Interview of Larry Summers. He assesses “double-dip” recession risk and advocates more government spending in order to generate enough demand. With 90% government debt-to-GDP ratio, Summers seems to ignore the fact that the US government is seriously constrained in engaging in further fiscal expansion.

Allan Meltzer, in his piece “Four Reasons Keynesians Keep Getting It Wrong“, offers a classic rebuttal:

Why is the economic response to increased government spending so different from the response predicted by Keynesian models? What is missing from the models that makes their forecasts so inaccurate? Those should be the questions asked by both proponents and opponents of more government spending. Allow me to suggest four major omissions from Keynesian models:

First, big increases in spending and government deficits raise the prospect of future tax increases. Many people understand that increased spending must be paid for sooner or later. Meanwhile, President Obama makes certain that many more will reach that conclusion by continuing to demand permanent tax increases. His demands are a deterrent for those who do most of the saving and investing. Concern over future tax rates is one of the main reasons for heightened uncertainty and reduced confidence. Potential investors hold cash and wait.

Second, most of the government spending programs redistribute income from workers to the unemployed. This, Keynesians argue, increases the welfare of many hurt by the recession. What their models ignore, however, is the reduced productivity that follows a shift of resources toward redistribution and away from productive investment. Keynesian theory argues that each dollar of government spending has a larger effect on output than a dollar of tax reduction. But in reality the reverse has proven true. Permanent tax reduction generates more expansion than increased government spending of the same dollars. I believe that the resulting difference in productivity is a main reason for the difference in results.

Third, Keynesian models totally ignore the negative effects of the stream of costly new regulations that pour out of the Obama bureaucracy. Who can guess the size of the cost increases required by these programs? ObamaCare is not the only source of this uncertainty, though it makes a large contribution. We also have an excessively eager group of environmental regulators, protectors of labor unions, and financial regulators. Their decisions raise future costs and increase uncertainty. How can a corporate staff hope to estimate future return on new investment when tax rates and costs are unknowable? Holding cash and waiting for less uncertainty is the principal response. Thus, the recession drags on.

Fourth, U.S. fiscal and monetary policies are mainly directed at getting a near-term result. The estimated cost of new jobs in President Obama’s latest jobs bill is at least $200,000 per job, based on administration estimates of the number of jobs and their cost. How can that appeal to the taxpayers who will pay those costs? Once the subsidies end, the jobs disappear—but the bonds that financed them remain and must be serviced. These medium and long-term effects are ignored in Keynesian models. Perhaps that’s why estimates of the additional spending generated by Keynesian stimulus—the “multiplier effect”—have failed to live up to expectations.