Every four years, presidential candidates tout their economic proposals as sure-fire job creators. The unstated assumption, of course, is that these jobs projections are based on strong, or at least defensible, economic analyses. But are they always?

Last February, President Bush's economic team was criticized when it forecast job growth of 3.8 million by the end of the year. Critics said the number was implausibly high. And indeed, it implied 3 percent job growth, which the country hasn't seen since 1994. The criticism struck home, and the administration eventually disowned its jobs-creation forecast.

Sen. John Kerry's economic team risks similar criticism over its jobs numbers. Since late March, his campaign has been highlighting his package of personal tax increases and corporate tax changes, saying it would add 10 million jobs by the end of 2008. How do we know it would? The number can be traced to a memorandum on the senator's Web site.

In the memo, Harvard economists Lawrence Katz and Elisabeth Allison claim the economy "is capable of a much stronger performance than it has exhibited over the last few years." By employing "sound economic policies," they say, we can get back to the unemployment rate we saw in 1999 and 2000: 4.1 percent. Getting it back to this figure would result in ... 10 million new jobs.

So Sen. Kerry's 10-million-job-creation claim rests entirely on whether the plan actually can reduce the unemployment rate  to 4.1 percent and keep it there. Unfortunately, it presents no economic modeling work and no economic theory to support this claim. In fact, Professors Katz and Allison, apparently unwilling to go out on a limb, are quite careful to avoid stating that Kerry's plan will create even a single job.

We now know that several fortuitously timed events kept the economy flying high back in 1999-2000, when it could have slid either toward recession or high inflation. These included a sustained acceleration in worker productivity, low petroleum prices during much of the period and the late '90s stock market illusion. This was, after all, right before the dot-com bubble burst and various corporate scandals came to light.

Back in 2000, President Clinton's own Council of Economic Advisors understood its good fortune and admitted that 1999's unusually low unemployment rate probably was unsustainable. In a February 2000 report, the council projected that the year's unemployment would match 1999's 4.2 percent but eventually would rise to 5.2 percent by 2003 and stay there through at least 2006.

They also freely admitted the critical importance of non-policy-related factors in preventing 1999's exceptionally low unemployment rate from triggering the inflation that usually heralds the onset of recession. These factors included "spare manufacturing capacity, new efficiencies in the labor market from expanded use of temporary help workers and Internet job search resources, higher-than-expected productivity growth and declining import prices."

President Clinton's economists assumed that these factors were temporary, which explains why they forecast higher unemployment in future years.

The non-partisan Congressional Budget Office agreed. In its January 2000 11-year forecast, CBO predicted a rise in unemployment to 4.7 percent by 2003 and 5.2 percent by 2010. In making these projections, CBO was required to assume that President Clinton's "sound" economic policy would stay in place throughout the 11 years. So even under the Clinton policies, the 4.1 percent unemployment rate was expected to rise by a full 1 percent over the next decade.

Sen. Kerry hasn't yet backed away from his own job-creation number, indicating that, perhaps, he'll soon release a serious analysis of his plan and its job-creation benefits. Otherwise, Kerry may be forced to take President Bush's lead and disown his own apparently untenable job-creation number.

After all, both candidates' claims about the economy seem overly optimistic. The only way they'll come to pass is if our economy sees a remarkable spasm of good luck, such as President Clinton enjoyed during his final two years. And economic policy should be based on facts -- not on hope that we'll hit the lottery.

Al Goyburu is a policy analyst in the Center for Data Analysis at The Heritage Foundation, a Washington-based public policy research institute.