Increasingly, though, budget night has a somewhat anti-climactic air about it. Speculation, rather than news, is what keeps the media excited, so budgets are about managing expectations as well as money. Key details will have been issued in advance, so that unpleasant surprises on the night are kept to a minimum. Most of us feared that the most recent federal budget would be horrific when it actually appeared, it was much milder than we had expected.

It would be a serious mistake, though, to place too much store on the numbers put forward on budget night. Programs foreshadowed to commence down the track may never materialise or, where additional or special legislation is required, may be defeated in the Senate. A budget is a financial work-in-progress, not a reality. Even the numbers relating to the budget year itself may be subject to rapid and substantial change, both on the revenue and the expenditure sides.

In May, the federal government has a fairly good idea of the fiscal outcome from the soon-to-be completed year. But often there is not much resemblance between that outcome and the one that was projected in the previous year's budget. This is not because the government is deceiving us. It is because economies are constantly changing and no one, certainly no one in the Treasury, knows what is going to happen. (If they did, they would not be in the Treasury, but making squillions in the private sector).

We might agree that changing economic conditions will affect revenue and expenditures in ways that are hard to predict. But what about the government's policy-related choices: the new programs it is going to fund, and the old ones it is going to axe? Surely it is reasonable to expect that these will be presented (and projected) accurately in the budget?

It is true that the government tells us on budget night about all those policies it knows about. But politics being the business that it is, new policies keep popping up right through the year. And in the lead-up to elections, the spending can become frenetic. In times past, the full implications of these emerging choices were not publicly known until the next budget appeared (accompanied, if the government changed, by much hand-wringing about the ineptitude of the previous lot). At least now (through the Charter of Budget Honesty that the Howard government legislated in 1998) we are told about these new policies and their financial implications when the mid-year economic and fiscal outlook is issued in December.

For example, in 2006, in what proved to be the last full year of the Howard government, $2 billion worth of new spending was decided upon between May, when the budget was presented, and December, when the Outlook was issued. The Rudd Government, not to be outdone, brought in $3 million of new policies in the wake of its first budget. Coincidentally, or perhaps not, these amounts almost exactly balanced the spending opportunities that had arisen because of so-called parameter variations increases in revenue or decreases in expenditure that occur as a result of unantic-ipated economic and other changes.

Far from being an instrument of longer term planning, the budget is held hostage, not only by politics, but by its role in macro-economic management. When spending is ramped up in times of recession, and pruned back in times of higher growth, it is difficult to maintain a steady course in relation to public investment, particularly when borrowing is frowned upon (except in emergencies). So we have the spectacle of the Government spending up big on infrastructure-related projects in order to stimulate the economy, after years of grimly accumulating surpluses.

At the same time, when more money comes in than was expected (as was the case before the financial crisis hit) the temptation to spend it on politically favourable projects is almost overwhelming. All this is made possible by the glorious flexibility that budgets provide. Provided that spending is properly authorised (that is, there is an appropriation for it), governments do not have to abide by their budgetary intentions. Indeed, there is no formal requirement for them to have a budget at all.

So, what is all the fuss about? Why the enormous effort to produce budget papers that go into the most minute detail of agency activity? Twenty years ago the Commonwealth budget papers numbered a mere 600 pages. Now they are 10,000 and growing. Much of this additional documentation is directed towards making it clearer at agency level how much money is being spent, on what, and to what effect. The idea was to enhance both the managerial capacity and the accountability of government.

Both objectives have proved elusive. It is a bit hard to manage well when your financial parameters fluctuate wildly from one year to the next. Accountability for results becomes problematic as well. How can public servants be held responsible for outcomes when their power to influence them is so qualified? Perhaps this is why, after years of requiring agencies to report their spending in terms of outputs (that is, the impacts or effects of their activities), in the most recent budget departments are obliged, once again, to report on the basis of the actual programs for which appropriations are made.

So, after almost a decade of continuous change, we are now going back (sort of) to the situation that obtained from roughly the early to the late-1990s, when agencies produced program-based budget statements. These were much easier to use and to understand than the more elaborate and ambitious productions that succeeded them, so the hardy tribe of budget-paper aficionados will be pleased. Has all the effort that has gone into improving the way government finances are reported upon been worth it? It seems to me that, as they labour unremittingly to achieve perfection in budget-related reporting, the Finance people are a little like librarians who agonise endlessly over the perfect bibliographic description of a new book, when all the users want to know is where to find it.

Dr Jenny Stewart is Associate Professor of Public Policy at the University of Canberra.