Published in July 2010 issue of Prospect Magazine
How do you set fiscal policy? The aim is not to allow government debt to get too large relative to GDP. One obvious way is for the government to plan its expenditure over the next year or two (or beyond) and then set taxes so that projected revenues are close to expenditure.
The key problem with this is the difficulty of making projections. Projecting expenditure is fairly straightforward because it can be managed by government departments under direction from the treasury. But projecting tax revenues is a different matter. First, it requires forecasting the future path of the economy. (This is not too difficult over a stable period like the early 2000s, and indeed the treasury had a reasonable record in those years.) Second, it is necessary to calculate actual tax revenues given the predicted state of the economy. In practice, despite the fact that all tax rates are known, this is a tricky business requiring judgement and guesswork—with plenty of room for politics to intrude.
To see how hard it is, consider the period from 2000 to 2006. At this time, growth was steady and predictable—in fact, stable growth continued up to the financial year 2007-08. Looking at the treasury’s projections of expenditure, tax revenues and the deficit (public sector net borrowing), we can now see that in every budget from 2001 to 2006, projected tax revenues both one year and two years ahead over-predicted the actual outcome. The average over-prediction was 0.9 per cent of GDP for one-year projections and 1.5 per cent of GDP for two-year projections.
In the light of this, it is no surprise that, for one-year predictions, the deficit projections under-predicted the actual outcome in every year bar one and, for two-year projections, in every single year. The result was a steady rise in the ratio of government debt to GDP throughout this period despite the buoyancy of the overall economy.
Whatever the reason for these faulty predictions, they suggest that Brita…