I haven't said much about the (most recent) recent flare up over Ricardian equivalence. Why? The answer's simple, the empirical evidence does not support it. Why argue about something when we already know it fails to adequately explain the data? Making the Ricardian equivalence assumption might be okay as a first approximation for some questions -- though I'd argue that it mostly isn't -- but in any case the theory does not adequately capture economic behavior.
But let me turn the microphone over to one of the architects of the modern version of the theory, Robert Barro. In the following interview with the Minneapolis Fed (from 2005), Barro emphasizes the point Krugman makes here, i.e. that Ricardian equivalence says nothing about the effectiveness of fiscal policy as a stimulus for the economy (a point that IS worth noting since this point is often confused in discussion of this topic. As Barro tries to make clear, "It's never part of Ricardian equivalence that the level of government expenditure doesn't matter.":
Region: The Ricardian equivalence hypothesis, which you brought to prominence in 1974, might be taken to suggest that deficit spending isn't inherently harmful since rational people, expecting to pay higher taxes in the future to pay off government debt, will save more, so private savings will balance out the public deficit.
Does that imply that concerns about "irresponsible" levels of debt are unfounded? And is it puzzling to you that the Ricardian equivalence hypothesis isn't a mainstream belief in macroeconomics?
Barro: Let me say first that I think the Ricardian equivalence idea is basically right as a first-order proposition. However, people get confused as to exactly what it says. Before I say what that is, I should mention that, although the proposition is not mainstream in the sense of being fully accepted by most economists, the idea has had tremendous influence on the way economists think about this issue.
Analysis often begins with Ricardian equivalence as a first-order proposition and then goes on to investigate why there are deviations from precise equivalence. Thus, like the Modigliani-Miller theorem on corporate finance, Ricardian equivalence has become a common starting point for the way people think about budget deficits. This situation is vastly different from what it was before the mid 1970s.
To illustrate the potential pitfalls in what Ricardian equivalence says and does not say, one can consider the famous quote attributed to Vice President Cheney to the effect that President Reagan proved that budget deficits don't matter. The Cheney quote is often interpreted to mean that the level of government expenditure does not matter, and that surely is not what Ricardian equivalence says. The Ricardian proposition is about the consequences of paying for a given amount of public expenditure in different ways. Specifically, does it matter—or does it matter a lot—whether the government pays for its spending with current taxes or with current borrowing, which entails higher future taxes?
So, a central part of the proposition is that the amount of public expenditure—today and tomorrow—is being held constant. It's never part of Ricardian equivalence that the level of government expenditure doesn't matter. As [University of Chicago economist] Milton Friedman put it, the costs or benefits of government outlays depend on the amount and nature of what the government spends—there is no free lunch about paying for that spending. So whether you pay for it now or later is secondary.
As a first-order proposition, it is right that it matters little whether you pay for government spending with taxes today or taxes tomorrow...
[For a more academic discussion of this topic, see this discussion from David Romer's graduate macro text. Romer explains (contra Barro) why ""there is little reason to expect Ricardian equivalence to provide a good first approximation in practice."]