(Inspired by the John Quiggin article, "The big issues in macroeconomics: the fiscal multiplier.")
It strikes me that the neo-classical school is once again arguing with the data, which is unproductive in science. No, not just unproductive. Unethical. OK, sure, maybe if the discussion of multipliers was a new thing there might be reasons to argue for measurement error. But there is decades, perhaps even centuries, of data, and to deny well-substantiated data is unethical in science—it strikes at the very heart of the scientific enterprise.
A year and a half before I wrote the article which Krugman quoted, I wrote: “This is ideological, rather than rational, opposition. [...] The field of economics, I think, needs ethics to keep it track, so that it is a scholarly discipline rather than one that produces extensively rationalized propaganda. [...] Yet it seems to me that without ethics, scientific research is not possible. Without basic honesty, without the ability to admit error, without the tools of criticism and review, there is no way to arrive at scientific truth. I believe that economics went off-track partly because of highly-rewarded status-seeking behavior on the part of many economists.”
Say I, read the whole thing. :-)
It is wholly unscientific to argue that multipliers are zero or less, just as it was unscientific for Wegener’s critics to reject his case for continental drift, despite the well-organized evidence he presented. There is another parallel with Wegener: his data was rejected because geologists of his times could not see a mechanism by which the data were possible. This is parallel to the demand for microfoundations. It would be much easier to make the case for Keynesian economics if a clear explanation for the behavior of labor markets could be found in economic data. But that is not required for it to be true; thermodynamics, the understanding of heat in bulk materials, began to emerge centuries before statistical mechanics, the understanding of energy of molecules in large numbers, was invented to explain it.
Perhaps economists have been looking to the wrong physical sciences for their models.
Valid microfoundations, it seems to me, would require the ability to take models and measurements of individual psychology and translate them into valid explanations of macro-economic behavior, and so far, the effort to do this has not been successful on any large scale. Molecules, it has emerged, obey simple physical laws, and no-one expects a molecule to suddenly decide that, oh, instead of going this way, it will go that way. But people produce unexpected behavior all the time. The only time human economic behavior is strongly constrained is when basic needs are involved: we expect hungry people to eat if they can, for instance. Even then, we can be surprised: people do sometimes fast, cannot eat due to health difficulties, or even die, which is another thing that molecules do not do.
It is only in subsistence economies that microfoundations are so far accessible. In the economies people prefer, they are not constrained by basic needs; people eat to live rather than live to eat. We understand some of human needs, but valid microfoundations, it seems to me, require an understanding of human desire, and this so far does not exist.