Sunday, June 9, 2019

Modern Monetary Theory 3


There’s a lot in at least some versions of MMT which overlaps with old Keynesian theory (and some, but rather less, which overlaps with new Keynesianism).  Let’s see what that overlap is (with the caveat that this is my interpretation of some MMT writings and that there are undoubtedly MMT theorists who would reject it, not to mention those MMT theorists who don’t believe in the validity of formal modelling at all).  I’ll stick key qualifiers in, but keep in mind that this is an outsider’s interpretation.  For links to various MMTers statements of their understanding of the theory, see here.


The starting point for many of the  MMT expositions which I’ve seen is a very Keynesian AD/AS diagram in Price/Output space.  The Aggregate Demand curve seems to be fairly conventional.  The Aggregate Supply curve is very much along the lines which Keynes discussed (MMT is a branch of post-Keynesian economics, remember) with a horizontal segment representing a region of output so far below potential (or full employment) output that increases in aggregate economic activity don’t result in price increases.  This differentiates the Keynesian AS curve from the neoclassical short run AS curve which is generally taken to be positively sloped everywhere, so that reductions in economic activity are always accompanied by reductions in the aggregate price level and increases in output are always accompanied by increases in aggregate prices. 

The other end of the Keynesian AS curve is a vertical segment at full employment, which corresponds to the vertical Long Run Aggregate Supply curve of the neoclassical macro model.  Arguably the only real gain from having a SAS and LRAS pair as opposed to having a single AS curve is in the representation of situations in which the economy is temporarily operating above potential (i.e. normal full employment) output.  In the neoclassical model we can slide along the SAS curve to a point beyond the LRAS curve, at which point intense competition for labour will drive wages up, shifting the SAS curve up, causing prices to rise and output to fall until we’re at a point of intersection between the SAS and LRAS curves.

Getting back to the Keynesian, single AS curve, some versions are strictly backward-L in shape, with the AS curve flat until potential output is reached, at which point it abruptly becomes vertical.  Other versions have a more gradual transition, with the AS curve entering a region of positive slope before potential output is reached.  Judging from exchanges in the Journal of Post-Keynesian Economics,  this difference seems to be a major point of contention between pro- and anti-MMT theorists.  I’ve got to admit that it strikes me as a dispute over how many angels can dance on the head of a pin, but I’m not a Post-Keynesian.

So some of the older MMT arguments which I’ve seen start from this pretty standard Keynesian AD/AS framework.  Equilibrium output is at the intersection of AD and AS, and if that intersection is at an output level well below (to the left of) potential output, you have a Keynesian unemployment equilibrium.

MMTers seem generally to be of the view that monetary policy by itself is pretty ineffectual as a tool for remedying unemployment.  They argue that conventional models argue that the interest rate operates mainly through the investment function investment, and that in a significant downturn investment is pretty inelastic with respect to interest.  On empirical grounds it’s pretty hard to dispute this (although that’s a gut feeling – to move beyond the level of rigor of a blog post would require estimating an investment function over a relatively long horizon) and MMTers can cite authority – Keynes reached the same conclusion, which was why he argued in favour of government spending in the General Theory.  Basically, MMT seems to argue that reductions by the central bank of the rate of interest will have minimal effect so far as shifting the AD curve to the right is concerned, so they come down in favour of government spending as the preferred tool for stimulus.
And this is where things start to go wonky.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.