Malcolm Turnbull and the many economists supporting his plan for a 10-year phased cut in the rate of company tax have failed to make the case that this expensive measure would deliver a significant increase in "growth and jobs".
Economists are meant to be too hard-headed to believe in the existence of a magic bullet – the single measure that will work wonders in solving our problems.
Yet economists throughout the world have come close to seeing a cut in company tax as almost magical in its ability to improve the economy's prospects. Ask them to name the one reform likely to do most good and this is what they'd pick.
A lot of studies have found that economies grow faster after they've cut their tax on companies' profits. It's a move that the International Monetary Fund and the Organisation for Economic Co-operation and Development have urged on us in successive reports on our economy.
It's a move recommended by the Henry tax review of 2010, and one high on Treasury's to-do list.
And now Turnbull has promised to do it, cutting the rate from 30 per cent to 25 per cent, with the phase-down starting this July and ending in 2026-27. For big companies, however, the phasing doesn't start until July 2024.
The phase-in will have a cumulative cost to the budget of $48.2 billion and, when completed, an ongoing annual cost of $8.3 billion (according to modelling by Independent Economics).
So the company tax cut is the centrepiece of this year's budget and the centrepiece of Scott Morrison's "economic plan for jobs and growth".
The first reservation comes from a factor peculiar to Australia: our system of "dividend imputation", introduced by Paul Keating in 1987 to eliminate the "double taxation" of the dividends people receive from the companies in which they hold shares.
To match their dividends, shareholders receive a "franking credit" set at the same rate as the company tax rate – presently 30 per cent – which has the effect of returning to them the company tax paid on their dividend and so ensuring their dividend is taxed only at their marginal income tax rate, like any other income.
Imputation has encouraged Australian companies to pay out a high proportion of their after-tax profits as dividends, rather than retain them for re-investment in the business.
Reserve Bank research shows companies are paying out more than 80 per cent of their underlying earnings.
Trick is, only Australian shareholders in our companies receive franking credits; foreign shareholders don't. In consequence, Australian shareholders (who include everyone with superannuation savings) have little to gain from the cut in company tax. It gives them no reason to change their behaviour.
Imputation has turned our company tax into little more than a tax on foreign shareholders. Which means almost all the benefit from the company tax cut goes to foreigners.
This is intended to induce them to greatly increase their investment in Australia – which is where the jobs and growth are supposed to come from – though they'll pay less tax on their present Aussie share investments whether or not they decide to invest more.
In modelling the effects of that assumed increase in foreign investment, it's natural for Treasury and others to look at the effect on real gross domestic product. Natural, but misleading.
It seems to have taken rival modelling by econometricians at the Centre of Policy Studies at Victoria University to oblige Treasury and its allegedly independent consultants to acknowledge that, because so much of the benefit of the cut goes to foreigners, and because foreigners will own all the extra share investment (and consequent dividend income) the cut induces them to make, GDP significantly overstates the expected benefit to Australians.
That's because GDP measures the value of goods and services produced in Australia, not how the income generated by that production is shared between Australians and foreign investors. The right measure is thus gross national income.
A good feature of Treasury's modelling is that it didn't just model the tax cut in isolation. It acknowledged that the cut would leave the government short of revenue and thus needing to fill the gap somehow. Its least unrealistic scenario is that it would be filled by increasing income tax – say, by allowing years of bracket creep.
Treasury's modelling of this scenario finds that the cut in company tax would cause the level of real GDP to be 1 per cent higher than otherwise in the "long term" (taken to be about 20 years).
But the level of real GNI would be only 0.6 per cent higher than otherwise. Daley and Coates put that into context by noting that if GNI per person increases by 1.5 per cent a year (as the budget papers routinely assume), over 25 years it will have risen by 45.1 per cent.
Cut company tax (not in one go, as the modelling assumes, but over 10 years) and incomes rise over the period not by 45.1 per cent, but by 45.7 per cent. Wow.
For reasons only an economist would believe (because it comes from their theory, not from empirical evidence), most of this gain would go to wage earners. After 20 or 25 or 30 years, the level of real after-tax wages will be 0.4 per cent higher than otherwise.
And get this: although voters are encouraged to believe that "jobs and growth" is really about jobs, the Treasury modelling finds that the level of employment in 20 or 30 years' time will be just 0.1 per cent higher than otherwise.
If you wanted to create jobs, cutting the tax on foreign investors isn't the way to do it.