General Motors wanted to boost sales this year, so it cut prices.

When the automaker offered its cars to the public at the same prices it charges employees, cars started flying out of showrooms. In fact, the deals were so good that other American automakers also had to slash prices in order to keep pace.

It's simple, really. People respond to financial incentives. Cut prices and they'll buy more.

The same rule applies to tax rates. Cut them, lowering the price of work, saving and investment, and people will earn more. That will improve the economy and end up boosting tax revenues.

Don't take my word for it. Ask Uncle Sam. Just five months ago, the White House was predicting that this year's federal deficit would be $427 billion. Now the shortfall is projected to be $333 billion. Huge, yes, but substantially lower than expected.

The projected deficit is dropping because the government has collected $87 billion in what The Washington Post called "unanticipated tax receipts." Unanticipated by some, perhaps -- but precisely what we at The Heritage Foundation predicted in 2003. President Bush cut tax rates two years ago, and because of those lower rates, people and businesses have more incentive to save and invest.

As I wrote in May 2004, "Once lower tax rates go into effect, businesses start to invest more, which means more jobs and higher wages. GDP starts to increase. Good news starts to pile on top of good news until you witness an economic recovery.

If we knew cutting taxes would improve the economy, why didn't the government see this windfall coming? Because it insists on using an outdated and discredited method to make its projections.

Both the White House and the Congressional Budget Office rely on what are called "static" predictions. That means if we cut tax rates, say, 10 percent, they simply assume federal revenues will be 10 percent lower in years to come. They never take into account the extra economic activity that lower tax rates generate. But as the GM example and this year's soaring tax receipts demonstrate, incentives matter.

GM never would have offered employee prices if it had simply assumed it would make less profit on each car and sell only as many cars as it had sold the year before. The company realized that offering lower prices would mean less profit per car, but it also knew those lower prices would allow it to sell more cars.

We call this "dynamic scoring," and that complicated name hides a simple reality: Actions have consequences.

When the federal government projects its budget, it ought to use a computer model such as the one Heritage economists used in May 2003 to analyze that year's tax cut. We predicted that GDP would grow almost 4 percent in 2004. The actual gain was an even-better 4.4 percent. We predicted the tax cuts would lead to 800,000 new jobs in 2004. In fact, there were 2.2 million. Also, we foresaw that the unemployment rate would fall, as it has throughout 2004 and this year.

We're not trying to gloat. Clearly, our numbers were somewhat on the conservative side. But they're far closer than "static" projections, which simply subtracted the revenue "lost" from the tax cuts and left it at that. This explains why the government's projections for this year are already off by some.

The government's actions affect everyone. After all, you might not be planning to change jobs or invest some money this year, but you'll surely be influenced by federal taxes. So it's critical that federal forecasters are as accurate as possible.

It's time for lawmakers to bring dynamic scoring out of the showroom -- and put it to work in Washington.

Ed Feulner is president of the Heritage Foundation.