Wednesday, June 24, 2015

Monkeying around with Multipliers

Over in a comment thread at mainly macro, I am engaged in a long discussion with the extremely polite, erudite and smart commenter weareastrangemonkey

I am putting my latest contribution here, because it is related to a continuing debate and contains links which required a few minutes of googling

we do agree that in 2013 there was fiscal consolidation and that only part of it was sequestration. I don't really have a serious objection to the practice of calling tax increases starting January 1 2013 part of sequestration.

I think it is important that sequestration did not correspond to a change in the trend of G (federal + state + local consumption plus investment). An estimate of the G multiplier using US data from the recovery is very imprecise but it happened to be 1.64 using data through 2014q3 .

(please do click the link -- these are the data which Sumner asserts proves Keynesians wrong (partly because he uses the word "sequester" sloppily in contrast to you)).

To the small extent that this is evidence of anything, it very much supports the ex ante guess of around 1.5 made by Krugman, C Romer and a whole bunch of Keynesians.

Prof Wren-Lewis does not accept the grudging relative praise, but I note that new as opposed to old Keynesian models imply that the correct fiscal variable is G not any sort of deficit. It is notably the fiscal variable Krugman discussed in his theory post on optimal fiscal policy in a liquidity trap (note the date).

However there were tax increases 2 months before sequestration. I am an ultra paleo Keynesian and I admit that they were not followed by the change I would have expected

I also note that this is a calculation with one data point (one payroll tax increase).

Overall considering overall US government consumption plus investment and Federal government taxes and transfers, the data look Keynesian, in spite of that one data point (which definitely stands out from the scatter)

1 comment:

Philip said...

The following post shows clearly that the Keynesian multiplier during recessions is negative.

And the first graph in this post shows clearly that the marginal propensity to consume seven years after the crash is not more than zero.

The marginal propensity to consume is the change in consumption to a given change in income. And the graph shows that the consumption remains absolutely constant despite income increases in some months. So any increase in income goes straight to saving. For example, the spike in saving rate at the end of 2012 is the result of the spike in income being entirely saved.