We can expect some scary numbers when the Treasury updates its statement on the long-term fiscal position this year. This is an exercise in projecting forward over the next 50 years what the Government's accounts will look like if current policy settings remain unchanged.
It will show the relentlessly mounting cost of health care, superannuation and debt servicing as the population ages.
Hopefully it will also make suggestions about how to bend these curves down to something less ruinous and less likely to send the young flocking to airport departure lounges.
But the savings industry's national body, the Financial Services Council, suspects the numbers will be less scary than they should be, that they will be sugar-coated by putting over-optimistic assumptions into the Treasury's model.
An informed debate on the options has to start from a cleared-eyed appreciation of the baseline, business-as-usual track, the council says.
It is critical of three key assumptions officials seem to be making: about longevity, productivity growth and tax revenue.
Life expectancy clearly affects the numbers of people eligible for superannuation.
We can expect there to be fewer than half as many workers per superannuitant by 2060 as there are now, but the economy's ability to bear the cost of an ageing population also depends on how much more productive those workers will be.
And the tax take, relative to the size of the economy, will affect public debt levels, among other things, and therefore how much of a future tax dollar is pre-empted by interest costs.
At a conference on all this in December the Treasury's chief economist, Dr Girol Karagaoglu, outlined what its initial modelling was showing if policy kept to its current course.
By 2060 the cost of New Zealand Superannuation will have climbed from 4.5 per cent of gross domestic product now to 8 per cent.
Public health costs will have climbed from 7 to 11 per cent of GDP, and burgeoning deficits will have driven the Government's interest bill to 11 per cent of GDP, from just over 1 per cent now.
"We all know in reality [that] will never happen," he said. "Crown debt levels will not be permitted to rise above 100 per cent or 200 per cent of GDP.
But even those projections rest on assumptions the Financial Services Council regards as dubious.
The statisticians have a history of being surprised by improvements in life expectancy. The council points out projections in 1991, when there was a major revision of retirement income policy, underestimated the actual population over 65 by about 100,000.
"At the current New Zealand Superannuation rate this equates to about $2 billion of additional cost each year."
The International Monetary Fund warned last year that governments and fund managers are likely to be underestimating life expectancy. The IMF said if people lived three years longer than expected, which was in line with underestimations in the past, the already large costs of ageing to governments could increase by a further 50 per cent.
Statistics New Zealand revised up its longevity projections in 2011.
The median projection assumes that future gains in life expectancy will be smaller than those observed between the mid-1970s and mid-2000s. It does, however, offer a projection based on "very low mortality" which assumes that the same rate of improvement continues.
If it does, the old age dependency ratio - the number of people over 65 divided by those aged between 15 and 64 - would be 51 per cent rather than 44 per cent.
That would push the cost of super by 2060 up from 8 per cent of GDP to 9 per cent. To calibrate the scale of that difference, we currently spend less than 1 per cent of GDP on defence.
The Treasury's modelling also assumes trend growth in labour productivity of 1.5 per cent a year.
Labour productivity growth is a very slippery thing to forecast. A lot of things affect it.
But the average over the past 40 years has been closer to 1.1 per cent.
The latest numbers from Statistics New Zealand put average labour productivity growth between 2000 and 2008 at 1.3 per cent and since 2008 just 0.6 per cent.
That covers about 80 per cent of the economy; it excludes public-sector activities such as health and education where productivity is especially hard to measure but which are normally assumed to dilute the economy-wide number down.
A report by Infometrics economist Dr Adolf Stroombergen for the Financial Services Council says, "There is nothing to indicate that the higher rate depends on some policy change, nor that it is the anticipated result of result policy. It may not be wrong, but there is no case made against adopting the historical average."
Stroombergen notes that both the Congressional Budget Office in the United States and the Office for Budget Responsibility in Britain assume the average rate observed over the past half-century. It can be argued that the future fiscal cost of an ageing population is not particularly sensitive to productivity growth.
While higher productivity would mean a bigger tax base, it would also mean higher health-sector wages and larger transfer payments indexed to wage growth, including superannuation.
But small differences in growth rates can make a cumulatively big difference if you are looking 50 years out.
Which brings us to the third major assumption, about tax revenue.
The projections assume that revenue (mostly tax) is kept at a stable ratio of 32.5 per cent of GDP from 2020 on.
That would be up from 30 per cent now.
It raises the suspicion that the Treasury is relying on an extended period, perhaps nine or 10 years, of fiscal drag to boost revenues.
That is where nominal wage growth combines with the progressivity of the income tax scale to push more and more taxpayers into a higher tax bracket and increase the tax take as a proportion of incomes.
If so, it would be at odds with other advice from the Treasury, which suggests adjusting the tax thresholds in 2015, to offset what by then would be five years of fiscal drag and leave an extra $1.5 billion in the hands of taxpayers.
The Treasury has had the benefit of some academic work on the fiscal savings to be made if tax thresholds were to be adjusted not to offset rises in wages, but by the average of wage inflation and consumer price inflation, which would be a lower amount.
Not for the first time, thinking about what is good for the economy and what is good for the Government's books give different answers.