Christina Romer changes the rules mid-stream

My argument is that a stimulus will always have a multiplier greater than 1.0 if it is stimulus. Romer decided this is no longer the case. Her argument is basically if government grows than the stimulus is successful, even if the private sector shrinks. Multipliers less than one are perfectly fine. My critics will say, wait, the economy performed worse than expected. Still, even in a worse than expected system, a small stimulus would still be linear and still be greater than one, under her theory.

My logic is: The economy has had nine months to figure out what is expected nominal GDP growth would be after the shock. Any further computation by the economy is likely to be in the noise because the financial system has about a six month adaptive time constant. At the beginning of the second quarter, nominal GDP was settled, and any good effect from a Keynesian stimulus would have surprised us with an even modest increase if private sector GDP. Instead we get a 3% increase total governments share of GDP and a 1% decline in private sector GDP.

The chart above is Romer's effort to show correlation. About as poor a science as I have seen from the multiplier advocates. If you remove the Asian outliers, we get a negative correlation.

The stimulus theory is a fraud. The current distribution of multipliers for government spending is going to be severely distorted by the cause of the depression, which hits both government and non-government sectors. There is no justification to believe the general rise in government spending helps move the economy toward equilibrium if the multipliers are so distorted.