Be sure your dodgy modelling will find you out. I’m starting to think economists have become so used to pretending to know more about the economy than they really do that they don’t notice the way they mislead the rest of us.
The Productivity Commission has proposed a radical change in the way companies are taxed which, it tells us, would improve the economy’s productivity and leave us better off. It has commissioned modelling that, it implies, supports its case for change.
But when you read its report – and add some knowledge of how “computable general equilibrium” models of the economy work – you’re left with nothing but doubts.
The proposal involves cutting the present 30 or 25 per cent rate of tax on company profits to 20 per cent for all companies except the 500 or so with annual turnover (total sales) of more than $1 billion.
But it also involves introducing a 5 per cent tax on the annual net cash flow of all companies. This new tax would include an allowance for the cost of companies’ equity capital, and an immediate write-off of the cost of newly purchased assets, but no allowance for interest paid on the companies’ borrowing.
The commission paid for two sets of modelling of the proposed changes, one from Chris Murphy, Australia’s leading commercial modeller, and the other from the leading academic modelling outfit, the Centre of Policy Studies (CoPS) at Victoria University.
The commission’s report compares the results of the two modelling exercises for just the first proposed change, cutting the rate of company tax to 20 per cent for all but the top 500 companies.
According to the Murphy model, this would cause companies to increase their investment in new equipment by 1.4 per cent, improve “productivity” by 0.4 per cent, increase real gross domestic product by 0.4 per cent and increase real before-tax wages by 0.6 per cent.
The CoPS modelling results are similar in some respects. It expects a smaller increase in business investment of 0.6 per cent, but improved productivity of almost as much, and the same increase in before-tax wages, even though GDP increases by only 0.2 per cent.
Does the modelling provide reasonably strong support for cutting company tax to make the economy bigger and better? Well, no, not really. Those results are shockingly small.
Economists have gone for years making their modelling results seem grander than they are by, in this instance, letting the rest of us conclude that the estimated increases of 1.4 per cent, 0.4 per cent and 0.6 per cent represent annual increases in the rate of growth in business investment, productivity, GDP and before-tax wages.
Wrong. What the people waving modelling results around rarely bother to make sure the punters understand is that these are once-only increases in the levels of investment, productivity, GDP and before-tax wages. What’s more, they’ll come about only “over the long run”.
And how long is the long run? They rarely bother to tell us – especially as it can vary with the modeller. But if you dig deep you can find out. CoPS sets it at five years, I’m told, but it’s more usually thought of as about 10 years. And the commission’s report seems to be saying that its comparison of the two modelling exercises is what they estimate will be the story in 2050.
Get it? We’re considering a hugely expensive cut in the rate of company tax in the belief that this will cause real GDP to be between 0.2 and 0.4 per cent greater in five to 25 years’ time.
Really? Its modelling shows the benefit from cutting the rate of company tax would take years to materialise, and still be trivial, but the commission thinks we should do it anyway.
See what this is saying? Even the economists who commissioned this modelling don’t take its results seriously. Why not? Well, for a start, they know how primitive and grossly oversimplified these modelling exercises are. It’s as though the economy they’ve been able to model is one inhabited by stick figures, not humans.
As modelling is such a dodgy exercise, economists know they don’t have to believe any results they don’t fancy – because, in truth, economics is based more on religious belief than scientific inquiry. What figures largest in the thinking of economists is the model of the economy they’ve been carrying around in their heads since about second year uni.
The model in their head tells them taxes discourage and distort economy activity, meaning lower taxes are always better. So if econometric modelling tells them a rate cut would make little difference, they’re undeterred.
Speaking of taxes, one reason the effects of a cut in company tax are so modest is the standard assumption that the lost government revenue has to be covered by tax increases somewhere else. The modellers here have assumed it’s covered by a “non-distorting lump-sum tax” (which doesn’t exist in the real world) or by bracket creep (a hidden, lasting increase in personal income tax).
Significantly, this would be why, under Murphy, the real wage increase of 0.6 per cent before tax, turns into an after-tax increase of zero. Really? We want to improve productivity to raise our material standard of living, but real after-tax wages would be unchanged under Murphy – or, under CoPS, would actually fall by 0.5 per cent. Great idea, eh?
Finally, economists at the Australia Institute reveal that what the commission chooses to call “productivity” is actually “output per worker”, which ain’t quite the same thing.
It turns out that, according to the modelling, national output per worker increases not because any worker becomes more productive, but because the company tax cut’s reduction in the after-tax cost of capital causes our capital-intensive mining industry (which thereby has higher output per worker) to expand at the expense of the labour-intensive health and education sectors.
And this would be progress, would it? The sad truth is that modelling is used to help sell policy changes someone thinks we should make, not to improve our understanding of what works and what doesn’t.
