There
has been a bit of a flutter of memes recently about new Keynesian
economics. Noah Smith digging out Robert
Barro’s 1989 harangue and Lars Syll’s comment that new Keynesian = new
Freidman.

I just
thought I would add my thoughts into the mix. To understand this you really
need a grasp of what has been going on in macroeconomics for the last 50
years. There is Keynes’s economics (the
stuff in the General Theory and lots of other writings) and “Keynesian”
economics that was developed mostly after Keynes had died. Keynesian economics was the stuff in
textbooks like IS/LM and AS/AD that dominated macroeconomic thinking until the
end of the 1970’s. Don’t forget, some of the key new classical results such as
policy neutrality and the neutrality of systematic monetary policy were derived
in this “Keynesian” framework by simply adding rational expectations (this
happened in the second half of the 1970’s with the “rational expectations
revolution”). Indeed, Milton Freidman said that “we are all Keynesians now”. For nearly all macroeconomists, there was
short-run nominal rigidity (IS/LM), but in the long-run all wages and prices
were flexible and you end up at the natural rate. This was called the neoclassical synthesis,
which I for one always associate with Don Patinkin’s masterpiece “Money,
interest and prices”.

So,
there was an academic consensus about the basic analytic framework for modelling things. There was certainly disagreement: at the time
lots of people did not like rational expectations. Also, there were many who thought that the
IS/LM framework left out a lot that was essential to Keynes’s original thought:
the title of Axel Leijonhufvud’s book “of Keynesian economics and the economics
of Keynes” says it all. I could also reference Robert Clower and Hyman Minsky.
Robert Lucas was putting forward a rather different view that we could
understand everything in terms of competitive equilibrium and Keynesian
concepts such as involuntary unemployment were just hokum. However, I think that it is fair to say that
up until the late 1970s, most macroeconomics, whether monetarist or Keynesian,
adopted the same basic setup.

Things
changed in the 1980s. The idea that you could explain economic fluctuations
using competitive equilibrium developed into the Real Business Cycle
approach. As the name implies, it was a
real model and did not even have money in the model or any concept of
inflation. In this period there was the first group of work that was at the
time designated new Keynesian. The key idea here was the introduction of
imperfect competition into macroeconomics (much as the young Paul Krugman was introducing
imperfect competition into the new international trade models), which made
possible the explanation of nominal rigidity based on an optimizing theory of pricing behavior
based on menu costs or bonded rationality (along with other ideas).

In the mid 1990s the new neoclassical synthesis bought these two strands together: the new Keynesian models of nominal
rigidity were combined with the RBC model of household behaviour (Euler
equations and all that). Now of course, whilst some people stressed the Keynesian
aspects of this (for example David Romer in his J.Econ perspectives paper in
1993), others saw it as a neo-monetarist exercise: for example Miles Kimball in
his 1996 paper “The Quantitative Analytics of the Basic Neomonetarist Model”.

So,
there has always been an underlying idea that the prevailing macroeconomic framework,
whether 60’s Keynesian, 80’s new Keynesian or 90’s first generation
neoclassical synthesis models are really “Keynesian”. Yes, they have nominal rigidity. However, there are lots of elements of the
economics of Keynes that have been left out. It is exactly the same now. The first generation new Keynesian models
became the second generation models (as exemplified by the Smets and Wouters
model, for example).

I
am not a historian of economic thought, but certainly from my recollection, the
criticisms made of the so-called Keynesian/new-Keynesian models have always
been much the same: they include amongst other things the inadequate treatment of uncertainty, money,
the assumption of a unique equilibrium. So no change there then.

Whether
you call the existing crop of models “new Keynesian” or “neo monetarist” is
really a matter of taste. However, one thing
is for sure: whilst the choice of name might have political or ideological
overtones, it sure as hell makes no scientific difference as to how good the
models are!