstep". To imagine that government can raise wages with no negative consequences, without doing anything to boost productivity, is magic money tree thinking.
Even if this was somehow untrue, the NLW rise does almost nothing to help people on low wages. The Institute for Fiscal Studies (IFS) found that even if the NLW had no impact on GDP or employment it would offset just 27% of the drop in household incomes – but stresses that the NLW is likely to hit both GDP and employment, so it will probably do even less than that.
When he was challenged about this earlier today in an interview with Guardian editor Katherine Viner, the chancellor was evasive. So too have been other Conservative MPs throughout the conference – many outright denying that the IFS's numbers are true, or simply trying to change the subject.
Of course cuts need to be made. But cutting one of the few forms of welfare that encourages work and distorts incentives the least is going to be counterproductive. Abandoning the 2% real terms rise to pensions would be one way of making up the money; allowing construction on the green belt to reduce the housing benefit bill would be another.
To imagine that the NLW will offset these cuts is a fantasy. If we want to boost low-paid workers' wages, we can't just pass a law commanding their employers to pay them more – we have to actually find the money from somewhere, and give it to them. Despite what George Osborne would like us to think, Labour is not the only party that thinks it has a magic money tree.
Sam Bowman is deputy director of the Adam Smith Institute