It is Budget day and attention is inevitably focused on the spending side of the Government's books.
After all, that is under its control.
Revenue forecasts, by contrast, are taken as given, when in fact they are very spongy numbers, reflecting the inherent difficulty of forecasting the underlying economic growth which drives the size of the tax base.
Treasury soothsayers in recent years have tended to err on the side of optimism in their economic forecasts, particularly for the out years.
Take growth in the year to March 2012.
In Budget 2009 they thought it would be 2.9 per cent. It was revised up to 3.1 per cent in Budget 2010 but then cut to 1.8 per cent in last year's Budget.
In the pre-election update this forecast was boosted to 2.3 per cent, but by the time of the Budget Policy Statement in February it had been revised back down again, to 1.9 per cent.
We won't get an official number from the statisticians for another month, but 1.1 per cent, unchanged from calendar 2011, looks more like it at this point.
In today's dollars, 1 per cent of GDP is a bit over $2 billion.
Even forecasts for the year which has already begun when a Budget is presented tend to be materially wide of the mark.
Budget 2009 forecast the economy to contract, on an annual average basis, by 1.7 per cent in the year to March 2010. In fact, Statistics New Zealand now reckons it shrank by 1.2 per cent.
Budget 2010 forecast growth of 3.2 per cent in the year to March 2011, twice the actual rate of 1.6 per cent. The miss cannot be laid at the door of the February earthquake.
The Treasury's forecast last February for growth in the year ahead, to March 2013, at 2.7 per cent was in line with consensus, but its pick of 3.8 per cent the following year is at the high end of the range of forecasters, as surveyed by NZIER.
The Reserve Bank's latest survey of expectations, released on Tuesday, found an average pick of 2.4 per cent for growth two years out.
Inflation can also be a source of error in the forecasts.
For the year ended March, inflation came in at 1.6 per cent instead of the 2.8 per cent the Treasury expected in the pre-election update.
Less inflation means a smaller nominal GDP - a proxy for the tax base.
Small revisions to the growth outlook in the near term can have a major effect on the Crown's bottom line and borrowing requirement over the following years.
The Budget Policy Statement in February shaved a cumulative percentage point off growth forecasts for the March 2012 and 2013 years.
The effect on the Crown accounts was bigger deficits in the near term and smaller surpluses further out - adding up to a $5.5 billion deterioration over the five years to June 2016 and an equivalent rise in debt.
Effectively, it swallowed up the estimated proceeds of asset sales.
The Government has subsequently scrapped the $800 million allowance for net new spending which had been pencilled in for the coming fiscal year, signalling instead a second zero Budget.
None of this is to say the Treasury is especially bad at forecasting, just that it is really difficult to do, particularly when major shocks like global financial crises and earthquakes come along.
A review last year of its forecasting performance found it compared well with a sample of 12 forecasters, ranking either first or second for forecasts of GDP growth and inflation over the 2000 to 2010 period, on average, for current year and one-year-ahead forecasts.
Even so, when it came to forecasting tax, the average difference between the Treasury's one-year-ahead forecasts and actual tax revenue was 4.4 per cent a year during the five boom years to June 2007.
Those were positive surprises.
The difficulty of figuring how much of that was cyclical froth and how much structural beer undoubtedly contributed to what in hindsight was excessive fiscal largess in Labour's last term.
In 2009 the Treasury set itself the target of reducing tax forecast error to plus or minus 3 per cent, which is still more than $1.5 billion either way.
Three quarters of the way through the current fiscal year the tax take was running 3.8 per cent below forecast.
So, if Finance Minister Bill English announces this afternoon, as expected, that the Government is on track to return to surplus in three years' time we should remember that he has also ruefully admitted that, "the fact that Treasury forecasts something doesn't mean it will happen".
From the Government's point of view, directing our attention to the dubious prospect of a surplus in the 2014/15 fiscal year serves the useful purpose of distracting us from a bleaker outlook in the near term, at least compared to its previous forecasts.
And distract us from its own hand in that.
Westpac chief economist Dominick Stephens says public sector demand has swung from being a significant contributor to a small outright drag on GDP growth.
"We expect government spending to remain a drag on growth for at least the next year, although not enough to prevent a faster overall pace of growth as the Canterbury rebuild gathers pace."
He makes the point that the extent to which the deficit shrinks over the next few years will make the tightening seem more swift than it actually is. That is because in the accruals-based accounting system, the Government's share of the costs of the February 2011 earthquake was largely booked in the 2010/11 fiscal year, but in cash terms most of that money has yet to be spent.
ANZ chief economist Cameron Bagrie says the fiscal stance will be tight - "contractionary in the order of 1 to 2 per cent of GDP".
That would allow the Reserve Bank to keep interest rates lower for longer but nonetheless represents a headwind to the economy.
The central bank, in its March forecasts, still has the Government in deficit in the 2014/15 year.