Time to factor in some realistic pessimism

Illustration: Eric Lobbecke

“Prediction is very difficult, especially about the future,” wrote Niels Bohr, a Danish physicist. That’s certainly proved true for Australian government budgeting. Every federal budget for a decade has proved wildly optimistic, pencilling in billions in tax revenues that have never materialised. And tomorrow night’s federal budget will once again prompt laser-like focus on budget outcomes that will never occur.

The Treasury will assume the fastest lift in revenues as a share of the economy for almost 20 years — “tax on steroids” according to Chris Richardson, a director of Access Economics. Company tax, for instance, is expected to bring in more as a share of the economy by 2020 than at any time since the late 1970s. Income tax collections are forecast to rise 7 per cent a year — more than double the sum of inflation and population growth — over the next three years.

Despite sluggish wage growth, anaemic inflation and a highly uncertain outlook for the global economy, the government will once again assume the economy goes gangbusters for five years in a row. And once again, it probably won’t. It is only because of these rosy forecasts the government will be able to present a return to a (tiny) surplus by 2021.

In truth, rather than a surplus, in 2021 the government is far more likely to have a budget deficit of $37 billion, which is the amount pencilled in for this financial year.

It’s not only optimism that undermines the budget, it’s the spurious precision. For example, in the previous budget update the government pencilled in a deficit of $19.7bn in two financial years from now. This is ridiculous; the level of uncertainty around these figures is huge. It would be better to specify a range, which the Reserve Bank has started to do in its forecasts for inflation and economic growth.

The combination of optimism and contrived precision is intellectually questionable; it has also had real and damaging consequences for taxpayers. It has fuelled unnecessary growth of government, making it easier for successive governments to avoid necessary decisions to either lift tax or cut spending. It has undermined respect for politicians, and economists too.

Perhaps it’s no surprise the Labor opposition doesn’t protest too much about the assumptions underpinning the budget. It tolerated exactly the same in government. And, after all, they are convenient for the political class: the prospect of a return to surplus, however dubious, makes promises to spend public money politicians don’t have more palatable.

Former treasurer Wayne Swan infamously uttered in his 2012 budget speech “the four years of surpluses I announce tonight”. If he had instead promised “four years of massive deficits”, which is what occurred, it would have been much more difficult for the Gillard government to legislate the National Disability Insurance Scheme, which will ultimately make Medicare look cheap. Likewise, the Turnbull government’s opportunistic promise this week to increase school funding by almost $19bn. This is in effect a promise to increase tax by $19bn. If you discounted the future revenues and expenses of the Commonwealth government back into today’s money, allowing for some indexation of the income tax thresholds, expenditure would dwarf revenues. In other words, the government is insolvent without significant tax increases. As Mr Richardson said last week, we’ve chosen to be a high-spending and low-taxing nation. Australia isn’t alone in this regard; the US government’s taxes fall significantly short of its forecast expenses, as US economist Lawrence Kotlikoff has shown.

The Reserve Bank in 2012 in effect conceded its economic growth forecasts were of limited value. “Forecasts of the unemployment rate outperform a random walk only for a few quarters ahead.” A random walk is the trajectory of a stumbling drunk. “Like many other forecasters, the RBA forecasts explain very little of the variations in GDP growth, medium-term changes in unemployment, or the medium-term deviations of underlying inflation from the target,” the authors added.

After six years of large forecasting errors, in 2014 the Treasury changed its formulae for making budget projections. No longer would the unemployment rate drop suddenly to 5 per cent after two years, it would return to 5 per cent gradually over a few years instead. Yet it kept the same overarching framework: the economy converges back to its long-run growth path over five years. That means the economy is assumed to grow above 3 per cent a year for five years in a row! In reality the economy hasn’t put in such a good performance for at least a decade, and it might not ever again.

This Treasury framework is based more on Keynesian theory of economies rather than actual experience. It assumes the “slack” in the economy that has emerged since the financial crisis, manifested in such things as lower participation in the job market, will dissipate. But it might not. Ironically Keynes himself projected that by 2030 the job market would be eviscerated by technology, as progress galloped ahead of capitalism’s ability to find jobs for the bulk of people.

To be sure, to make forecasts some assumptions must be made. But the facts should be tortured with as little theory as possible.

A more intellectually honest and reliable rule would be to assume that what is happening now continues to happen. Treasury could simply assume last year’s outcome, be it a growth rate or a level depending on what is appropriate in the given case, is what will happen this year, adjusted for population growth and inflation. This strategy would have given vastly more reliable and accurate results over the past decade. It would have predicted ongoing huge deficits. And if it were employed tomorrow it would see big dollops of red ink far beyond 2021.