In a long and extremely persuasive piece for National Affairs, Harvard Economics professor Greg Mankiw argues that, even if you're sympathetic to Keynesian economic theory, there's very little clear evidence to support the claim that government spending makes the best fiscal stimulus. Moreover, despite the efforts of the Obama administration to track the effects of last year's spending-heavy stimulus package—through self-reports and through modeling—the data gathered don't really tell us much about the actual overall effects of the stimulus. "The fiscal-policy decisions of the past year and a half have not been implausible or inexplicable," writes Mankiw, "but they have also not been empirically shown to work. The data point to other approaches."
On the stimulus, Mankiw is unquestionably right. The stimulus reports that have come from the administration and the Congressional Budget Office don't rely on actual measurements of job-creation outputs. They measure the dollar value of the inputs, and then feed those measurements into their models. But their models are obviously unreliable, because the predictions they've made have turned out to be wrong. So we don't know what the effects of the stimulus are, and we don't have a good way to measure them.
And that feeds into larger point that Mankiw makes, which is that macroeconomics is a relatively young science and there's still a lot that its practitioners don't really know. As a result, he argues, economists dealing with macro effects ought to be far more humble about their work and their conclusions. I'd actually take this a little further and argue that pundits, wonks, politicians, and policymakers of all stripes ought to be more humble about what we know and don't when it comes to public policy—and beyond that, about what public policy can actually accomplish. At times, certain politicians will let slip the truth that what they're doing is, at best, pretty speculative—see Chris Dodd's admission that, on the new financial bill, "no one will know until this is actually in place how it works"—but in general, there far too much certainty in Washington, both about how things work and what the government can actually achieve.