The set the government promotes to the public has a healthier bottom line: a $318 billion deficit in 2005. The set the government doesn’t talk about is the audited financial statement produced by the government’s accountants following standard accounting rules. It reports a more ominous financial picture: a $760 billion deficit for 2005.
…
The audited financial statement - prepared by the Treasury Department - reveals a federal government in far worse financial shape than official budget reports indicate, a USA Today analysis found. The government has run a deficit of $2.9 trillion since 1997, according to the audited number. The official deficit since then is just $729 billion.
…
The new Medicare prescription-drug benefit alone would have added $8 trillion to the government’s audited deficit. That’s the amount the government would need today, set aside and earning interest, to pay for the tens of trillions of dollars the benefit will cost in future years.

Standard accounting concepts say that $8 trillion should be reported as an expense. Combined with other new liabilities and operating losses, the government would have reported an $11 trillion deficit in 2004 - about the size of the nation’s entire economy.

The federal government also would have had a $12.7 trillion deficit in 2000 because that was the first year that Social Security and Medicare reported broader measures of the programs’ unfunded liabilities. That created a one-time expense.

In the domain of the stimulus packages, these results suggest that the method of delivering them (individual tax relief in the form of tax rebates, money toward retirement saving, gift certificates, pre-paid debit cards, etc.) could have large consequences on its effectiveness.

The next question, of course, is which delivery method to select. Here behavioral economics has been instructive as well. In particular, years of research have demonstrated over and over that our intuitions about the relative effectiveness of different approaches are often wrong. Given that the method of delivery could make a large difference, and given that our intuitions about their relative effectiveness could be wrong, what should we do?

One answer is to conduct an experiment, as this is the only method we have for testing what really works and what is likely to fail. In the same way that we force drug companies to test the efficacy of their drugs before rolling them onto the market, shouldn’t we ask the government to first test their ideas before they invest billions of dollars of our tax money on some stimulus packages?